title_sr
stringlengths
11
267
title_en
stringlengths
0
918
abstract_sr
stringlengths
36
21.4k
abstract_en
stringlengths
43
16.7k
full_abstract
bool
2 classes
Prognostički značaj elektroencefalografskog nalaza kod novorođenčadi sa epileptičkim napadima
Prognostic significance of electroencephalographic findings in neonates with epileptic seizures
Epileptički napadi su najčešći neurološki poremećaj kod novorođenčadi i često prvi znak neurološke disfunkcije. Oni su predmet mnogih naučnih i kliničkih ispitivanja zbog njihovog snažnog prediktivnog uticaja na kognitivna i razvojna oštećenja. Etiologija neonatalnih napada je veoma heterogena, uglavnom akutna reaktivna. Veoma je važno odrediti uzroke neonatalnih napada, jer etiologija ima značajan uticaj na prognozu i ishod i utiče na dalje terapijske strategije. Ciljevi: Procena dugoročnog ishoda i utvrđivanje prediktora prognoze kod dece sa neonatalnim napadima. Materijal i metode: Retrospektivna studija je obuhvatila novorođenčad sa epileptičkim napadima koja su se lečila u Institutu za zdravstvenu zaštitu majke i deteta Srbije “Dr Vukan Čupić” u Beogradu od 1. januara 2005. godine do 31. decembra 2015. godine. Podaci su dobijeni iz istorija bolesti i baze podataka elektroencefalograma (EEG). Procenjena je prediktivna vrednost sledećih parametara: karakteristike pacijenata (pol, gestacijska starost, telesna masa na rođenju, ocean na rođenju, primena mehaničke ventilacije), etiologija, karakteristike napada (tip, vreme početka, farmakorezistencija) i EEG nalaz (osnovna aktivnost, paroksizmalne promene). Ishod neonatalnih napada je procenjen na kraju perioda praćenja i kategorisan je kao smrtni ishod, neurološki nalaz, razvojni količnik i epilepsija. Univarijantna i multivarijantna logistička regresiona analiza korišćene su za procenu prediktora ishoda neonatalnih napada. Rezultati: Ispitivanjem je obuhvaćeno 168 novorođenčadi (109 muškog pola, 59 ženskog pola) sa epileptičkim napadima, prosečnog uzrasta 4 godine (SD 3.4) na kraju praćenja (opseg 1-12 godina). Normalni neurološki razvoj bez epilepsije zabeležen je kod 131 pacijenta (78%), neurološke abnormalnosti kod 31 (19,0%), nepovoljan razvojni količnik kod 28 (17,2%), epilepsija kod 12 (7,4%) i smrtni ishod kod 7 pacijenata (4,17%). Zaključak: Dugotrajni ishod kod dece sa neonatalnim napadima je povoljan. Novorođenčad sa patološkom osnovnom EEG aktivnošću, rezistencijom na antiepileptičke lekove i niskom ocenom na rođenju su u velikom riziku od lošeg ishoda. Novorođenčad ženskog pola su u većem riziku od smrtnog ishoda nego novorođenčad muškog pola.
Neonatal seizures are the most common neurological disorder in newborns, and often the first sign of neurologic dysfunction. The seizures in neonatal period are the subject of many scientific and clinical investigations due to their powerful predictor value of long-term cognitive and developmental impairment. The etiology of neonatal seizures is very heterogeneous, mostly acute reactive. It is very important to determine causes of neonatal seizures, since etiology has significant impact on prognosis and outcome and influences further therapeutic strategies. Objectives: Evaluation of long-term outcome and assesment of predictors of prognosis in children with neonatal seizures. Matherial and methods: The retrospective study included children with neonatal seizures treated in the Mother and child health care Institute “Dr Vukan Čupić” in Belgrade, from January the 1st 2005 to December the 31st 2015. The data were collected from medical charts and database of electroencephalograms (EEG). The predictive value of following parameters was evaluated: characteristics of the patients (gender, gestational age, birth body weight, Apgar score, artificial ventilation), etiology, characteristics of seizures (type, time of onset, resistance to tretment) and EEG (background activity, paroxysmal discharges). The outcome of neonatal seizures was assessed at the end of follow-up period and was categorized as lethal outcome, neurological abnormalities, intellectual disability and epilepsy. Univariate and multivariate logistic regression analyses were used to assess predictors of neonatal seizures outcome. Results: The study included 168 (male 109, female 59) children with neonatal seizures, mean aged 4 (SD 3.4) years at the end of the follow up (range 1-12 years). There was normal neurological development without epilepsy in 131 patients (78%), neurological abnormality in 31 (19.0%), intellectual disability in 28 (17.2%), epilepsy in 12 (7.4%) and lethal outcome in 7 patients (4.17%). Conclusion:.Long-term outcome in children with neonatal seizures is favorable. The newborns with anbormal background activity on EEG, resistant seizures to antiepileptic drugs and low Apgar score are at a high risk of poor outcome. Females are at higher risk of lethal outcome than males.
true
Uticaj kongresnog turizma na razvoj destinacije
Influence of the meetings industry on destination development
Kongresni turizam je važan činilac performansi koje se ostvaruju na nivou destinacije, pogotovo imajući u vidu da se poslednjih nekoliko godina beleži trend povećavanja broja međunarodnih događaja. Aktivnim nastupom na međunarodnom kongresnom tržištu, etablirane i nove destinacije nastoje da privuku što veći broj kongresa, jer se time doprinosi ukupnim performansama destinacije, ali i promociji lokalne ekspertize u oblasti iz koje se kongres organizuje. Da bi bile uspešne u tome, destinacije, nezavisno od nivoa razvoja, moraju da razumeju kompleksnost tržišta kongresnog turizma, kao i da implementiraju odgovarajuće tehnike menadžmenta destinacije. U disertaciji se analizira kongresno tržište i specifičnosti menadžmenta destinacije u kongresnom turizmu i istražuje uticaj kongresnog turizma na destinaciju. Istraživanje se odnosi na sistematizaciju aktivnosti menadžmenta destinacije na nivou kongresnog biroa i definisanje značaja aktivnosti kongresnog biroa za organizatore događaja. Disertacija ima za cilj utvrđivanje povezanosti između kongresnih performansi destinacije sa jedne strane i turističkih performansi, hotelskih performansi i ekonomskim rastom i razvojem destinacije, sa druge strane. U istraživanju se polazi od toga da je uspeh kongresne destinacije uslovljen destinacijskim menadžmentom. Pregled literature pokazuje da se menadžmentu destinacije pristupa sa strateškog aspekta, dominantno u sferi odmorišnog turizma, dok je pitanje menadžmenta destinacije u kongresnom turizmu predmet analize manjeg broja autora. Uvažavajući specifičnosti kongresnog turizma i analizirajući okvir menadžmenta destinacije, u disertaciji se predlaže sistematizacija aktivnosti menadžmenta destinacije na nivou kongresnog biroa koja se dalje detaljno analizira. Disertacija u obzir uzima i činjenicu da je kongresni turizam postavljen kao „quick win“ strateški proizvod za turizam Srbije. U disertaciji se istražuje činjenica da u prethodnim godinama kongresni turizam u Srbiji beleži značajne promene, zbog čega se posebno analiziraju kongresne performanse Srbije i razvoj kongresnih kapaciteta. 4 Predmetom naučno-istraživačkog rada i ostvarenjem ciljeva, disertacija doprinosi izučavanju kongresnog turizma i destinacijskih menadžment organizacija u Srbiji i regionu i to kroz analizu modela i alata baziranih na praktičnim poslovnim veštinama primenljivih na međunarodnom tržištu.
Meetings industry (congress tourism) is the inevitable component of performances achieved on the destination level, especially having in mind that in last several years the trend of increase of the number of international meetings has been recorded. By being active on international meetings market, established and new destination are trying to attract larger number of events, since it is influencing the total destination performances, as well as promotion of local expertise from the specific field in which event is being organised. In order to be successful in this, destination, no matter the development level, need to understand the meetings market complexity, and to implement appropriate techniques of destination management. Dissertation analyses meetings markets and specifics of destination management in the congress tourism (meetings industry) and researches the influence of congress tourism on destination. Dissertation is focused on systematisation of destination management activities from the perspectives of convention bureaus and on defining importance of convention bureau activities for event planners. Dissertation aims at defining relations among destination congress performances on one side, and tourism performances, hotel performances and economic growth and development, on the other side. Research starts from the facts that success of congress destination is being influenced by destination management. Literature review shows that destination management is being research from strategic perspective, foremost in the field of leisure travel, while aspect of destination management in congress tourism is being researched by smaller number of authors. Respecting specifics of congress tourism and by analysing destination management framework, dissertation suggests systematisation of convention bureau destination management activities that is further detailed analysed. Dissertation takes into account the fact that congress tourism is seen as “quick win” strategic product for Serbian tourism. In dissertation the fact, that in last period congress tourism in Serbia reached strong changes, is researched, and special analyse of congress performances in Serbia and congress infrastructure is also included. 6 With the defined subject of scientific research and by accomplishing defiend aim, dissertation supports further academic research of congress tourism and destination management organisations in Serbia and region, especially through models and tools that are based on practical business skills possible for implementation on international market.
true
Modelling the electron and hole states in semiconductor nanostructures by the multiband k.p theory
Modelovanje elektronskih i šupljinskih stanja u poluprovodničkim nanaostrukturama pomoću višezonske k.p teorije
Rezime: Disertacija predstavlja analizu elektronskih i šupljinskih stanja u različitim nanostrukturama korišćenjem k . p formalizma. Razmatrane su Si=SiO2 kvantne jame, GaAs/(Al,Ga)As kvantne jame, prstenolike GaAs/(Al,Ga)As kvantne tačke formirane metodima kapljične epitaksije i (In,Ga)As/GaAs kvantne tačke proizvedene epitaksijom u Stranski-Krastanov modu. Dat je pregled tehnika fabrikacije koje se obično koriste za proizvodnju razmatranih nanostruktura. Osnovna svojstva elektronske strukture silicijuma i III-V jedinjenja su takođe ukratko razmatrana. Zatim su predstavljeni osnovi k . p teorije. Prvo je prikazan koncept ove teorije, a zatim je k . p formalizam primenjen na izvođenje jednozonske jednačine efektivnih masa. Zatim je predstavljena k . p teorija degenerisanih energetskih nivoa, što je slučaj u valentnoj zoni poluprovodnika sa dijamantskom i sfaleritnom rešetkom. Izveden je Dresselhaus- Kip-Kittelov model stanja u valentnoj zoni u odsustvu spin-orbitne interakcije korišćenjem vremenski nezavisne teorije perturbacija za degenerisana stanja.Potom je uzeta u obzir spin-orbitna interakcija u okviru šestozonskog Latindžer-Konovog modela. Stanja u provodnoj i valentnoj zoni u poluprovodnicima sa širokim energetskim procepom obično se mogu dobro opisati jednozonskom teorijom efektivnih masa i šestozonskim Latindžer- Konovim modelom, respektivno. Međutim, primena ovih modela je ograničena na stanja sa malim vrednostima talasnog vektora elektrona k, dok se ne mogu primeniti za opis elektronske strukture daleko od G tačke. Potreba za dobrim opisom disperzije elektronskih zona daleko od centra Briluenove zone postoji u silicijumu, gde niskoenergetska elektronska stanja imaju veliku vrednost k. Pogodna teorija za ovaj slučaj je 30-zonski k . p model. Ovaj metod omogućuje računanje elektronske strukture u celoj prvoj Briluenoj zoni, stoga se može koristiti kako za direktne, tako i za indirektne poluprovodnike. Zatim je ukratko izložena aproksimacija anvelopnih funkcija, koja predstavlja metod za...
The dissertation presents a study of the electron and hole states in various semiconductor nanostructures using the k p formalism. Si/SiO2 quantum wells, GaAs/ (Al,Ga)As quantum wells, GaAs/(Al,Ga)As ringlike quantum dots grown by droplet epitaxy, and ringlike (In,Ga)As/GaAs quantum dots formed by epitaxy in the Stranski-Krastanov mode are all considered. An overview of the techniques which are used to fabricate the analyzed nanostructures is first given. The most basic properties of the electronic structure of silicon and III-V compounds are briefly described. The fundamentals of the k p theory are explained, and then applied to derive the equations of the single-band effective-mass theory. The kp theory for degenerate energy levels, which is the case in the valence band of diamond and zinc-blende semiconductors, is then presented. By using the perturbation theory, the three-band Dresselhaus-Kip- Kittel model of the valence-band states is derived for the case of absent spin-orbit interaction. The spin-orbit interaction is taken into account in the 6-band Luttinger-Kohn model, whose derivation is also briefly described. The conduction and valence band states of wide band-gap direct semiconductors and their nanostructures are usually well described by the single-band effective-mass theory and the 6-band Luttinger-Kohn model, respectively. The mentioned k p models are well suited for low k states, whereas they cannot be straightforwardly adopted to model states which are far away from the G point. Such a situation arises in silicon, where low energy states in the conduction band have large k. The model which has recently been proposed to handle such a situation is the 30- band k p theory. It is the full-zone approach, which provides an accurate description of the electronic structure of both direct and indirect semiconductors in the whole Brillouin zone. The envelope function approximation, which is a way to apply the k p theory to nanostructures, is also briefly discussed. Also, it is demonstrated how effects of magnetic...
false
Finansijsko izveštavanje i vrednovanje performansi entiteta javnog sektora
Financial reporting and performance evaluation of public sector entities
Predmet ovog rada je istraživanje uticaja savremenih reformi javnog sektora na računovodstvo i finansijsko izveštavanje entiteta javnog sektora. Osnovni cilj rada je da se, kritičkom analizom efekata primene različitih osnova računovodstva na vrstu i informacioni obuhvat finansijskih izveštaja entiteta javnog sektora, oceni u kojoj meri se, prelaskom sa računovodstva i finansijskog izveštavanja zasnovanog na gotovinskoj osnovi na računovodstvo i finansijsko izveštavanje zasnovano na obračunskoj osnovi, može uticati na unapređenje efikasnosti i efektivnosti aktivnosti entiteta javnog sektora i način demonstriranja njihove javne odgovornosti za efekte korišćenja poverenih im javnih resursa. Dopunski cilj istraživanja je identifikacija eventualnih slabosti u normativnoj regulativi i sistemu finansijskog izveštavanja entiteta javnog sektora u Republici Srbiji i formulisanje predloga za njihovo umanjenje ili potpuno eliminsanje. Polazeći od razlika u karakteristikama između entiteta javnog i privatnog sektora, primarnim korisnicima njihovih finansijskih izveštaja i iz njih proistekle razlike u ciljevima njihovog finansijskog izveštavanja, u radu su prezentovani specifični koncepti relevantni za finansijsko izveštavanje entiteta javnog sektora. Pored toga, analizirani su međunarodni izveštajni okviri (statistički i računovodstveni) za entitete javnog sektora i aktivnosti na njihovoj harmonizaciji. Posebna pažnja posvećena je prezentovanju konceptualnih karakteristika računovodstvenih sistema zasnovanih na gotovinskoj, obračunskoj, modifikovanoj gotovinskoj i modifikovanoj obračunskoj osnovi, sa akcentom na karakteristike gotovinske i obračunske osnove. Za gotovinsku i obračunsku osnovu računovodstva prikazani su izveštajni modeli usaglašeni sa zahtevima Međunarodnih računovodstvenih standarda za javni sektor. ii Komparacijom karakteristika finansijskih informacija koje produkuje finansijsko izveštavanje zasnovano na računovodstvu na gotovinskoj i na računovodstvu na obračunskoj osnovi ocenjena je njihova podobnost za ostvarenje ciljeva finansijskog izveštavanja. Preciznije, ocenjena je upotrebljivost finansijskih informacija u procesima vrednovanja performansi entiteta javnog sektora sa funkcionalnog, finansijskog i aspekta usaglašenosti sa relevantnom regulativom, javno proklamovanim politikama i usvojenim profesionalnim standardima. Od izuzetnog značaja je činjenica da je u radu izvršeno istraživanje karakteristika normativne osnove za finansijsko izveštavanje entiteta javnog sektora u Republici Srbiji i njenog uticaja na sposobnost, njome uslovljenog, izveštajnog modela da generiše finansijske informacije koje su podobne za ocenu efikasnosti i efektivnosti aktivnosti entiteta javnog sekora. Na osnovu toga urađena je evaluacija trenutnog stanja i koncipirani su predlozi za unapređenje normativne osnove finansijskog izveštavanja u našoj zemlji. U sklopu izlaganja koja se odnose na praksu finansijskog izveštavanja entiteta javnog sektora u Republici Srbiji, prezentovani su rezultati anketiranja profesionalnih računovođa i menadžera zaposlenih u javnom sektoru koji su poslužili kao osnova za izvođenje zaključaka o njihovoj tekućoj osposobljenosti da prihvate predložene reforme računovodstva i finansijskog izveštavanja.
This thesis examines the impact of the contemporary public sector reforms on the accounting and financial reporting of public sector entities. The main objective of the thesis is to perform a critical analysis of the effects on the type and information coverage of financial statements of public sector entities which are produced by implementing different accounting bases, in order to determine to what extent can the transition from cash based accounting and financial reporting to the accrual based accounting and financial reporting lead to enhanced efficiency and effectiveness of the public sector entities, as well as to improved manner of demonstrating their public accountability for the results of usage of public funds allocated to them. The additional objective of the research is to identify potential weaknesses in the underlying legislation and financial reporting system of public sector entities in the Republic of Serbia and to formulate the proposal for their reduction or removal. Starting from the differences in main features of public and private sector entities, in primary users of their financial reports and the resulting divergence in the objectives of their financial reporting, the thesis presents specific concepts relevant for financial reporting of public sector entities. Additionally, internationally recognized reporting frameworks (statistical and accounting) for public sector entities are analysed and activities aimed at their harmonisation are elaborated. Special attention is paid to presenting the conceptual characteristics of accounting systems based on cash, accrual, modified cash and modified accrual bases, with particular focus on cash and accrual methods. For cash and accrual accounting, the iv thesis presents reporting models that comply with the requirements of International Public Sector Accounting Standards. By comparing the features of financial information produced by financial reporting under the cash and accrual basis of accounting, their suitability for achieving financial reporting objectives is assessed. More precisely, the usability of financial information for the public sector performance evaluation is examined from the functional and financial perspective, as well as regarding their compliance with relevant legislation, publicly proclaimed policies and adopted professional standards. It is particularly important to note that the thesis examines the normative basis of public sector financial reporting in the Republic of Serbia and its impact on the ability of the reporting model, conditioned by the applicable legislation, to generate financial information suitable for assessing the efficiency and effectiveness of activities performed by public sector entities. As a result, the present situation is evaluated and proposals for improving normative basis for financial reporting in our country are formulated. Presentation of the practice in the financial reporting of public sector entities in the Republic of Serbia also contains results of survey conducted among professional accountants and managers employed in the public sector, which served as a basis for drawing conclusions on their capability to accept the proposed reforms in accounting and financial reporting.
true
Migracije visokoobrazovanih lica iz Srbije od 1991. godine u Kanadu i Sjedinjene Američke Države
Predmet istraživanja ovog rada su migracije visokoobrazovanih lica koja su diplomirala na nekom od univerziteta u Srbiji, emigrirala su iz Srbije 1991. godine i kasnije, i sada žive u SAD i Kanadi. Glavni cilj predložene doktorske disertacije je da doprinese boljem razumevanju fenomena migracija visokoobrazovanih lica i stručnjaka, a naročito problema ubrzanog iseljavanja ovih struktura radne snage iz Srbije od 1991. godine. U radu su razmatrana teorijska stanovišta o činiocima savremenih medjunarodnih migracija, i ulozi države, i iz perspektive posmatranih zemalja prijema i Srbije kao zemlje porekla. Data je detaljna analiza statističkih podataka o gradjanima Srbije u inostranstvu, njegovih osnovnih karakteristika prema području porekla i državi prijema. Opsežno anketno istraživanje predstavlja pokušaj da se iz perspektive samih migranata unaprede postojeća znanja iz oblasti migracija visokoobrazovanih kadrova iz Srbije. Iako se njegovi rezultati ne mogu smatrati reprezentativnim, oni obuhvataju različite aspekte emigracije visokoobrazovanog stanovništva, a saznanja do kojih se došlo u prvom redu se odnose na upotpunjavanje slike o razlozima emigriranja, integrisanosti na tržištu rada i društvu zemlje prijema, odnosu prema zemlji porekla, kao i upoznavanju sa planovima za budućnost. Rezultati teorijskih razmatranja upućuju na potrebu kompleksnog pristupa i dinamičke perspektive proučavanja u analizi spoljnih migracija. Analiza statistike o migracijama ukazuje i na neophodnost kombinovanja podataka iz statističkih izvora zemlje porekla i zemalja prijema. Jedna od najvećih prepreka za bolje definisanje kompleksnih međunarodnih migracionih tokova i procesa je što ne postoji statistička osnova koja istovremeno, sveobuhvatno i metodološki usaglašeno može da prati migraciona kretanja u zemljama porekla i zemljama prijema. Rezultati nekoliko poslednjih popisa stanovništva Srbije, kao izvora koji pruža najviše informacija o karakteristikama gradjana Srbije u inostranstvu, ukazuju da se u protekle 3-4 decenije na odredjenim područjima zemlje snažno razvijaju migracione veze s najvažnijim evropskim imigracionim zemljama. Takodje, postoje značajne razlike u udelima visokoobrazovanih po zemljama destinacije, naročito u popisima iz 2002. i 2011. SAD i Kanada su zemlje u kojima IV fakultetski obrazovana lica iz Srbije čine polovinu ukupnog broja lica iz Srbije u tim zemljama. Rezultati ankete pokazuju da se na vrhu značajnosti razloga za emigraciju nalaze motivi koji se odnose na ekonomske i političke prilike u zemlji emigracije. Veliki uticaj krize kao razloga za emigraciju naročito je prisutan kod ispitanika u Kanadi, koji su u najvećoj meri emigrirali zajedno sa drugim članovima porodice, i to u vreme najdublje krize u Srbiji. Nakon početnog perioda privikavanja na novu sredinu, učesnici su bili uglavnom zadovoljni stepenom integracije, kako na tržištu rada, tako i u društvu prijema u širem smislu. Medjutim, prisutne su i odredjene razlike po polu u odnosu na stavove o integrisanosti na tržištu rada. Muškarci su u svim posmatranim stavovima izrazili nešto viši stepen zadovoljstva nego žene, i to u obe posmatrane države prijema. Pokazano je i da emocionalne veze sa sredinom iz koje potiču ostaju jake čak i kod ispitanika koji su emigrirali na početku posmatranog perioda, tj. početkom 1990-ih. Takodje, značajno su upućeni i na sunarodnike u sredinama u kojima žive, ne samo u smislu održavanja prijateljskih i kolegijalnih odnosa, već i u smislu razvijanja lokalnih mreža solidarnosti i pomoći. Većina ispitanika smatra da su šanse za povratak veoma male ili čak da ne postoje. To upućuje na zaključak da građenje “mostova” sa našom intelektualnom dijasporom prvenstveno treba da se razvija kroz pronalaženje adekvatnih načina za razmenu i cirkulaciju znanja, odnosno saradnje „na daljinu. Ispitanici u Kanadi su u znatno većoj meri zainteresovani za učešće u privrednim projektima, dok je za ispitanike u SAD-u na prvom mestu učešće u oblasti naučne delatnosti. Jačanje veza, stvaranje i kontinuirano održavanje mrežnog povezivanja različitih struktura naše intelektualne dijaspore, kao mogućih prenosilaca novih znanja i modernih tehnologija bi moglo da proizvede značajan podsticaj promenama u Srbiji, kao zemlji porekla. To se ne odnosi samo na privredni, nego i razvitak u širem društvenom smislu. Medjutim, osnovni preduslov za to je da se mere države ubrzano i kontinuirano usmeravaju ka politikama koje podstiču pozitivnu klimu za naučno-istraživački rad, inovacije i poslovnu saradnju.
The subject of this paper is the migration of highly educated persons graduated from a university in Serbia, who emigrated from Serbia in 1991 and later, and are now living in the United States and Canada. The main objective of the proposed doctoral dissertation is to contribute to a better understanding of the phenomenon of migration of highly qualified persons and professionals, in particular the problem of rapid emigration of these labour force structures from Serbia since 1991. The paper considers theoretical aspects of contemporary international migration factors, and the role of the state, both from the perspective of the observed receiving countries and Serbia as a country of origin. It provides a detailed analysis of statistical data on citizens of Serbia abroad, their main characteristics according to the area of origin and the receiving country. A comprehensive survey is an attempt to improve the existing knowledge in the field of emigration of highly educated people from Serbia from the perspective of migrants themselves. Although its results cannot be considered representative, they include various aspects of emigration of highly educated population, and its findings primarily relate to a full picture of the reasons for emigration, integration in the labour market and society of the receiving country, relations to the country of origin, as well as their plans for the future. The results and theoretical considerations suggest the need for a complex approach and dynamic perspective to researching the analysis of international migration. The migration statistics analysis also points to the necessity of combining data from statistical sources of the country of origin and receiving countries. One of the biggest obstacles to a better definition of the complex international migration flows and processes is that there is no statistical basis that can simultaneously, comprehensively and methodologically monitor migratory movements in the countries of origin and receiving countries. Results of the last several population censuses in Serbia, as a source that provides the most information about the characteristics of Serbian citizens abroad, indicate that in the past 3-4 decades in certain areas of the country migration links with the most important European immigration countries have been vigorously developed. In VI addition, there are significant differences in the shares of highly educated people by destination countries, particularly in population censuses of 2002 and 2011. The USA and Canada are countries where university-educated persons from Serbia make up half the total number of persons from Serbia in these countries. The survey results show that the most important reasons for emigration are the motives related to the economic and political conditions in the country of emigration. The large impact of the crisis as a reason for emigration is especially present among respondents in Canada, who largely emigrated together with other family members, at the time of the deepest crisis in Serbia. After an initial period of adjustment to the new environment, the respondents are generally satisfied with the degree of integration in both the receiving labour market and society in a broader sense. However, there are also some differences by gender in relation to attitudes towards integration in the labour market. In all the observed attitudes, men expressed somewhat higher level of satisfaction than women, in the both observed receiving countries. It has also shown that emotional connection with the area of origin remains strong even in the respondents who emigrated at the beginning of the observed period, i.e. in the early 1990s. Moreover, they also significantly rely on their compatriots in the communities where they live, not only in terms of maintaining friendly and collegial relations, but also in terms of developing local networks of solidarity and assistance. Most respondents believe that the chance to return is very small or even non- existent. This suggests that building "bridges" with our intellectual diaspora should be primarily developed through finding adequate ways for the exchange and circulation of knowledge, i.e. outsourcing. The respondents in Canada are much more interested to participate in economic projects, while those in the US would primarily participate in the area of science. Strengthening ties, the creation and continuous maintenance of networks that link different structures of our intellectual diaspora as potential carriers of new knowledge and modern technologies could produce significant impetus to changes in Serbia, as the country of origin. This applies not only to economic development but also to the development in the wider social sense. However, the basic precondition for this is that the state measures are rapidly and continuously directed towards policies that foster a positive climate for scientific research, innovation and business cooperation.
true
Elektrohemijsko ponašanje bakra u kiselom rastvoru natrijum-sulfata u prisustvu organskih inhibitora
Electrochemical behaviour of copper in acidic sodium sulphate solution in the presence of organic inhibitors
Elektrohemijsko ponašanje bakra u ovom radu ispitivano je u kiseloj sulfatnoj sredini u prisustvu organskih jedinjenja: purina, cisteina, 4(5)-metilimidazola i 2-merkapto- 1-metilimidazola. Na onovu rezultata dobijenih ispitivanjem ovih jedinjenja može se konstatovati da njihovo prisustvo uzrokuje smanjenje gustine struje, odnosno da se pomenuta jedinjenja ponašaju kao inhibitori korozije bakra u kiseloj sulfatnoj sredini. Najveća vrednost efikasnosti inhibicije prilikom istraživanja dobijena je kada je elektroda bila izložena najvećoj koncentraciji inhibitora. Vrednost potencijala otvorenog kola i korozioni potencijal, u zavisnosti od inhibitora, pomeraju se i ka negativnijim i ka pozitivnijim vrednostima, ali kako ovo pomeranje nije značajno, može se reći da su inhibitori mešovitog tipa. Inhibitorski efekt organskih jedinjenja javlja se kao posledica intreakcije njihovih molekula i površine bakra tj., građenja kompleksa sa bakrom ili adsorpcija molekula inhibitora na površini bakra. Svi korišćeni inhibitori pokoravaju se Lengmirovoj adsorpcionoj izotermi. Dobijene vrednosti Gibsove slobodne energije adsorpcije čija je maksimalna vrednost dobijena za cistein -47,055 kJ mol-1, a minimlna vrednost za 4(5)- metilimidazol -29,220 kJ mol-1, ukazuju na spontanu adsorpciju koja se može okrakterisati uglavnom kao fizička adsorpcija sa učešćem hmijske, osim u slučaju cisteina gde se može reći da je adsorpcija molekula po karkteru hemijska adsorpcija...
Electrochemical behaviour of copper was investigated in acidic sulpahte media in presence of organic compounds: purine, cisteyne, 4(5)-methylimidazole and 2-mercapto-1- methylimidazole was studied. From the obtained results gathered by examination of these compounds it could be concluded that their presence causes decrease of current dencity. Mentioned compounds act as copper corrosion inhibitors in acidic sulphate media. The biggest value of inhibition eficiency obtained during researching was when electrode were exposed to highest concentration of inhibitor. Value of open cisrcuite potential and corrosion potential depends of inhibitor and shifted toward more negative or toward more pozitive values, but this shift is not so significant it can be said that those inhibitors act as a mixed type. The inhibitory effect of organic compounds occurs as consequence of interaction of their moleculs and copper surface, that is formation of complex with copper or adsorption of inhbitor molecules on copper suraface. All used inhibitors obey Langmuir adsorption isotherm. Obtained values Gibbs free energy of adsorption with maximum value for cisteyne -47,055 kJ mol-1 and minimum value for 4(5)-methylimidazole -29,220 kJ mol-1, point to spontaneous adsorption. It can be characterized as a strong physical adsorption with slight chemisorption. But the exception is observed when cisteyne was used as inhibtor when adsorption of inhibitor moleculs can be characterized as a chemisorption. ...
false
Obnavljanje pogona distributivnih sistema u okviru razvoja inteligentnih elektroenergetskih mreža
Smart grid: service restoration in distribution power systems
Prilikom rešavanja problema obnavljanja pogona distributivnih sistema u postojećoj literaturi su uglavnom upotrebljavane ili modifikovane već poznate metode rešavanja problema iz drugih oblasti elektroenergetike. U radu je prezentovan pregled i opis postojećih načina rešavanja obnavljanja pogona U ovoj disertaciji je razvijen i verifikovan novi način rešavanja problema obnavljanja pogona. U disertaciji se razmatra problem obnavljanja pogona distributivnih sistema i u svetlu aktuelnih i globalnih napora kojim bi se postojeće elektroenergetske mreže transformisale u inteligentne elektroenergetske mreže. Iako postoji određeni broj konvencionalnih metoda i načina rešavanja obnavljanja pogona u distributivnim sistemima ipak novi koncept inteligentnih mreža zahteva usavršavanje postojećih kao i pronalaženje novih načina rešavanja ovog problema. Efikasno rešavanje obnavljanja pogona distributivnih sistema se u predloženoj temi ostvaruje pomoću algoritma koji je zasnovan na Primovom algoritmu, odnosno njegovoj modifikaciji. Primov algoritam, uključujući navedenu modifikaciju, je matematički detaljno opisan a zatim je primenljivost modifikovanog algoritma na rešavanje obnavljanja pogona distributivnih sistema sveobuhvatno analizirana. Iz toga proističe novi način rešavanja obnavljanja pogona distributivnih sistema koji se efikasno uklapa u širi koncept inteligentne mreže. Novi način rešavanja obnavljanja pogona je sastavni deo kompletno razvijenog algoritmom koji uključuje i proračune tokova snaga, rasterećenje preopterećenih elemenata prilikom rešavanja obnavljanja pogona kao i generisanje potrebnih prekidačkih akcija. Na kraju je razvijeni algoritam testiran na modelu složene distributivne mreže. Cilj istraživanja i doprinosi Cilj istraživanja i konkretan doprinos u predloženoj disertaciji je razvoj logike modifikovanog Primovog algoritma sa zadatkom nalaženja novog načina rešavanja obnavljanja pogona distributivnih sistema. U tom svetlu je detaljno prikazana modifikacija Primovog algoritma u kontekstu predloženog načina rešavanja obnavljanja pogona distributivnih sistema. Ključni doprinos je formiranje efikasnog, novog algoritma za obnavljanje pogona distributivnih sistema. Verifikacija ovog doprinosa je moguća samo postizanjem zadovoljavajućih performansi u postupcima realnih aplikacija na složene zadatke obnavljanja pogona i to je takođe dodatni zadatak u disertaciji. Naravno time je stvorena mogućnost za dalja istraživanja i usavršavanja postupaka obnavljanja pogona distributivnih sistema i provera primenljivosti na rešavanje drugih problema kako u oblasti elektroenergetike tako i u drugim tehničkim disciplinama. Predloženi način rešavanja je ilustrovan za slučaj kada je primarni cilj obnavljanja pogona minimizacija neisporučene električne energije potrošačima koji su ostali bez napajanja posle identifikacije i izolacije kvara na distributivnoj mreži. Sekundarni cilj je minimizacija broja prekidačkih akcija potrebnih za realizaciju obnavljanja pogona. Takođe, predloženi način rešavanja omogućuje definisanje željenog uklopnog stanja mreže nakon realizacije obnavljanja pogona u nenapojenom području. Ovi ciljevi zajedno sa unapređenjima kod njihove realizacije zaokružuju doprinose u disertaciji. Zaključak U doktorskoj disertaciji je razmatran problem obnavljanja pogona u distributivnim sistemima uključujući ulogu obnavljanja pogona u inteligentnim elektroenergetskim mrežama. Predložen je novi način rešavanja obnavljanja pogona u distributivnim sistemima. Glavni doprinosi ove disertacije su:  Povećan broj potrošača kojima je obezbeđeno obnavljanje napajanja električnom energijom posle poremećaja u distributivnom sistemu.  Smanjenjen broja prekidačkih akcija neophodnih da se dobijeno rešenje obnavljanja napajanja električnom energijom realizuje na terenu.  Mogućnost definisanja željene pogonske strukture oblasti pogođene kvarom. Željena pogonska struktura se definiše pre početka nalaženja rešenja obnavljanja pogona.  Smanjenjeno prosečno vreme trajanja prekida napajanja električnom energijom u toku godine.  Smanjeni troškovi obnavljanja pogona.  Poboljšane performanse distributivnog sistema. Predloženi način rešavanja može biti dobra osnova za dalja istraživanja obnavljanja pogona prilikom teških kvarova u gradskim distributivnim sistemima, kao i doprinos daljem razvoju problematike obnavljanja pogona distributivnih sistema, uključujući samostalno obnavljanje pogona sistema koje treba da bude jedna od karakteristika inteligentnih mreža.
Methods and approaches for the solving of the service restoration problem in the existing literature are based on well known methods that are also used for the solving of other problems in electrical power systems. Review of the methods and approaches is presented in this dissertation. New approach for the solving of the service restoration in distribution systems is developed and verified in this dissertation. The relation between the service restoration problem and transformation of conventional electrical power networks into smart grid are addressed in the dissertation. Although there is a number of known methodologies and approaches for solving the service restoration of the distribution power systems, an improvement of existing methodologies and approaches or research for new ones in order to accommodate the smart grid requirements are required. An effective service restoration approach proposed in the dissertation is based on modified Prim’s algorithm. Prim’s algorithm including its modification is mathematically described in details followed by comprehensive analysis of the modified Prim’s algorithm implementation to the solving of the service restoration. The new approach for the solving of the service restoration problem in distribution networks that can be incorporated into smart grid concept is derived from the analysis. Proposed new approach for the solving of the service restoration problem is incorporated in the developed algorithm. The algorithm also contains other typical distribution power system analysis tools such as load flow, load transfer from elements that are overloaded during the service restoration calculations, and the filtering of the required switching actions. The developed algorithm is tested on the complex model of the distribution network. Objectives and scope of work The objective of the dissertation is to define a logic of modified Prim’s algorithm in order to develop a new approach for the solving of the service restoration of distribution systems. In a context of the proposed new service restoration approach the modification of the Prim’s algorithm is presented in details. Major contribution of the dissertation is the generation of the new effective algorithm for the service restoration of the distribution system. In addition, the objective is to verify the algorithm performance on complex service restoration tasks. The proposed approach can be implemented in further research of the service restoration problem in distribution systems including the implementation of the approach for the solving of other problems in electrical power systems analysis but also in other technical areas. The proposed service restoration approach is illustrated with the algorithm where the primary objective is the minimisation of the unsupplied load to the de- energized consumers. Secondary objective is the minimization of the switching actions required for the implementation of the calculated service restoration solution on the network. In addition, the preferred network operational configuration of the de- energized area after the restoration can be defined by the proposed approach. Conclusion The service restoration problem of distribution systems in light of smart grid concept is considered in the dissertation. The new approach for the service restoration of distribution systems is proposed in the dissertation. Major contributions of the proposed algorithm in the dissertation are listed as follows:  The power supply is provided to maximal number of de-energized consumers.  The number of switching actions required for the implementation of the service restoration solution is decreased.  The preferred network operational configuration of the de-energized area after the restoration can be defined.  System Average Interruption Duration Index is decreased.  The service restoration costs are decreased.  The overall performance of the distribution system is improved. The proposed approach for the solving of the service restoration problem can be a base for further research of the service restoration problem including selfhealing that is one of the major characteristics of the Smart Grid concept.
true
Višekriterijumsko odlučivanje u upravljanju i optimizaciji ekonomije obrazovanja
Multiple criteria decision aiding in economics of education management and optimization
U ovoj disertaciji analiziran je teorijski značaj i mogućnost primene višekriterijumskog odlučivanja u ekonomiji, detaljno se fokusirajući na ekonomiju obrazovanja. U radu je prikazana metodologija ELECTRE familije metoda i ukazano je na njenu važnu ulogu u ekonomskoj teoriji i praktičnoj primeni. Posebna pažnja posvećena je ELECTRE MLO metodi rangiranja, pri čemu su detaljno analizirane njene prednosti i nedostaci, prevashodno se fokusirajući na rešenje problema ciklova koje je ova metoda donela defnisanjem apsolutnog praga značajnosti, i redefinisanjem indeksa saglasnosti i nesaglasnosti. Bazirajući se na metodološkoj osnovi metode, u disertaciji je dokazano da apsolutni prag značajnosti predstavlja, u opštem slučaju, donju granicu nakon koje se neće pojavljivati ciklovi u grafu. ELECTRE MLO metoda primenjena je za procenu atraktivnosti srednjoškolskih obrazovnih profila u Srbiji. Centralni deo teze predstavlja kreiranje više-etapnog modela komparativne analize. Savremena akademska literatura posvetila je pažnju traženju najboljeg primera u praksi radi poređenja, dok su pitanja kako učiti i kako se poboljšati zanemarena. Upravo pomenuta dva pitanja su osnovna motivacija modela ove disertacije. Novokreirani model više-etapne komparativne analize pretpostavlja da kao podlogu ima bilo koju metodu višekriterijumskog odlučivanja koja daje parcijalni pretporedak kao finalno uređenje. U disertaciji je korišćena ELECTRE MLO metoda, ali metodologija modela više-etapne komparativne analize konstruisana je tako da je robusna na promene metodološke osnove za finalno uređenje alternativa. ELECTRE MLO metoda kao i više-etapni model komparativne analize primenjeni su radi rangiranja evropskih obrazovnih sistema, kao i dalje analize putanja razvoja koje će iz tog ranga proizići. Bazirano na ciljevima Evropske unije i zacrtanim ishodima u obrazovanju kao kriterijumima analize, kreiran je rang obrazovnih sistema Evrope. Ovaj rang osnova je za primenu više-etapnog modela komparativne analize. Kao rezultat modela prezentovana je putanja razvoja za Srbiju, iz koje je generisan predlog reformi srpskog obrazovnog sistema.
UNIVERZITET U BEOGRADU EKONOMSKI FAKULTET Mladen S. Stamenković VIŠEKRITERIJUMSKO ODLUČIVANJE U UPRAVLJANJU I OPTIMIZACIJI EKONOMIJE OBRAZOVANJA Doktorska disertacija Beograd, 2016. UNIVERSITY OF BELGRADE FACULTY OF ECONOMICS Mladen S. Stamenković MULTIPLE CRITERIA DECISION AIDING IN ECONOMICS OF EDUCATION MANAGEMENT AND OPTIMIZATION Doctoral dissertation Belgrade, 2016 Podaci o mentoru i članovima komisije: Mentor: dr Zoran Popović, vanredni profesor, Ekonomski fakultet, Univerzitet u Beogradu Članovi komisije: _________________________________________ _________________________________________ _________________________________________ _________________________________________ _________________________________________ _________________________________________ _________________________________________ _________________________________________ Datum odbrane: _______________________ Višekriterijumsko odlučivanje u upravljanju i optimizaciji ekonomije obrazovanja Rezime U ovoj disertaciji analiziran je teorijski značaj i mogućnost primene višekriterijumskog odlučivanja u ekonomiji, detaljno se fokusirajući na ekonomiju obrazovanja. U radu je prikazana metodologija ELECTRE familije metoda i ukazano je na njenu važnu ulogu u ekonomskoj teoriji i praktičnoj primeni. Posebna pažnja posvećena je ELECTRE MLO metodi rangiranja, pri čemu su detaljno analizirane njene prednosti i nedostaci, prevashodno se fokusirajući na rešenje problema ciklova koje je ova metoda donela defnisanjem apsolutnog praga značajnosti, i redefinisanjem indeksa saglasnosti i nesaglasnosti. Bazirajući se na metodološkoj osnovi metode, u disertaciji je dokazano da apsolutni prag značajnosti predstavlja, u opštem slučaju, donju granicu nakon koje se neće pojavljivati ciklovi u grafu. ELECTRE MLO metoda primenjena je za procenu atraktivnosti srednjoškolskih obrazovnih profila u Srbiji. Centralni deo teze predstavlja kreiranje više-etapnog modela komparativne analize. Savremena akademska literatura posvetila je pažnju traženju najboljeg primera u praksi radi poređenja, dok su pitanja kako učiti i kako se poboljšati zanemarena. Upravo pomenuta dva pitanja su osnovna motivacija modela ove disertacije. Novokreirani model više-etapne komparativne analize pretpostavlja da kao podlogu ima bilo koju metodu višekriterijumskog odlučivanja koja daje parcijalni pretporedak kao finalno uređenje. U disertaciji je korišćena ELECTRE MLO metoda, ali metodologija modela više-etapne komparativne analize konstruisana je tako da je robusna na promene metodološke osnove za finalno uređenje alternativa. ELECTRE MLO metoda kao i više-etapni model komparativne analize primenjeni su radi rangiranja evropskih obrazovnih sistema, kao i dalje analize putanja razvoja koje će iz tog ranga proizići. Bazirano na ciljevima Evropske unije i zacrtanim ishodima u obrazovanju kao kriterijumima analize, kreiran je rang obrazovnih sistema Evrope. Ovaj rang osnova je za primenu više-etapnog modela komparativne analize. Kao rezultat modela prezentovana je putanja razvoja za Srbiju, iz koje je generisan predlog reformi srpskog obrazovnog sistema. Ključne reči: višekriterijumsko odlučivanje, ELECTRE, ekonomija obrazovanja, više-etapni model komparativne analize Naučna oblast: Ekonomija, operaciona istraživanja Uža naučna oblast: višekriterijumsko odlučivanje, ekonomija obrazovanja UDK: 37.014:330]:519.816(043.3) Multiple Criteria Decision Aiding in Economics of Education Management and Optimization Abstract The aim of this dissertation is to analyze the impact and the potentials of multiple criteria decision aiding in economics, especially focusing on its application in the field of economics of education. Detailed methodology of ELECTRE methods is presented, and the significance and importance of these methods are discussed. Special attention was addressed to ELECTRE MLO ranking method. Strong features and weaknesses of the method are presented, primarily focusing on the problem of the cycles in ELECTRE methodology. ELECTRE MLO method resolved this issue, by defining the Absolute Significance Threshold and by redefining concordance and discordance indices. Furthermore, there has been shown that the Absolute Significance Threshold is the lower bound under which there will be no cycles in the relation graph. ELECTRE MLO method is ultimately applied to evaluate on the evaluation of the attractiveness of the secondary educational profiles in Serbia. The essential part of the thesis is newly developed stepwise benchmarking model. Modern academic literature is focused on finding the best exemplar for comparison, whilst questions such as how to learn and how to improve, have been neglected. These two questions are the focus of the stepwise benchmarking model derived in thesis. Foundation of the model is partial pre-order by any ranking method, while ELECTRE MLO was used in the thesis. ELECTRE MLO method and stepwise benchmarking model are used to rank European educational systems, as well as to convey further analysis of the development paths, derived from the model. Criteria for the analysis are based on European Union goals given by EU2020 agenda. Partial pre-order of European educational system has been used to generate development path of Serbia, from which we have defined suggested measures for the reform of the Serbian educational system.
true
Direktno upravljanje asinhronim motorom
Induction motor direct torque control
U radu je dat prikaz metode za direktno upravljanje momentom asihnronog motora, bez upotrebe davača pozicije rotora. Predložena metoda uključuje algoritam za estimaciju i upravljanje trenutnom vrednošću statorskog fluksa, čija je referenca proračunata tako da se obezbedi rad motora sa konstantnom amplitudom rotorskog fluksa. Momentom se upravlja varijacijama trenutne vrednosti statorske učestanosti, što je moguće realizovati u pogonima koji rade sa konstantnim rotorskim fluksom.
A novel control strategy for the induction motor drive, based on the field acceleration method, is presented. The torque is controlled through variations of the stator flux angular velocity. The stator flux is controlled by using a feedforward control scheme, with the stator flux reference vector adjusted so as to obtain the fixed rotor flux amplitude. The applied controller assures a fast torque response, low torque ripple in the steady state, and drive operation with a constant switching frequency. The algorithm includes the improved stator and rotor flux estimation that guarantees the stable drive operation in all operating conditions, even at low speeds. The experimental tests verify the performance of the proposed algorithm, proving that good behaviour of the drive is achieved in the transient and steady-state operating conditions.
false
Razvoj algoritma za povećanje energetske efikasnosti sistema tračnih transportera na površinskim kopovima
Development of the algorithm for energy efficiency improvement of belt conveyor system on open pit mines.
Potrošnja električne energije je u porastu zbog stalne potrebe za poboljšanjem kvaliteta života. Energetska efikasnost je danas jedan od ključnih elementa u energetskoj politici svih razvijenih zemalja sveta, jer doprinosi unapređenju ekonomije u celini, i produžava vek korišćenja neobnovljivih izvora energije, iz kojih se danas dobija najveći deo električne energije. U mnogim zemljama najveći deo električne energije dobija se iz uglja, npr. u Srbiji je 62% svih kapaciteta za proizvodnju električne energije je u termoelektranama na ugalj, a 65% proizvedene električne energije dobija se iz ovih elektrana. U narednim decenijama težište daljih povećanja kapaciteta za proizvodnju električne energije biće na termoelektranama, što će zahtevati povećanje proizvodnje uglja. Povećani zahtevi za proizvodnjom uglja vode ka povećanju količine uglja i jalovine koju treba transportovati u okviru rudnika sistemom tračnih transportera, zbog čega se povećava njihova dužina i kapacitet, a time i instalisana snaga njihovih pogona. Zbog značajnih snaga pogona i po pravilu 24-časovnog rada od posebnog interesa je unapređenje energetske efikasnosti tračnih transportera. U različitim granama industrije, u kojima se proizvodi ili koristi rastresiti materijal, za transport se koriste različiti tipovi tračnih transportera, sa različitim konstruktivnim i pogonskim karakteristikama. Bez obzira na vrstu i organizaciju površinskog kopa, elektromotorni pogoni na rudarskoj mehanizaciji zbog velikih snaga, složenosti rudarske opreme i vrlo teških uslova rada predstavljaju izazov za primenu savremenih metoda pokretanja i upravljanja, u cilju povećanja efikasnosti rada, kao i energetske efikasnosti. Kod konvencionalnih sistema transportera, rastresiti materijal se transportuje stalnom, nazivnom brzinom. Pri nazivnom kapacitetu sistema površina poprečnog preseka materijala na traci je optimalna (maksimalna), odnosno traka je maksimalno popunjena. Međutim, istraživanja i statističke analize pokazuju da ovakvi sistemi zbog tehnologije kopanja bagera vrlo retko rade sa punim kapacitetom, odnosno da se veliki deo radnog vremena na trakama nalazi manja količina materijala od maksimalno moguće, pa čak i da prazan hod čini značajan deo radnog vremena sistema. Potrebna snaga za pokretanje sistema traka zavisi od količine transportovanog materijala, ali i od brzine kretanja traka. Imajući u vidu navedene činjenice, nameće se zaključak da je moguće transportovati materijal manjom brzinom od nazivne kada se na trakama nalazi manja količina materijala od maksimalno moguće, čime bi se smanjila trenutna snaga pogona, odnosno ostvarila značajna ušteda energije. Potrebna brzina kretanja traka se mora određivati u funkciji trenutne količine materijala na njima. Predmet ove doktorske disertacije je razvoj i primena algoritma za generisanje referentne brzine pojedinih transportera složenog sistema transportera. Na osnovu merenja površine poprečnog preseka materijala na ulazu u sistem tračnih transportera može se odrediti potrebna brzina kretanja traka koja će obezbediti stalnu optimalnu popunjenost svake trake u sistemu. Međutim, zbog značajnih varijacija količine dolazećeg materijala u sistem u vremenu ovakav način rada ne obezbeđuje energetski efikasan način rada, a i može dovesti do prekomernog habanja elemenata konstrukcije. Primenom različitih algoritama može se kontrolisati ubrzanje i usporenje trake, tako da se u okvirima ograničenja koja nameće sistem obezbedi minimalna potrošnja energije. U disertaciji su kombinovane teorijske, simulacione i eksperimentalne metode u cilju razvoja novog, originalnog algoritma kojim bi se ostvarilo povećanje energetske efikasnosti sistema, povećanje vremenskog i kapacitivnog iskorišćenja sistema, njegove pouzdanosti, produženje radnog veka opreme i smanjenje troškova održavanja. Pokazalo se da upravljanje primenom fazi logike predstavlja adekvatno rešenje, zbog svoje robusnosti i adaptivnosti, jednostavne primene i lakog podešavanja. Dokaz ispravnosti predložene ideje izveden je egzaktnim merenjima na postojećem postrojenju. Imajući u vidu navedene prednosti predloženog rešenja i dokazanu mogućnost uspešne primene, očekuje se da će se algoritam upravljanja sistemom tračnih transportera primeniti na sličnim postrojenjima u našoj zemlji, ali i u svetu, na površinskim kopovima sa kontinualnim načinom rada.
Electrical energy consumption is increasing globally in order to keep improving our quality of life. Energy efficiency is today one of the key elements in energy policy of all developed countries in the world, because it contributes to the improvement of economy globally and extends the lifetime of conventional energy sources, since the greatest part of total electrical energy is generated by fossil fuels. In many countries, most electrical energy comes from coal, for example in Serbia, 62% of all capacities for energy production is in thermal power plants which use coal, and they produce 65% of total electrical energy production. In the following decades, the focus will be on thermal power plants to further increase electrical energy capacities, requiring the rise of coal production. The growing demand for coal production leads to increased amounts of coal and overburden to be transported within the mine with belt conveyors systems. The length and the capacity are increasing causing the installed power of these systems to increase. Due to the large power of these drives and existing requirement for continuous operation, 24 hours per day, the improvement of energy efficiency of belt conveyors is naturally of significant interest. In various branches of industry where bulk materials are produced or used, various types of belt conveyors are used for the transport of materials, with different mechanical and drive characteristics. Despite of the type and the organization of the open pit mine, electrical drives on mining machines present a unique challenge to the application of modern techniques of driving and control, due to very large installed power, complexity of mining equipment and demanding environmental conditions, in order to increase the operational and energy efficiency. Conventional systems of conveyors transport bulk material at the constant, rated speed. When the system is at rated capacity, the cross section of material on the belt is at the maximum value, i.e. the belt is fully loaded. However, investigations and statistic analyses show that systems like this, due to technology of excavation very often work at reduced capacity. Most often the instantaneous cross section area of material on the belt is less than the rated value, and even no-load operation makes up a significant part of the system’s operational time. The necessary power for driving the system of belt conveyors depends on the quantity of transported material, as well as on the speed of the belts. Considering the aforementioned facts, the following conclusion can be made: since the belt conveyor often operates at a decreased capacity, the same quantity of material can be transferred at a lower than rated speed, leading to instantaneous power reduction, along with significant energy savings. The required speed of the belt must be determined as a function of the instantaneous quantity of material on the belt. The subject of this dissertation is the development and application of an algorithm for generating the reference speed of the belt conveyor drive in a complex system of belt conveyors. Based on measurements of instantaneous cross section area of material at the beginning of the system of belt conveyors, the required belt speed can be determined to provide that each belt in the system of belt conveyors will be constantly and optimally loaded. The instantaneous capacity changes quite frequently and sporadically, meaning that the speed should be increased and decreased in the same manner as the instantaneous capacity changes. These dynamic processes would be unfavorable for the mechanical assemblies of a belt conveyor, especially for the belt, and could lead to increased energy consumption. With the application of different control strategies, the acceleration and deceleration of a belt can be controlled in order to ensure minimum energy consumption under the existing technical constraints of the system. The dissertation combines theoretical, simulation and experimental methods, with the goal of developing a new, original algorithm to improve the energy efficiency of the system, capacitive utilization of the system, system reliability, extension of the service life of the equipment and reduction of costs for system maintenance. During the development of the control strategy, fuzzy logic control appeared to be a suitable solution for improving the energy efficiency of the system, due to its robust and adaptive characteristics, simple implementation and appropriate tuning procedure. The validation of the proposed idea is achieved through experimental results recorded on an existing system of belt conveyors in an open pit mine. Considering the indicated advantages of the proposed algorithm, as well as the proven opportunity of its successful implementation, it is expected that the algorithm for generating speed reference of the system of belt conveyors will be applied on similar systems in our country, as well as abroad, in open pit mines with continuous mining.
true
Kontakti engleskog i srpskog jezika u oblasti bankarstva i finansija
The English-Serbian language contacts in the field of banking and finance
Cilj ove disertacije jeste da istraži kvantitativne i kvalitativne aspekte uticaja engleskog jezika na srpski jezik bankarstva i finansija. Naše sinhronijsko kontaktno-kontrastivno proučavanje uključuje, kako mikrolingvističku, tako i makrolingvističku analizu kontakata dva jezika struke i nauke. Analiza je zasnovana na korpusu novijih anglicizama koji dolaze iz izvora širokog opsega. U radu koristimo sledeće metode: deskriptivnu, kontaktno-kontrastivnu, analitičku, komparativnu i statističku metodu. Rad polazi od pregleda teorijskih osnova istraživanja jezika u kontaktu, kao i našeg unapređenja i proširenja postojećih teorija, pre svega kroz uvođenje termina „kontaktonisema“ i „kontaktoterminema“, kao glavnih kontaktnih jedinica jezika za posebne namene, odnosno jezika struke i nauke. Analiza korpusnih anglicizama izvršena je na grafološko-fonološkom, morfološkom i semantičko-pragmatičkom nivou. Posebna pažnja posvećena je leksičkim i semantičko-pragmatičkim segmentima analize, kao savremenom pristupu u analizi terminologije jezika za posebne namene (bankarstva i finansija). U disertaciji smo opisali nove vrste jezičkih kontakata i pozajmljivanja koje smo otkrili. Takođe smo istražili postojeće i ponudili naše kriterijume za formiranje i vrednovanje termina. Jedno od najznačajnih dostignuća ovoga rada jeste Selektivni rečnik kontaktoterminema srpskog jezika bankarstva i finansija. Ovo je rečnik identifikacije, deskripcije i klasifikacije koji uključuje preko 800 kontaktoterminema srpskog jezika bankarstva i finansija. Svaka od njih je analizirana u okviru pedesetak različitih kategorija. Rezultati analize korpusa pokazuju da je više onih srpskih kontaktema koje su nastale preuzimanjem iz engleskog jezika uz preoblikovanje, nego onih koje su nastale prevođenjem sa engleskog jezika. Engleski jezik jeste lingva franka svetske bankarsko-finansijske zajednice. Nasuprot njemu, srpski jezik ima značaj isključivo u komunikaciji lokalnog tipa, pod snažnim je uticajem engleskog jezika, odlikuje se nesređenom i nestandardizovanom terminologijom i nebrigom od strane svojih korisnika. Zato smo u našem rečniku dali predloge za oko četiri stotine standardizovanih termina srpskog jezika bankarstva i finansija, kao i naše viđenje sistematizovanog pristupa standardizaciji ovog zaista vrlo važnog i prisutnog jezika struke i nauke. Naši predlozi za stvaranje terminologije srpskog jezika bankarstva i finansija su zasnovani na načelu davanja prvenstva srpskom autohtonom jezičkom materijalu, zatim internacionalizmima, uz integraciju objektivno potrebnih i segregaciju objektivno nepotrebnih anglicizama. Naposletku, smatramo da se na ovaj način neposrednim putem čuvaju i unapređuju srpski jezik i srpska ćirilica.
The aim of this dissertation is to examine the quantitative and qualitative aspects of the influence of the English language on the Serbian language for Banking and Finance. Our état de langue contact-contrastive research includes microlinguistic as well as macrolinguistic analysis of the contacts of the two Languages for Occupation and Science. The analysis is based on a corpus of contemporary anglicisms which come from the wide range of sources. In our research, we have used the following methods: descriptive, contactcontrastive, comparative, analytical and statistical method. The thesis sets out with an overview of the theoretical framework of language contact studies, as well as our extention and improvement of the existing theories, primarily through the introduction of the terms “EOS contacteme” and “terminological contacteme” as the basic units of contact of Languages for Specific Purposes, i.e. Languages for Occupation and Science. The analysis of corpus anglicisms is done within graphological-phonological, morphological and semantic-pragmatic language levels. Special attention is paid to lexical and semantic-pragmatic segments of the analysis because they represent the basis for contemporary approach to the analysis of terminology of the Language for Specific Purposes (Banking and Finance). In our dissertation we have described new types of language contacts and linguistic borrowing we have discovered. We have also explored the existing, and offered novel criteria for formation and evaluation of terms. One of the most significant achievements of the thesis is The Selective Dictionary of Terminological Contactemes of the Serbian Language for Banking and Finance. This is a Dictionary of identification, description and classification, which includes more than 800 terminological contactemes of the Serbian Language for Banking and Finance. Each of the contactemes is analysed within around fifty different categories. The results of the analysis show that there are more Serbian contactemes which are formed within the process of borrowing and adaptation of the terms from English than those which originate in translation from English. It is a well-known fact that English is a lingua franca of the world’s bankingfinancial professional community. As opposed to English, the Serbian Language for Banking and Finance is of local importance only, being under strong influence of English. The main features of the Serbian Language for Banking and Finance are disorganized, nonstandardized terminology and indifference of its users for its proper formation and usage. For this reason, in our Dictionary we have offered suggestions for more than four hundred standardized terms of the Serbian Language for Banking and Finance, as well as our view of the systematic approach to the standardization of this very present and important Language for Occupation and Science. Our suggestions for Serbian for Banking and Finance terminology formation are based on the principle of giving priority to the Serbian autochthonous language potential, followed by internationalisms, including integration of justified anglicisms and segregation of the injustified ones. Finally, it is our opinion that in this way the Serbian language and the Serbian Cyrillic alphabet are directly being fostered and improved.
false
Primena Kurnoovog modela konkurencije na oligopolskim tržištima u uslovima ograničenih kapaciteta
A use of Cournot's competition model in oligopolistic markets in the terms of limited capacities
Cilj ovog rada je da teorijski utemelji primenu Kurnoovog modela kao skraćene forme dvostepene konkurencije na oligopolskim tržištima u uslovima ograničenih kapaciteta. Razvijen je model koji predstavlja prečicu za analizu potencijalnih efekata na blagostanje horizontalnih spajanja preduzeća. Model mogu upotrebiti komisije za zaštitu konkurencije kao simulacioni model pri kontroli horizontalnih spajanja preduzeća. Pokazano je da se ishod modela dvostepene konkurencije, gde se preduzeća obavezuju na određene nivoe kapaciteta, pre konkurencije cenama, poklapa sa ishodom Kurnoovog modela količinske konkurencije. Šta više, ispostavlja se da ovaj rezultat važi kako za tržišta homogenih proizvoda tako i za tržišta diferenciranih proizvoda. To protivreči gotovo standardnom stavu da je količina strateška varijabla isključivo na tržištima homogenih proizvoda. Robusnost ovog rezultata proverava se kako za simetrični oligopol, što predstavlja standardno polazište svih modela iz ove klase, tako i za okolnost asimetričnog oligopola. Asimetrija se uvodi kroz internalizaciju konkurencije izazvanu spajanjem dva simetrična preduzeća na tržištu diferenciranih proizvoda. Komisijama za zaštitu konkurencije je predložen algoritam za primenu modela simulacija. U skladu sa definisanim koracima, sprovedene su simulacije spajanja na realnom primeru, primenom Kurnoovog modela konkurencije u uslovima ograničenih kapaciteta. Na osnovu rezultata simulacija, komisijama su date praktične preporuke za sprovođenje kontrole koncentracija.
The objective of the thesis is to examine the Cournot model as a reduced form of the two-stage competition in oligopolistic markets with limited capacities. A short-cut model for analysing welfare effects of potential horizontal mergers is developed. The model can be used by Anti-trust commissions as a simulation model in the process of horizontal merger assessment. It is shown that the outcome of the two-stage competition, where firms had been conditioned by their level of capacities even before the price competition has started, coincide with the outcome of the Cournot quantity competition model. Moreover, it turns out that this result is valid for both the homogeneous product markets and the markets of differentiated products. This contradicts an almost universal belief that the quantity is a strategic variable exclusively within the homogeneous product markets. The robustness of this result is examined both for symmetric oligopoly, which is a starting point for all models in the class, and for the asymmetric oligopoly. The asymmetry is introduced through the internalisation of competition due to a merger of two symmetric firms in the differentiated products market. The implementation algorithm of the simulation model is suggested to the Anti- trust commissions. In accordance with defined steps, merger simulation can be carried out in the real world by using the Cournot model of competition with limited firm capacities. Based on the simulation results, practical recommendations for the merger control policies are offered to the commissions.
true
Samoubistva u Srbiji u kontekstu savremenih demografskih i društvenoekonomskih promena
Suicides in Serbia in the context of contemporary demographic and socio-economic changes
Samoubistvo predstavlja veoma složen, dinamičan i heterogen fenomen, uslovljen međudejstvom niza psiholoških, socijalnih i bioloških faktora, ali ta pojava postaje sve bitnija i sa demografske tačke gledišta. Prema podacima Svetske zdravstvene organizacije u svetu je početkom 2010-ih godina usled samoubistva godišnje umiralo oko 800 hiljada lica ili oko 11 na 100.000 stanovnika. U pogledu vrednosti stope samoubistva postoje veoma naglašene regionalne razlike, a stope se kreću u intervalu od ispod 0,3 do blizu 40 na 100.000. U Srbiji, u razdoblju od ranih 1950-ih do sredine druge decenije 21. veka, ukupan broj samoubistava premašio je 75 hiljada. Srbija se sredinom 2010-ih sa oko 16 samoubistava na 100.000 stanovnika nalazi znatno iznad svetskog i neznatno iznad evropskog proseka. Predmet istraživanja doktorske disertacije su samoubistva u Srbiji od 1950. do 2014. godine. To 65-godišnje razdoblje podeljeno je na dva potperioda – prvi od 1950. do 1990, i drugi od 1990. do 2014. Pažnja je fokusirana na novije razdoblje (1990-2014), ne samo zbog toga što je to skorašnji period, već i zbog dostupnosti mnogo sadržajnije i kvalitetnije statističke građe. Osnovi cilj disertacije je da se kompleksno, ali prvenstveno iz demografskog ugla istraži fenomen samoubistva koji, pored toga što je prvorazredni problem javnog zdravlja, ima i značajne socijalne, ekonomske, a takođe i demografske implikacije. Ostvarenje tog cilja podrazumeva utvrđivanje osnovnih tendencija u dinamici tog fenomena, ali i ukazivanje na perspektive budućih promena. Jedan od važnih ciljeva je i sagledavanje uticaja opštih, društvenoekonomskih uslova na pojavu i karakteristike samoubistava od sredine 20. veka, a posebno imajući u vidu da je za Srbiju poslednja decenija 20. veka bila razdoblje velikih političkih previranja i ratova na prostoru bivše Jugoslavije, kao i eskalacije ekonomske krize. S tim u vezi, izvršeno je poređenje uočenih trendova s onim u drugim zemljama kako bi se sagledale sličnosti i eventualno izdvojile specifičnosti koje se odnose na stanovništvo Srbije. Takođe, cilj je i da se ukaže na najvažnije teorijske nalaze iz sve brojnije naučne literature koja obrađuje fenomen suicida, njegove faktore rizika i zaštite. Prvenstveno je korišćen standardni, tzv. normativni naučnoistraživački metod. Primenjena je demografska analiza mortaliteta i to uglavnom transverzalnog metoda, a izračunate su vrednosti najvažnijih indikatora smrtnosti usled samoubistva (opšta stopa samoubistva, specifične stope suicidnog mortaliteta po starosti i polu, standardizovana stopa samoubistva). Primenjena je i široka lepeza metoda statističke analize, kao što su regresiona i korelaciona analiza, statističke tehnike analize vremenskih serija, statistički metodi ocene ii cikličnih varijacija (D2-metod, Edwardsova procedura), kao i metodi za merenje relativnog rizika. Uvedena su i dva nova pokazatelja: indeks magnitude konsekutivnih tromesečnih varijacija (MtMV) i starosni indeks metoda suicida (SIMS). Opsežna analiza je pokazala da u Srbiji, kao i u ogromnoj većini zemalja u svetu, među licima koja su izvršila samoubistvo postoji naglašena diferenciranost po polu, samoubistva muškaraca su znatno češća. Takođe, da postoji više starosnih modela suicida, a da je u Srbiji, kao i u većini zemalja, najveća rasprostranjenost samoubistva među starima. Potvrđeno je da su razlike u nivou suicidnog mortaliteta po bračnosti umrlog lica velike, i da je rizik od samoubistva najmanji kod osoba koje su u braku. Uticaj sociokulturnih faktora je posmatran preko varijabli školske spreme i etničke pripadnosti. Izveden je zaključak da je između stopa samoubistva i školske spreme odnos obrnuto proporcionalan, kao i da se široki spektar faktora koji su karakteristični za određenu nacionalnu zajednicu ogleda i u nivou stopa samoubistva njenih pripadnika. U tezi su takođe analizirani uticaji političkih, ekonomskih i ratnih kriza na učestalost samoubistva, naročito među muškarcima. Posebna pažnja u disertaciji je posvećena dvema temama koje u dosadašnjim istraživanjima suicida u Srbiji nisu bile dovoljno zastupljene, a to su metod izvršenja samoubistva i sezonalnost suicida. Ustanovljen je tromodalni model metoda suicida, i diskutovane prisutne starosne i polno/rodne razlike. Takođe, potvrđena je pretpostavka o sezonskim varijacijama suicida u Srbiji, a one su naglašene po mesecima i po danima u nedelji. Rezultati doktorske disertacije doprinose obogaćivanju saznanja o suicidnom mortalitetu stanovništva Srbije u jednom dugom, višedecenijskom, razdoblju, a posebno u uslovima turbulentnih društvenih kretanja i dugoročne ekonomske krize. Naučni doprinos disertacije je dopunjen ukazivanjem na najnovije rezultate istraživanja fenomena suicida u svetu, ponajviše s aspekta faktora rizika i zaštite Saznanja do kojih se došlo u disertaciji bi mogla da podstaknu dodatno animiranje javnosti za bolje razumevanje problema suicida i razvijanje senzibiliteta šire društvene zajednice prema fenomenu suicidnog ponašanja na koji se, svakako, može preventivno delovati. U disertaciji je ponuđeno više potvrda polazne pretpostavke da je prevencija suicida potrebna, i moguća. Ukazano je na mogućnosti za prevenciju u Srbiji, sprovođenjem već usvojenih strategija, donošenjem novih mera, kao i definisanjem nacionalne strategije za prevenciju samoubistva.
Suicide is a very complex, dynamic and heterogeneous phenomenon, caused by interactions of a range of psychological, social and biological factors, but the phenomenon is also becoming increasingly important from the demographic point of view. According to the World Health Organization, at the beginning of 2010s, about 800 thousand persons in the world or about 11 per 100,000 population died due to suicide each year. In terms of the suicide rates there are very pronounced regional differences, and the rates range from less than 0.3 to nearly 40 per 100,000. In Serbia, in the period from the early 1950s to the middle of the second decade of the 21st century, the total number of suicides exceeded 75 thousand. In the mid-2010s, with about 16 suicides per 100,000 Serbia was significantly above the world and slightly above the European average. The research subject of the doctoral dissertation is suicide in Serbia from 1950 to 2014. This 65-year period was divided into two sub-periods - the first from 1950 to 1990, and the second from 1990 to 2014. Attention is focused on the latest period (1990-2014), not only because it is recent, but also due to the availability of much more substantial and better quality statistical data. The basic aim of the dissertation is to investigate comprehensively, but primarily from the demographic perspective, the phenomenon of suicide, which besides being one of the greatest public health problems, also has significant social, economic, as well as demographic implications. The achievement of this aim involves establishing basic tendencies in the dynamics of this phenomenon, but also pointing to the prospects of future changes. One of the important aims is also to assess the influence of the general, socio- economic conditions on the occurrence and characteristics of suicide starting from the mid- 20th century, especially bearing in mind that for Serbia the last decade of the 20th century was a period of great political turmoil and wars in the former Yugoslavia and the escalation of the economic crisis. In this regard, comparisons of the observed trends were made with those in other countries in order to identify similarities and possibly single out the specificities relating the population of Serbia. In addition, the aim is to point to the most important theoretical findings from the growing scientific literature that deals with the phenomenon of suicide, its risk and protective suicide factors. Primarily, the standard, the so-called normative research method was used. The demographic analysis of mortality was applied, mainly of transversal method, and the values of the most important indicators of mortality by suicide were calculated (suicide rate, age-specific suicide rate by sex, age-standardised suicide rate). A wide range of statistical analysis methods was also used, such as regression and correlation analysis, statistical iv techniques of time series analysis, statistical methods of cyclic variation assessment (D2- test, Edwards' test), as well as methods for measuring relative risk. Two new indicators were also introduced: the magnitude of trimester variations of /12/ moving consecutive months index (MtMV index) and suicide method age index (SIMS). Extensive analysis has shown that in Serbia, as well as in the vast majority of countries in the world, there is a strong differentiation by sex among those who have committed suicide, suicide in men being much more common. In addition, it has shown that there are several age patterns of suicide, and that in Serbia, like in most countries, the highest prevalence of suicide is among the elderly. It has been confirmed that the differences in the level of suicide mortality by marital status of the deceased are large, and that the risk of suicide is smallest with married people. The influence of socio-cultural factors was observed through the variables of education and ethnicity. The conclusion is that the relation between the suicide rate and educational attainment is inversely proportional, and that a wide range of factors typical for a particular ethnic community is also reflected in the level of suicide rates of its members. The dissertation also analyses the impact of political, economic and war crises on the incidence of suicide, especially among men. Special attention is devoted to two themes that have been underrepresented in the current research on suicide in Serbia, the suicide method and suicide seasonality. The tree-modal model of suicide method was established, and the present age and gender differences were discussed. Furthermore, the assumption about the seasonal variations of suicides in Serbia was confirmed, highlighted by months and days of the week. Results of the doctoral dissertation contribute to the enrichment of knowledge about suicidal mortality in Serbia in a decades-long period, especially during turbulent social events and long-term economic crisis. The scientific contribution of the dissertation is supplemented by pointing to the latest results of research on the phenomenon of suicide in the world, most notably in terms of risk factors and protection. The findings obtained in the dissertation could be additionally incentive for the public to better understand the problem of suicide and develop sensibility of the society towards the phenomenon of suicidal behaviour. The dissertation offers more confirmation of the initial assumptions that the prevention of suicide is necessary, and possible. It also points to the possibility for prevention in Serbia by implementing already adopted, introducing new measures, as well as defining the national strategy for suicide prevention.
true
Izbor modela nabavke u strateškom pozicioniranju proizvodnog preduzeća u prehrambenoj industriji u Republici Srbiji
Choice of model purchasing in strategic positioning of production company in food industry in the Republic of Serbia
Tokom poslednjih decenija došlo je do promene paradigme o ulozi nabavke u mnogim preduzećima. Nabavka je evoluirala od jednostavne kupovne funkcije do strateške funkcije. Tradicionalni koncepti kao što su nabavka i upravljanje materijalima postepeno nestaju dok na znaĉaju dobijajaju novi koncepti kao sto je strateška nabavka. Ono što je novo, jesu novi pristupi: globalna nabavka, nabavka kao centar profita a ne centar troškova, nabavka kao osnov strateških partnerstava i brojne druge inovativne strategije zasnovane na strateškom pristupu. Zbog toga je nabavka postala izvor dodatne vrednosti u preduzećima sa strateškim pristupom. Predmet ovog rada odnosi se na istraţivanje strateške funkcije nabavke u proizvodnim preduzećima u prehrambenoj industriji u Republici Srbiji, izbor superiornog modela nabavke, a zatim i koncepata i strategija koji se u njemu primenjuju. Istraţivanje je fokusirano na oblast prehrambene industrije zbog velikog znaĉaja nabavke za poslovanje u ovoj industriji, zbog razvijene prakse nabavke u prehrambenoj industriji ali i zbog znaĉaja koji ova industrija ima u privrednom razvoju Republike Srbije, a posebno u izvozu gde otvara mogućnost za znaĉajno poboljšanje spoljnotrgovinskog bilansa i smanjenje deficita. Veća konkurentnost na domaćem i svetskom trţištu preduzeća iz ove industrije, zahteva nuţno i strateški pristup nabavci. S obzirom na sve veći naglasak na izgradnji i upravljanju odnosom kupac-dobavljaĉ kao osnove odrţive konkurentske prednosti, uraĊeno je sistematsko istraţivanje mogućnosti i stepena u kojem strateška nabavka doprinosi unapreĊenju poslovnih performansi i boljem pozicioniranju proizvodnog preduzeća na globalnom trţištu. Korišćenjem strukturnog modela, empirijski su istraţeni odnosi izmeĊu strateške nabavke, upravljanja kanalima nabavke, reakcije dobavljaĉa i poslovnih performansi kupca. Osim toga, istraţivanje je dokumentovalo u kojoj meri strateška nabavka, podstiĉe organizacione sposobnosti neophodne za efikasno upravljanje preduzećem. Strukturni model i hipoteze pruţaju osnovu za izbor superiornog modela i empirijsko potkrepljivanje ovih tvrdnji. Kljuĉni rezultat ovoga rada je izbor strateškog modela nabavke za integralnu primenu u pozicioniranju proizvodnih preduzeća na globalnom trţištu. U modelu su definisana rešenja vi za uspešnu primenu strateške nabavke što će biti primenljivo u budućem radu proizvodnih preduzeća u Republici Srbiji.
Over the past decades, there has been a change in the paradigm of the role of purchasing in many companies. Purchasing has evolved from a simple purchasing function to a strategic function. Traditional concepts, such as purchasing and material management, gradually disappear while new concepts, such as strategic purchasing, are gaining in importance. We have new approaches: global purchasing, purchasing as a profit center rather than a cost center, purchasing as the basis for strategic partnerships and many other innovative strategies based on a strategic approach. Therefore, purchasing has become a source of added value in enterprises with a strategic approach. The subject of this paper relates to the research of the strategic purchasing function in manufacturing enterprises in the food industry in the Republic of Serbia, the selection of the superior purchasing model, and then the concepts and strategies that apply in it. The research is focused on the field of food industry because of the great importance of purchasing in the food industry, but because of the great importance this industry has in the economic development of the Republic of Serbia, and especially in exports, where it opens the possibility for significant improvement of the foreign trade balance and deficit reduction. Greater competitiveness on the domestic and global markets of enterprises in this industry indispensably requires a strategic approach to purchasing as well. Considering the increasing emphasis on building and managing the customer-supplier relationship as a basis for a sustainable competitive advantage, a systematic research into the possibilities and extent in which strategic purchasing contributes to the improvement of business performance and better positioning of a manufacturing company in the global market has been done. Using the structural model, relationships between strategic purchasing, supply chain management, supplier response, and business performance of the buyer have been empirically explored. In addition, the research has documented the extent to which strategic purchasing is encouraging the organizational skills necessary for effective corporate governance. The structural model and hypotheses provide the basis for selection of a superior model and empirical reinforcement of these claims. The key result of this paper is the choice of a strategic purchasing model for integral application in the positioning of production companies in the global market.
false
Indikatori održivog razvoja i mogućnosti njegovog merenja
Indicators of sustainable development and the possibilities of measurement
Uobičajen način merenja privrednog uspeha predstavlja merenje privrednog rasta, kao jedinog cilja ekonomske politike bazirane na postulatima neoklasične ekonomije. Isključiva posvećenost privrednom rastu na značaju dobija naročito u periodima krize kada se jedini izlaz vidi isključivo u povoljnim indeksima rasta. Rast, kao uvećanje materijalne baze, je neophodan kako bi se povećala zaposlenost, ostvarili preduslovi za napredak i tehnološka dostignuća. U siromašnom svetu rast predstavlja jedini izlaz iz bede i sveopšte nemaštine. Društveno blagostanje, zagađenje okoline, pa i socijalno zadovoljstvo, postaju lako zanemarljivi zarad rasta. Za merenje privrednog rasta razvijeni su i u upotrebi su značajni pokazatelji, od kojih su najzastupljeniji makroekonomski agregati kojima se sumiraju privredne aktivnosti, među kojima prednjači bruto domaći proizvod. Iako predstavlja meru proizvodnje, a ne dohotka, bruto domaći proizvod se veoma često koristi kao aproksimativna mera blagostanja, odnosno životnog standarda stanovništva koje nije monetarna kategorija. Lakoća primene i dostupnost podataka učinila je ovaj pokazatelj gotovo nezamenljivim. Održivi razvoj podrazumeva rast kao izlaz iz siromaštva i putanju ka razvijenom stanju koje se može održati na veoma dugačak period. U efikasnoj privredi, rast ili razvoj pretpostavlja odvajanje resursa od potrošnje trenutne generacije zarad investicija koje će povećati produktivnost buduće generacije. Da bi razvoj bio održiv putanja razvoja privrede mora biti u okviru ekoloških i tehnoloških ograničenja. Pitanje načina izražavanja održivosti izuzetno je bitno i nekoliko parametara je do sada već definisano, poput: konstantne stope štednje, konstantnog odnosa kapitala i ii autputa, uravnoteženog rasta, konstantnog zadovoljstva, odnosno korisnosti i slično. Cilj rada je potraga za adekvatnim instrumentarijem iskazivanja stepena ekonomskog razvoja privreda, a s tim u vezi i kritički prilaz postojećim merama održivog razvoja. Odabrani cilj diktira i metode koje su primenjene, prvenstveno metod ekonomske analize i sinteze i komparativna analiza. Donošenje suda o korektnosti i svrsishodnosti postojećih indikatora baziran je na dinamičkoj optimizaciji. Održivost se može definisati na različite načine, ali u neoklasičnim modelima mogu se identifikovati tri grupe modela, odnosno tri grupe pretpostavki na kojima ovi modeli počivaju. Jedan od poznatijih i češće primenljivih kriterijuma je maksimin kriterijum, na bazi Roulsovog stanovišta da bi korisnost trebalo održati na maksimalno održivom nivou. Održivost se može naslutiti i iz političke retorike o ograničavanju korisnosti tokom vremena, gde su moguće varijante konstantna korisnost, neopadajuća korisnost i zadržavanje trenutnog nivoa korisnosti na sadašnjem maksimalnom nivou održivosti. Održivost se može definisati i kao konstantna ili nepromenjena mogućnost tokom vremena. U ovom slučaju kao ograničenje nameće se neopadajuće bogatstvo ili ukupni kapital, a ne korisnost. Tretman međugeneracijskog blagostanja predstavlja i etičko pitanje i u modelima održivog razvoja posvećuje mu se velika pažnja. U zvisnosti od etičkog poimanja jednakosti i adekvatne intertemporalne raspodele bogatstva zavisi i postavka samog modela. Osnovni uslov prilikom konstrukcije indikatora održivosti mora biti smislenost. Pod smislenim pokazateljima održivosti mogu se smatrati samo indeksi koji omogućavaju nedvosmislenu manipulaciju relevantnim „situacijama“ tokom vremena, nezavisno od mernih jedinica u kojima su izražene promenljive koje opisuju date situacije. iii Uslov validnosti indikatora odživosti mora biti i uska povezanost sa konceptom održivosti. Kao što je prethodno već naglašeno, održivi razvoj kao termin još uvek nije nedvosmisleno preciziran, pa shodno tome možemo očekivati i različite formulacije pojma održivosti i prilikom kreiranja indikatora. Kod razumevanja indikatora razvoja bitno je naglasiti da li je model na osnovu koga se indikator razvija posvećen: optimizaciji, opstanku ili održivosti. Često se prave previdi u smislu poistovećivanja održivosti sa jednim od ova dva pojma. Neki modeli održivost će posmatrati kao bilo koje stanje ili kretanje iznad granice opstanka odabranog cilja (rasta, zaliha resursa ili potrošnje i slično), dok ostali održivost poistovećuju sa optimizacijom određene promenljive, iako optimizacija, uglavnom, podrazumeva maksimizaciju sadašnje diskontovane vrednosti korisnosti, što može rezultirati zanemarivanjem bitnih aspekata održivosti posvećenih prirodnim resursima i nivou zagađenja. Najveći problem konstrukcije indikatora održivog razvoja predstavlja dostupnost podataka, što ograničava kako broj, tako i karakter promenljivih koje će se obuhvatiti u proračunima. Naučno utemeljeni indikatori podrazumevaju adekvatno odabran metod normalizacije (svođenja podataka na uporediv nivo), ponderisanja (određivanje „ispravne“ međuzavisnosti varijabli odnosno pokazatelja) i agregacije (uspostavljanje „prave“ funkcionalne povezanosti). Analizirajući indikatore koji su posvećeni proceni održivosti izuzetno je teško presuditi koji od analiziranih je dovoljno blizu merenja onoga za šta se izdaje. Preispitujući postavke modela održivosti sa etičkim kriterijumima koje nameće teorija društvenog izbora, uočeno je da nijedan kriterijum međugeneracijske pravičnosti ne ispunjava sve nametnute etičke uslove. Drugim rečima, ne postoji nijedan potpuni, tranzitivni i kontinualni set iv društvenih preferencija koji zadovoljava uslove jednakosti. Navedeni zaključak potvrđuje konflikt između jednakosti i efikasnosti. Održivi razvoj posvećen je složenim problemima sadašnjice, koji proizilaze iz pokušaja harmonizacije često konfliktnih zahteva ljudskog razvoja, zaštite životne sredine i održavanja mogućnosti budućih generacija na nivou sadašnjih. Pokretanje ideje o mogućnostima realizacije jednog stabilnijeg, tehnološki naprednijeg, socijalno uravnoteženog i humanog društva, i pri tome još u skladu sa ekološkim načelima cilj je današnjih privreda okrenutih održivom razvoju. Iako je saglasnost, u načelu, postignuta u teoriji, u praski je izuzetno teško obuhvatiti sva tri aspekta održivosti u jednom indikatoru. Stoga, ozbiljna analiza održivosti privreda zahteva analizu po više kriterijuma i time iskazivanje kroz više naučno utemeljenih indikatora, podrazumevajući pre svega kvalitetnu bazu podataka.
The standard way to measure economic success is by measuring economic growth, as the sole goal of the economic policy based on the principles of neoclassical economics. Exclusive orientation towards economic growth gains special importance in times of crises, when favourable growth indexes are seen as the only solution. Growth, together with the enlargement of the material base, is inevitable in order to increase employment and provide preconditions for development and technological advancement. In underdeveloped parts of the world, growth is the only way out of crisis and general poverty. Social welfare, environmental pollution and even social satisfaction are easily neglected in favor of growth. There are significant indicators available for measuring economic growth; the most widely used being macroeconomic aggregates that represent a sum of economic activity, with gross domestic product as the most prominent one. Even though it measures production rather than income, gross domestic product is often used as a proxy for welfare, i.e. the standard of living although a non-monetary value. Its ease of use and data availability are what makes this indicator almost indispensable. Sustainable development refers to growth as a means to ending poverty and taking the trajectory to a developed state that can be obtained for a very long period of time. In an efficient economy, growth or development means taking the resources from the current spending, in order to make investments that will increase the productivity of future generations. For development to be sustainable, the trajectory of economic development must remain within the limits of environmental and technological constraints. The problem of presenting the sustainability is of utmost importance, with several parameters having been defined to that effect, such as constant savings rate, constant capital to output ratio, balanced growth, constant pleasure, utility and so on. vi The goal of this paper is to search for an adequate instrument for expressing the degree of economic growth of economies, critically assesing existing sustainable development measures. The goal thus defined dictates the selection of methods, above all the method of economic analysis and synthesis, as well as comparative analysis. Evaluation of the correctness and appropriateness of the existing indicators is based on dynamic optimization. Sustainability can be defined in a variety of ways, but in the neoclassical model three groups of models can be identified, or more precisely, three groups of assumptions on which these models are based. One of the most familiar and frequently used criteria is the maximin criterion, based on the Rawls’s belief that utility should be kept at the maximum sustainable level. Sustainability can also be intuited from the political rhetoric aimed as a constraint on utility over time, with such potential variations as constant utility, non-declining utility or sustaining current utility level over time. In this case the constraint is the non-decreasing wealth or total capital, rather than utility itself. The treatment of cross-generational welfare also poses an ethical question, which has been given substantial attention in the sustainable development models. The very basis of this model depends on the ethical concept of equality and adequate intertemporal distribution of wealth. The main precondition for the construction of a sustainability indicator is undoubtedly its meaningfulness. Only the indexes that permit unambiguous manipulation of relevant “situations” over time, regardless of the measuring units in which the variables describing these situations are given, can be considered meaningful. Another precondition for the validity of an indicator is its close relation to the concept of sustainability. As it has already been pointed out, sustainable development as a term has yet to be unambiguously defined, so we can expect varied definitions of this term in the process of creating an indicator. For understanding an indicator it is necessary to specify whether the model on which the indicator is constructed is based vii on optimization, subsistence or sustainability. Some models will treat sustainability as any state or trend above the subsistence line of the selected goal (growth, resources, consumption and so on), while the others will equate sustainability with the optimization of a certain variable, despite the fact that optimization usually entails the maximization of a discounted utility, which can result in the neglect of important aspects of sustainability in terms of natural resources and the level of pollution. The main problem in the construction of sustainable development indicators is data accessibility, which limits not only the number but also the type of potential variables to be included in the analysis. Scientifically based indicators require an adequately chosen normalization method (reduction of data to a comparable level), weighting (determining the “correct” interdependence of variables i.e. indicators) and aggregation (establishing a “right” functional relationship). By analyzing the indicators created for sustainability assessment it is extremely difficult to determine which of them is sufficiently close to the measurements of what it claims to show. Questioning the assumptions of the sustainability model with the ethical criteria imposed by the social choice theory, it has been observed that none of the criteria of cross-generational justice meets all the ethical preconditions imposed. In other words, there is no such things as a single, complete, transitive and continuous set of social preferences satisfying the conditions of equality. This conclusion thus attests to the existence of a conflict between equality and efficiency. Sustainable development is oriented towards the complex problems of our times stemming from the attempts to bring harmony into the often conflicted requirements of human development, environmental protection and the preservation of the opportunities of future generations at their current level. Promoting the idea of the possibilities for creating a more stable, technologically advanced, socially balanced and humane society, together with sound environmental policies, is the goal of the modern economies oriented towards sustainable development. Even though a general consensus has been reached in theory, on a practical level it is extremely difficult to viii encompass all the three aspects of sustainability by one single indicator. Therefore, thorough analysis of the sustainability of economies requires a multi-criteria analysis presented through an array of scientifically based indicators, a good database being one of the essential prerequisites.
true
Regulacija struje elektrolučne peći za jednosmernu struju prediktivnom metodom bez korišćenja modela električnog luka
A predictive method for dc arc furnace current control without using a dc electric arc model
Elektrolučne peći su potrošači veoma velike instalisane snage, pri čemu trenutna snaga peći može da varira u vremenu sa velikim amplitudama. Kao posljedica ovih velikih fluktuacija aktivne i reaktivne snage, elektrolučne peći velikih instalisanih snaga mogu da izazovu poremećaje u funkcionisanju prenosne i distributivne mreže, pogotovo u slučaju slabijih mreža kada su ovi poremećaji izraženiji. Najizraženiji poremećaji koji se pojavljuju u mreži kao posljedica rada elektrolučnih peći su pojava flikera (Flicker) i pojava neželjenih viših harmonika i interharmonika koji se unose u prenosnu mrežu. U predmetnoj disertaciji je analiziran uticaj predloženog prediktivnog regulatora struje električnog luka na poremećaje koji dovode do pojave flikera. S obzirom na to da se elektrolučne peći sve češće primjenjuju u industriji, njihovi negativni uticaji na funkcionisanje prenosnih mreža, na koje su priključene, postaju sve izraženiji. Zbog toga je neophodno upravljati radom peći tako da se pomenuti negativni uticaji potisnu u što većoj mjeri, a da se pri tome obezbijedi stabilno i pouzdano funkcionisanje peći. Elektrolučne peći za jednosmjernu struju danas se najčešće napajaju tiristorskim ispravljačima, pri čemu se struja električnog luka reguliše isključivo PI regulatorom kojim se reguliše vrijednost ugla upravljanja tiristorima. Napon luka reguliše se podešavanjem rastojanja između katode i šarže elektromehaničkim putem ili pomoću hidrauličkih cilindara upravljanih servohidrauličkim ventilima.Sa druge strane, regulacija struje i napona električnog luka kod peći za naizmjeničnu struju vrši se elektromehanički: promjenom prenosnog odnosa transformatora i podešavanjem rastojanja između katode i šarže. Poseban problem kod elektrolučnih peći za naizmjeničnu struju predstavljaju tranzijenti do kojih dolazi prilikom uključenja transformatora (nakon punjenja peći), koji su posljedica magnećenja Regulacija struje elektrolučne peći za jednosmernu struju prediktivnom metodom bez korišćenja modela električnog luka transformatora. Kao i u slučaju peći za naizmjeničnu struju, i kod peći za jednosmjernu struju postoji mehanička regulacija pozicije elektrode (kod peći za jednosmjernu strju uglavnom postoji jedna pokretna elektroda–katoda). S obzirom na to da mehanička regulacija pozicije elektrode ne može da se vrši dovoljno brzo, jedini način za smanjenje pomenutih negativnih uticaja elektrolučnih peći za jednosmjernu struju je implementacija brzog, pouzdanog i robusnog regulatora struje električnog luka...
Electric arc furnace (EAF) is a high-power electric load, whose active and reactive power varies significantly during the normal furnace operation. Due to the large active and reactive power variations, electric arc furnace can induce significant unwanted disturbances into the power system network in the form of voltage flicker and the high-order voltage and current harmonics and interharmonics. An influence of the proposed predictive arc current controller on the generated flicker is analyzed in this dissertation. Considering the increasing number of the dc arc furnaces installed, their negative influence on the power system networks becomes an important problem that needs to be addressed. In order to mitigate the unwanted disturbances caused by large electric arc furnaces, the control of the furnace operation needs to be improved. Currently, most of the dc arc furnaces are supplied through high-power thyristor rectifiers, and the arc current control is performed by adjusting the thyristor delay angle. The arc voltage is controlled electromechanically (or with the hydraulic valves) by means of electrode position control. In the case of the ac arc furnaces, both the arc current and the arc voltage are controlled by means of electrode position control, and by changing (online or offline) the furnace transformer tap number. Since the arc furnace voltage control is significantly slower than the current control, the only way to improve the control of the furnace operation is to improve the arc current control by implementing a fast, reliable and robust arc current controller. Since the filtering of the feedback current signal (in order to obtain the dc component of the arc current) produces a time-lag of the feedback signal, the classical PI current controller cannot compensate fast enough for the current overshoots occurring after frequent electrode short-circuits during the bore-in and the melting periods in the furnace cycle. On the other hand, the implementation of the fast modelbased predictive current controller is not feasible, because a precise electric arc model, which at the same time has to be suitable for circuit simulations, is not yet developed. The inability to prevent the current overshoots fast enough and to reduce its amplitudes is the main reason why the electric arc furnaces are still the major source of disturbances leading to the generation of flicker. Therefore, there is a need for the fast current controller which would enable the reduction or the complete elimination of the induced disturbances. The thorough survey of the available literature shows that such a controller is not yet developed. A basic idea behind the predictive arc current controller presented in this dissertation is that the arc current can be controlled by knowing the dc-side inductor voltage-time product, and the instantaneous arc current...
false
Uticaj kvaliteta institucija na privredni razvoj i efekte privatizacije u Srbiji
Influence of quality of institutions on economic development and effects of privatization in Serbia
Predmet istraživanja ovog rada je međusobni odnos institucija i privrednog rasta i razvoja. Analiza se posebno fokusira na ovaj odnos tokom procesa privatizacije u Srbiji, kao i na uticaj kvaliteta institucija na sam proces privatizacije. U skladu sa tim, teorijski okvir ove disertacije predstavlja institucionalna ekonomija, koja kao teorijski koncept, u prvi plan ističe ulogu i značaj institucija za privredni rast i razvoj svake zemlje. Institucije kreiraju „pravila igre“, odnosno predstavljaju normativni okvir date ekonomije, kao i mehanizme koji utiču na njenu efikasnost. Prema ovom teorijskom konceptu, ekonomska analiza se ne može svesti samo na ispitivanje tržišta i tržišnih odnosa, već mora obuhvatiti i institucije koje uređuju tržište i omogućavaju njegovo efikasno funkcionisanje. U radu se najprije razmatraju teorijske osnove institucionalne ekonomije i daju osnovni elementi njene geneze (stari institucionalizam i nova institucionalna ekonomija). Na toj osnovi se definišu osnovni pojmovi institucionalne ekonomije i određuje njihov metodološki smisao za potrebe analize koja se vrši u ovom radu. Tako se posebno razmatra uloga pravila i ugovora, te svojinska prava i transakcioni troškovi, prvenstveno sa aspekta ekonomske efikasnosti. Primjenom navedenih osnovnih pojmova institucionalne ekonomije analizira se i osnovno pitanje ovog rada – uticaj kvaliteta institucija na proces privatizacije i njegove efekte u Srbiji, a posledično i na njen privredni rast i razvoj. Različit ishod institucionalnih reformi i odnos prema njima je u zemljama u tranziciji rezultirao različitim nivoom njihovog ekonomskog razvoja i u značajnoj mjeri je odredio karakter ekonomskog sistema i kvalitet njegovih institucija. Nejednak razvoj u tim zemljama je posledica različitih institucionalnih kapaciteta, uslovljenih političkim i ekonomskim determinantama. Razvijenije zemlje u tranziciji su imale i kvalitetnije institucije, što im je omogućilo da sveukupne reforme sprovedu brže i efikasnije. Sa druge strane, manje razvijenije zemlje su se našle pred neodložnim institucionalnim reformama za koje nisu imale odgovarajuću ekonomsku podlogu, usled čega je njihov ekonomski razvoj bio znatno sporiji. Nisu postojale adekvatne i kvalitetne institucije. Utvrditi u kakvom ekonomskom i institucionalnom okruženju je tekao proces privatizacije u Srbiji predstavlja važan element ovog rada.
The subject of this dissertation is the mutual relationship between institutions and economic growth and development. The analysis focuses especially on this relationship during the privatization process in Serbia, as well as the influence of institutional quality on the process of privatization. Accordingly, the theoretical framework of this thesis represents institutional economics, which as a theoretical concept, points out the role and importance of institutions for economic growth and development of each country. The institutions create the "rules of the game", that represent the normative framework of given economy, as well as the mechanisms that affect its efficiency. According to this theoretical concept, economic analysis can not be reduced to an examination of the market and market relations, but must also include the institutions that regulate the market and allow its efficient functioning. The dissertation first discusses the theoretical bases of institutional economics and provide the basic elements of its genesis (old institutionalism and new institutional economics). On this basis, the basic concepts of institutional economics are defined and their methodological meaning for the purposes of the analysis are determined in this dissertation. That is why it is specifically discussed the role of rules and contracts and property rights and transaction costs, particularly in terms of economic efficiency. Using the above basic concepts of institutional economics is analyzed and the fundamental question of this dissertation - the influence of institutional quality on the privatization process and its effects in Serbia, and consequently its economic growth and development. Different outcome of institutional reforms and attitude towards them is in transition countries resulted in different levels of their economic development and to a large extent determined by the character of the economic system and the quality of its institutions. Uneven development in these countries is due to the different institutional capacities, conditioned by the political and economic determinants. Developed countries in transition had the better of the institution, which has allowed them to implement the overall reform faster and more efficiently. On the other hand, the less developed countries find themselves in front of urgent institutional reforms that did not have adequate economic base, resulting in the potential for economic growth was much slower. There were no adequate and quality institutions. Determine what kind of economic and institutional environment did the process of privatization in Serbia is an important element of this dissertation.
false
Jedna klasa trofaznih ispravljača sa sinusoidalnom ulaznom strujom koji koriste ubrizgavanje viših harmonika struje
A class of harmonic current injection based three-phase rectifiers with sinusoidal input current
Ova disertacija sadrži analizu teorijskih principa rada i praktičnu realizaciju klase trofaznih ispravljača baziranih na principu ubrizgavanja struje. Metod ubrizgavanja struje zahteva procesiranje malog dela snage ispravljača, ne generiše elektromagnetne interferencije i ima visok stepen iskorišćenja u poređenju sa kompetitivnim rešenjima, budući da nema komutacionih gubitaka. Izvršeno je detaljno razmatranje metode ubrizgavanja trećeg harmonika i analiza pojava koje se dešavaju u trofaznom diodnom ispravljaču pri ubrizgavanju trećeg harmonika struje različitih amplituda i faznih stavova. Na osnovu analitički dobijenog izraza za ukupno harmonijsko izobličenje ulazne struje je izveden zaključak da je optimalna vrednost amplitude ubrizganog trećeg harmonika 1,5 puta veća od jednosmerne struje kojom je opterećen izlaz ispravljača, uz pretpostavku da je potrošač strujni ponor. Najmanje totalno harmonijsko izobličenje koje se može postići na taj način iznosi 5,125% i pri tome mreža za ubrizgavanje struje apsorbuje 8,571% ulazne snage. Zatim su prikazane neke od mogućih realizacija optimalnog ubrizgavanja trećeg harmonika i daje se poređenje nekih tipičnih mreža za ubrizgavanje. Analizirana je mogućnost ubrizgavanja idealnog talasnog oblika struje i postavljeni problem je rešen u petom poglavlju disertacije, putem razmatranja tokova viših harmonika ubrizgane struje i to kako na naizmeničnoj tako i na jednosmernoj strani, na primeru trofaznog diodnog ispravljača koji je opterećen sa dva boost DC/DC konvertora. Prikazani su različiti putevi ubrizganih parnih i neparnih harmonika struje. Izvedene su jednačine koje daju talasni oblik ubrizgane struje sa kojim se postiže idealni sinusni talasni oblik struje koju ispravljač uzima iz napojne mreže. Zatim je prikazano nekoliko praktičnih realizacija trofaznih diodnih ispravljača sa optimalnim ubrizgavanjem struje, sa eksperimentalnim rezultatima. Pri tome je prvo analizirano optimalno ubrizgavanje samo neparnih harmonika i pokazano je da se tako može postići totalno harmonijsko izobličenje ulazne struje ispod 5% a ako se na jednosmernoj strani ubrizgavaju i parni harmonici, postiže se totalno harmonijsko izobličenje od 0%. Pokazuje se da u svakom slučaju mora da bude disipirano na otporima (ili procesirano kroz konvertore koji aktivno emuliraju otpornost) 8,81% ulazne snage i to 8.65% u mreži koja ubrizgava neparne harmonike a 0,16% u mreži koja ubrizgava parne harmonike na jednosmernoj strani. Osnovni koncept optimalnog ubrizgavanja struje je primenjen na slučaj šestofaznog diodnog ispravljača. Pri tome su fazni naponi pomereni za 30 električnih, čime se postiže generisanje kružne struje i njeno ubrizgavanje u ulaznu faznu struju ispravljača, odnosno nije potrebno koristiti neutralnu tačku. Tako je ceo problem postizanja sinusoidalnog talasnog oblika ulaznih struja trofaznog diodnog ispravljača redukovan na ubrizgavanje optimalnih talasnih oblika struje na jednosmernoj strani. I u ovom slučaju je potrebno da mreža za ubrizgavanje apsorbuje približno 2,35% izlazne snage.
This thesis presents the theoretical explanation of the current injection technique applied in three phase diode rectifiers as well as practical rectifier circuits using that technique. The current injection technique application needs to process a small amount of the total rectified power through current injection network; it does not generate EMI and has high conversion efficiency due to absence of the switching losses. The detailed analysis of the third harmonic injection technique is performed, regarding different amplitude and phase values of the injected current. By using analytically derived exact expression for input current THD, the optimal amplitude value of the injected current is found to be 1.5 times higher than the output rectifier current. In that case the minimum input current THD value achieved is 5.125% at the price of processing the 8.571% of the input rectifier power through current injection network. Some of the possible third harmonic current injection networks are explained and compared. The next problem tackled and the solution presented in the 5th chapter is the analysis of the possibility of the ideal current waveshape injection.. The optimal solution is derived by analyzing the current flow of higher harmonics. The base rectifier circuit which is used for the analysis is three phase diode rectifier loaded with two boost converters. Different flow paths for odd and even current harmonics are clarified and explained. To achieve the most sinusoidal waveshape of the rectifier input current, the optimal injected current waveshape is derived based on the previous analysis Besides that some practical current injection networks are presented. By current injection of the odd harmonics only, the THD achieved can be below 5%. Power flow in optimal waveshape current injection case is such that 8.81% of input rectifier power has to be processed through current injection network (8.65% in odd harmonics current injection network and 0.16% in even harmonics current injection network). In the 6th chapter the optimal current injection technique is applied on the six phase diode rectifier case. The two sets of three phase input voltages are displaced for 30 electrical degrees, thus enabling circular current generation and injection of the same current into input rectifier current. In this way all the necessary current injection is performed on DC rectifier side and the power processed through the current injection network in this case is only 2.35% of the rectifier output power.
false
Simulacioni model za procenu izlazne snage gsm bazne stanice
Simulation model for estimation of GSM base station output power
U okviru doktorske disertacija predstavljeni su teorijski postupci koji omogućuju vrlo precizno određivanje zavisnosti izlazne snage GSM bazne stanice od sledećih činilaca: gustine korisnika u ćeliji, udaljenosti korisnika od bazne stanice, slabljenja signala i ponašanja korisnika. Takođe su razvijeni postupci za proveru dobijenih teorijskih rezultata. Razvijeni su simulacioni modeli za procenu vrednosti izlazne snage složenih modela GSM ćelije kao što su: ćelija sa velikim udelom unutarćelijskog saobraćaja, ćelija sa velikim uticajem ograničenog broja izvora saobraćaja, ćelija sa složenom zavisnošću raspodele gustine korisnika. Na osnovu napravljenih modela, izvršena je procena efikasnosti pojedinih metoda za uštedu energije u GSM mreži, i omogućeni su uslovi za definisanje strategija za uvođenje novih mera uštede. Značaj istraživanja je izražen kroz nova saznanja o mogućnostima uštede energije u mobilnim mrežama.
In this doctoral dissertation, theoretical methods that allow very accurate dependence estimation of GSM base station output power from several factors such as user's density in radio cell, distance betwen user and base station, signal attenuation and users behavior, are presented. Also, methods for verification of obtained theoretical results are presented. Simulation models for estimation of base station output power in very complex GSM scenarios, are developed. Special cases, such as cell with large part of intracell traffic, cells with large influence of limited number of traffic sources and cells with very complex dependence of user's density, are analyzed in detail. Based on developed models, efficiency estimation of particular methods for power saving is performed. Therefore, basic conditions to define strategies for introduction of new power saving measures, are provided. Importance of research is pronounced throw new knowledge about power saving in mobile networks.
true
Koncentracije privrednih subjekata u Republici Srbiji : sa osvrtom na pravo Evropske unije
Concentrations of undertakings in the Republic of Serbia with a review on European Union law
Predmet doktorske disertacije jeste kontrola koncentracija privrednih subjekata u Republici Srbiji sa osvrtom na kontrolu koncentracija u Evropskoj uniji. Autor se bavi analizom odredaba Zakona o zaštiti konkurencije iz 2009. godine i drugih propisa kojima se uređuje kontrola koncentracija, i upoređuje ih sa pravilima koja postoje u pravu konkurencije EU. Istraživanje se vrši kako bi se predstavili i kritički analizirali određeni problemi koji se javljaju prilikom primene pravila o kontroli koncentracija, i efekti primene ovih pravila na zaštitu konkurencije i na povezivanje učesnika na tržištu. Kontrola koncentracija predstavlja veoma važan, treći stub politike zaštite konkurencije u Republici Srbiji, jer se primenom propisa o kontroli koncentracija sprečavaju koncentracije kojima se stiče ili jača dominantan položaj na tržištu ili olakšava međusobno usaglašavanje učesnika na tržištu, a kojima se na taj način može narušiti konkurencija. Koncentracije mogu narušiti konkurenciju stvaranjem nepovoljnih tržišnih struktura koje predstavljaju osnove za nekonkurentna tržišna ponašanja (zaključivanje restriktivnih sporazuma i zloupotreba dominantnog položaja), tako da je kontrola koncentracija značajan instrument za očuvanje konkurentnih tržišnih struktura koje jedine mogu da generišu povoljne efekte po ekonomsku efikasnost i time po društveno blagostanje. Sprovođenjem koncentracija ostvaruju se brojne koristi za učesnike na tržištu i za privredu, budući da koncentracije olakšavaju reorganizaciju i povezivanje učesnika na tržištu, jačaju njihovu konkurentnost i unapređuju njihove performanse, a disciplinovanjem uprave, unapređuju i korporativno upravljanje. Ipak, koncentracije mogu negativno uticati na tržište i narušiti konkurenciju, na štetu potrošača, posebno kada se koncentracijom stiče ili jača dominantan položaj na tržištu. iii Imajući u vidu negativne efekte koncentracija na konkurenciju, osnovna polazna hipoteza u radu je da je uprkos velikim potencijalnim koristima koncentracija potrebno vršiti njihovu kontrolu i kada postoje razlozi, predviđeni zakonom, usloviti ili zabraniti njihovo sprovođenje. Zato je osnovni cilj ovog istraživanja da ukaže na ona pravila o kontroli koncentracija čijom primenom se može valjano zaštiti konkurencija na tržištu, a kojom se istovremeno mogu stimulisati koncentracije učesnika na tržištu. Da bi se postigao ovaj cilj, u radu se koriste normativni, uporednopravni i istorijski metod. Na osnovu izvršenog naučnog istraživanja i analize, autor zaključuje da su u propisima Republike Srbije prihvaćeni najvažniji principi na kojima se zasniva kontrola koncentracija u EU i da ne postoji neko značajnije odstupanje od rešenja prava EU. Rešenja Zakona o zaštiti konkurencije uglavnom su usklađena sa pravilima koja važe u EU i Zakon predstavlja dobru osnovu za zaštitu konkurencije. Ipak, autor ukazuje na to da Zakon i ostali propisi ne sadrže određene odredbe kojima bi se moglo unaprediti odlučivanje u postupcima kontrole koncentracija i time obezbediti bolja zaštita konkurencije na tržištu. U disertaciji su navedena pojedina rešenja za koje autor smatra da predstavljaju nepotrebno administrativno opterećenje za učesnike na tržištu i za Komisiju za zaštitu konkurencije, a osim toga, autor se zalaže za veću primenu znanja iz ekonomije i ekonometrijskih testova i za dalji razvoj regulative, kako bi se uvećala pravna sigurnost i predvidljivost u radu Komisije za zaštitu konkurencije. U disertaciji je predloženo da se uklone suvišne barijere za učesnike na tržištu i da se obezbedi veća konzistentnost prilikom odlučivanja Komisije za zaštitu konkurencije. Na taj način bi se doprinelo i procesnoj efikasnosti i ekonomičnosti, jer bi se rasteretio rad Komisije za zaštitu konkurencije usled smanjenog obima postupanja po prijavama koncentracija koje nemaju uticaj na konkurenciju i time joj se omogućilo da se usredsredi na slučajeve koji zahtevaju veću pažnju. Izneseni predlozi se čine opravdanim, jer imaju za cilj smanjenje administrativnog opterećenja za učesnike na tržištu i stimulisanje investicija u Republici Srbiji. iv
The subject of the doctoral dissertation is the control of concentrations of undertakings in the Republic of Serbia with a review on EU merger control regime. The author explores the provisions of the Law on Protection of Competition of 2009 and other regulations concerning to the control of concentrations, and compares those provisions with EC competition law, not in order to give a simplified overview of competition law of the Republic of Serbia, but to present and critically analyse certain specific problems related to the application of those provisions and its effect on protection of competition. Control of concentrations is the third pillar of competition policy of the Republic of Serbia, which purpose is to prevent concentrations leading to the creation or reinforcement of a dominant position and to avoid the creation of a market structure that would significantly facilitate coordination of market behaviour between undertakings. Concentrations can impede effective competition by altering the market structure in such a way that undertakings are more likely to coordinate or abuse their market power to the detriment of consumers, which means that control of concentrations is very important instrument for the maintenance of a competitive market structure that is capable of delivering the benefits from competition in market of the Republic of Serbia. Concentrations offer a number of benefits for undertakings and for economy, because they facilitate restructuring and consolidation of undertakings, improve the quality of management and performance of undertakings, discipline management and enable undertakings to compete effectively on market. Nevertheless, concentrations may negatively affect market and impede competition to the detriment of consumers, especially in the case when concentration leads to the creation or strengthening of a dominant position in the market. vi Considering potential anticompetitive effects of concentrations, the basic dissertation hypothesis is that it is necessary to control concentrations of undertakings, even though other aspect of the concentrations can yield efficiency gains. Therefore, the aim of this study is to point out the rules whose application enables proper protection of competition on the market and stimulates concentrations of undertakings, at the same time. In order to achieve this aim, the dissertation uses normative, comparative and historical method. Based on the given scientific research and analysis, the author concludes that the control of concentrations in the Republic of Serbia is for the most part congruent with the key principles of the EU merger control regime and there are no immediate areas of concern. The rules laid down by the 2009 Law are generally in line with EU law standards and the Law can be seen as good basis for protection of competition. In addition, the author also indicates that the Law and other regulations do not contain some provisions that should improve the quality of decision-making process and consequently protect and develop competition in market. Throughout the dissertation, the author identifies some shortcomings, which are considered as unnecessary burden for the undertakings and for the work of the Commission for protection of competition (’Commission’), and advocates further development of economic methodology and legal standards in order to increase legal certainty and predictability of the merger control. It is suggested that it should be more consistency and it would be useful to further remove regulatory burden from undertakings’ transactions. Such improvements would be also in the interest of procedural efficiency, because they should reduce investigations of concentrations that are harmless to competition and enable the Commission to focus its resources on those cases that require more attention. Those proposals would appear justified, because their objective is a reduction in the burden of the administrative procedure and they could stimulate investments in economy of the Republic of Serbia. vii
true
Interakcija čovek-računar u integrisanom okruženju digitalnih TV prijemnika, mobilnih uređaja i interneta
Human-computer interaction in an integrated environment combining digital tv, mobile devices and internet.
Broj multimedijalnih uređaja koji su dostupni u dnevnoj sobi i imaju pristup internetu, kao što su digitalni TV prijemnici, tablet uređaji i mobilni telefoni, je u konstantnom porastu. Ove uređaje je na jednostavan način moguće povezati putem lokalne kućne mreže ili interneta. Takođe, u poslednje vreme, digitalni TV prijemnici se sve češće opremaju i operativnim sistemima koji se mogu naći na mobilnim uređajima što omogućava i lak razvoj distribuiranih aplikacija. Aplikacije razvijene u ovako integrisanom okruženju digitalnih TV prijemnika, mobilnih uređaja i interneta mogu korisnicima da pruže nove scenarije interakcije čovek-računar u odnosu na one koje pružaju aplikacije razvijene nezavisno za svaki uređaj. Proučavanjem sistema u oblasti interakcije čovek-računar u okruženju dnevne sobe sa digitalnim TV prijemnicima i mobilnim uređajima uočeno je da se postojeći sistemi razvijaju nezavisno, da su ograničene fleksibilnosti i da im je fokus uglavnom na pojedinačnim scenarijima interakcije čovek-računar. U ovom radu predstavljena je SHARP softverska platforma koja pruža programsku podršku za efikasan i uniforman razvoj distribuiranih aplikacija u integrisanom okruženju digitalnih TV prijemnika, mobilnih uređaja i interneta. Cilj razvijene platforme je da pomogne istraživačima u oblasti interakcije čovek-računar da na jednostavan i brz način kreiraju nove aplikacije koje nude različite kombinacije scenarija interakcije čovek-računar u integrisanom okruženju. Kako bi se postigao navedeni cilj na početku rada je napravljen pregled postojećih sistema u oblasti interakcije čovek-računar u dnevnoj sobi sa digitalnim TV prijemnikom i mobilnim uređajima. Nakon toga, pronađene su zajedničke karakteristike sistema i identifikovani su sledeći scenariji interakcije koje aplikacije razvijene u integrisanom okruženju mogu da pruže korisnicima: podela prikaza na više ekrana (na mobilnim uređajima se prikazuje dodatni sadržaj kao prateći onome što je na velikom TV ekranu), upotreba mobilnog uređaja za upravljanje akcijama na centralnom ekranu, detekcija pokreta korišćenjem senzora, dodatne čulne povratne informacije (haptički odgovor, svetlosno obaveštenje), pasivno gledanje televizije obogaćeno paralelnim interaktivnim aktivnostima, korisničko iskustvo poboljšano pristupom podacima iz digitalnog video toka, socijalnim mrežama i bazama metapodataka i interakcija grupe ljudi u okviru jedne prostorije sa grupom ljudi na drugim lokacijama...
The amount of digital multimedia devices in a modern day household capable of connecting to the internet has dramatically increased over the last years, including mobile devices such as smart phones and tablets as well as digital TV sets and set-top boxes. Since these devices are readily available and allow customization through software, development of distributed applications is greatly simplified. Applications developed in an integrated environment can offer users novel human-computer interaction scenarios compared to the scenarios that a single device could offer. By investigating existing systems in the field of Human-Computer Interaction (HCI) that combine digital TV and mobile devices in a living room environment it was observed that existing systems have ad-hoc implementations with limited flexibility and lack the ability to explore different combinations of HCI scenarios. This dissertation presents the SHARP development framework allowing efficient implementation of distributed systems in an integrated environment consisting of digital TV, mobile devices and internet. The purpose of the framework is to assist researchers in the field of HCI to create innovative applications that would offer different combinations of HCI scenarios in the living room with less effort. At the start of the dissertation the results of a survey providing an overview of existing HCI systems targeting a living room environment are presented. Common characteristics of these systems are extracted and a list of the following HCI scenarios that applications, developed for an integrated environment, can offer are identified: sharing content on multiple screens (private content showing on the mobile device screen and public content showing on the central screen), using the mobile device to control actions on the central screen, motion detection using sensors, supplemental sensory feedback (haptic feedback, background lighting), combining passive viewing of television with parallel interaction, enriching the user experience with access to various information services (such as DVB, social media and metadata databases), combining interaction between co-located and non-co-located people. After identifying the scenarios a proposal for the SHARP development framework is outlined along with details on its design and implementation. The design of the SHARP framework allows developers to create applications that combine the identified scenarios...
false
Relacione norme, zavisnost i elementi kvaliteta odnosa između učesnika u kanalima marketinga u Republici Srbiji
Relational norms, dependence and the elements of relationship quality between participants in marketing channels in the Republic of Serbia
Uspešno upravljanje odnosima u kanalima marketinga, odnosno ostvarivanje neophodnog nivoa koordinacije i saradnje, prepoznato je kao ključni faktor kreiranja vrednosti i ostvarivanja konkurentske prednosti, posebno imajući u vidu trendove razvijanja dugoročnih i sloţenih aranţmana izmedju učesnika. Perspektive analize transakcionih troškova i teorije relacionih ugovora sugerišu veliki značaj implementacije odgovarajuće organizacione strukture i mehanizama upravljanja procesima razmena. Upravljački mehanizmi moraju obezbediti neophodan nivo efikasnosti i prilagodljivosti sistema, odnosno zaštitu idiosinkratičkih investicija, posebno u uslovima neizvesnosti. Relacioni pristup, koji je konceptualno uobličio Ian Roderick Macneil, predvidja da su ponašanja i percepcije aktera odredjeni normama koje se vremenom razvijaju u odnosima razmene. U tom kontekstu se relacione norme posmatraju kao oblik implicitnog upravljačkog mehanizma, a koji je posebno relevantan u uslovima inherentne nekompletnosti formalnih ugovora u dugoročnim odnosima. U ovom radu se analizira uticaj odnosa zavisnosti na razvoj relacionizma u kanalima marketinga. Takodje, u fokusu su upravljački kapaciteti relacionih normi u ostvarivanju neophodnog nivoa kvaliteta odnosa, oličenog u konstruktima poverenja, oportunizma i konflikata izmedju učesnika u razmenama. Relacionizam je, u ovom istraţivanju, odredjen nivoom relacionih normi solidarnosti, reciprociteta, fleksibilnosti, integriteta uloga i uzdrţavanja od upotrebe moći. Kako pokazuju relevantna prethodna istraţivanja, odnosi izmedju učesnika u kanalima marketinga i veze izmedju ključnih konstrukata su uslovljeni, izmedju ostalog, i kulturnim i institucionalnim faktorima, kao i kontekstom organizacije konkretnog kanala. Na osnovu toga se ukazuje potreba za istraţivanjem u specifičnim uslovima privrede Republike Srbije, kao i kanalu marketinga koji je bliţi tradicionalnoj organizaciji aktivnosti. Krajnji cilj analize predstavlja kreiranje platforme za donošenje v odgovarajućih odluka o organizaciji aktivnosti razmene, odnosno implementaciju efikasnog miksa upravljačkih mehanizama i odgovarajućih strategija razvoja odnosa. Kvantitativno istraţivanje je zasnovano na podacima dobijenim putem ankete 30 maloprodajnih kompanija u Republici Srbiji, a koji tretiraju odnose sa njihovim glavnim dobavljačima mleka i mlečnih proizvoda. U postupku modeliranja strukturnih jednačina, ocena parametara je vršena na osnovu metoda maksimalne verodostojnosti, ali i primenom Bajesove statistike, što obezbedjuje dodatnu pouzdanost zaključaka. Kako pokazuju rezultati istraţivanja, viši nivoi medjuzavisnosti i ravnoteţe u odnosima zavisnosti izmedju učesnika predstavljaju povoljan kontekst za razvoj relacionih normi. Sa druge strane, viši nivo relacionizma obezbedjuje uslove za rast poverenja, odnosno umanjuje rizik od oportunističkog ponašanja i štetnih konflikata u kanalima marketinga. Rezultati istraţivanja pruţaju empirijsku podršku pretpostavci da odnosi zavisnosti predstavljaju strateški elemenat u menadţmentu interorganizacionih odnosa. Takodje, potvrdjena je primenjivost koncepta relacionog upravljanja u kontekstu odnosa izmedju učesnika u kanalima marketinga u Republici Srbiji, kao i u sistemima koji su organizovani na više konvencionalnim osnovama. Relacionizam, ako se pravilno implementira, moţe biti moćno orudje u stimulisanju razvoja i odrţavanja uspešnih dugoročnih partnerskih odnosa, odnosno unapredjivanju kapaciteta za kreiranje vrednosti i obezbedjivanje odrţive konkurentske prednosti.
Successful management of marketing channel relationships, i.e. attaining the necessary level of coordination and cooperation, has been recognized as the key factor of value creation and achieving competitive advantage, especially having in mind trends of the development of long-term and complex arrangements among participants. As suggested by both the perspectives of Transaction Cost Analysis and Relational Contract Theory, the implementation of a suitable organizational structure and mechanisms for governing exchanges is of great importance. The purpose of these governance mechanisms is to ensure the necessary level of a system efficiency and adaptability, as well as the protection of idiosyncratic investments, particularly in conditions of uncertainty. Relational approach, conceptualized by Ian Roderick Macneil, envisages that participants’ behaviours and perceptions are determined by norms which in time develop in exchange relations. In that context, relational norms are seen as a form of implicit governance mechanism, particularly relevant considering an inherent incompleteness of formal contracts in long-term relations. The impact of dependence relations on development of relationalism in marketing channels is being analyzed in this paper. Moreover, the focus is on the governing capacity of relational norms in the achievement of the necessary level of relationship quality, represented by the constructs of trust, opportunism, and conflicts among participants in exchanges. In this research, relationalism is determined by the level of relational norms of solidarity, reciprocity, flexibility, role integrity and restraint in the use of power. As indicated by relevant previous researches, relations between marketing channels participants, as well as between key constructs, are determined among other, by cultural and institutional factors, also by the context of the organization of a particular channel. Having that in mind, there is a need to conduct a research in the specific conditions of the economy of the Republic of Serbia, as well as in the marketing channel which is vii closer to traditional organization of exchange activities. The final purpose of the analysis is to create a platform for decision-making on the organization of exchanges, moreover the implementation of an efficient mix of governance mechanisms and appropriate relationship development strategies. The quantitative research is based on data obtained through a survey of 30 retail companies in the Republic of Serbia, which describe the relations with their principal milk and dairy products suppliers. The estimation of parameters of SEM model was done using the maximum likelihood method, but also through application of Bayesian statistics, which ensures the additional reliability of conclusions. As the results of the study show, the higher levels of interdependency and dependence symmetry in relationships represent a favourable context for relational norms development. On the other hand, the higher level of relationalism ensures the conditions for increase of trust, and lowers a risk of opportunistic behaviour and damaging conflicts in marketing channels. Research results provide empirical support to the assumption that dependence relations represent a strategic element in the management of interorganizational relationships. Moreover, these results confirm the applicability of the relational governance concept within the context of relationships among participants in the marketing channels in the Republic of Serbia, as well as in more conventionally organized systems. Relationalism, being properly implemented, may be a powerful tool for the purpose of stimulating development and maintaining of succesful long-term partnerships, as well as value creation capacities enhancement and providing sustainable competitive advantage.
true
Uticaj institucionalne infrastrukture na dinamiku privrednog rasta i korupciju u zemljama u tranziciji
The influence of institutional infrastructure on the economic growth dynamics and corruption in the transitional countries
Predmet istraživanja ove disertacije je način na koji odnos formalnih i neformalnih institucija koje čine institucionalnu infrastrukturu, utiče na pojavu različitih režima privrednog rasta. Dodatno je analizirao da li razlike u kvalitetu formalnih i neformalnih institucija utiču na mehanizam delovanja korupcije na privredni rast. Uzorak na kojem je vršeno istraživanje čine 111 razvijenih i zemalja u razvoju u periodu 1985-2014. godine.
Тhe research subject of this dissertation is the way in which the interaction of formal and informal institutions constituing institutional infrastructure, affects the emergence of different economic growth regimes. Additional analyse was conducted concerning the question whether the differences in the quality of formal and informal institutions affects the mechanism of corruption and economic growth relationship. The research was conducted on the sample that consists of 111 developed and developing countries in the period 1985-2014.
false
Promena biomehaničkih varijabli zaveslaja pod uticajem veslanja 2000m maksimalno mogućom brzinom na veslačkom ergometru
Changes of biomechanical variables of the stroke under influence of rowing 2000m with maximum speed on rowing ergometer
Istraživanje je realizovano na uzorku od 73 veslača iz šest veslačkih klubova iz Srbije, prosečne starosti 16 ± 2godine, prosečne telesne mase 74 ± 10kg i prosečne telesne visine 184 ± 8cm. Za prikupljanje podataka korišćen je veslački ergometar „Concept II –d“. Za merenje varijabli zaveslaja korišćen je instrument „FITROROWER“ firme „WEBA SPORT“, a za obradu podataka program „SOFTWARE EXPERT“. Praćeno je 18 biomehaničkih varijabli zaveslaja: ukupno vreme na 2000m (s), ukupan broj zaveslaja na 2000m (zav), prolazno vreme na 100m (s), brzina na veslačkom ergometru (m/s), trajanje aktivne faze (s), trajanje pasivne faze (s), ritam (%), frekvencija zaveslaja (zav/min), trajanje zaveslaja (s), pređeni put rukohvata (cm), prosečna brzina rukohvata (m/s), prosečna sila zaveslaja (N), maksimalna sila (N), gradijent sile (s), odnos prosečne i maksimalne sile (%), odnos prosečne sile sa telesnom masom veslača (%), prosečna snaga (W), izvršeni rad (J). Pored razlika apsolutnih vrednosti, zabeležene su i razlike u varijabilnosti varijabli zaveslaja, kao i korelacije između samih varijabli kod različitih takmičarskih kategorija i nivoa kvaliteta veslača. Maksimalnu brzinu na veslačkom ergometru kao krajnji cilj, moguće je ostvariti na različite načine, sa različitim uticajem i doprinosom pojedinih varijabli zaveslaja. Moguće je definisati redosled stabilizacije uticaja određenih varijabli zaveslaja u doprinosu ostvarivanja visokih vrednosti brzine. Definisani modeli predstavljaju stanja i odnose između testiranih grupa veslača u Srbiji za vreme prelaznog perioda, sa ograničenom moći uopštavanja. Pri zadavanju načina ostvarenja brzine, (većeg ili manjeg uticaja pojedinih biomehaničkih varijabli zaveslaja) kod svakog pojedinog veslača na treningu i takmičenju treba uzeti u obzir njegove specifičnosti u smislu funkcionalnih, motoričkih i tehničko – taktičkih sposobnosti. Pri definisanju zadataka za pojedine treninge potrebno je biti posebno oprezan kod kadeta i juniora, gde pojavne oblike retardanata treba tretirati strpljivo, poštujući specifična pravila i ritam njihovog razvoja sportskog rezultata, ne odbacujući ih. Greške u veslačkoj tehnici moguće je definisati i po udelu određenih biomehaničkih varijabli zaveslaja u ostvarivanju prosečne brzine. Veslanje pri visokim vrednostima tempa, na 2000m maksimalnom brzinom na veslačkom ergometru, definiše sve takmičarske kategorije i nivoe kvaliteta veslača, kao izrazito homogene skupove po skoro svim biomehaničkim varijablama zaveslaja.
Research was done on a sample of 73 rowers from 6 serbian rowing clubs, of average age 16 ± 2 years, average weight 74 ± 10kg, and average height 184 ± 8cm. Rowing machine „Concept II-d“ was used for gathering basic data. Stroke variables in real time was measured by „FITROROWER“ instrument „WEBA SPORT“, and they were processed on programe „SOFTWARE EXPERT“. Eighteen biomechanical variables of the stroke were measured: overall time on 2000m (s), overall number of the strokes for 2000m, lap times on 100m (s), speed on rowing ergometer (m/s), drive time (s), recovery time (s), rhythm (%), frequency (stroke/min), duration of the stroke (s), length of the handle movement (cm), average handle velocity (m/s), average force (N), force peak (N), time to force peak (s), average force and maximum force ratio (%), average force and rowers weight ratio (%), average power (W) and performed work (J). Diferences of absolute values, variability and corelation between stroke variables were defined between different competition catogories and quality level of rowers. Maximum speed on rowing ergometer as a main goal, is possible to achive on different ways, with different influence and contribution of specified variables of the stroke. It is possible to define order of stabilization on specified variables contribution of high speed. Defined models describe conditions and relations between tested groups of rowers in Serbia, during transition training period, with limited power of conclusion. During defining ways of achieving speed (higher or lower influence of some biomechanical variables) for every individual rower on training and competition, specific functional, motorical and techniqual-tactical capabilities, must be take into the consideration. During defining training tasks for rowers under 16 and juniors, special caution must be in power, where defined retardants must be treated with patience, and respect of their slower rhythm of sport result development, whithout rejection. Errors in rowing technique could be define also with influence specific stroke variables on achieving average speed. High frequency during rowing 2000m, defines extremly homogenous sets for all competition categories and quality level of rowers almost for all variables of the stroke.
true
Procena sličnosti procedura u binarnom kodu
Estimation of procedure similarity in binary code
U ovom radu predložen je pristup za ubrzavanje otkrivanja upotrebe softverske biblioteke koja je dostupna u obliku izvornog koda i koja je bespravno ugrađena u binarni kod. Pretpostavka je da su iz binarnog koda odstranjene sve dodatne informacije koje bi mogle pomoći prilikom otkrivanja upotrebe softverske biblioteke, što je čest slučaj u namenskim uređajima. Stoga je u radu korišćena ARM arhitektura koja je jedna od najzastupljenijih u namenskim uređajima. Pristup je zasnovan na softverskim metrikama i sastoji se od dve faze. U prvoj fazi računa se sličnost između binarnog oblika tražene procedure (procedura iz biblioteke) i svake od procedura binarnog koda za koji se sumnja da koristi biblioteku. U drugoj fazi se rangiraju procedure iz posmatranog binarnog koda u skladu sa izračunatom sličnošću prema traženoj proceduri. Na kraju se izdvaja prvih N procedura i dalje analizira ručno, pri čemu ekspert bira vrednost N u skladu sa raspoloživim vremenom i važnošću otkrivanja neovlašćene upotrebe biblioteke. Određivanje sličnosti između dve binarne procedure počinje prikupljanjem vrednosti softverskih metrika. U radu je predloženo ukupno 19 različitih softverskih metrika. Vrednosti softverskih metrika se porede jednim od tri predložena komparatora, čime se dobijaju parcijalne mere sličnosti upoređenih procedura. Parcijalne mere sličnosti se potom transformišu kako bi se pokušale dobiti vrednosti koje vernije predstavljaju sličnost. Transformisane parcijalne mere sličnosti se dalje kombinuju pomoću jedne od 7 predloženih formula kako bi se izračunala jedna vrednost koja predstavlja sličnost upoređenih procedura. Pored opisanog pristupa predloženo je i pet tehnika koje treba da poboljšaju rezultate korigujući primećene anomalije. Deo tehnika odstranjuje instrukcije i procedure koje pri poređenju unose šum i dovode do pogrešnih zaključaka. Drugi deo tehnika prikuplja dodatne informacije koje su izostavljene jednostavnim metrikama koje su predložene u osnovnom pristupu. Evaluacija je pokazala da je primenom predloženog pristupa i tehnika moguće postići odziv od 53% u slučaju testova baziranih na STAMP okruženju i 56% u slučaju testova baziranih na BusyBox paketu. Navedeni rezultati su za slučaj kada se kao uspeh posmatra pronalazak tražene procedure na prvoj poziciji. Odziv koji postiže predloženi pristup je skoro duplo veći odziv od drugog najboljeg testiranog alata, koji postiže svega 27% ukoliko se kao uspeh posmatra pronalazak procedure na prvoj poziciji. U pogledu F i F2 mera, predloženi pristup zajedno sa tehnikama postiže znatno veće vrednosti od svih drugih testiranih alata.
This work proposes an approach to accelerate the discovery of a case when a software library, available in the source code, is used in a binary code without an appropriate permission. It is assumed that the binary code does not contain any additional information that could help in detecting the use of the software library, which is often the case in embedded devices. This study uses ARM architecture which is one of the most commonly found architectures in embedded devices. Approach is based on software metrics and consists of two phases. The first phase calculates similarity between the searched binary procedure originating from the library, and each of the procedures found in the binary code that is suspected to use the library. In the second phase of the proposed approach, the procedures from the binary code are ranked in accordance with the similarity to the searched procedure. After sorting, expert conducts manual analysis of the top N procedures, where N is selected in accordance with the available time of expert and the importance of detecting that the library is used. In order to determine similarity between the two binary procedures, values of the proposed software metrics are collected. The paper proposes a total of 19 different software metrics. The values of software metrics are compared with one out of the three proposed comparators, thus obtaining partial similarity measures between compared procedures. Partial similarity measures are then transformed in such a way that more closely represent the similarity. Furthermore, they are combined using one of the 7 proposed formulas in order to calculate a single value representing the similarity between procedures. In addition to the described approach, a set of 5 techniques is proposed with the aim to improve results by correcting the observed anomalies. One set of these techniques removes instructions and procedures that introduce noise and lead to misleading conclusions. The other set of techniques collects additional information that is omitted by using simple metrics proposed in the basic approach. The evaluation shows that using the proposed approach and techniques, a recall of 53% in the case of tests from STAMP environment and 56% in the case of tests of BusyBox package can be achieved if the goal is to find the searched procedures in the first position. The proposed approach achieves almost double recall when compared to the second-best tested tool, which achieves a recall of only 27% observing the first position only. With respect to the F and F2 measures, the proposed approach, along with techniques, greatly outperforms all other tested tools.
true
Prostorna i vremenski razložena spektroskopska dijagnostika laserski indukovane plazme na čvrstoj meti u vazduhu na atmosferskom pritisku
Space and time resolved spectroscopic diagnostics of laser induced plasma on solid target in air at atmospheric preassure
LIBS ili LIPS (Laser Induced Plasma Spectroscopy) - spektroskopija laserski indukovane plazme je relativno nova tehnika za analizu materijala, čiji je razvoj uslovljen tehnološkim napretkom na polju lasera, spektrometara i detektora. Tehnika omogućava multi-elementarnu analizu svih vrsta materijala u realnom vremenu, u raznim sredinama transparentnim za lasersko zračenje, bez zahteva za pripremom uzorka. To je čini pogodnom za mnoge primene, poput kontrole kvaliteta, zaštite čovekove okoline, recikliranje materijala i analize u nedostupnim sredinama. Kvantitativna merenja pomoću LIBS-a, međutim, i dalje pokazuju nižu granicu detekcije i generalno slabiju tačnost u odnosu na druge analitičke metode, iako je postignut značajan napredak u pouzdanosti i preciznosti LIBS merenja za različite tipove materijala. Razlog je nedovoljno poznavanje složene interakcije laserskog zračenja sa ispitivanim uzorkom, koja zavisi od karakteristika lasera (talasna dužina, vreme trajanja impulsa, energije) i svojstva materijala ispitivanog uzorka, kao i nedovoljna proučenost, prostorno i vremenski zavisne interakcije plazme i čestica. Kako su kvantitativna merenja zasnovana na povezivanju (integralnog) intenziteta spektralne linije sa gustinom svih emitujućih čestica u plazmi, i preko toga sa koncentracijom elemenata u uzorku, razumevanje fizičkih procesa i njihove povezanosti u laserski indukovanoj plazmi od suštinskog je značaja za analitičke primene. Najveći deo informacija iz LIBS-a dobija se analizom zračenja plazme. Emisioni spektar je diskretan i sastoji se od atomskih, jonskih i molekulskih linija elemenata prisutnih u plazmi (metala, oksida, stena, legura, itd.). čijim ispitivanjem se može izvršiti detaljna analiza i dijagnostika plazme. Međutim, nehomogenost i vremenske promene oblika LIBS plazme, praćene promenom parametara plazme i emisijom materijala nastalog ablacijom u plazmenom oblaku, znatno otežavaju primenu spektroskopske dijagnostike. Ovaj rad obuhvata analizu tehnika, brze fotografije i optičke emisione spektroskopije, potrebnih za prostornu i vremenski razloženu dijagnostiku laserski indukovane plazme na čvrstoj meti u vazduhu na atmosferskom pritisku. Tehnikom brze fotografije određeno je optimalno vreme posmatranja plazme za merenje oblika i intenziteta spektralnih linija prisutnih u plazmi, kako bi se izbegle teškoće vezane za nehomogenost i vremensku promenljivost LIBS plazme. Pored toga uzeto je u obzir dejstvo samoapsorpcije na oblik i intenzitet spektralnih linija. Zatim je primenjena procedura Abel-ove inverzije za određivanje radijalne raspodele intenziteta linija korišćenih za merenje novih Stark-ovih parametara pre svega ne-vodoničnih linija koje su dominantne u ispitivanoj plazmi. U radu je izvršeno i nezavisno određivanje elektronske gustine i temperature što predstavlja uslov potreban za precizno određivanje Stark-ovih parametara. Dobijeni eksperimentalni rezultati predstavljaju značajan doprinos za dijagnostiku i modelovanje laserski indukovanih plazmi a time i bolje razumevanje složenih fizičkih procesa, što vodi do ostvarivanje uslova za još širu primenu LIBS-a u različitim oblastima.
LIBS (Laser Induced Breakdown Spectroscopy) or LIPS (Laser Induced Plasma Spectrosropy) is a relatively new technique for material analysis, recently developed due to advances in laser technology, spectrometers and detectors. This technique enables a multi-element real time analysis of all kinds of materials without previous sample preparation, making it useful for various applications, like quality control, environmental protection, material recycling and analysis in unavailable environments. Even with advances in reliability and precision of LIBS measurements of different kinds of materials, quantitative measurements that are based on the connection between integral intensity of the spectral line and density of emitting particles in plasma, and via that with concentration of elements in sample, are still challenging for most of the materials and suffer from lower accuracy and detection limits compared to other analytical techniques. The main reasons for that is the complex nature of the laser- sample interaction process which is dependent on several laser parameters (wavelength, pulse duration, energy) and on the characteristics of material under investigation, and space and time dependent plasma-particle interactions. Thus, understanding the physical processes and their interactions in laser induced plasma is essential for analytical applications. Most of the information obtained from LIBS measurements come from plasma radiation. Emission spectrum is discreet and includes atomic, ionic and molecular lines of all the elements present in the plasma (metals, oxides, alloys, etc.). Analysis of these spectral lines shapes is a part of analysis and diagnostics of plasma. Difficulties in applying spectral diagnostics in LIBS plasma originate from plasma inhomogeneity, changes in a plasma shape that occur over time followed by changes in plasma parameters and emission of materials due to ablation in the plasma cloud. This work includes the analysis of techniques, fast photography and optical emission spectroscopy, needed for space and time resolved diagnostics of laser induced plasma on a solid target in air at atmospheric pressure. Using the technique of fast photography, optimal time of plasma spectral observation is determined needed for recording line shapes of elements which are present in plasma. Influence of self- absorption on spectral line shapes is taken into account during analysis. Abel inversion procedure is applied to determine radial intensity distributions of spectral lines used for measurements of new Stark broadening parameters of, in the first place, non-hydrogenic spectral lines. These lines are predominant in this kind of plasma. Independent determination of electron density and temperature is done which is necessary condition for measurements of Stark broadening parameters. Obtained experimental results represent significant a contribution to diagnostics and modeling of laser induced plasmas and to better understanding of complex physical processes. This leads to achieving a broader application of LIBS technique in different areas.
true
Dual-arm robotic manupulation inspired by human skills
Dvoručna manipulacija inspirisana ljudskim veštinama
Cˇ ovekova želja da stvori svet po svojoj zamisli (ili viziji) dovela je do razvoja velikih tehniˇckih dostignu´ca. Inspiracije za stvaranje brojnih mašina istraživaˇci su pronašli posmatranjem bioloških sistema i procesa u prirodi. Najve´ci izazov za istraživaˇca je razvoj mehaniˇckog sistema nalik ˇcoveku koji ima karakteristike ljudskog bi´ca. Mehaniˇcka struktura ljudskog tela u velikoj meri je mapirana u model humanoidnih robota. Sadašnji napori istraživaˇca u oblasti humanoidne robotike i kognitivnih nauke su da transformišu modele ljudskog ponašanja i procesa u model humanoidnih robota. Modeliranje ljudskog pokreta široko je prouˇcavano i istraživano u literaturi sa ciljem da se dizajnira humanoidni robot i definiše njegovo upravljanje inspirisano karakteristikama ljudskog pokreta u svakodnevnim ljudskim aktivnostima. Razne discipline nauke analiziraju ljudsko kretanje na razliˇcite naˇcine. Postoje brojne tehnike i strategije koje imitiraju i analiziraju ljudsko kretanje, kao što su imitacioni proces, uˇcenje iz imitacije i optimizacioni pristupi. Cilj ove teze je: – da unapredi konverzione procese za imitaciju pokreta gornjeg dela tela ˇcoveka sa humanoidnim robotom za zadatke sa ili bez kontakata izmed¯u ruku i opreme; – da definiše objektivnu funkciju koja predstavlja optimizirano ljudsko kretanje sa ciljem prenošenja ljudskih veština na humanoidne robote; i – da humanoidni robot izvršava pokret na isti naˇcin kao ˇcovek minimiziranjen istih kriterijuma S obzirom da je model ˇcoveka kompleksan i obuhvata mnogo stepeni slobode, naš cilj je da se ljudski pokret i veštine kretanja prenesu na kretanje humanoidnog robota sa manje stepeni slobode. U tu svrhu, kinematski model robota ROMEO je koriš´cen za predstavljanje kinematskog model ljudskog tela. Pokreti dvoruˇcne manipulacije koji ukljuˇcuju kontakt sa okruženjem su predmet istraživanja u tezi. Ova vrsta kretanja nije dovoljno istražena u literaturi. U prvom delu teze predložen je konverzioni proces za imitaciju ljudskog kretanja sa humanoidnim robotom. Ovaj konverzioni proces poboljšava postoje´ce tehnike i razvijen je za imitaciju ljudskog kretanja sa humanoidnim robotom za izvod¯enje zadataka sa ili/i bez kontakta izmed¯u ruku i opreme...
Man’s desire to shape the world to fit his ideas (or vision) has led to major technological achievements. Many machines have been inspired by biological systems and processes observed in nature. The greatest challenge for the research engineer is to develop a human-like mechanical system that has the skills of a human being. The shape of the human body has largely inspired the mechanical structure of humanoid robots. Current efforts of researchers in humanoid robotics and cognitive sciences are aimed at translating models of human behavior to humanoid robots. Modeling of human movement has been extensively studied and explored in the literature, to design and control a humanoid robot inspired by human movement in daily human activities. Various scientific disciplines analyze human movement in different ways. There are many techniques and strategies for imitating and analyzing human motions, such as the motion imitation process, imitation learning, and optimization approaches. The objectives of this thesis are to: – improve the conversion process for the imitation of human upper-body movement by a humanoid, for a task with and/or without contact between hands and equipment; – define an objective function optimized by human movement, in order to transfer human skills to humanoid robots; and – execute humanoid robot motion similar to human motion, by minimizing the same criteria. Since the human model is complex and has many degrees of freedom, the goal is to imitate human motion and transfer human skills to a humanoid robot with fewer degrees of freedom. To this end, a kinematic model of the robot ROMEO was used as the kinematic model of a human body. The motions of a human being that uses both arms to perform a task, including contact with the environment, are analyzed in the thesis. This type of motion has not been sufficiently explored in the literature. A conversion process for the imitation of human motion by a humanoid robot is proposed in the first part of the thesis. This conversion process enhances existing techniques and was developed to imitate human motion and have a humanoid robot perform a task with and/or without contact between hands and equipment...
false
Uticaj monetarne politike na izbor valutnog režima u dvovalutnom sistemu
Monetary policy impact on the choice of currency regime in the dual currency system
Pri izboru cilja monetarne politike centralna banka, uvažavajući zakonska rešenja, može da se odluči za targetiranje monetarnih agregata, tagretiranje deviznog kursa, targetiranje inflacije kao i da kombinuje neki od prethodno nabrojanih režima. Svaki izabrani režim monetarne politike je neophodno uskladiti sa režimom deviznog kursa. Adekvatno izabran režim deviznog kursa je neophodan, ali ne i dovoljan uslov za uspešan razvoj privrede jedne zemlje. Ni najbolje postavljen sistem neće dati zadovoljavajuće rezultate ukoliko se javljaju slabosti i poremećaji u domaćim privrednim tokovima, dok sa druge strane nedostaci u deviznom sistemu mogu biti uzroci takvih poremećaja. Ukoliko se sa teorijskog aspekta analizira usklađenost cilja monetarne politike i valutnog režima uviđa se ispravnost odluke monetarnih vlasti u Srbiji da implicitno od 2006. godine, a eksplicitno od 2008. godine primenjuju inflatorno targetiranje, kao cilj monetarne politike, i u skladu sa tim izaberu rukovođeno plivajući devizni kurs. Pored toga pretpostavka implementacije inflatornog targeriranja je da se obezbedi da se stopa inflacije kreće u targetiranim okvirima, a da se referentna stopa menja u zavisnosti od kretanja stope inflacije i inflatornih očekivanja. Međutim, ukoliko se sagleda praktična primena inflatornog targetiranja u Srbiji rezultati nisu zadovoljavajući. Međugodišnja stopa inflacija, od kada se primenjuje inflatorno targetiranje, je najčešće van postavljenog targeta, dok se referenta stopa u dužem periodu držala na visokom nivou u poređenju sa niskim inflatornim očekivanjima. Ovakvo stanje u monetarnom sistemu ukazuje ili na mogućnost dopune targeta ili na njegovu promenu.
In the selection of targets of the monetary policy, taking into account legal decisions, central bank may decide to target monetary aggregates, targeting exchange rate, inflation targeting and to combine some of the previously mentioned modes. Each elected regime of monetary policy is necessary to harmonize with the exchange rate regime. Adequately chosen exchange rate regime is a necessary, but not sufficient condition for successful economic development of a country. The best set up of the exchange rate system will not give satisfactory results if occurring weaknesses and disorders in domestic economic flows, while on the other hand, the disadvantages in the foreign exchange system can be the causes of such disorders. If from theoretical aspect analyzes the compliance of the objectives of monetary policy and currency regime, it can be recognized the validity of decisions of the monetary authorities in Serbia to implicitly since 2006, and explicitly since 2008 apply inflation targeting as a monetary policy objective, and accordingly choose managed floating exchange rate. Additionally, assumption of implementation inflation targeting is to ensure that the inflation rate is moving in the targeted framework and that the reference rate changes depending on the movement of inflation and inflationary expectations. However, if we examine the practical application of inflation targeting in Serbia, the results are not satisfactory. Inflation rate, since inflation targeting is applied, is usually beyond the set targets, while the reference rate had holded at a high level in comparison with low inflationary expectations. This situation in the monetary system indicates either the option to amendment target or on its change.
true
Optimizacija protokola i metrika rutiranja za višekanalne višeinterfejsne bežične mesh mreže
Optimization of routing protocols and metrics for multi-channel multi-interface wireless mesh networks
Bežične mesh mreže – WMN (Wireless Mesh Network) predstavljaju noviji trend u bežičnim komunikacijama kojim se obezbeđuje veća fleksibilnost, pouzdanost i bolje performanse nego standardnim bežičnim lokalnim mrežama. Imajući u vidu sve veće potrebe korisnika za brzim i efikasnim prenosom podataka, kao i prenosom multimedijalnih (govornih i video) sadržaja, povećanje protoka, ostvareno uvođenjem višekanalnih višeinterfejsnih – MCMI (Multi Channel Multi Interface) WMN, i vođenje računa o kvalitetu servisa, QoS (Quality of Service), predstavljaju glavne izazove u projektovanju WMN. Kako bi se ostvarile što bolje mrežne performanse neophodno je definisati protokole i metrike rutiranja kojima se određuje optimalna putanja između izvorišnog i odredišnog čvora. S obzirom da WMN karakteriše velika dinamika signala, posebno u kompleksnim okruženjima (sa stanovišta radio pokrivanja), kakva su zatvoreni objekti, u okviru doktorske disertacije definisane su metrike rutiranja koje mogu da isprate brze promene uslova propagacije koje se dešavaju na svakom pojedinačnom linku u mreži. Predložene su tri nove metrike: powerETX (Expected Transmission Count), powerWCETT (Weighted Cumulative Expected Transmission Time) i powerMIC (Metric of Interference and Channel-switching), koje u cenu linka, osim parametara definisanih u okviru postojećih metrika ETX, WCETT i MIC, respektivno, uključuju i parametar koji zavisi od nivoa snage signala na prijemu. Nove metrike obezbeđuju veće protoke korisničkih podataka i manje kašnjenje s kraja na kraj u poređenju sa postojećim, a njihova praktična realizacija moguća je bez dodatnih hardverskih ulaganja, i uz minimalne softverske izmene. Imajući u vidu da postojeći algoritmi dodele kanala ne uključuju QoS podršku u MCMI-WMN, i da se prenos multimedijalnih sadržaja obavlja uz neprihvatljiva kašnjenja, predložen je novi protokol za sloj linka, QoS-HMCP (QoS-Hybrid Multi Channel Protocol), koji uvodi podršku QoS kroz EDCA (Enhanced Distributed Channel Access) mehanizam pristupa, i koristi dobre karakteristike hibridnog algoritma dodele kanala definisane HMCP protokolom za sloj linka. Predložene su četiri varijante QoS-HMCP protokola koji u obzir uzimaju različite kriterijume za promenu kanala i Optimizacija protokola i metrika rutiranja za višekanalne višeinterfejsne bežične mesh mreže vii procedure nadmetanja za medijum. Rezultati poređenja varijanti predloženog protokola sa osnovnim HMCP protokolom po ostvarenom protoku i kašnjenju u mreži dokazale su opravdanost uvođenja podrške QoS u MCMI mreže. Na kraju, na osnovu međusobnog poređenja predloženih varijanti protokola, doneti su zaključci koja varijanta QoS- HMCP protokola daje optimalne rezultate (najveći protok i najmanje kašnjenje), za koje saobraćajne kategorije i pod kojim mrežnim uslovima. Pored navedenog, doprinos disertacije ogleda se i u detaljnom i sistematičnom pregledu postojećih protokola rutiranja, strategija dodela kanala, i metrika rutiranja za višekanalne višeinterfejsne WMN.
Wireless mesh networks (WMNs) are a new trend in wireless communications which provide greater flexibility, reliability and better network performances than standard wireless local area networks. Having in mind the growing demand for the fast and efficient data transfer as well as multimedia transmission (voice and video), the major challenges in designing WMN are increasing throughput (by introducing multi- channel multi-interfaces (MCMI) approach into the WMN), and taking into account the quality of service (QoS). To achieve the best network performances, it is necessary to define protocols and routing metrics that determine the optimal path between the source and the destination node. Since the WMN is characterized by high dynamics of the signal, especially in complex environments (from the radio coverage point of view), such as those found indoors, routing metrics that can track fast changes in propagation conditions at each link in the network, are proposed in this dissertation. The link costs of the three new metrics, powerETX (Expected Transmission Count), powerWCETT (Weighted Cumulative Expected Transmission Time) and powerMIC (Metric of Interference and Channel-switching), besides the parameters defined in the existing ETX, WCETT and MIC metrics, respectively, include the parameter that depends on the received signal level. In comparison with existing metrics, these new metrics not only provide a greater data throughput and less average end-to-end delay, but also can be realized without any additional hardware and with minimal software changes. Keeping in mind that existing channel assignment (CA) algorithms do not include QoS support in MCMI-WMNs and that the multimedia transmission can suffer from unacceptable delays, a new protocol for link layer QoS-HMCP (QoS Hybrid Multi-Channel Protocol) which introduces QoS support through EDCA (Enhanced Distributed Channel Access) mechanism, and uses the good characteristics of the hybrid CA algorithm defined by HMCP protocol for link layer, is proposed in this dissertation. Four versions of the QoS-HMCP protocols, which take into account different criteria for channel switching and procedure for medium competition, are proposed. These versions are compared with basic HMCP protocol with regards to the throughput of user data and Optimizacija protokola i metrika rutiranja za višekanalne višeinterfejsne bežične mesh mreže ix delay through the network, and the results justify introducing QoS support into the MCMI network. Finally, based on a side-by-side comparison of the proposed versions, it is concluded which version of the QoS-HMCP protocol gives the best results (the greatest throughput with the smallest delay), and for which traffic categories and network conditions. Furthermore, this dissertation makes an additional contribution by providing a detailed and systematic review of existing routing protocols, channel assignment strategies, and the routing metrics for multi-channel multi-interface WMNs.
true
Merenje rizika pri utvrđivanju solventnosti neživotnih osiguravača
Risk measurement in determination of solvency of non-life insurers
Osiguranje pruža zaštitu od rizika koji ugrožavaju imovinu i lica kroz isplatu naknade štete, odnosno osigurane sume, nakon nastupanja osiguranog slučaja, u zamenu za unapred utvrđenu i naplaćenu premiju osiguranja. Iz primarne funkcije osiguranja, koja se ogleda u obezbeđenju ekonomske i socijalne sigurnosti osiguranika, proizilazi značaj solventnosti, kao dugoročne platežne sposobnosti osiguravajućih kompanija. U uslovima vremenske nepodudarnosti uplata i isplata iz osiguravajućeg fonda, na osnovu podataka iz prošlosti, primenom matematičko-statističkih i aktuarskih metoda, osiguravači procenjuju iznos premija i tehničkih rezervi koje su potrebne za pokriće očekivane vrednosti šteta.
Insurance provides protection against risks that threaten lives and priority through the payment of compensation or the sum insured after the occurrence of the insured event, in exchange for a previously determined and collected insurance premium. The importance of solvency as long-term debt-paying ability of insurance companies stems from the primary function of insurance, which is reflected in ensuring economic and social security of policyholders. In terms of temporal mismatch between receipts and payments from the insurance fund, based on data from the past, applying mathematical, statistical and actuarial methods, insurers estimate the amount of premiums and technical reserves needed to cover the expected value of claims.
false
Adaptibilnost modela i uticaj povratne logistike na konkurentnost zelenog lanca snabdevanja
Model adaptability and influence of reverse logistics onto the green supply chain competitiveness
Složena i dinamična poslovna filozofija upravljanja lancem snabdevanja predmet je preokupacije stručne i naučne javnosti od momenta njenog koncipiranja, tri decenije unazad. Dokaz plodotvornosti ove filozofije je mnoštvo poslovnih koncepata koji su razvijeni pod njenim okriljem, a zatim i uspešno primenjeni u praksi. Jedan od najaktuelnijih koncepata te vrste jeste koncept zelenog lanca snabdevanja. Bazirajući koncept zelenog lanca snabdevanja jednako na ekološkim, socijalnim i ekonomskim principima, usmerenost ka ostvarivanju konkurentnosti ovakvog lanca omogućena je putem aktivnosti njegove ključne dimenzije- povratne logistike. Utvrđivanje međuzavisnosti koncepta zelenog lanca snabdevanja i povratne logistike, kao njegove ključne dimenzije i faktora koji utiču na potencijal te veze, polazište je svih istraživanja u disertaciji. Celokupna disertacija podeljena je u tri glavne celine. U okviru prve celine tretiraju se pitanja vezana za promenu paradigme u lancu snabdevanja i prelazak sa tradicionalnog na zeleni lanac- glavni koncepti koji nastaju kao posledica te promene, globalni regulatorni okvir za odabrane zemlje koji se odnosi na ovo pitanje, osnovne dimenzije zelenog lanca snabdevanja i način ostvarivanja njegove konkurentnosti. Druga celina posvećena je analizi aktivnosti povratne logistike- njenom profilisanju, utvrđivanju povratnog logističkog procesa i njegovih entiteta, kao i razmatranju modela ključnih faktora primene povratne logistike. Konačno, poslednja celina vezana je za empirijsku proveru da li su globalno afirmisani modeli ključnih faktora primene povratne logistike adaptibilni uslovima srpske poslovne prakse, te od čega zavisi uticaj povratne logistike na konkurentnost zelenog lanca snabdevanja u srpskoj poslovnoj praksi i da li je taj uticaj uvek jednoobrazan. Adaptibilnost modela i uticaj povratne logistike na konkurentnost zelenog lanca snabdevanja (Doktorska disertacija) |ms Veljko. M. Mijušković 6 Sprovedenom analizom u disertaciji teži se ostvarenju više ciljeva: 1. Isticanju značaja koncepta zelenog lanca snabdevanja i njegove ključne dimenzije, globalno i na nivou srpske poslovne prakse; 2. Proveri i utvrđivanju da li najvažniji faktori primene povratne logistike, definisani i modelirani u okviru relevantnih teorijskih i praktičnih međunarodnih istraživanja, odgovaraju i u kom stepenu, faktorima primene ovih aktivnosti identifikovanih od strane menadžera u Srbiji i 3. Proveri da li je uticaj povratne logistike na konkurentnost zelenog lanca snabdevanja u srpskoj poslovnoj praksi opredeljen modalitetom tretiranja vraćenog proizvoda i da li postoji jednoobraznost tog uticaja. Uspešna realizacija svakog od navedenih ciljeva ponaosob, predstavlja značajan pomak u vršenju naučno-istraživačke delatnosti u ovoj oblasti.
A complex and dynamic business philosophy of supply chain management is the subject of preoccupation of the scientific and expert public since the moment of it's formation, three decades ago. The proof of the fruitfulness of this philosophy is a great number of business concepts which have been developed under it's wing, and then successfully implemented in practice. One of the newest concepts of that kind is the concept of green supply chain. Establishing the concept of green supply chain equally on ecologic, social and economic principles, the focus of achieving competitiveness of this chain is enabled by the activities of it's key dimension- reverse logistics. Determining the relationship between the concept of green supply chain and reverse logistics, as it's key dimension, as well as factors which influence the potential of this relationship, is the starting point of all research within the dissertation. The entire dissertation is divided into three main parts. Within the first part, issues connected with the supply chain paradigm change and the transfer from the traditional onto the green supply chain are discussed: main concepts derived as the consequence of the change, global regulatory framework for selected countries which refers to this issue, basic dimensions of the green supply chain, as well as the way of achieving it's competitiveness. The second part is dedicated to the analysis of reverse logistics activities- it's profiling, determining the reverse logistics process and it's entities, as well as the revision of reverse logistics key factors of implementation models. Finally, the last part is connected with the empirical analysis aimed to see whether the globally recognized models of reverse logistics key factors of implementation are adaptable to the conditions of Serbian business practice and what determines the influence of reverse logistics onto the competitiveness of the green supply chain in Serbian business practice and is that influence always the same. Adaptibilnost modela i uticaj povratne logistike na konkurentnost zelenog lanca snabdevanja (Doktorska disertacija) |ms Veljko. M. Mijušković 8 By conducting the analysis within the PhD dissertation we tend to achieve multiple goals: 1. To point out the significance of green supply chain and it's key dimension, both globally and within the Serbian business practice; 2. To check and determine whether the most important factors for the usage of reverse logistics, defined and modeled within the relevant theoretic and practical international surveys, suit and to what degree, the factors of usage of these activities identified by managers in Serbia and 3. To check whether the influence of reverse logistics onto the competitiveness of green supply chain in Serbian business practice is determined by mode of treating the returned product and whether there is a uniformity of that influence. A successful realization of each of the stated goals, represents a significant break through in performing the scientific and research activities within this field.
true
Jaka formulacija metode konačnih elemenata za kvazistatičku analizu vodova
The strong formulation of the finite elemet method for quasi-static analysis of transmission lines
Pojava brzih računskih mašina dovela je poslednjih decenija do pojave niza numeričkih metoda za proračun, odnosno modelovanje elektromagnetskog (EM) polja. Istraţivanje novih numeričkih metoda je konstantno aktuelno jer postoji neprestana potreba za poboljšanjem efikasnosti, tačnosti i brzine računanja. Jedna od često i multidisciplinarno primenjivanih metoda je metoda konačnih elemenata (FEM). U skoro čitavoj raspoloţivoj literaturi prikazana je tzv. slaba FEM formulacija. Ona se zasniva na funkcijama bazisa koje su izvan domena operatora koji opisuje analizirani problem, a koji je tipično diferencijalni operator drugog reda. Osim toga, najveći broj prikazanih metoda se zasniva na aproksimaciji niskog, odnosno prvog reda. Jaka FEM formulacija je malo zastupljena u literaturi i njoj je posvećen mali broj radova. Prvi radovi iz ove oblasti su pokazali da jaka formulacija moţe imati odreĎene prednosti u odnosu na slabu. Ove prednosti se ogledaju u koncepcionoj jednostavnosti i jednostavnijem i prirodnijem definisanju graničnih uslova. Kako je razvoj jake FEM formulacije u začetku, prirodno je bilo da se ona prvo razvije za dvodimenzionalne statičke i kvazistatičke probleme. Tipičan primer ovakvih problema su oklopljeni vodovi. Osnovni cilj istraţivanja ove doktorske disertacije je teorijsko i praktično zasnivanje jake FEM formulacije (Galerkinova varijanta FEM-a) za zatvorene dvodimenzionalne elektromagnetske probleme, tj. probleme kod kojih su granični uslovi zadati eksplicitno. Osim toga, pojedinačni ciljevi bili su kreiranje softvera na osnovu zasnovane numeričke metode, proučavanje karakteristika nove formulacije i primena nove formulacije na proračun vodova sa TEM talasom. Najpre je dat pregled u literaturi poznatih čvorno zasnovanih funkcija bazisa. Zatim su, polazeći od u literaturi poznatih jednodimenzionalnih funkcija bazisa formirane jake dvodimenzionalne funkcije bazisa. One zadovoljavaju granične uslove prve i druge vrste na granicama pravougaonih konačnih elemenata na koje je izdeljen računski domen. Funkcije bazisa su definisane tako da čine hijerarhijski skup, pri čemu je posvećena paţnja koncepcionoj jednostavnosti. Uočen je najniţi red aproksimacije, odnosno minimalan red jakih funkcija bazisa. Proučene su funkcije bazisa za homogene, nehomogene, izotropne, anizotropne i biizotropne sredine. Zatim je jaka FEM formulacija primenjena u kvazistatičkoj analizi oklopljenih vodova, što je ilustrovano na većem broju jednostavnijih primera za koje u literaturi postoje ili analitička rešenja, ili rešenja visoke tačnosti. Dobijeni rezultati su uporeĎeni sa rezultatima dobijenim drugim numeričkim metodama, kao i sa rezultatima dobijenim slabom FEM formulacijom. UporeĎivanjem rezultata dobijenih jakom FEM formulacijom sa rezultatima iz literature, uočeno je da postoji odlično slaganje. Pri tome se zadovoljavajuća tačnost postiţe za nizak red funkcija bazisa, obično minimalan, a konvergencija rezultata je stabilna. Kako prikazane funkcije bazisa nisu čvorno zasnovane, povećanje stepena aproksimacije (reda funkcija bazisa) ne unosi dodatne čvorove. To znači da se dodatno ne komplikuje mreţa konačnih elemenata, što predstavlja veliku prednost nad čvorno zasnovanim funkcijama bazisa. UporeĎivanjem rezultata dobijenih jakom formulacijom sa rezultatima dobijenim slabom formulacijom pokazalo se da jake funkcije bazisa, za istu tačnost rešenja, zahtevaju daleko manji broj nepoznatih. Dalje je izvršen pregled i kritička analiza formulacija FEM-a za otvorene probleme. Zatim je predloţena nova, hibridna FEM-MoM metoda za otvorene statičke probleme zasnovana na specijalnim funkcijama bazisa na granici računskog domena i primeni Green-ovog identiteta. Specijalne funkcije bazisa su iz skupa jakih funkcija bazisa. Predloţena metoda ne zahteva, barem ne eksplicitno, uvoĎenje novih nepoznatih. Na kraju su analizirani mogući dalji pravci istraţivanja.
The advent of powerful computing machines has in recent decades led to the emergence of numerous numerical methods for electromagnetic (EM) field computation and modeling. The research of new numerical methods perpetually constitutes a mainstream activity, since there is a continuous need for improvement of efficiency, accuracy and computation speed. One of frequently and multi-disciplinary used methods is the finite element method (FEM). In almost all available literature, the so called weak FEM formulation is described. It utilizes basis functions that are outside the domain of the operator describing the analyzed problem, which is typically a second-order differential operator. Moreover, the largest number of existing methods is based on the low order, i.e., first-order or second-order, approximations. Strong FEM formulation is underrepresented in the literature, and it is pursued in a small number of papers. The first papers in this field showed that the strong formulation may have certain advantages over the weak formulation, manifested in conceptual simplicity and easier and more natural definition of boundary conditions. Because the development of strong formulations is in its beginning, a natural choice was to start with a two- dimensional static and quasi-static problems. Typical examples of these problems are shielded transmission lines. The main aim of this PhD thesis is theoretical and practical establishment of the strong FEM formulation (Galerkin version of FEM) for closed two-dimensional electromagnetic problems, i.e., problems for which the boundary conditions are given explicitly. Additional objectives were to create the software based on the established numerical methods, study the features of the new formulation, and apply the new formulation in analyses of TEM-wave transmission lines. Firstly, an overview of the literature of the known node-based basis functions is given. Then, starting from the known literature on one-dimensional basis functions, two-dimensional basis functions are created. They satisfy the boundary conditions of the first and second kind on rectangular finite element borders into which the computational domain is segmented. The basis functions are defined as a hierarchical set, whereby attention is given to their conceptual simplicity. The lowest order, i.e., the minimal order for the strong formulation, approximation is observed. Basis functions for homogeneous, inhomogeneous, isotropic, anisotropic, and biisotropic media are studied. Afterwards, the strong FEM formulation is applied to the quasistatic analysis of shielded cables, which is illustrated by a number of simple examples for which analytical solutions or solutions of high accuracy exist in the literature. The obtained results are compared with numerical results obtained by other methods, as well as with the results obtained by the weak FEM formulation. Good agreement of the results obtained by the strong FEM formulation and the results from the literature is observed. At the same time, satisfactory accuracy is achieved using low-order basis functions, usually minimal, and good convergence is obtained. As the proposed basis functions are not node-based, increasing the degree of approximation (increasing the order of basis functions) does not introduce additional nodes. Hence, no additional complexness is introduced in the finite element mesh, which is a big advantage over the node-based basis functions. Comparing the results obtained by the strong FEM formulation with the results obtained by the weak FEM formulation, it is proved that strong basis functions require far fewer unknowns for the same accuracy. Subsequently, an overview and critical analysis are given of the FEM formulation for open problems. Then, a new hybrid FEM-MoM method for open static problems is proposed, based on special basis functions on the border of both computational domains and application of Green’s identity. These special basis functions belong to a set of strong basis functions. The proposed method does not require introduction of new unknowns, at least not explicitly. In the end, possible directions for future research are analyzed.
true
Filtri sa površinskim akustičkim talasom za hemijske gasne senzore
Surface acoustic wave filters for chemical vapor sensors
Ova disertacija razmatra važne probleme u analizi i modelovanju senzora na bazi elemenata sa površinskim akustičkim talasima (PAT) kao i projektovanje i realizaciju PAT filtara koji su osnova za izradu senzora. Kod ovih elemenata je zastupljena specifična vrsta procesiranja signala, procesiranje u realnom vremenu: konverzijom električnih signala visokih učestanosti u mehanički talas, njegovu modulaciju merenom neelektričnom veličinom i ponovnu konverziju u električni signal. Predloženi su nov model i metod analize transverzalnih PAT hemijskih senzora, zasnovan na metodi ekvivalentnih električnih kola, koji se može lako primeniti na bilo koju podlogu, polimer i gas. Prezentirana je takođe primena predloženog modela za određivanje brzine površinskih akustičkih talasa u višeslojnoj podlozi što je neophodno za određivanje koncentracije gasa kao i primena razvijenog metoda za analizu hemijskih PAT senzora na višeslojnim podlogama. Predložen je algoritam za projektovanje PAT filtara sa uniformnim pretvaračima sa konstantnim grupnim kašnjenjem i realizovan prototip filtra koji je baza za izradu hemijskog senzora. U prvom delu se izlažu osnovi teorije površinskih akustičkih talasa. Sažeto se predstavljaju svojstva površinskih akustičkih talasa i odgovarajuća analiza, odnosno fizički i matematički modeli. Objašnjava se funkcionisanje PAT elemenata utemeljeno na prostiranju elastičnih mehaničkih talasa visokih učestanosti duž površine monokristalne piezoelektrične podloge (supstrata). Zatim se izlaže princip rada elemenata sa površinskim akustičkim talasima, a potom načini analize i modelovanja interdigitalnih pretvarača kao i celog PAT elementa. Princip rada, modelovanje i implementacija transverzalnih senzora koji su zasnovani na elementima sa površinskim akustičkim talasom se prvo obrazlažu u drugom delu. Sažeto se prikazuje originalni metod u modelovanju PAT biosenzora koji omogućuje određivanje funkcije prenosa i ulazne admitanse PAT elementa sa uniformnim pretvaračima u konačnoj formi za opšti slučaj opterećenja. Upotrebom ovog metoda, gde se ceo filtar predstavlja kao četvoropol a IDP-i i bezelektrodni deo admitansnim matricama, moguće je vršiti analizu i predikciju u svakom pojedinačnom slučaju biosenzora kao i u slučaju PAT transpondera. U ovom delu je predložen nov model i metod analize transverzalnih hemijskih gasnih senzora sa površinskim akustičkim talasima. Problem određivanja promene izlaznog električnog signala u funkciji koncentracije hemijskog agensa je rešen korišćenjem elektromehaničkog ekvivalentnog modela. U predloženom modelu gasni senzor se predstavlja kao četvoropol, interdigitalni pretvarači se predstavljaju mrežom sa tri para krajeva, a linija za kašnjenje prekrivena osetljivim slojem mrežom sa dva para krajeva. Određuje se promena izlaznog signala (napona ili učestanosti) u funkciji koncentracije gasa. Analizira se i uticaj osobina realnog PAT senzora, kao što je slabljenje usled prostiranja, koje se u postojećim poznatim metodama analize zanemaruje. Izvedeni su eksplicitni analitički izrazi u zatvorenom obliku za estimaciju koncentracije gasa. Izrazi direktno povezuju koncentraciju gasa, parametre podloge, i centralnu učestanost. Oni omogućuju neposredan uvid u uticaj parametara projektovanja senzora na performanse senzora kao i veoma efikasnu predikciju naponskog i frekvencijskog pomaka zbog koncentracije gasa u okruženju. Prikazuju se rezultati simulacije, zasnovani na predloženom modelu, i poređenja sa eksperimentalnim rezultatima dobijenim merenjem. Predstavlja se slučaj predikcije masene osetljivosti kao funkcije učestanosti, izračunavanje koncentracije gasa trihloretilena (TCE) kada je osetljivi sloj polimer PVDT, kao i primer u kome se simulira slabljenje usled prostiranja na višim učestanostima, a za slučaj gasnog senzora sa kvarcnom podlogom za otkrivanje nedozvoljene koncentracije gasova koji emuliraju bojne otrove. Posebno su razmatrane mogućnosti primene novog metoda za analizu i modelovanje savremenih hemijskih PAT senzora na višeslojnim podlogama, koji su ili deo složenog elektronskog sistema u MEMS ili CMOS tehnologiji ili se pak radi o senzoru sa dijamantskim slojem na površini silicijumskog integrisanog kola. Prezentirana je procedura koja se sastojala iz dva koraka. Pošto je polazni parametar u analizi centralna učestanost PAT elementa koja je funkcija brzine prostiranja površinskih akustičkih talasa u materijalu, u prvom koraku je određena efektivna brzinu talasa u višeslojnoj podlozi, što je neophodno za određivanje koncentracije gasa. Izveden je izraz za brzinu talasa u složenoj PAT strukturi u konačnoj formi. Brzina je određena primenom predloženog modela, formiranjem ekvivalentne elektro-mehanička šeme linije za kašnjenje bez osetljivog sloja i korišćenjem analogije između mehaničkih i električnih veličina. Izraz eksplicitno povezuje brzinu talasa, parametre slojeva (gustina, debljina) i talasnu dužinu koja odgovara centralnoj učestanosti i koja je osnovni prametar u analizi. Pokazano je da se prikazanim načinom brzina izračunava veoma efikasno i tačno. Efikasnost znači mnogo brže i jednostavnije određivanje brzine nego uobičajeno korišćenim metodom koji zahteva složena numerička izračunavanja uključujući parcijalne diferencijalne jednačine. U narednom koraku je prezentirano određivanje koncentracije detektovanog gasa PAT senzorom sa višeslojnom strukturom korišćenjem izraza izvedenih u predloženom metodu analize. Pokazano je da se primenom preloženog metoda analize i modelovanja na PAT hemijske gasne senzore sa višeslojnom strukturom postižu značajne prednosti: (1) realizacija senzora sa manjim dimenzijama uz višu radnu učestanost, posebno na dijamantu, što povećava osetljivost senzora, (2) razvoj senzora postavljenog na podlozi od Si odnosno SiO2/Si što je savremena tehnologija, (3) dobija se kompaktan senzor (integrisani) na manjoj površini podloge. Predloženi algoritam omogućuje rešavanje posebnog problema: određivanje koncentracije gasa u slučaju senzora na višeslojnoj podlozi, zahvaljujući izvedenim formulama i jedinstvenom postupku određivanja brzine talasa u podlozi sa višeslojnom strukturom. Projektovanje i realizacija transverzalnih PAT filtara propusnika opsega učestanosti namenjenih za primenu u hemijskim senzorima su razmatrani u trećem delu i predložen je algoritam pogodan za CAD − sintezu filtara sa površinskim akustičkim talasima sa uniformnim pretvaračima. Kao teorijski osnov za razvoj algoritma za projektovanje filtara sa površinskim akustičkim talasima navedeni su principi sinteze filtra sa površinskim akustičkim talasom i uspostavljena je analogija između blok šeme idealnog transverzalnog filtra i filtra sa površinskim akustičkim talasom. Na osnovu uspostavljene analogije predstavlja se projektovanje filtra sa površinskim akustičkim talasom na osnovu opšte postojeće teorije i metoda sinteze transverzalnih filtara. Analiziran je uticaj odstupanja realizacije filtra u odnosu na željene performanse i konstrukcija pretvarača da bi se nesavršenosti kompenzovale. Posebno je prikazano projektovanje realnih PAT filtara uzimajući u obzir opterećenje i njegov uticaj na karakteristike filtra kako bi se korigovala geometrija pretvarača, pošto bi karakteristike dobijenog filtra znatno odstupale od željenih. Izložen je algoritam za kompletno projektovanje transverzalnih RF PAT filtara sa uniformnim pretvaračima sa maksimalnim propusnim opsegom i minimalnim slabljenjem i nov softver, pod imenom, IDPMPAT ver. 1.0, koji služi za projektovanje maske interdigitalnih pretvarača odabrane vrste PAT filtara. Softver IDPMPAT izračunava geometrijske dimenzije interdigitalnih pretvarača i u izlaznom izveštaju generiše crtež maske interdigitalnih pretvarača i celog filtra. Prikazana je primena predloženog algoritma i softvera IDPMPAT: projektovan je filtar sa dva identična neapodizovana pretvarača i izrađen laboratorijski prototip PAT filtra propusnika opsega učestanosti sa minimalnim slabljenjem PAT FPO, koji je osnova za izradu hemijskog senzora nanošenjem osetljivog sloja polimera između pretvarača. Ispunjenje zadatih specifikacija je provereno merenjem parametara rasejanja proizvedenog filtra. Izloženi algoritam bi se mogao upotrebiti u budućnosti za projektovanje optimalnih filtara s obzirom na raspoloživu tehnologiju i izabrani tip podloge prilikom razvoja novih senzora.
This dissertation examines important issues in the analysis and sensor modeling based on the surface acoustic waves (SAW) devices as well as the design and realization of SAW filters which are the basis for the development of sensors. A specific type of signal processing is present in these elements − the processing in real time: converting high frequency electrical signals into mechanical wave, its modulation by measured a non-electric quantity and re-conversion into electric signal. A new model and a method of analysis of transversal SAW chemical sensors are proposed; they are based on the method of equivalent electrical circuit which can be easily applied to any surface, polymer and gas. In addition, the application of the proposed model for determination of the velocity of SAW in multilayer substrate is presented what is necessary for the determination of gas concentration as well as the application of the developed method for the analysis of chemical SAW sensors on multilayer substrates. An algorithm for a complete design of filters with uniform transducers with constant group delay is proposed. A prototype filter, which is the base for the development of chemical sensors has been designed and made with proposed algorithm. The first section presents the basis of the theory of surface acoustic waves. The properties of surface acoustic waves and the corresponding analysis and physical and mathematical models are concisely presented. The functioning of the SAW devices, which is based on propagation of the high frequency elastic mechanical waves along the surface of a single crystal piezoelectric substrate, is explained. Then, the principle of opreation of the surface acoustic wave devices is presented, and methods of analysis and modeling of the interdigital transducers as well as the entire SAW device. The principle of operation, modeling and implementation of transversal sensors, which are based on the surface acoustic wave devices, are the first to be explained in the second part. The original method of modeling SAW biosensors, which allows the determination of the transfer function and the input admittance of SAW devices with uniform transducers in closed form for the general case load, is summarized. Using this method, where the entire filter is represented as an equivalent two port and IDTs and sensing part with the admittance matrix, it is possible to perform the analysis and prediction in each case of biosensors as in the case of SAW transponder. In this part a new model and method of analysis of transversal surface acoustic wave chemical vapor sensors are proposed. The problem of determining the change in the output electrical signal, as a function of the chemical agent concentration, is solved using electromechanical equivalent model. In the proposed model, the vapor sensor is modeled as a two-port device, transducers as three-port devices which have one electrical and two mechanical ports, and a delay line that is covered with a sensing layer as two-port network. The change in the output signal (voltage or frequency) in function of the gas concentration is determined. The influence of the properties of real SAW sensor, such as propagation losses, which in the existing well-known methods of analysis are neglected, is also analyzed. Explicit analytical expressions in closed form for the estimation of vapor concentration are obtained. Expressions directly connect the gas concentration, the parameters of the substrate, and the central frequency. They enable direct insight into the influence of the sensor design parameters on the sensor performance and very efficient prediction of voltage and frequency shift due to the gas concentration in the surrounding environment. The simulation results, based on the proposed model, and a comparison with experimental results, obtained by the measurement, are shown. The case of predicting the mass sensitivity as a function of frequency is represented, calculation of gas concentration of trichlorethylene (TCE) when a sensitive layer is polymer PVDT, and an example in which attenuation due to propagation at higher frequencies is simulated, for the case of a gas sensor with a quartz substrate for the detection of illicit concentration gases which emulate warfare agents. Especially are considered the possibilities of applying new method for analysis and modeling of modern chemical SAW sensors on multilayer substrates, which are either part of a complex electronic system in the MEMS or CMOS technology, or a sensor is with a diamond layer placed on the surface of silicon integrated circuits. The used procedure consisted of two steps. Since the initial parameter in the analysis is the central frequency of the SAW device which is a function of the velocity of propagation of surface acoustic waves in the material, the first step was to determine the effective velocity of waves in a multilayered substrate, which is necessary for the determination of the gas concentration. An expression for the velocity of waves in a complex SAW structure was derived in a closed-form. The velocity is determined by applying the proposed model, by forming electro-mechanical equivalent scheme of the delay line without sensitive layer and using the analogy between mechanical and electrical quantities. The relation explicitly connects the velocity of waves, the layer parameters (density, thickness) and the wavelength corresponding to the central frequency and which is the essential parameter in the analysis. It is shown that with the presented method velocity is calculated very efficiently and accurately. Efficiency means much faster and easier velocity determination than in the commonly used method that requires complex numerical calculations including partial differential equations. The next step presents the determination of the concentration of gas detected by SAW sensor with multi-layer structure using the expressions derived from the proposed method of analysis. It is shown that application of the proposed method of analysis and modeling on SAW chemical gas sensors with multi-layer structure provides significant advantages: (1) realization of the sensor with a smaller size with higher operating frequency, particularly on diamond, what increases the sensitivity of the sensor, (2) the development of the sensor placed on the surface of Si or SiO2/Si as modern technology, (3) a compact sensor (embedded) in a small area of the substrate is obtained. The proposed algorithm allows solving specific problem: the determination of the gas concentration in the case of multilayered sensor, thanks to its derived formulas and unique procedure for determining the velocity of waves in substrate with multilayer structure. The design and realization of transversal SAW bandpass filter intended for use in chemical sensors are considered in the third section and the algorithm suitable for CAD − surface acoustic wave filters synthesis with uniform interdigital transducers is proposed. As a theoretical basis for the development of an algorithm for the design of SAW filters the principles of surface acoustic wave filter synthesis are mentioned. The analogy between the block diagram of ideal transversal filter and the surface acoustic wave filter is established. Based on the existing analogies, the design of a filter with a surface acoustic wave is presented based on the existing theories and general method for the synthesis of transversal filters. The effect of discrepancies in the realization of the filter in relation to the desired performance is analyzed and design of transducers in order to compensate for imperfections. It particularly shows the design of real SAW filters taking into account the load and its influence on the characteristics of the filter in order to correct the geometry of the transducers, as the characteristics of the obtained filter would significantly deviate from the desired ones. The algorithm for the complete design of transversal uniform transducers RF SAW filters with maximal bandwidth and minimum attenuation and new software, named, IDPMPAT ver. 1.0, which is used for designing masks of interdigital transducers of selected kinds of SAW filters, are presented. IDPMPAT software calculates the geometric dimensions of interdigital transducers and in the output report generates a drawing of a mask of interdigital transducers and the entire filter. The application of the proposed algorithm and software IDPMPAT is shown as follows: the filter is designed with two identical unapodised transducers and a laboratory prototype of SAW bandpass filter with minimal attenuation PAT FPO is made, which is the basis for the development of chemical sensors by applying a sensitive polymer layer between the transducers. The fulfillment of the given specification is checked by measuring the scattering parameters of the produced filter. The presented algorithm could be used in the future for the design of optimal filters for the given available technology and the selected type of substrate, intended for the development of new sensors.
true
Optimizacija neblokirajućih paketskih mreža upotrebom praktičnog protokola za rutiranje sa balansiranjem saobraćaja
Optimization of Non-blocking Packet Networks Using the Practical Routing Protocol with Load Balancing
Kvalitetan odabir strategije rutiranja u paketskoj mreži u velikoj meri utiče na pouzdanost mreže i kvalitet prenosa podataka u njoj. Odabrana strategija rutiranja utiče i na stepen iskorišćenja različitih mrežnih resursa. Loše odabrano rutiranje može prouzrokovati stvaranje uskih grla u mreži i dovesti do blokade određenih linkova. Takođe, od primenjene strategije rutiranja zavisi kašnjenje pri prenosu paketa, kao i cena mreže. U cilju odabiranja ispravne strategije rutiranja za neku mrežnu topologiju i date korisničke zahteve, vrši se optimizacija rutiranja u mreži. Proces optimizacije rutiranja za cilj ima određivanje putanja paketa tako da se postigne što bolja vrednost neke od mrežnih performansi. U zavisnosti od problema koji se rešava, optimizacija rutiranja može imati za cilj maksimizaciju propusne moći mreže, minimizaciju cene mreže, minimizaciju kašnjenja u prenosu i slično. U ovome radu bavićemo se problemima optimizacije rutiranja u cilju maksimizacije propusne moći i minimizacije cene mreže. Pri tome, želimo da komunikaciona mreža bude neblokirajuća. To znači da mreža mora garantovati servis za sve korisnike koji generišu ili primaju saobraćaj u okviru dozvoljenih granica. Nezavisno od trenutnog stanja saobraćaja u mreži, korisnički zahtevi za slanje ili prijem paketa moraju biti usluženi, ukoliko ne izlaze iz okvira dozvoljenih vrednosti. Okosnicu ovog rada predstavlja predlog nove šema rutiranja zasnovane na balansiranju saobraćaja i rutiranju po najkraćoj putanji, koju nazivamo rutiranje sa balansiranjem saobraćaja (RBS). Predložena šema rutiranja može se optimizovati tako da se ostvari maksimalna propusna moć mreže ili minimalna cena mreže, a za optimizaciju se koristi linearno programiranje. Glavnu prednost predložene strategije rutiranja predstavlja realna mogućnost njene implementacije. Naime, rutiranje po najkraćoj putanji je u širokoj primeni, a predložena šema rutiranja je modifikacija koja se relativno lako realizuje. Dimenzija linearnog programa koji koristimo je prihvatljiva za praktičnu primenu, kao i vreme neophodno za njegovo izvršavanje. Primenom predložene šeme može se ostvariti servis za veći broj korisnika nego u slučaju kada se primenjuje regularno rutiranje po najkraćoj putanji, a da pri tom nema potrebe za promenama postojeće mrežne topologije. Takođe, u već postojećoj mreži, predložena šema omogućava servisiranje određenog broja korisnika po nižoj ceni u odnosu na cenu servisiranja istog skupa korisnika u mreži koja koristi rutiranje po najkraćoj putanji. U ovom radu biće prvo definisana strategija rutiranja sa balansiranjem saobraćaja. Zatim će biti data analiza performansi predložene strategije rutiranja i izvršeno njeno poređenje sa običnim rutiranjem po najkraćoj putanji. Konkretno, biće analizirana dva slučaja. U prvom slučaju vršićemo optimizaciju rutiranja sa balansiranjem saobraćaja, tako da se ostvari maksimalna propusna moć mreže. Propusna moć u kojoj se primenjuje rutiranje sa balansiranjem saobraćaja zatim će biti poređena sa propusnom moći mreže u kojoj je primenjeno rutiranje po najkraćoj putanji. Biće pokazano da se primenom balansiranja saobraćaja može ostvariti značajno povećanje propusne moći mreže, uz korišćenje već postojećih resursa. U drugom slučaju, biće vršena optimizacija rutiranja sa balansiranjem saobraćaja, tako da se ostvari minimalna cena mreže koja može da podrži neki zadati saobraćaj. Cena mreže koja koristi rutiranje sa balansiranjem saobraćaja zatim će biti poređena sa cenom mreže koja koristi obično rutiranje po najkraćoj putanji. Biće pokazano je da je cena mreže koja može da podrži neki zadati saobraćaj manja ukoliko se primenjuje predloženo rutiranje sa balansiranjem saobraćaja. Analiza će biti vršena kako za regularne, tako i za realne mrežne topologije. Takođe, optimizacija rutiranja biće vršena i za slučajeve u kojima se zahteva da mreža bude pouzdana. Pouzdana mreža mora biti u stanju da i u slučaju otkaza nekog resursa (linka ili rutera) omogući nesmetanu komunikaciju korisnika. U slučaju otkaza nekog resursa, veće opterećenje se postavlja na one resurse koji su ostali u funkciji. Zato se u ovakvoj mreži garantovani saobraćaj svakog korisnika mora dodatno ograničiti (čime se smanjuje propusna moć mreže), odnosno kapacitet svakog linka dodatno povećati (čime se povećava i cena mreže). Biće pokazano da su performanse predloženog rutiranja sa balansiranjem saobraćaja u svakom od analiziranih slučajeva iste ili bolje od performansi rutiranja po najkraćoj putanji, koje je danas u praksi široko rasprostranjeno.
The choice of the routing strategy in a packet network has the great impact on the reliability and the overall quality of data transmission. The chosen routing strategy also has an impact on the usage of different network resources, and bad routing decisions can cause bottlenecks in the network, and limit the throughput. Additionally, network latency as well as network cost also depend on the chosen routing scheme. In order to determine the best routing strategy for the given network and expected traffic, routing optimization is performed. The goal of routing optimization is achieving the best possible value for the chosen network performance metric. Depending on the problem that is being solved, optimization can be performed to maximize the network throughput, minimize the network cost, minimize the latency, etc. In this paper, we will address the problems of network throughput maximization, and network cost minimization. We will focus on the non-blocking networks, which guarantee the service to all users that are generating and receiving the traffic within the defined boundaries, regardless of the current traffic distribution in the network. In this paper, we will propose the novel routing scheme, based on load balancing and shortest path routing. We will refer to the proposed scheme as load balanced routing (LBR). The proposed routing protocol can be optimized using linear programming, in order to minimize the cost, or maximize the network throughput. The proposed routing scheme can easily be deployed in practice. Namely, shortest path routing is widely used, and the proposed routing is a modification that can relatively easy be implemented. The dimension of the linear program is acceptable, as well as the time needed to perform the optimization. With the proposed routing, it is possible to guarantee the service to the greater number of users, then in case of the regular shortest path routing. Also, the cost of servicing a given number of users is smaller if the proposed routing is applied instead of the shortest path routing. First, we will introduce the proposed load balanced routing, and the linear program for its optimization. Then, we will analyze the performance of the proposed routing scheme, and compare it with the performance of the shortest path routing. In particular, we will focus on two cases. In the first one, we will optimize the load balanced routing to achieve the maximum network throughput. It will be shown that the proposed load balanced routing can significantly increase the guaranteed network throughput, when compared to the regular shortest path routing. In the second case, we will optimize the load balanced routing in order to achieve minimum cost of the network that can support given user demands. It will be shown that the cost of the network with load balanced routing is lower than the cost of the network using shortest path routing. Analysis will be performed for both regular and realistic network topologies. Also, the analysis will be performed for the networks in which service reliability is demanded, in the case of single resource (link or router) failure. It will be shown that the performance of the proposed load balanced routing is better than the performance of shortest path routing in all of these cases.
true
Metode određivanja raspoloživog prenosnog kapaciteta u visokonaponskim prenosnim mrežama elektroenergetskog sistema
Methods for determination of availabile transfer capacity in high-voltage transmission network
Trend uvođenja tržišta u elektroenergetski sektor se proširio po celom svetu. U ovakvom okruženju potrošači teže da smanje svoje račune za električnu ener- giju izborom snabdevača koji najviše odgovara njihovim potrebama. Sa druge strane, od prenosnih mreža se zahteva da obezbede dovoljan kapacitet za prenos energije u različitim smerovima pri tome obezbeđujući adekvatan nivo pouzda- nosti i kvaliteta snabdevanja svih potrošača. U prošlosti su prenosni sistemi planirani da se susreću i rade samo sa malim brojem unapred dogovorenih tran- sakcija između tačno definisanih lokacija (učesnika). Pojavom deregulisanog okruženja broj bilateralnih transakcija značajno raste i operatori prenosnog sistema će morati da održavaju pouzdan i siguran rad sistema u prisustvu ovih transakcija. Zbog ovoga je potrebno razviti nove metode i alate koji će pomoći operatorima prenosnog sistema da procene uticaj različitih bilateralnih transakcija na prenosni sistem. Jedan od glavnih zahteva koji se nameće prenosnim mrežama je omogućavanje nediskriminatornog pristupa svim učesnicima tržišta električne energije. Da bi postojala zdrava konkurencija na tržištu električne energije potrebno je omogućiti i trgovinu iz susednih elektroenergetskih sistema. U cilju ostvariv- anja ovakve trgovine potrebno je obezbediti ali i objavljivati vrednosti raspo- loživog prenosnog kapaciteta (ATS) za svaki interval na kome se vrši trgovi- na. Statička vrednost ATC-a predstavlja maksimalnu snagu transakcije koja se može preneti kroz sistem u određenom smeru i preseku a da se pri tome ne naru- še relevantna ograničenja. Metode koje se bave izračunavanjem vrednosti ATS-a se dele na linearne (koje izračunavaju približnu vrednost) i iterativne metode. Obezbeđivanjem dovoljne veličine ATS-a omogućiće se slobodna i ekonomična trgovina na širem području koje može da obuhvata više susednih elektroener- getskih sistema. v Jedan od ciljeva ove disertacije je razvoj nove procedure za izračunavanje raspo- loživog prenosnog kapaciteta između dva čvora ili oblasti. Osnovna prednost razvijene procedure je što omogućava jednostavno uključivanje različitih vido- va transakcija, kao i uvažavanje svih ograničenja koja se mogu iskazati matetati- čkim putem. Razvijena procedura se zasniva na upotrebi više različitih pristupa i kombinovanjem njihovih dobrih osobina. Nalaženje vrednosti ATS-a se može postaviti i kao optimizacioni problem koji je u ovoj disertaciji rešavan pomoću metaheurističkih metoda optimizaci- je. Korišćene su dve metaheurističke metode optimizacije, genetski algoritam i algoritam crne rupe, za rešavanje ovog problema. Genetski algoritam je izab- ran zato što predstavlja jedan od standardnih metoda optimizacije. U ovoj disertaciji genetski algoritam je dodatno unapređen poboljšavanjem operacije mutiranja hromozoma. Na početku procesa operaciju mutiranja je moguće vršiti na svim bitovima, što za posledicu ima širu pretragu prostora rešenja, dok je na kasnijim iteracijama mutacija dozvoljena samo na bitovima koji imaju niži težinski faktor. Drugo poboljšanje se ogleda u implementaciji auto korekcije najveće moguće snage transakcije. Treće poboljšanje se ogleda u upotrebi više metoda selekcije i ukrštanja hromozoma čime se poboljšava proces pretrage prostora rešenja i izbegavaju se loše osobine svake pojedinačne metode selek- cije ili ukrštanja. Zbog ovih poboljšanja povećana je efikasnost i brzina izra- čunavanja vrednosti ATS-a u odnosu na klasični genetski algoritam. Nova procedura za nalaženja vrednosti ATS-a je testirana na dva realna inži- njerska problema. Prvi problem se odnosi na određivanje vrednosti ATS-a bilateralnih i multilateralnih transakcija u prenosnoj mreži. U okviru ovog problema primenom predloženog algoritma rešavana su i zagušenja u prenosnoj mreži koja nastaju kao posledica postojanja više istovremenih bilateralnih transakcija. Posmatran je uticaj redosleda propuštanja bilateralnih transak- cija u cilju što većeg prenosa električne energije između dve oblasti. Uz male modifikacije predloženi algoritam je moguće koristiti za određivanje opti- malne procentualne preraspodele snage multilateralne transakcije po prodav- cima i kupcima u cilju postizanja najveće razmene električne energije između dve oblasti. Proračun je izvršen na standardnoj IEEE 30 sabirničkoj mreži, za vi jedno bazno stanje. Na prenosnoj mreži Srbije i Crne Gore testiran je predlože- ni algoritam za izračunavanje raspoloživog prenosnog kapaciteta između ovih oblasti. Način na koji operatori prenosnog sistema određuju prekogranični kapacitet je poslužio za formiranje kriterijumske funkcije i upravljačkih promenljivih optimizacione metode. Drugi problem se odnosi na nalaženje optimalne veličine i lokacije PV elektrane. Primenom razvijenog algoritma i upotrebom kriterijumske funkcije sa težinskim faktorima nalažena je lokac- ija za izgradnju PV elektrane tako da vrednost ATS-a iz odabranog čvora ka osta- lim čvorovima u mreži bude maksimalna. Kritrijumska funkcija sa težinskim faktorima je korišćena u svrhu rešavanja ovog problema.
The introduction of the competitive market into the electrical power industry has spread all over the world. In this setting, it is expected that customers can save money on their electricity bills by selecting the adequate suppliers for their needs. On the other hand, transmission companies are required to provide sufficient capacity for power transmission in both ways while maintaining the acceptable level of reliability and quality of supply to all customers. In the past, transmission systems were planned and operated to handle a limited number of wheeling transactions between the members of the interconnected power system. In a deregulated environment, the number of bilateral transactions will grow rapidly and system operators are expected to operate their systems securely and reliably in order to accommodate these transactions. New methods and tools will then be required to help the system operators evaluate the impact of different bilateral transactions on the operation of the systems. One of the main requirements imposed on the transmission networks is to provide non- discriminatory access to all of the participants of the electricity market. In order to have fair competition in the electricity market it is essential to enable the trade between the neighboring power systems. With the aim of achieving this kind of trade it is necessary to provide and post the value of available transfer capacity (ATC) for each interval in which the trade is conducted. The static value of available transfer capacity is the maximum power that can be transmitted through the system in a certain direction without violating any of the system constraints. Methods that calculate the value of ATC are divided into linear (which calculate the approximate value) and iterative methods. Ensuring a sufficient amount of ATC will enable the undisturbed and economical trade in the wider area that can include more neighboring power systems. The aim of this dissertation is to introduce a new procedure for the calculation of the available transfer capacity between two nodes or areas. The main advantage of this method is that it enables simple inclusion of various types of transactions, including consideration of all constraints that can be expressed mathematically. The developed viii procedure is based on the use of several different approaches and on the combination of their good features. Finding the value of ATC can be set as an optimization problem which, in this dissertation, is solved by using metaheuristic methods of optimization. For the purpose of solving this optimization problem, two of metaheuristic optimization methods are used, genetic algorithm and black hole algorithm. Genetic algorithm was chosen because it represents one of the standard optimization methods. In this dissertation, genetic algorithm is further improved by introducing new operation of chromosome mutation. At the beginning of the process, operation of mutation can be performed on all bits, which results in a wider search space, while in the latter iterations mutation is allowed only on the bits that have lower weighting factor. The second improvement is reflected in the implementation of auto correction of maximum possible power of the transaction. The third improvement is reflected in the use of multiple methods of selection and crossover of chromosomes, thereby improving the process of search within the search space, and avoiding the bad qualities of each method of selection or crossover. Because of these improvements the efficiency and speed of calculation of the ATC value are increased compared to classical genetic algorithm. The new procedure for finding the value of ATC is tested on two real engineering problems. The first problem relates to the determination of the ATC value of bilateral and multilateral transactions in the transmission network. By using the proposed algorithm congestion in the transmission network can be solved, which arise as a consequence of the existence of multiple bilateral transactions. The effect of the order of establishing the transaction is observed in the order of higher electricity transmission between the two areas. With minor modifications, the proposed algorithm can be used to determine the optimal percentage of redistribution of power of multilateral transactions by sellers and buyers in order to achieve the maximum power exchange between the two areas. The calculation is done on IEEE 30 bus system, for one base case. On the transmission network of Serbia and Montenegro the proposed algorithm was tested for calculating the available transmission capacity between these areas. The way in which the transmission system operators determine the cross-border capacity is used for forming the criterion function and control variables of this optimization method. The second problem is related to the determination of optimal size and location ix of PV power plants with the aim of maximizing the value of ATC from the selected node to other nodes in the network. Aiming to solve this optimization problem the proposed algorithm was used with the criteria function that uses weight factors.
true
Vizuelna reprezentacija kriptografskih algoritama
Visual representation of cryptographic algorithms
U ovoj disertaciji je opisan novi sistem za vizuelnu reprezentaciju kriptografskih algoritama (COALA). Glavni cilj razvijenog sistema je da se pomogne studentima koji su novi u oblasti zaštite podataka da lakše i bolje savladaju gradivo iz oblasti kriptografskih algoritama koje predstavlja osnovu na koju se nadovezuju složenije teme. Na Elektrotehničkom fakultetu u Beogradu postoji kurs pod nazivom Zaštita podataka na kome se najpre izučavaju kriptografski algoritmi na kojima se zasnivaju sigurnosni protokoli i aplikacije koje služe da obezbede sigurnosne servise, a nakon toga se izučavaju aplikacije koje obezbeđuju mrežnu sigurnost, dok se na kraju ulazi u problematiku sigurnosti računarskih sistema (hakeri, virusi, itd.). Praćenjem rezultata studenata na ispitu primećeno je da studenti imaju problema da savladaju gradivo iz oblasti kriptografskih algoritama, što se kasnije negativno odražava na njihov sveukupan uspeh na ispitu. Analizom razloga koji su doveli do ovakvog stanja došlo se do zaključka da bi uvođenje sistema za vizuelnu reprezentaciju kriptografskih algoritama u okviru laboratorijskih vežbi na predmetu moglo biti jedno od rešenja problema. Korišćenje ovakvih sistema je uobičajeno na predmetima iz oblasti računarske tehnike i informatike koji izučavaju softver. Pre razvoja novog sistema, napravljena je analiza postojećih rešenja kako bi se utvrdilo da li već postoji sistem koji bi ispunio potrebe predmeta i kako bi se ustanovilo koje su to karakteristike koje bi taj sistem trebao da ima da bi se uspešno koristio u nastavi. Ustanovljeno je da iako postoji veliki broj sistema za vizuelnu reprezentaciju algoritama, relativno mali broj njih omogućava vizuelnu reprezentaciju kriptografskih algoritama, dok među analiziranim sistemima ne postoji sistem za vizuelnu reprezentaciju kriptografskih algoritama koji u potpunosti zadovoljava potrebe predmeta Zaštita podataka na Elektrotehničkom fakultetu u Beogradu. Na osnovu sistematizacije postojećih sistema za vizuelnu reprezentaciju algoritama sa posebnim osvrtom na sisteme koji se mogu koristiti za vizuelnu reprezentaciju kriptografskih algoritama definisana je metodologija razvoja ovakvih sistema. Analizom literature koja se bavi sistemima koji se koriste za pomoć u učenju definisan je način korišćenja razvijenog sistema u nastavi. Realizovani sistem omogućava praćenje izvršavanja svih kriptografskih algoritama koji su objašnjeni na časovima predavanja i vežbi, prikazuje sve detalje izvršavanja za svaki od podržanih algoritama, za pojedine algoritme gde je to od važnosti sistem radi sa realnim veličinama parametara algoritma koje se koriste u praksi i svi parametri algoritama su konfigurabilni kako bi mogli lako i brzo da se prikažu različiti primeri izvršavanja algoritama. U okviru disertacije napravljen je pregled oblasti sistema za pomoć u učenju, pregled oblasti vizuelne reprezentacije algoritama, kao i pregled oblasti vizuelne reprezentacije kriptografskih algoritama. Dat je opis kriptografskih algoritama koji su podržani u COALA sistemu. Sistem ima mogućnost vizuelne reprezentacije pet vrsta kriptografskih algoritama: supstitucionih algoritama (Cezar algoritam, monoalfabetski algoritam, Playfair algoritam i Vigenere algoritam), transpozicionih algoritama (Rail Fence algoritam i Row Transposition algoritam), produkcionih algoritama (Rotor Machine algoritam), simetričnih blok algoritama (DES algoritam i AES algoritam) i algoritama sa javnim ključem (Diffie-Hellman algoritam i RSA algoritam). Poseban akcenat je stavljen na objašnjenje detalja vezanih za vizuelnu reprezentaciju algoritama korišćenu u COALA sistemu. Prikazan je i detaljan funkcionalni opis COALA sistema za vizuelnu reprezentaciju kriptografskih algoritama. U radu su objašnjeni i detalji vezani za softversku implementaciju COALA sistema. Dat je pregled tehnologija korišćenih za softversku implementaciju, objašnjena je struktura softverskog rešenja i dati su detalji napravljenih proširenja Java i Swing API-ja, kao i način izvršavanja algoritama podržanih u COALA sistemu. Način korišćenja realizovanog sistema za vizuelnu reprezentaciju kriptografskih algoritama u okviru predmeta Zaštita podataka na Elektrotehničkom fakultetu u Beogradu opisan je kroz pregled osmišljenih laboratorijskih vežbi u okviru kojih se koristi COALA sistem. U školskoj godini 2013/2014. uvedena je još jedna novina na predmetu, a to je korišćenje sistema Moodle za ispitivanje i ocenjivanje studenata u okviru laboratorijskih vežbi, pa je i to opisano. U radu je prikazana i evaluacija COALA sistema sa aspekta efikasnosti prilikom primene na predmetu Zaštita podataka na Elektrotehničkom fakultetu u Beogradu. Sistem se koristi u nastavi u prethodne tri školske godine (uveden je u nastavu 2011/2012. školske godine). Rezultati primene su mereni za prve dve godine korišćenja sistema, s obzirom da je treća školska godina u kojoj se sistem koristi još uvek u toku. Evaluacija je podeljena na tri dela. Najpre su mereni efekti uvođenja COALA sistema u nastavu na osnovu uspešnosti studenata na predmetu u prvoj školskoj godini, zatim su upoređeni rezultati korišćenja iz prve i druge godine, a treći deo analize obuhvatio je utisak studenata na osnovu upitnika koji su popunjavali. Rezultati evaluacije su pokazali da su najveću korist od uvođenja novog sistema u nastavu imali studenti koji inače odlažu trenutak u kome polažu ispit za kasnije ispitne rokove u toku školske godine kako bi imali više vremena da se pripreme za ispit i koji ne uspevaju da polože ispit ili ga polože ali sa niskom ocenom, kao i najbolji studenti kojima je korišćenje sistema pomoglo da poboljšaju svoje ocene. Analiza ankete koju su studenti popunjavali je pokazala da je COALA sistem uspešno ispunio osnovni cilj, a to je da studenti bolje razumeju gradivo i detalje algoritama koji su pokriveni ovim sistemom.
In this dissertation a novel system for visual representation of cryptographic algorithms (CryptOgraphic Algorithms visuaL simulAtion – COALA) is presented. The main goal of the developed system is to help students who are new to the field of data security to easier and better understand the material that covers cryptographic algorithms, which represents the foundation on which the more complex topics rely on. At the School of Electrical Engineering, University of Belgrade there is a course called Data Security that covers topics from cryptographic algorithms, through security protocols and applications, to system security. By following student’s results on the exam it was noticed that students tend to have problems to cope with the material covering the field of cryptographic algorithms, which later had negative impact on their overall success on the exam. After analyzing the reasons that lead to the described situation, the conclusion was drawn that the introduction of a system for visual representation of cryptographic algorithms within the laboratory exercises in the course might be a solution to the problem. Usage of similar systems is quite usual at courses that teach software related areas in the field of computer engineering and information systems. Before a new system was developed an analysis of existing solutions in the field was made in order to determine whether a system that covers the needs of the Data Security course already exists and to define which characteristics should such system have in order to be successfully used as an educational aid. It was found that even though there is a large number of systems for visual representation of algorithms (AV systems), there is only a small number of systems that can be used for visual representation of cryptographic algorithm and none of these systems fully covers the needs of the Data Security course. Based on the systematization of the existing systems for visual representation of algorithms with special attention on those that can be used for visual representation of cryptographic algorithms a methodology for the development of such system is proposed. By analysis of the literature that is concerned with the eLearning tools an efficient way of using the developed system in education process is defined. The developed system enables detailed analysis of the execution of all cryptographic algorithms that are explained at the lecture classes and the problems classes. It presents all details of the execution of supported algorithms, where it is significant for understanding the system works with real world lengths of the algorithm parameters, and all input parameters in the algorithms are configurable in order to make it possible to easily and quickly show execution of algorithms on different examples. As a part of the dissertation, several overviews were made: an overview of the eLearning tools field, an overview of the field covering systems for visual representation of algorithms (AV systems), and an overview of the field concerned with systems for visual representation of cryptographic algorithms. The description of cryptographic algorithms that are supported in the COALA system is presented. The system supports five different types of cryptographic algorithms: substitution algorithms (Caesar, monoalphabetic, Playfair, and Vigenere algorithms), transposition algorithms (Rail Fence and Row Transposition algorithms), production algorithms (Rotor Machine), symmetric block algorithms (DES and AES), and public-key algorithms (Diffie-Hellman and RSA). Special emphasis is placed on the explanation of details about the visual representation used in the COALA system. The detailed functional description and details about software implementation of the COALA system are presented in the dissertation. Overview of the technologies that are used, the structure of the software solution, and the modifications of the standard Java API and the Swing API, and details about the implementation of the algorithm execution are given. The way in which the developed system was used within the Data Security course is described through the presentation of the designed laboratory exercises which make use of the COALA system. In the school year 2013/2014 another novelty was introduced to the course and that is the use of the Moodle platform for examination and assessment of students at the laboratory exercises and this is also explained in the dissertation. The dissertation also shows an evaluation of the COALA system from the aspect of efficiency in the application in education. The system has been in use for the past three school years (introduced in school year 2011/2012). The results of the system use at the course have been measured for the first two years, since the third year is still in progress. The evaluation is divided in three parts. In the first part the effects of the COALA system introduction to the course is measured based on the results of students on the exam. In the second part the results from two consecutive years of the system usage have been compared. In the third part the opinion of students who used the system was analyzed through a survey conducted in the second year of usage. The evaluation results showed that the introduction of the COALA system brought benefit to all students, especially to those students who previously could not pass the exam in the first examination period of a school year and to some of the best students to improve their grades. The analysis of the survey conducted among the students who used the system showed that the COALA system successfully fulfilled the main goal which was to help students to better understand the material and details of the algorithms that are covered by the system.
true
Softverski sistem za učenje i primenu algoritama veštačke inteligencije
Software system for learning and application of artificial intelligence algorithms
U ovoj doktorskoj disertaciji opisan je softverski sistem za učenje i primenu algoritama veštačke inteligencije nazvan SAIL (eng. System for Artificial Intelligence Learning). U sistemu je realizovano 30 standardnih algoritama pretraživanja i algoritama teorije igara, algoritama rezonovanja i reprezentacija znanja, algoritama za rešavanje problema, algoritama za rad u neizvesnom okruženju i algoritama mašinskog učenja. Ideja da se realizuje ovakav softverski sistem proistekla je pre svega iz nastave, na predmetu Inteligentni sistemi, gde su studenti zbog raznolikosti ovih algoritama vrlo teško razumeli rad tih algoritama i koncepte koji su predavani. Značaj ovakvog sistema je što on takođe osim u nastavi, može da se primeni i u rešavanju praktičnih i realnih problema, pa su pokazane i njegove primene kod problema nalaženja slobodnog mesta na pametnom parkingu sa bežičnom senzorskom mrežom, kod sistema za odlučivanje u medicinske svrhe i kod sistema za obradu velike količine podataka. Istraživanje tokom disertacije obuhvatilo je pregled postojećih softverskih sistema koji se koriste u ovoj oblasti na različitim svetskim univerzitetima, primenu takvih sistema u elektronskom i mobilnom učenju, samostalnom učenju i učenju na daljinu, definisanje zahteva za izradu potpuno novog softverskog sistema zasnovanog na rezultatima analize postojećih sistema, realizaciju novog višeplatformskog sistema zajedničkog korisničkog interfejsa za sve simulacije i njegovu primenu. iv Softverski sistem SAIL je modularno projektovan, iterativnom metodologijom uz delimičnu primenu agilnog pristupa i razvoj testova tokom procesa implementacije. Sistem je osmišljen tako da bude lak za implementiranje, testiranje i nadogradnju. U prvoj verziji realizovana je računarska aplikacija sa algoritmima pretraživanja, kojom je pokazan značaj vizuelizacije algoritama, visok nivo interakcije između korisnika i sistema, izvršavanje simulacije algoritma korak po korak i prikaz simulacije u realnom vremenu. Osim toga realizovano je učitavanje postavke predefinisanih problema i zadataka iz datoteke i njihovo izvršavanje u okviru sistema, mogućnost pravljenja sopstvenih primera, primera sa promenljivim ulaznim parametrima simulacije, snimanje tekućih primera sa svim parametrima i učitavanje ranije snimljenih primera. U drugoj verziji sistema su dodate simulacije algoritama teorije igara, algoritmi rezonovanja, produkcioni sistemi, GPS i STRIPS algoritam, a zatim su u narednim verzijama dodati algoritmi rada u neizvesnom okruženju i strategije mašinskog učenja. U disertaciji je diskutovano i za koje probleme i zadatke je najbolje upotrebiti koju grupu algoritama. Raznolikost među realizovanim algoritmima uticala je da se pre početka realizacije organizuje zajednički grafički korisnički interfejs za sve simulacije, sa intuitivnim zajedničkim komandama. Svaka simulacija dodatno ima svoje specifične komande u traci sa alatima i korisničkom meniju. Sistem podržava i opciju izvoženja simuliranih primera u PDF datoteku i višejezičnost, pa u sistemu trenutno postoje srpski jezik ćiriličnog pisma, srpski jezik latiničnog pisma i engleski jezik, ali je moguća vrlo laka nadogradnja i za druge jezike. Jedan od bitnih ciljeva istraživanja bio je da ovaj softverski sistem bude višeplatformski, odnosno da osim računarske aplikacije bude realizovana i mobilna aplikacija, koja bi našla primenu u samostalnom učenju uz pomoć mobilnih uređaja. Implementirane klase iz računarske aplikacije iskorišćene su prilikom projektovanja mobilne platforme uz generisanje novog izgleda aplikacije prilagođenog ekranima na mobilnim uređajima. Uz softverski sistem SAIL koji je realizovan kao računarska i mobilna aplikacija, u okviru istraživanja prikazana je i mogućnost njegove nadogradnje kao kompjuterske igre i pokazana je primena igara u edukaciji (eng. gamification). Igre su napravljene tako da korisnik igre može u svakom trenutku da vidi izvršavanje algoritma uz graf/stablo v pretraživanja, neki drugi grafički ili tekstualni prikaz i da ima mogućnost da napravi uporednu analizu algoritama tokom njihovog izvršavanja. Na kraju istraživanja prikazane su primene ovog softverskog sistema u nastavi i u praksi, i dati su rezultati kroz evaluaciju. Evaluacija je pokazala da primena ovakvog sistema u nastavi ima jako veliki značaj i za predavače i za studente, jer su rezultati primene ovog sistema u intervalu od šest godina pokazali da studenti bolje razumeju algoritme obuhvaćene ovim softverskim sistemom i da je njihov uspeh, kao i znanje koje su stekli, veće nego u vremenskom periodu pre primene ovog softverskog sistema. Primeri iz realnih sistema su pokazali da takvi sistemi zasnovani na algoritmima veštačke inteligencije imaju veoma široku primenu u mnogim oblastima i danas ne postoji softverski sistem koji ne koristi neki aspekt veštačke inteligencije.
The aim of this doctoral thesis is to present a software system for learning and application of the artificial intelligence algorithms SAIL (System for Artificial Intelligence Learning). Within the system 30 standard algorithms have been realised: search and game theory algorithms, reasoning and knowledge representation algorithms, problem solving algorithms, uncertain environment algorithms and machine learning algorithms. The idea to realise this system was generated during the teaching of the course Intelligent Systems, where students had difficulties to understand these algorithms and related concepts due to their complexity. The importance of this system is in its applications – in addition to its educational purposes, the system can be applied in solving practical and real problems. Its applications in finding free parking spot within the smart parking with wireless sensor network, within the system for decision-making for medical applications and within the big data processing system have been demonstrated. This research work included the review of the existing software systems that are used in this area at universities across the world, application of such systems in electronic and mobile learning, individual and distance learning, requirements definition for the development of completely new software system following the analysis of the existing systems, realisation of the new multiplatform system of common user interface for all the simulations and applications. Software system SAIL was modularly designed, using iterative methodology and partially using agile approach and testing during its implementation. The system was designed to be easy to implement, test and upgrade. In its first version, a computer application with search algorithms was vii realised, which demonstrated the importance of algorithm visualisation, high level of interaction between the user and the system, execution of the algorithm simulation step-by-step and real-time simulation. Furthermore, loading of existing predefined problems and exercises from files and their execution within the system have been realised, the possibility to define new examples, examples with changed input parameters, saving of examples and loading of saved examples. The second version of the system added game theory algorithms simulations, reasoning algorithms, production systems, GPS and STRIPS algorithm, and the following versions added the uncertain environment algorithms and machine learning strategies. This research work discusses which problems and exercises are most suitable for certain groups of algorithms. Diversity among realised algorithms influenced the choice of the common graphical user interface for all the simulations, with intuitive commands. In addition, every simulation has its own specific commands in the toolbar and user menu. The system supports exporting of simulated examples into a PDF file and it has multilingual support, it currently supports Serbian Cyrillic, Serbian Latin and English, with an option to easily add other languages. One of the important aims of this research work was to create a multiplatform system and in addition to a computer application, a mobile application was realised, which finds use in self- learning using mobile devices. The implemented classes from the computer application were used during the design of the mobile platform and the new application was adjusted to mobile screens. In addition to the software system SAIL realised as computer and mobile application, this research work presented the possibility of its upgrade as a computer game and the concept of gamification was demonstrated. Games are designed in a way that allows user to see running of the algorithm with a graph/search tree, other graphic or textual representation, with the possibility to compare algorithms during their execution. The final part of this research presents the applications of this software system in education and in practice, and it includes the evaluation of the results. The evaluation demonstrated that the application of this system in education is of great importance for both instructors and students, as the results of the application of this system during six years of use show that students understand the included algorithms much better and that their success and acquired knowledge are greater when compared with the period before this software system was applied. viii Examples from real systems have shown that such systems based on artificial intelligence algorithms have a very wide application in many areas and today there is no software system that does not use some aspect of artificial intelligence.
true
Projektovanje asinhronih mašina sa malom inercijom i velikom brzinom obrtanja
Design of high speed, low inertia induction machines
Glavni cilj ovog rada je razvoj novog pristupa projektovanju i optimizaciji asinhronih mašina male inercije i velike brzine obrtanja. U radu su detaljno razmotrene postojeće koncepcije rotora asinhronih mašina sa malom inercijom od kojih je izdvojen koncept čašastog rotora zbog svoje jednostavnosti i jako male inercije. Zatim je analiziran uticaj povišenih učestanosti napajanja visokobrzinskih mašina na izbor materijala i proceduru projektovanja namotaja statora. Predložena je modifikovana procedura za projektovanje statora koja rezultuje geometrijom koja zadovoljava sva ulazna ograničenja i specifičnosti visokobrzinskih mašina. Posebna pažnja u radu je posvećena modelovanju i optimizaciji čašastog rotora u cilju dobijanja geometrije koja rezultuje maksimalnim stepenom iskorišćenja snage mašine. U svrhu analize rada motora i mogućnosti za optimizaciju čaše korišćen je prilagođeni analitički model koji ne uzima u obzir površinski efekat u rotoru. Rezultat analize su ulazna ograničenja za proces optimizacije u 2D FEM softveru. Na osnovu rezultata 2D FEM simulacija dobijena je optimalna geometrija rotora koja je zatim verifikovana kroz nekoliko 3D FEM simulacija. Poređenjem je potvrđena validnost optimizacije u 2D FEM i dobijeni odgovarajući korekcioni faktori. Rad takođe obrađuje i uticaj materijala rotorske čaše na performanse motora i optimalnu geometriju rotora. Poseban deo rada je posvećen analizi termičkih i mehaničkih aspekata čašastog rotora gde se razmatra praktična primenljivost predloženog rešenja. Poslednji deo rada opisuje postupak formiranja eksperimentalne postavke koja je iskorišćena za validaciju predloženog pristupa projektovanju i optimizaciji rotora. U okviru ovog dela obrađen je i postupak projektovanja šestofaznog namotaja i data je kratka analiza pogodnosti njegove primene umesto trofaznog. Modifikacijom konvencionalnog mrežno napajanog asinhronog motora dobijen je šestofazni asinhroni motor sa čašastim rotorom čije su glavne dimenzije određene predloženim procesom optimizacije. Motor je zatim ispitan u praznom hodu i pod opterećenjem za...
The aim of this work is to present the new design and optimization procedure for high speed asynchronous machines with very low rotor inertia. The work contains an overview of different rotor concepts for high speed asynchronous machines. From the found solutions, drag-cup rotor is distinguished as a concept with simple construction and very low inertia. After that, the impact of high supply frequency in high speed machines on materials selection and winding design procedure is investigated. Modified procedure for stator design is proposed which results with geometry which satisfies all input limitations and specific issues of high speed electrical machines. The main attention is paid to modeling and optimization of drag-cup rotor as very specific structure in order to find the optimal design which results with maximum degree of efficiency of machine in defined operation regime. For the purpose of analysis of motor characteristics and its potential for optimization the analytical model for drag-cup machine is adopted which doesn’t take into account only the skin effect in drag-cup. The main results of this analysis are the input boundaries for optimization procedure in 2D FEM software. The results of 2D FEM simulations are the optimized rotor geometry which was then verified with few 3D simulations. Comparison of the results verifies the validity of the 2D optimization procedure and gives the necessary correction factors. The work also deals with impact of drag-cup rotor material on machine performance and optimal geometry of the rotor. Special part of the work is dedicated to analysis of thermal and mechanical aspects of drag-cup rotor design with the main accent to the applicability of proposed rotor concept for high speed applications. The last part of the work describes the development of experimental setup which is used for verification of proposed design and optimization procedure. This part also contains the description of design procedure for asymmetrical six phase stator winding which emphasizes its advantages over the three phase winding. Prototype of six phase drag-cup motor is built by modification of conventional asynchronous motor. Main dimensions of drag-cup rotor for the prototype are determined with proposed optimization procedure. The motor is then tested in no-load and on-load conditions. Obtained motor characteristics are...
false
Fraktalna i multifraktalna karakterizacija 3D video formata
Fractal and multifractal characterization of 3D video formats
Trodimenzionalni (3D) video sadržaj postaje sve prisutniji u saobra- ćaju žičanih i bežičnih mreža. Kako 3D video karakteriše velika količina podataka, njegovo prisustvo u mreži predstavlja značajano opterećenje za saobraćaj mreže, pa je neophodno poboljšanje u načinu prenosa ove vrste signala. Predmet ovog rada je fraktalna i multifraktalna karakterizacija 3D video formata sa ciljevima pokazivanja da 3D video ima fraktalne i multifraktalne osobine i opisiva- nja ovih svojstava, poređenja svojstava za različite video formate i mogućih primena u usavršavanju modelovanja saobraćaja 3D video formata i razvijanju boljih tehnika za prenos, smoothing i statističko multipleksiranje videa. Izvršeno je kreiranje algoritama i softverska realizacija u programima Matlab i Python, u okviru matematičke metode iz oblasti kompleksnih sistema i haosa, za određivanje fraktalnih osobina (metoda agregacije varijanse (aggregated variance method), R/S statistički metod (R/S statistic method) i metod više skala (multi- scale method) za određivanje Hurst-vog indeksa) i multifraktalnih osobina (metoda momenta za multifraktalne spektre i procenu generalizovane dimenzije, metoda hi- stograma za multifraktalne spektre i inverznu multifraktalnu analizu). Na osnovu dobijenih rezultata potvrđene su fraktalne i multifraktalne osobine 3D video formata i izvršeno njihovo poređenje, uzimajući u obzir format 3D videa, strukturu, tipove frejmova, kvantizacione parametre kodera, agregaciju frejmova i metode emitovanja.
Three-dimensional (3D) videos are becoming increasingly present in wireland and wireless networks, and due to the high data content they present significant load in the network. This impact of 3D video on network traffic causes a need of improving means to handle this type of data. Subject of this thesis is fractal and multifractal characterization of 3D video formats with the main goals to show that 3D video has fractal and multifractal properties, to describe them, to compare these properties for different 3D formats, for possible applications in traffic modeling of 3D video formats and for the development of improved transport techniques for smoothing and statistical multiplexing of video. Algorithms and software realizations for mathematical methods in the area of complex systems and chaos were developed in Matlab and Python, for determina- tion of fractal properties (the aggregated variance method, the R/S statistic method, and the multiscale method) and multifractal properties (the method of moments for multifractal spectra and estimation of the generalized dimensions and the histogram method for multifractal spectra and inverse multifractal analysis). Based on the ob- tained results, fractal and multifractal properties of 3D video formats are confirmed, their comparison is performed, for various 3D video formats, video structures, frame types, values of the encoder quantization parameters, aggregation of the frames, and streaming method of video.
true
Nove primene procesa nagrizanja silicijuma u vodenom rastvoru TMAH u izradi MEMS senzora
New applications of silicon wet etching using tmah water solution for fabrication of mems sensors.
U ovom radu prikazani su razvoj i usavršavanje tehnološkog procesa vlažnog anizotropnog hemijskog nagrizanja silicijuma orijentacije (100) u vodenom rastvoru tetrametil-amonijum hidroksida (TMAH) koncentracije 25 tež. % na temperaturi od 800C. Razvoj maskless tehnike nagrizanja silicijuma u vodenom rastvoru TMAH je odredio kristalografske ravni koje se pojavljuju tokom ove vrste nagrizanja. Preostale kristalografske ravni su određene dodatnim nagrizanjem silicijumskih struktura definisanih kvadratnim likovima sa stranicama projektovanim u različitim kristalografskim pravcima. Određene su brzine nagrizanja uočenih kristalografskih ravni merenjem odgovarajućih parametara. Na osnovu brzina nagrizanja i pojavljivanja i nestajanja uočenih kristalografskih ravni tokom nagrizanja objašnjen je mehanizam nagrizanja u vodenom rastvoru TMAH koncentracije 25 tež. % na temperaturi od 800C. Mehanizam anizotropnog nagrizanja silicijuma u vodenom rastvoru TMAH definisao je ograničenja tehnološkog procesa. Razvijene su tehnike kompenzacije konveksnih i konkavnih uglova i tehnika maskless nagrizanja čiji je cilj prevazilaženje uočenih ograničenja. Prikazane su primene osvojenih tehnika nagrizanja u vodenom rastvoru TMAH u izradi različitih trodimenzionalnih silicijumskih struktura, u izradi novih MEMS senzora i poboljšavanju postojećih senzorskih struktura.
This dissertation presents the development and improvement of the technological process of wet anisotropic chemical etching in a 25 wt % TMAH water solution at а temperature of 800C of a (100) silicon substrate. The development of a maskless etching technique in TMAH water solution has determined all the silicon crystallographic planes that appeared during this type of etching. The remaining crystallographic planes were determined by additional etching of silicon structures that had been defined by square islands with the sides designed along various crystallographic directions. The etch rates of all exposed planes have been calculated by measuring the time dependence of the appropriate parameters. Various silicon crystallographic planes have different etch rates and some planes disappear while others appear during etching. Based on these facts, a mechanism of wet silicon etching in 25 wt % TMAH water solution at a temperature of 800C is explained. The mechanism of anisotropic etching in TMAH water solution has determined the limitations of this technological process. Convex corner compensation, concave corner compensation and maskless techniques were developed to overcome the observed limitations. As examples of the developed techniques in TMAH water solution, various 3D silicon structures, new and improved MEMS sensors structures have been fabricated.
true
Određivanje temperature i otpornosti namotaja rotora asinhrone mašine na osnovu analize statorskih napona i struja
Estimation of Rotor Resistance and temperature of induction machine Based on Analysis of terminal voltages and currents
Merenje temperature asinhronih mašina u pogonu je od velike važnosti za postizanje optimalnih uslova eksploatacije. Zagrevanje mašina iznad određenih temperatura može dovesti do skraćivanja veka mašina pa čak i do uništenja. Postojeće tehnike za merenje temperature mašina svode se na direktno ili indirektno merenje temperature. Direktno merenje temperature statora moguće je fizičkim postavljanjem senzora na stator dok se temperatura rotora veoma teško meri direktno. Indirektno merenje uključuje određene algoritme koji koriste ili termičke ili električne parametre kako bi izračunali temperaturu. Opravdanost korišćenja termičkih modela se javlja tek razvojem digitalnih signal procesora koji se koriste za upravljanje pretvaračima za napajanje mašina, međutim uslovi rada veoma utiču na preciznost takvih metoda. Merenje temperature korišćenjem električnih veličina najčešće se svodi na injektovanje test signala u mašinu kako bi se izvršila estimacija temperaturno zavisnih parametara na osnovu koje se može izračunati temperatura. Takve tehnike su opravdane samo u slučajevima gde se ne očekuju visoke performanse pogona jer se injektovanjem test signala stvaraju oscilacije u elektromagnetskom momentu. Skoriji razvoj digitalnih signal procesora je doveo do toga da su brzine rada procesora i periferijskih jedinica takve da mogu odgovoriti veoma zahtevnim aplikacijama u kojima se zahteva brzina i visoka tačnost. Ovim se pre svega misli na radni takt koji prelazi frekvenciju od 300 MHz, brzinu odabiranja analogno digitalnim konvertorima od preko 2 MS i sa rezolucijom konverzije i do 16 bita kao i uključivanjem DMA1 jedinica u okviru periferijskih jedinica. Takođe, sve je veći broj primena u kojima se zahtevaju veoma visoki propusni opsezi u regulacionim petljama struje, brzine i pozicije te se projektuju motori sa veoma malim rasipnim induktivnostima. Energetski pretvarači koji se koriste za napajanje ovakvih pogona imaju sve više frekvencije komutacija kako tehnologija proizvodnje silicijumskih 1 Direct Memory Access iii komponenata napreduje. Frekvencije komutacija se sve češće postavljaju i iznad 10 kHz što se do skora nije praktikovalo u industriji. U ovoj tezi predstavljen je metod za merenje temperature rotora neinvazivnim putem i bez korišćenja dodatnih senzora. Oslanja se na okolnost da modulisani naponi, koje daju energetski pretvarači, sadrže u spektru određene komponente na povišenim frekvencijama usled svog impulsnog karaktera. To omogućava da se ovakva inherentna perturbacija posmatra kao test signal koji već postoji u mašini i da merenjem napona i struja na povišenim frekvencijama postoji mogućnost za ocenu parametara. Prisustvo struja od nekoliko kHz do nekoliko desetina kHz stvara značajno veći uticaj skin efekta na rotorskom namotaju nego na statorskom što rezultuje uvećanjem rotorske otpornosti čak i nekoliko desetina puta. Na povišenim frekvencijama karakteristike magnetskih materijala postaju veoma nelinearne. Zbog toga je neophodno analizirati pojave u mašini i na osnovu toga načiniti određene linearizacija kako bi predloženi model bio primenjiv. Temperatura se može odrediti poređenjem temperaturno zavisnih parametara mašine na različitim temperaturama. Poređenjem otpornosti ulazne impedanse na 10 kHz može biti vrlo nepouzdano usled fazne greške u odabiranju napona i struja. Analizom elektromagnetskih pojava u frekvencijskoj oblasti sa jakim prisustvom skin efekta, zaključuje se da otpornost i reaktansa postaju jednake. Ovo važi samo ako su frekvencije manje od onih na kojima se parazitni efekti ne mogu zanemariti. To omogućava određivanje temperature poređenjem modula impedanse pa se na taj način izbegava greška usled fazne greške u merenju. Poznavanjem temperature može se u realnom vremenu vršiti korekcija rotorske otpornosti kao parametra koji se koristi u regulacionim petljama. U tezi se nalazi opsežna analiza postojeće literature na osnovu koje se utvrđuju polazne hipoteze na osnovu kojih se formira metod. Eksperimentalnim rezultatima pokazuje se da je opisani metod i to korišćenjem 4 različita asinhrona motora kao i 2 različita pretvarača za napajanje koji daju veoma različite spektre izlaznog napona.
Monitoring of temperature of induction motor drives during operation is of great importance for achieving optimal conditions. Machine overheating may result in reduction of the machine lifetime or, in the worst case in destruction of the machine. Existing techniques for measuring temperature of induction machine are based on direct or indirect measuring. Direct measurement of stator temperature can be obtained by using a physical sensor placed in stator slots where direct measurement of rotor is very difficult to achieve. Indirect measurement is based on usage of certain algorithms that need correct thermal or electrical parameters in order to accurately calculate temperature. Usage of machine’s thermal model demands high computing speed and can be very accurate, but is prone to error caused by working conditions that can greatly affect the accuracy of these methods. Temperature measurements using electrical quantities usually use a test signal injection method in order to measure temperature-dependant parameters of a machine, which can further be used to calculate the temperature. Such techniques can be used only in cases where the high performance of drive is not required because test signal injection can generate oscillations in torque. Recent development of digital signal processors has created a new application segment, where response time and accuracy of high speed processors and peripheral units satisfies requirements certain of very demanding applications. This primarily refers to clock frequencies exceeding 300 MHz, and sampling speed of analog to digital converters of over 2 MS per second and conversion resolution of up to 16 bits as well as the inclusion of DMA engines within the peripheral units. Also, increasing number of applications requiring very high bandwidth of regulation current loop, and speed and position loops resulted in design of machines with very low leakage inductance. Power converters used to power these drives have higher frequency switching technology following the advances in silicon components manufacturing. Switching frequency has risen above 10 kHz, which until recently was a rarity. v In this thesis a non-invasive method for measuring rotor temperature without using additional sensors is presented. It relies on the fact that modulated voltages supplied by power converters include a specific component in the spectrum at higher frequencies, due to impuls nature of output voltage of converter. This allows these inherent perturbations, that already exists in the machine, to take place as a test signal. Voltage and current at high frequencies, in order of several kHz to several tens of kHz resulting in much higher impact of skin effect in rotor circuit than in stator. This fact allows that the temperature obtained on the basis of terminal voltages and currents interpret as rotor temperature dominatly. In high frequency range properties of core materials are geting very nonlinear. It is necessary to analyze the electromagnetic phenomena in machine in order to select the frequency range in which certain linearisation can be made so described method could be applied. Temperatre can be obtained by comparing temperature dependant parameters from different temperatures. Comparing the resistance of input impedance at 10 kHz, can be very unreliable due to phase error in sampling voltages and currents. Analysis of electromagnetic phenomena can be concluded that in a frequency area with high skin effect impact, resistance and reactance of rotor circuit becomes equal. This is valid only if frequency is lower than frequency where parasitic capacitance can not be neglected. This allows to calculate temperatures by comparing impedance modules so error in temperature calculation due to phase error can be avoided. Obtaining reliable information of rotor temperature can be used in real time to compensate for rotor resistance as parameter used in control loops or for protection. This thesis contains an extensive analysis of existing literature which leads to initial hypothesis of described method. Experimental results shows that described method can be applied on 4 different type of induction machines and 2 different types of converters with different voltage modulations.
true
Automatska konfiguracija distribuiranih mernih sistema korišćenjem elektronskih specifikacija senzora
Automatic configuration of distributed measurment systems based on transducer electronic data sheets
Savremeni merni sistemi se u novije vreme suočavaju sa sve većim izazovom da zadovolje rastuće potrebe za merenjem različitih fizičkih veličina uz korišćenje velikog broja senzora koji su često distribuirani na širokom geografskom prostoru. S obzirom na to da se za povezivanje senzora i aktuatora koristi veliki broj različitih mrežnih interfejsa i protokola, proizvođači i korisnici moraju uložiti izuzetan napor kako bi obezbedili odgovarajuću podršku za sve aktuelne mrežne interfejse. Sa druge strane, konfiguracija i održavanje složenih mernih sistema koji sadrže različite tipove i sisteme povezivanja senzora predstavlja skup i zahtevan proces, posebno ako se on sprovodi manuelno, imajući u vidu da je unošenje brojnih važnih parametara senzora dugotrajan proces a uz to i podložan greškama. Kao rešenje navedenog problema nameće se upotreba inteligentnih senzora, koji podržavaju Plug and Play koncept, čime bi se obezbedila automatizacija procesa povezivanja, konfiguracije i održavanja senzora. Uvođenjem Plug and Play koncepta za povezivanje senzora i aktuatora ostvarile bi se značajne uštede i omogućila potpuno automatizovana konfiguracija mernih i kontrolnih sistema. Imajući u vidu prednosti koje pruža upotreba mrežnih senzora i primena Plug and Play koncepta u njihovom povezivanju i konfigurisanju, organizacije za standardizaciju prepoznale su važnost ove tehnologije i predložile veći broj standarda u ovoj oblasti. Usvajanje i šira primena odgovarajućih standarda pomogli bi da se prevaziđu mnogi problemi vezani za kompatibilnost i interoperabilnost merne opreme različitih proizvođača. S obzirom da skup standarda IEEE 1451 predstavlja najprihvatljiviji okvir za mrežno povezivanje analognih pretvarača, u ovoj doktorskoj disertaciji detaljno su analizirane mogućnosti proširenja primene ove familije standarda u mrežnom povezivanju specifičnih klasa analognih senzora. Jedan od važnih rezultata predložene doktorske disertacije je doprinos uvođenju Plug and Play koncepta u oblast detekcije i monitoringa jonizujućeg zračenja. S obzirom da današnji sistemi za monitoring jonizujućeg zračenja ne koriste u potpunosti ovaj široko primenjivani koncept, posebna pažnja posvećena je primeni Automatska konfiguracija distribuiranih mernih sistema korišćenjem elektronskih specifikacija senzora iv inteligentnih analognih i kvazidigitalnih detektora jonizujućeg zračenja koji zadovoljavaju IEEE 1451.4 standard i na taj način obezbeđuju automatsku konfiguraciju u monitoring mrežu. IEEE 1451.4 standard definiše grupu standardnih šablona (template) za opis karakteristika senzora koji se najčešće koriste u industriji. Međutim, definisani standardni šabloni ne uključuju detektore jonizujućeg zračenja, što značajno otežava prihvatanje i primenu ovog standarda u mrežama za monitoring zračenja. Stoga je definisanje standardnog načina za specifikovanje i memorisanje karakteristika detektora jonizujućeg zračenja od ključnog značaja za povezivanje i automatsku konfiguraciju monitoring mreža primenom Plug and Play koncepta. Posebna pažnja u disertaciji posvećena je razvoju inteligentnih detektora jonizujućeg zračenja na bazi IEEE 1451.4 standarda, te je za opis karakteristika detektora predloženo više novih originalnih šablona kao i modifikacija postojećih standardnih šablona. Postupak specifikovanja karakteristika i izrade elektronskih specifikacija senzora jonizujućeg zračenja, izložen u ovoj doktorskoj disertaciji, prema uvidu u raspoloživu literaturu, predstavlja prvi pokušaj da se omogući potpuno automatizovano povezivanje senzora jonizujućeg zračenja po Plug and Play konceptu. U okviru disertacije predloženi su novi šablon za smeštanje elektronskih specifikacija detektora jonizujućeg zračenja za GM (Geiger-Müller) brojač i srodne detektore impulsnog tipa i odgovarajući šablon za jonizacionu komoru, koji omogućavaju integraciju ovih detektora u heterogene monitoring mreže. Osim toga, predložen je i šablon za korekcionu tabelu koji omogućava korekciju funkcije prenosa senzora u odnosu na specifikovani parametar. Detalji praktične realizacije inteligentnih detektora, kao i odgovarajućeg akvizicionog sistema, uključeni su u disertaciju. Iako su predložena rešenja ograničena na primenu u monitoringu gama zračenja, uz manje izmene, mogu se primeniti i na monitoring drugih tipova jonizujućeg zračenja.
Modern measurement systems have the important task to meet growing demands for measuring the increasing number of different physical quantities, using a large number of sensors often distributed over a wide area. With the existence of a multitude of incompatible networks and protocols, development efforts required to support variety of sensor interfaces and network protocols are enormous for both sensor producers and users. On the other hand, configuration and maintenance of various sensor types in complex measurement systems, is a demanding and expensive task, especially if it is carried out manually, bearing in mind that writing a number of important parameters of the sensor is a time-consuming process and very susceptible to errors. Solution to this problem appeared in the form of smart sensors that enable Plug and Play concept and provide automation of the process of sensor connecting, configuration and maintenance. Introducing the Plug and Play concept to connect sensors and actuators would be quite useful in measurement and control systems that will lead to self-configuring and automatic system set-up. Numerous benefits offered by networked sensors and application of Plug and Play concept for connecting these sensors are recognized by organizations for standardization that gave proposals for a series of standards. Adoption and wider application of appropriate standards would help to overcome many compatibility issues and to solve problems related to the interoperability of transducers from different vendors. Since the IEEE 1451 standard represents the most appropriate framework for connecting analog transducers in the network, possibilities to extend the application of this family of standards in network connection of specific classes of analog sensors were analyzed in detail in this thesis. One important goal of this thesis is to advance the Plug and Play concept in the field of ionizing radiation monitoring. Unfortunately, today’s radiation monitoring systems use little or no Plug and Play technology at all. Therefore, special attention is paid to the implementation of the analog and quasi-digital radiation sensors compatible to IEEE 1451.4 standard that enable the automatic configuration of the monitoring network. The IEEE 1451.4 defines standard transducer type templates for the Automatska konfiguracija distribuiranih mernih sistema korišćenjem elektronskih specifikacija senzora vi description of commonly used industrial sensors. Among these templates there is no standardized template that can be used for radiation detectors. The fact that standard templates do not include radiation sensors significantly complicates the acceptance and application of IEEE 1451.4 standard in radiation monitoring networks. Therefore, defining a standard way to store data on ionizing radiation detectors is crucial for their connection in the monitoring network enabling automatic configuration based on the Plug and Play concept. Special attention in the thesis is devoted to the development of intelligent detectors of ionizing radiation based on the IEEE 1451.4 standard, and therefore several new original templates as well as the modifications of the existing standard templates for description of the characteristics of the detectors are proposed. Based on the literature available to the author, there has been no attempt to use the Plug and Play concept in the interfacing of radiation detectors with built-in electronic sensor specifications. A new TEDS template suitable for counting types of radiation detectors, such as GM (Geiger-Müller) detectors, and a new template for ionizing chambers have been proposed in the thesis to enable integration and automatic configuration of the monitoring network containing heterogeneous sensors. Also, a new correction table template for the correction of sensor’s measurement transfer function, based on specified parameter, has been proposed. The implementation details of the proposed smart radiation detectors as well as appropriate data acquisition system are included in the thesis. Although the proposed transducer templates are limited to the monitoring of gamma radiation, with minor modifications, they can be applied to other types of ionizing radiation.
true
Zakrivljeni kontinualno nehomogeni i neizotropni konačni elementi višeg reda za velikodomensko elektromagnetsko modelovanje
Curved continuously inhomogeneous and anisotropic higher order finite elements for large-domain electromagnetic modeling
Savremeno društvo se u velikoj meri oslanja na elektronske servise, uređaje i sisteme. Za uspešno razumevanje načina rada postojećih elektronskih uređaja i sistema, kao i prilikom projektovanja novih, neophodno je sprovesti odgovarajuću elektromagnetsku (EM) analizu. U većini slučajeva EM analizu nije moguće sprovesti analitički, a do rešenja se tada mora doći na neki drugi način, najčešće numeričkom elektromagnetikom. Metoda konačnih elemenata (finite element method, FEM) predstavlja jednu od najpopularnijih tehnika numeričke elektromagnetike. Sa razvojem tehnike i tehnologije pred FEM se stalno postavljaju sve izazovniji zahtevi, kao što su: efikasno i tačno modelovanje kontinualno nehomogenih neizotropnih sredina, efikasna prostorna segmentacija sa visokom tačnošću rezultata, efikasno modelovanje otvorenih EM problema i implementacija prilagođena novim računarski sistemima. Kako bi se ostvarili savremeni zahtevi, osnovne formulacije i implementacije FEM metode, poznate od ranije, moraju se stalno unapređivati. U ovoj disertaciji opisano je numeričko modelovanje kontinualno nehomogenih neizotropnih sredina FEM metodom. Tenzori parametara kontinualno nehomogenih neizotropnih sredina dobijeni su primenom teorije transformacione elektromagnetike, a opisana je i teorija transformacione elektromagnetike u vektorskoj i tenzorskoj notaciji prilagođenoj inženjerima. Kako bi se proširio dijapazon problema koji se mogu rešiti FEM metodom, opisano je modelovanje otvorenih EM problema, a posebno su prodiskutovani detalji u vezi sa apsorpcionim graničnim uslovom (absorbing boundary condition, ABC). U cilju postizanja rezultata visoke tačnosti predloženo je novo prostorno mapiranje velikih zakrivljenih konačnih elemenata. Kako bi se bolje iskoristili trenutno raspoloživi računarski resursi, sprovedena je i paralelizacija delova računarskog koda FEM metode. ii Primenom prikazane teorije transformacione elektromagnetike konstruisana je sferna prekrivka za smanjenje radarskog poprečnog preseka (poznata od ranije), a zatim je konstruisana i kockasta prekrivka za smanjenje radarskog poprečnog preseka. Konstruisani su velikodomenski numerički modeli sferne i kockaste prekrivke za smanjenje radarskog poprečnog preseka. Analizom numeričkih rezultata zaključeno je da je velikodomenska FEM metoda višeg reda sposobna da efikasno i tačno modeluje ove prekrivke, kao i da tačno opisuje sve efekte prostiranja elektromagnetskog polja kroz kontinualno nehomogene i neizotropne sredine. Numeričkim modelima prekrivki postignuto je odlično detektovanje svih fenomena smanjenja radarskog poprečnog preseka, kao i odlično poklapanje sa analitičkim (kada postoje) i alternativnim numeričkim rešenjima. U odnosu na alternativna numerička rešenja, prikazani numerički modeli sa višim redom aproksimacije polja i velikim geometrijski fleksibilnim kontinualno nehomogenim konačnim elementima pokazali su se vrlo efikasnim i tačnim. Proširenjem rezultata i zaključaka iz ove disertacije moguće je konstruisati efikasne i tačne numeričke modele EM problema nastalih primenom teorije transformacione elektromagnetike u budućnosti, posebno ako te probleme nije moguće rešiti analitički. U disertaciji je zatim razmatrana mogućnost efikasnog modelovanja otvorenih EM problema FEM metodom. Predstavljen je simetrični (nerigorozno implementiran) ABC drugog reda primenjen na velikodomenske zakrivljene Lagranžove elemente. U okviru prikazanog ABC drugog reda nerigorozno je implementiran proračun divergencije rotorski-konformnih funkcija bazisa. Performanse simetričnog nerigoroznog ABC drugog reda ispitane su kroz nekoliko numeričkih primera i zaključeno je da je tačnost dobijenih rešenja uporediva sa drugim numeričkim i analitičkim rešenjima. U odnosu na ABC prvog reda, nerigorozni ABC drugog reda proširio je skup problema koji se mogu efikasno i tačno modelovati velikodomenskom FEM metodom. Pokazano je da zaključci o neophodnosti rigorozne implementacije simetričnog ABC drugog reda u malodomenskoj FEM-ABC metodi ne važe striktno i u velikodomenskoj FEM-ABC metodi. iii Iako se, posebno u poslednje vreme, velikodomenski način modelovanja intenzivno koristi u numeričkoj EM, određivanje optimalnih geometrijskih parametrizacija skromno je obuhvaćeno u trenutno dostupnoj literaturi. U disertaciji su razmatrane i pojedinosti geometrijskog modelovanja Lagranžovim elementima reda višeg od prvog. Predstavljena je geometrijska parametrizacija konstantnom brzinom i njena aproksimacija Lagranžovim velikim zakrivljenim konačnim elementima. Kroz nekoliko numeričkih primera pokazano je da je geometrijskom parametrizacijom konstantnom brzinom moguće postići značajno tačnije rezultate u odnosu na standardno korišćene geometrijske parametrizacije. U cilju proširenja skupa problema koje je moguće efikasno modelovati FEM metodom na postojećim personalnim računarima, sprovedena je paralelizacija računarski najzahtevnijih delova FEM računarskog koda. Paralelizacija računarskih kodova posebno je aktuelna u numeričkoj elektromagnetici, a sprovedenom paralelizacijom dodata je praktična vrednost postojećoj implementaciji velikodomenske FEM metode.
Modern society is heavily reliant on electronic services, devices and systems. When analyzing functionality of existing electronic devices and systems, as well as when designing a new one, it is necessary to perform adequate electromagnetic (EM) analysis. In most cases, the EM analysis cannot be performed analytically, i.e., in a closed form, and the solution must then be obtained in an alternative manner, usually employing computational electromagnetics. The finite element method (FEM) is one of the most popular computational electromagnetics techniques. With rapid development of technology, FEM is constantly confronted with increasingly challenging requirements such as: efficient and accurate modeling of continuously inhomogeneous and anisotropic media, efficient meshing for high accuracy results, efficient modeling of open EM problems and programming appropriate for the new computer architectures. In order to achieve these new requirements, basic FEM formulations and implementations, used in the past, must be constantly improved. In this dissertation, we propose, present and evaluate numerical modeling of continuous inhomogeneous and anisotropic media by the FEM. The parameters of continuously inhomogeneous and anisotropic media tensors in the presented models are obtained by transformation electromagnetics. In addition, the transformation electromagnetics is presented in a novel fashion, where vector and tensor notations are adapted for engineers. Moreover, to extend the range of problems that can be solved by the FEM technique, modeling of open EM problems is described with special emphasis on the absorbing boundary condition (ABC). In addition, novel spatial mapping of large curved finite elements for achieving high accuracy of results is also proposed and evaluated. Finally, for more efficient use of currently available computing resources, existing FEM computer code is partially parallelized. Utilizing the presented theory of transformation electromagnetics, we first construct the (formerly known) spherical cloak, after which we design a novel cubical cloak. For both the spherical and the cubical cloak we design higher-order large-domain numerical v models, which we evaluate by simulations. From the obtained numerical results, we conclude that the higher-order large-domain FEM technique is fully capable of efficient and accurate modeling of cloaks, as well as of accurate modeling of all the effects of the field propagation through continuously inhomogeneous anisotropic media. The developed models achieved excellent prediction of all phenomena regarding the reduction of the radar cross section using cloaks, as well as excellent agreement with analytical (when available) and other numerical results. Compared to the alternative numerical results, the presented higher-order large domain numerical models with geometrically flexible continuously inhomogeneous finite elements proved to be very efficient and accurate. By extending the results and conclusions from this dissertation, it is possible to construct efficient and accurate numerical models for EM problems arising in application of transformation electromagnetics in the future, particularly if the problems cannot be solved analytically. As a special class of problems, in this dissertation we also address efficient modeling of open EM problems by the FEM. We present a symmetric (non-rigorously implemented) second-order absorbing boundary condition (ABC), applied to large- domain curved Lagrange-type elements. Within the presented second-order ABC, computation of divergence of curl-conforming basis functions is implemented non- rigorously. We validate the accuracy of solutions obtained by employing the symmetric non-rigorous second-order ABC on a few numerical examples. We conclude that it is on a par with other existing numerical and analytical solutions. Thus, utilizing the non- rigorous second-order ABC instead of the first-order ABC, the set of problems that can be efficiently and accurately modeled by the large-domain FEM technique is expanded. It is shown that the conclusions for the necessity for rigorous implementation of the symmetric second-order ABC in small-domain FEM technique does not apply strictly to the large-domain FEM technique. Although, particularly recently, large-domain modeling has been extensively used in computational EM, determining the optimal geometric parametrization is still only modestly covered in the available literature. Additionally, we study details of geometrical modeling by large Lagrange elements with (geometrical) interpolation orders higher than one. Constant-speed parametrization and its approximation by the large curved Lagrange finite elements are presented. In few numerical examples, we vi show that utilization of the constant-speed parametrization can yield significantly higher accuracy of solutions compared to alternative standard parametrizations. In order to expand the set of problems that can be efficiently modeled using existing personal computers (PCs), we implement the parallelization of the computationally most demanding parts of the FEM code. Parallelization of computer codes is a mainstream topic in the computational EM community and the implemented parallelization adds significant practical value to used large-domain FEM code.
true
Manifestations of intrinsic and induced magnetic properties of graphene nanostructures
Испољавање сопствених и индукованих магнетских својстава графенских наноструктура.
Ova doktorska disertacija se može grubo podeliti u dva dela. U prvom delu, ona se bavi magnetskim svojstvima kvazi-nulto dimenzionih grafen- skih struktura, kao što su nanotačke i nanoprstenovi. Konkretno, kružna grafenska kvantna tačka je analizirana u poglav u 3 koristeći Dirak-Vejlovu jednačinu. Energetska stanja i optički apsorpcioni spektar su izračunati za slučaj primenjenog spoljašnjeg magnetnog polja. Rezultati su dobijeni za dva različita slučaja, primenom graničnog uslova beskonačne mase i takoz- vanog cikcak graničnog uslova, koji opisuju strukture sa različitim vrstama ivica. Pronađeno je da energetski spektar tačke za cikcak granični uslov podržava nultoenergetski nivo, bez obzira na vrednost magnetskog polja, dok se za granični uslov beskonačane mase nultoenergetska stanja pojavljuju samo za visoke vrednosti magnetskog polja i to u obliku nultog Landauovog nivoa. Analitički dobijeni rezultati su zatim upoređeni sa rezultatima dobijenim pomoću metoda jake veze, da bi se odredila oblast važenja kontinualnog mod- ela. Pronađeno je da se kontinualni model sa graničnim uslovom beskonačne mase dobro slaže sa metodom jake veze, što se donekle može pripisati mini- mizovanju uticaja ivica, što je posledica naizmeničnog potencijala. Nakon ovoga disertacija se bavi analizom formiranja antiferomagnetizma u heksagonalnim grafenskim kvantnim prstenovima sa cikcak unutrašnjom ivicom koristeći Habardov model u aproksimaciji srednjeg polja. Spoljna ivica prstena je ili cikcak ili foteljastog tipa. Pronađeno je da obe vrste struktura pokazuju veću sklonost ka antiferomagnetizmu nego heksagonalne kvantne tačke. Ova razlika se delimično može pripisati većoj dužini cik- cak ivica u prstenovima nego u tačkama. Pored toga, hibridizacija ivič- nih stanja lokalizovanih na unutrašnjoj strani prstena je drugačije prirode nego hibridizacija stanja na ivici tačaka, što dovodi do bitnih razlika mag- netizma prstenova u odnosu na tačke. Najveća magnetizacija je dobijena za slučaj kada je i spoljašnja ivica cikcak tipa. Ipak, vrlo uski prstenovi sa...
The dissertation can be roughly divided into two parts. In the first part it deals with magnetic properties of quasi-zero dimensional graphene structures, such as nanodots and nanorings. In particular, a circular graphene quantum dot is analyzed in Chapter 3 using the Dirac-Weyl equation. The energy and the optical absorption spectra are computed for the case of the present external magnetic field. The results are obtained for two distinct boundary conditions, namely infinite-mass and zigzag boundary conditions, which model different physics in the structure with different edges. It is found that the energy spectrum of a dot with zigzag boundary condition exhibits a zero energy band regardless of the value of the magnetic field, while for the infinite mass boundary condition the zero energy states appear only for high magnetic fields in the form of the zeroth Landau level. The analytical results are compared to those obtained from the tight-binding model in order to show the validity range of the continuum model. It is found that the continuum model with infinite mass boundary condition describes rather well its tight binding counterpart, which can partially be attributed to blurring of the mixed edges by the staggered potential. The mean-field Hubbard model is subsequently used to investigate the formation of the antiferromagnetic phase in hexagonal graphene quantum rings with inner zigzag edges. The outer edge of the ring is taken to be either zigzag or armchair, and it is found that both types of structures exhibit a larger antiferromagnetic interaction than hexagonal quantum dots. This difference could be partially ascribed to the larger number of zigzag edges per unit area in rings than in dots. Furthermore, edge states localized on the inner ring edge are found to hybridize differently than the edge states of dots, which results in important differences in the magnetism of graphene rings and dots. The largest staggered magnetization is found when the outer edge has a zigzag shape. However, narrow rings with armchair outer edge are found to have larger staggered magnetization than zigzag hexagons. The edge defects are shown to have the least effect on magnetization when the outer ring edge is armchair shaped.
false
Detekcija promena kadrova u video sekvenci primenom multifraktalne analize
Shot boundary detection in video sequence using multifractal analysis
Moderan i efikasan menadžment video sadržaja podrazumeva mogućnost pretraživanja videa na osnovu automatski generisanih oznaka i informacija dobijenih analizom sadržaja frejmova. Vremenskom segmentacijom videa generiše se hijerarhijska struktura video sadržaja, koja predstavlja osnovu za automatsku anotaciju videa. Prvi korak vremenske segmentacije predstavlja određivanje granica kadrova kao najniže hijerarhijske logičke strukture videa. Kadar predstavlja jedno kontinuirano snimanje kamerom između uključivanja i isključivanja kamere. Grupe kadrova kreiraju više hijejarhijske logičke strukture videa kao sto su: scene, sekvence i programi. Pri detekciji granica kadrova razlikuju se dva osnovna tipa tranzicije izmedju kadrova - nagle i gradijentne promene kadrova. Naglim promenama kadrova nazivaju se tranzicije između kadrova koje traju samo dva frejma, dok kod gradijentnih tranzicija prelazak između dva kadra traje više frejmova. Tema ove disertacije je realizacija algoritma za efikasnu detekciju naglih promena kadrova u video sekvencama različitog sadržaja, dinamike i nivoa postprodukcije. U tu svrhu, sadržaj frejmova video sekvenci opisan je obeležjima za boju i teksturu. Korišćeni su histogram boje i vejvlet obeležje kao elementarna obeležja izdvojena iz frejmova, na lokalnom i globalnom nivou. Odabrana obeležja obezbeđuju sa jedne strane dovoljnu osetljivost na nagle promene vizuelnog sadržaja na frejmovima. Sa druge strane, odabrana obeležja pokazuju određeni nivo tolerancije na promene sadržaja niskog intenziteta, kao što su: kretanje objekata, kamere i promene osvetljenja. Na obeležja frejmova primenjene su dve različite diferencijalne metrike, korelacija i multifraktalna analiza. Korelacija, kao diferencijalna metrika, kombinovanim pristupom je omogućila formiranje specifičnih artefakata koje je jednostavno detektovati u strukturi signala. Artefakti zbog dinamike promena unutar kadrova mogu biti degradirani i pomereni u odnosu na poziciju stvarne nagle promene kadra. Za lokalizaciju nagle promene sadržaja u vremenu primenjena je multifraktalna analiza, tako što je iskorišćena sposobnost ove tehnike da dobro opiše lokalnu strukturu signala. U domenu grubih Hölder-ovih eksponenata multifraktalne analize, nagla promena signala se prikazuje kao lokalni maksimum. Kombinacijom ove dve metrike (ponovi metrike) detektuju se pozicije kandidata naglih promena kadrova. Klasifikacija kandidata izvršena je primenom klasifikatora baziranih na fiksnim pragovima i primenom dobro definisanih pravila. Kao dodatna mera u cilju smanjenja osetljivosti sistema na promene osvetljenja, u algoritam je implementiran i flesh detektor. Realizovani algoritam za detekciju naglih promena kadrova testiran je na velikom broju frejmova (>1.500.000 frejmova), video sekvenci različitih sadržaja, žanrova i tehnologija proizvodnje. Algoritam je pokazao visoke performanse, odnosno tačnost i preciznost, među kojima se posebno ističe tačnost. Pokazalo se da preciznost algoritma zavisi od klase video sekvenci nad kojom je obavljeno testiranje. U disertaciji je prikazano poređenje realizovanog algoritma sa algoritmima za detekciju naglih promena kadrova visokih performansi iz literature. Poređenja su vršena na osnovu tačnosti, preciznosti i F-mere. Rezultati poređenja su pokazali da je predloženi algoritam porediv po performansama sa ostalim algoritmima, a za neke video sekvence pokazao je bolje rezultate od ostalih. Prikazani rezultati ukazali su na velike mogućnosti predloženog algoritma u pogledu daljeg razvoja.
Modern and efficient management of video content implies the possibility of video searching on the basis of automatically generated labels and information gained by the analysis of frame content. Temporal video segmentation generates hierarchical structure of video. This structure represents the basis for the automated annotation of video content. The first step of temporal segmentation is the determination of shot boundaries, as the lowest hierarchical logical structure of video. A shot refers to a continual recording with video camera between switch-on and switch-off of camera. Groups of shots create higher hierarchical logical structures of video, such as: scenes, sequences and programs. In the shot boundary detection there are two basic types of transitions between shots – abrupt and gradual shot changes. Abrupt shot changes refer to those transitions between shots which last only two frames, whereas in gradual shot changes the transitions between two shots last several frames. The topic of this thesis is the experimental realization of algorithm for efficient abrupt shot changes detection in video sequences of different contents, dynamics and postproduction levels. For this purpose, the frame content of video sequences is described with color and texture features. Color histogram and wavelet feature are used as basic features extracted from frames, on both local and global levels. On the one hand, the selected features provide sufficient sensibility regarding sudden changes of visual content on frames. On the other hand, the selected features show certain level of tolerance for the low intensity content changes, such as: moving of objects, camera and light changes. Correlation and multifractal analysis have been used on frame features as differential metrics. With combined approach, correlation as a differential metrics applied on frames features, has enabled the formation of specific artefacts which can be easily detected in signal structure. Due to the change dynamics within shots, artefacts can be degraded and shifted regarding the position of real abrupt shot changes. Multifractal analysis has been used for temporal localization of abrupt content changes due to its ability to provide a detailed description of signal local structure. In the domain of rough Hölder multifractal analysis exponents, abrupt signal change is shown as a local maximum. By combining these two metrics, correlation and multifractal analysis, the positions of abrupt shot change candidates are detected. Candidate classification has been carried out by means of classifiers based on fixed thresholds and predefined rules. As an additional precaution aimed at reducing system sensibility related to light changes, flash detector has been implemented into the algorithm. The implemented algorithm for the abrupt shot change detection has been tested on a large number of frames (>1.500.000 frames), video sequences of different contents, genres and production technologies. The algorithm has shown high performances, i.e. recall and precision, especially recall. It has been shown that precision of algorithm depends on the class of video sequence being tested. In this thesis, the comparisons between the proposed algorithm and the high- performance algorithms for the detection of abrupt shot changes in literature have been presented. Comparisons have been carried out on the basis of recall, precision and F- measure. The findings have shown that proposed algorithm is comparable by performances with other algorithms, showing even better results for some video sequences. The achieved results indicate great possibilities of the proposed algorithm for further development.
true
Procena raspoloživosti složenih komunikacionih mreža metodom linearne segmentne aproksimacije
Availability assesment of complex communication networks by method of linear segment aproximation
Razvoj modernog društva se u velikoj meri zasniva na kvalitetnim telekomunikacionim mrežama. U tom smislu, od velike je važnosti da telekomunikacione mreže imaju visoku raspoloživost čime se postiže i visoka raspoloživost servisa koje koriste krajnji korisnici. Danas se okosnice mreža najčešće realizuju sa velikim brojem čvorova (više desetina) zbog čega je problem direktno izračunati raspoloživost takve mreže, jer vremena izvršavanja algoritama za tačno računanje raspoloživosti eksponencijalno rastu sa porastom broja čvorova i linkova. Zbog toga je od velike praktične važnosti korišćenje aproksimativnih algoritama koji tokom procesa projektovanja takvih mreža omogućavaju definisanje početnih uslova za dizajniranje topologije mreže. Cilj ove doktorske disertacije je uvođenje novog metoda za aproksimativno računanje maksimalne all-terminal raspoloživosti složenih telekomunikacionih mreža. U cilju dobijanja vrednosti maksimalne all-terminal raspoloživosti, razvijena je metoda sinteze mreže. Kod metode sinteze mreže početna topologija je lanac. Kod određivanja maksimalne raspoloživosti mreže, u prvom koraku sinteze se formira prsten da bi se dalje dodavao po jedan link u mrežnoj topologiji tako da se svaki put dobije maksimalni priraštaj all-terminal raspoloživosti mreže. Dodavanjem jednog po jednog linka, gde svi linkovi imaju istu raspoloživost, vrši se sinteza mreže koja u svakom koraku ima maksimalnu raspoloživost za dati broj čvorova, linkova i datu raspoloživost linkova. Analizom dobijenih rezultata za maksimalnu all-terminal raspoloživost mreže definisane su funkcije aproksimacije maksimalne all-terminal raspoloživosti mreže za zadati broj čvorova, linkova i raspoloživost linkova. Na osnovu dobijenih aproksimativnih izraza može se izračunati granica raspoloživosti mreže u zavisnosti od topologije za zadate parametre. S druge strane, korišćenjem funkcija aproksimacija može se definisati donja granica za broj linkova da bi se mogla dobiti željena VI raspoloživost izborom odgovarajuće topologije mreže. Pokazano je da je za maksimalnu raspoloživost mreže potrebno da svi čvorovi imaju istu konektivnost ili konektivnost koja se razlikuje najviše za jedan. Polazeći od ovog pravila, i znajući prethodno određene granice za broj linkova u mreži, u procesu projektovanja može se u nekoliko iteracija doći do topologije mreže koja obezbeđuje zadatu all-terminal raspoloživost. Vreme izvršavanja funckija aproksimacije definisanih u okviru ove doktorske disertacije ne zavisi od veličine mreže. To omogućava veliki broj izračunavanja čime se omogućava analiza velikog broja kombinacija ulaznih parametara. Pokazano je da funkcije aproksimacije definisane u okviru disertacije imaju veću tačnost od funkcija koje su do sada publikovane u stručnoj literaturi. Izračunate aproksimativne vrednosti su upoređene i sa tačnim vrednostima raspoloživosti i pokazano je da je relativna greška aproksimacije značajno manja od 1% za mreže sa linkovima čija je raspoloživost veća od 0,75.
The development of modern society is largely based on high-quality telecommunications networks. In this regard, it is of great importance to telecommunication networks have high availability which achieves high availability of services used by end users. Today, the backbone networks are usually implemented with a large number of nodes (several dozen) why there is a problem to directly calculate the availability of such networks, since the execution time of algorithms for exact calculation of availability increase exponentially with the number of nodes and links. Therefore, it is of great practical importance of the use of approximation algorithms during the design of such networks to help in the definition of initial conditions for the design of the network topology. The aim of this PhD thesis is the introduction of a new method for approximate calculation of the maximum availability of complex telecommunications networks. In order to obtain the maximum values of the availability, it was developed a method of network synthesis. Starting network topology for this new method is a chain. In determining the maximum network availability, the first step of the synthesis is to form the ring and then continue adding one-by-one links in the network topology so that it delivers a maximum increase of all-terminal availability. By adding one link at a time, where all links have the same availability, perform the synthesis of the network in which each step has a maximum availability for a given number of nodes, links, and the availability of links. The approximation functions of the maximum network availability for a given number of nodes, links and the availability of links are obtained by the analysis of the results of network synthesis method for the networks with maximum availability. Limits of the availability of the network (which depends of the network topology) can be calculated based on the approximation functions. In addition, based on the approximation functions can be defined the boundaries in which they need to move a number of links to be able to get the desired availability of appropriate choice of the network topology. It VIII has been shown that for maximum availability of the network, all nodes should have the same connectivity, or differs by one at most. Starting from this rule and knowing predetermined limit on the number of links in the network, the network design process can be done in a few iterations to get to the network topology that ensures the requested availability. Running time, of approximation functions defined within this dissertation, does not depend on network size. This allows a large number of calculations which allows analysis of a large number of combinations of input parameters. It is shown that approximation functions, defined within the thesis, are more accurate than the functions that have been published in the literature. Calculated approximate values are compared with the exact values of availability and it is shown that the relative error of approximation is significantly less than 1% for the networks whose links have the availability greater than 0,75.
true
Digitalna zaštita energetskih transformatora od unutrašnjih kvarova
Digital protection of power transformers against internal faults
U ovoj doktorskoj disertaciji predstavljeni su algoritmi za digitalnu zaštitu energetskih transformatora (ET-a) od unutrašnjih kvarova. Predstavljeni algoritmi za zaštitu od kratkih spojeva namotaja ET-a i za zaštitu ET-a od zemljospojeva baziraju se na primeni digitalne fazne komparacije u vremenskom domenu. Digitalni fazni komparator ne računa direktno fazni pomeraj između signala koje poredi, već proračunava integral proizvoda dva signala na intervalu dužine polovine osnovne periode signala. Ispitana je mogućnost primene fazne komparacije primarne i sekundarne struje za detekciju unutrašnjih kratkih spojeva u ET-u. U normalnom režimu rada i tokom kvarova van zone zaštite struje primara i sekundara su u fazi, dok se u slučaju unutrašnjih kratkih spojeva pojavljuje fazni pomeraj između ove dve struje. U disertaciji je pokazano da fazni pomeraj između struja primara i sekundara tokom unutrašnjih kratkih spojeva zavisi kako od procenta kratkospojenih navojaka, tako i od parametara mreže i radnog režima koji je prethodio kvaru. Korišćenjem fazne komparacije primarne i sekundarne struje mogu se detektovati unutrašnji kratki spojevi. Zbog postojanja opadajuće jednosmerne komponente u struji kvara algoritam je proširen digitalnim filterom opadajuće jednosmerne komponente, a zbog postojanja viših harmonika digitalnim Fourier-ovim filterom. Na ovaj način poboljšane su performanse algoritma. Zatim je predstavljen algoritam koji koristi novi pristup u rešavanju problema nepotrebnog reagovanja diferencijalne zaštite ET-a usled kvarova van štićene zone praćenih zasićenjem strujnih transformatora (ST-a). Novi pristup se zasniva na faznoj komparaciji primarne i sekundarne struje i služi za indikaciju prisustva kvara van štićene zone. Prikazani rezultati potvrđuju da algoritam može da identifikuje prisustvo spoljašnjeg kvara i aktivira signal za blokadu pre nego što ST-i uđu u zasićenje. Njegovom primenom uspešno bi se moglo eliminisati nepotrebno delovanje diferencijalne zaštite i na taj način povećala njena sigurnost u radu. U nastavku je predstavljen algoritam za zemljospojnu zaštitu energetskog transformatora koji se bazira na faznoj komparaciji sume struja na priključcima ET-a i struje neutralnog provodnika. Prikazani algoritam predstavlja novo rešenje za realizaciju zemljospojne zaštite energetskog transformatora. Primenom predstavljenog algoritma eliminisalo bi se nepotrebno reagovanje konvencionalne zemljospojne zaštite ET-a bazirane na diferencijalnom principu tokom spoljašnjih kvarova i uključenja neopterećenog ET-a praćenih zasićenjem ST-a. Algoritam koristi signale koji se ne filtriraju digitalnom Fourier-ovom transformacijom. Takođe, algoritam ne koristi blokadu po drugom harmoniku što ima prednost u brzini formiranja signala reagovanja. Prikazano je da se dodatnim usrednjavanjem vrednosti indeksa usmerenosti povećava otpornost algoritma na oscilacije indeksa koje nastaju usled zasićenja ST-a. Svi algoritmi koji su predstavljeni u ovoj doktorskoj disertaciji su testirani pomoću signala koji su snimljeni u laboratoriji na fizičkim modelima jednofaznog i trofaznog energetskog transformatora i pomoću kompjuterski generisanih signala. Prednost algoritama, koji su predstavljeni u ovoj doktorskoj disertaciji, za primenu u zaštiti energetskog transformatora čini njihova jednostavnost, kratko vreme odziva, pouzdanost i sigurnost u radu. Predstavljeni algoritmi se lako mogu implementirati u već postojeće jedinice za zaštitu energetskog transformatora. Zaštita energetskog transformatora najčešće obuhvata nekoliko jedinica, među kojima su diferencijalni releji i zemljospojni releji. Za rad pomenutih jedinica vrši se merenje struja na priključcima transformatora i struje neutralnog provodnika. Tako da su u već postojećim jedinicama za zaštitu ET-a, signali potrebni za rad predloženih algoritma obezbeđeni. Na osnovu razvijenih algoritama za zaštitu energetskog transformatora kreiran je softverski alat pomoću koga su korisniku slikovito i jasno približeni principi rada tradicionalne diferencijalne zaštite i zemljospojne zaštite ET-a, i predstavljen je rad algoritama za zaštitu koji se baziraju na faznoj komparaciji u vremenskom domenu. Softverski alat omogućava da se algoritmi za zaštitu testiraju učitavanjem realnih signala koji su snimljeni u laboratoriji ili korišćenjem signala koji su generisani kompjuterskim simulacijama. Kreirani softverski alat se može koristiti u edukativne svrhe.
Algorithms for digital protection of power transformers (PT) against internal faults have been presented in this doctoral dissertation. The presented algorithms for protection against short circuits of PT windings and for protection against ground-faults are based on time domain digital phase comparison. The digital phase comparator does not calculate the phase displacement between the signals which are being compared, but calculates the integral of the product of two signals over one basic half-period of the signals. The possibility of applying the phase comparison of the primary and secondary currents aimed at detecting internal faults in PTs has been investigated. In normal operating regimes and during faults outside the protection zone, the primary and secondary currents are in phase, while in case of internal short-circuit faults a phase shift appears between these two currents. In this dissertation, it has been shown that the phase shift between the primary and secondary currents during internal short-circuits depends both on the percentage of the short circuited windings and the network parameters as well as operating regime prior to the fault. It is possible to detect internal short-circuit faults by utilizing the phase comparison of the primary and secondary currents. The algorithm has been expanded by adding a digital filter for the dc component due to the existence of a decreasing dc current component in the fault current; in addition to this, the algorithm has been extended by adding the digital Fourier filter due to the existence of higher harmonics. In this way, the algorithm performance has been enhanced. Next, the algorithm that uses a new approach to solve the problem of misoperation of the transformer differential protection due to faults outside the protected zone accompanied by current transformer (CT) saturation has been presented. The new approach is based on the phase comparison of the primary and secondary currents, and is used as indication of an external fault presence. The presented results confirm that the algorithm can identify an external fault presence and activates the blocking signal before CTs reach saturation. Its application would successfully eliminate an unnecessary operation of the differential protection and so increase its security. In the part that follows, an algorithm for restricted earth fault protection of power transformers has been presented based on the phase comparison of the sum of phase currents and the neutral conductor current. The presented algorithm represents a new solution for realization of restricted earth fault protection of power transformers. By applying this algorithm one would eliminate false tripping of the conventional restricted earth fault protection of PTs based on the differential principle during external faults or energization of an unladen PT accompanied by CT saturation. The algorithm utilizes the signals that are not filtered by means of the digital Fourier transform. Moreover, the new method does not make use of second harmonic blockade which results in a high speed of making trip decision. It has been shown that an additional averaging of the directional index values increases robustness of the algorithm to oscillations of the directional index caused by CT saturation. All algorithms presented in the thesis have been tested by the laboratory recorded signals using physical models of a single-phase or three-phase PT and also by computer generated signals. The advantage of the algorithms, which have been presented in this doctoral dissertation, in the domain of their application in PT protection lies in their simplicity, short response time, reliability and security. The presented algorithms can easily be implemented in already existing PT protection units. The PT protection usually requires several units, these being differential relays and ground fault relays. The phase currents at transformer terminals and current of the neutral conductor are measured with the aim of enabling proper function of these units. In such a way, the already existing PT protection units require signals necessary for the proposed algorithms’ functionality. On the basis of the developed PT protection algorithms a software tool has been created with the help of which the principles of operation of the traditional differential protection and restricted earth fault protection of power transformers have been brought closer to an end user in order to understand them visually and clearly. In addition to this, the performance of new protection algorithms that are based on time domain phase comparison has been presented. The software tool enables the protection algorithms to undergo testing by loading the genuine signals, recorded in the laboratory, or by means of computer simulation generated signals. The created software tool can also be used for educational purposes.
true
Spektralne karakteristike kvantnih crta u srednjoj infracrvenoj oblasti
Spectral properties of mid-infrared quantum dashes
U mnogim primenama fotonskih sistema od interesa su komponente koje funkcionišu u oblasti srednjeg i dalekog infracrvenog zračenja (3 – 30 μm). Neke od atraktivnih aplikacija vezane su za termalnu vizuelizaciju, spektroskopiju, astronomiju, gasne i bio-senzore, nedestruktivnu evaluaciju, industrijski i ekološki monitoring, medicinsku dijagnostiku, vojne primene, komunikaciju u slobodnom prostoru i slično. Ove primene su mahom omogućene postojanjem transparentnih atmosferskih prozora (3–5 μm i 8–13 μm) u srednjoj infracrvenoj oblasti, kao i činjenicom da veliki broj molekula u ovoj oblasti ima izražene karakteristične vibracione prelaze. Danas, kvantno-kaskadni laseri predstavljaju široko prihvaćeno rešenje za realizaciju koherentnih poluprovodničkih izvora zračenja za srednju infracrvenu i terahercnu spektralnu oblast. Međutim, još uvek postoji prostor za unapređenje njihovih performansi u pogledu struje praga i efikasnosti konverzije snage. Nedavno sprovedena istraživanja potvrdila su da uređaji za srednju-infracrvenu oblast koji koriste poluprovodničke nanostrukture u kojima je kretanje elektrona ograničeno u više od jednog pravca, mogu prevazići neka od ograničenja koja se javljaju u kvantnim jamama, prvenstveno kratko vreme života eksitovanih stanja koje je posledica izraženog rasejanja naelektrisanja u ravni jame. Samo-organizovane nanostrukture, kao što su trodimenzionalno konfinirane kvantne tačke, ili njihova izdužena verzija sa izraženim dvodimenzionalnim konfiniranjem, nazvana kvantne crte, obezbeđuju dodatno konfiniranje koje omogućava smanjenje rasejanja i poboljšanje jačine optičkih prelaza. Iz tog razloga, istraživanje elektronske strukture ovih nanostruktura predstavlja osnovu za dalje razumevanje problematike i optimizaciju procesa rasejanja nosilaca, njihovog vremena života, jačine optičkih prelaza i tipa polarizacije. Kvantne crte koje se fabrikuju mehanizmima samo-organizovanog narastanja, prvobitno su korišćene za realizaciju lasera i optičkih pojačavača za telekomunikacione talasne dužine u okolini 1.55 μm. Aktivni materijal na bazi kvantnih crta omogućio je ostvarivanje novih osobina, koje uređaji sa konvencionalnim dizajnom nisu mogli da ponude. Primera radi, realizovane su aktivne oblasti za širokopojasne lasere sa širinom spektra pojačanja od preko 300 nm za emisiju u opsegu od 1.4 do 1.65 μm. Međutim, i pored intenzivnih eksperimentalnih istraživanja usmerenih ka eksploataciji međuzonskih prelaza u kvantnim crtama, unutarzonski prelazi pogodni za emisiju u srednjoj i dalekoj infracrvenoj oblasti, do nedavno nisu privlačili značajnu pažnju. Ova disertacija daje opštu sliku spektralnih osobina kvantnih crta u srednjoj infracrvenoj oblasti, sa posebnim fokusom na razmatranje kako na spektar utiču koncentracija elektrona, fluktuacije veličina kvantnih crta u ansamblu i debljina vlažećeg sloja. Pre nego što pristupi razmatranju detalja matematičkih formulacija koje leže u osnovi fizičkog modela unutarzonske apsorpcije, disertacija daje pregled osobina kvantnih crta kroz odgovore na neka jednostavna pitanja: šta je kvantna crta, kako se formiraju ansambli kvantnih crta, koje su njihove osnovne geometrijske i optičke karakteristike i konačno, kako se kvantne crte mogu koristiti u poluprovodničkim fotonskim komponentama. Nakon ovih uvodnih poglavlja, izloženi su detalji teorijskog modela za izračunavanje zonske strukture i unutarzonske apsorpcije, kao i metoda korišćenih za numerička izračunavanja. Elektronska struktura izolovane kvantne crte se određuje rešavanjem jednozonske Schrödinger-ove jednačine u aproksimaciji anvelopne funkcije. Zonska struktura se određuje samosaglasno, u paraboličnoj aproksimaciji efektivne mase, a numerička implementacija je izvedena primenom metode konačnih elemenata. Pored toga, u disertaciji je predložen još jedan efikasan i jednostavan metod za određivanje elektronske strukture. Predloženi metod kombinuje koordinatnu transformaciju geometrije kvantne crte i barijernog materijala koji je okružuje u pravougaoni domen izračunavanja i rešavanje jednozonske Schrödinger-ove jednačine na pravougaonom domenu primenom metode konačnih razlika. Prilagodljivost ovog metoda potiče od velikog broja familija funkcija koje se mogu koristiti za fitovanje heterogranica kvantnih crta i definisanje domena izračunavanja pogodnog oblika i veličine. Za predložene koordinatne transformacije analizirana je primenljivost, robusnost i konvergencija metode promenom veličine diskretizacionog koraka i veličine domena izračunavanja. Metod obezbeđuje zadovoljavajuću preciznost, stabilnost i fleksibilnost u pogledu veličina i oblika kvantnih crta, a pre svega, obezbeđuje izuzetnu jednostavnost, što ga čini potencijalno pogodnim za rešavanje višezonskih Schrödingerovih jednačina. Izborom odgovarajuće periodične funkcija koja fituje heterogranicu kvantne crte, ovaj metod se može koristiti i za određivanje zonske strukture periodičnog niza kvantnih crta i njene karakterizacije u zavisnosti od relevantnih geometrijskih parametara. U nastavku, mini-zonska struktura InAs/GaAs niza kvantnih crta određena je u aproksimaciji koja ne uzima u obzir dopiranje barijere i uticaj konfiniranih elektrona na strukturu provodne zone. Izračunavanja su pokazala da je energija mini-zona jako zavisna od visine i širine kvantne crte, dok širina mini-zona zavisi od periode niza i širine kvantne crte. Za velike vrednosti perioda i određene kritične dimenzije kvantnih crta, između susednih pobuđenih mini-zona može doći do pojave nultog-energetskogprocepa i višestrukih anti-ukrštanja koja su praćena promenom karaktera talasnih funkcija vrha i dna susednih mini-zona. Kao posledica prisustva vlažećeg sloja, više mini-zone su grupisane u okolini vrha jame i formiraju neku vrstu “vezanog kontinuuma” koji može biti od značaja za kontrolu procesa zahvata i otpuštanja eksitovanih nosilaca. Međutim, u narednim razmatranjima, pretpostavlja se da se minizone, odgovorne za unutarzonske prelaze onda kada određivanje zonske strukture uzima u obzir uticaj jonizovanih donora i odgovarajućih elektrona, mogu aproksimirati diskretnim energetskim nivoima, te da svojim prisustvom ne utiču na perturbaciju izračunatog Hartree potencijala. Polazeći od zonske strukture, usrednjeni apsorpcioni spektar ansambla kvantnih crta izračunava se u dipolnoj aproksimaciji za polarizaciju upadne svetlosti u pravcu izraženijeg i u pravcu slabijeg konfiniranja u kvantnoj crti, uz pretpostavku da se fluktuacije dimenzija kvantnih crta u ansamblu mogu modelovati dvodimenzionalnom Gauss-ovom raspodelom. Izračunati apsorpcioni spektar je izražen za polarizaciju upadne svetlosti u pravcu slabijeg konfiniranja i u zavisnosti od materijalnog sistema i dimenzija kvantnih crta u ansamblu zauzima opseg talasnih dužina od 4 do 16 μm, što pripada srednjoj infracrvenoj oblasti. Sa povećanjem fluktuacija u dimenzijama, spektar ansambla postaje širi, zaravnjen i sve više asimetričan. Kada se uzme u obzir prisustvo vlažećeg sloja, koncentracija elektrona značajno utiče na apsorpciju, a spektar se može podešavati promenom debljine bafer slojeva. Povećanje debljine vlažećeg sloja dovodi do smanjenja apsorpcije i crvenog pomaka spektra, što je posebno izraženo za ansamble sa manjim fluktuacijama dimenzija. Konačno, pažljivim podešavanjem debljine bafer slojeva može se postići optimizacija spektra u smislu ostvarivanja maksimalne apsorpcije. U zavisnosti od materijalnog sistema, nominalne debljine InAs sloja i parametara narastanja, samo-organizovane kvantne crte mogu imati različite profile poprečnog preseka. U cilju ravnopravnog poređenja uticaja profila poprečnog preseka na apsorpcioni spektar, definisani su kriterijumi ekvivalencije uobičajenih geometrijskih oblika koji se koriste za aproksimaciju poprečnog preseka kvantnih crta (sočivasti, trapezni, trougaoni i pravougaoni oblik). Bez obzira na fluktuaciju u dimenzijama, spektri ansambla sa sočivastim i ekvivalentnim trapeznim poprečnim presekom su slični. Primena pravougaone aproksimacije sočivastog oblika je opravdana samo za relativno male fluktuacije u dimenzijama. Apsorpcioni spektri za ansamble sa ekvivalentnim trougaonim profilom su značajno crveno pomereni i sa nešto većom maksimalnom apsorpcijom nego za ansamble sa sočivastim profilom kvantnih crta. Konačno, može se zaključiti da uticaji debljine vlažećeg sloja, distribucije po veličinama i koncentracije elektrona u jami određene debljinom bafer sloja na spektar unutarzonske apsorpcije, pokazuju zanimljive trendove koje nije moguće intuitivno pretpostaviti. Iz tog razloga, rezultati prezentovani u okviru ove disertacije mogu biti od koristi u eksperimentalnim istraživanjima, za optimizaciju spektralnih karakteristika kvantnih crta sa primenom u novim poluprovodničkim uređajima za srednju infracrvenu spektralnu oblast.
Mid- and far-infrared regions (3 – 30 μm) are very attractive for many photonics applications such as thermal imaging, spectroscopy, astronomy, gas and biosensing, nondestructive evaluation, industrial and ecology monitoring, medical diagnostics, military applications, free-space communications, and etc. This is mainly due to the high transparency atmospheric windows (3–5 μm and 8–13 μm) in which the Earth’s atmosphere is relatively transparent, and the fact that a large number of molecules undergo strong characteristic vibrational transitions in this spectral range. Now-a-days quantum cascade lasers are well-established coherent sources of radiation in the midinfrared and terahertz regions. Even though in the mid-IR these devices work extremely well, they could still benefit from improved threshold currents and higher wall plug efficiencies. Recent studies have shown that mid-infrared photonic devices incorporating semiconductor nanostructures in which the electronic motion is confined in more than one spatial direction, may overcome some of the limitations of quantum well based devices, especially short excited-state carrier lifetime, caused by numerous in-plane scattering paths. Self assembled nanostructures, as three-dimensionally confined quantum dots or their elongated version with two dimensional confinement, called quantum dashes (QDHs), may provide the additional confinement required for reduction of scattering events or improvement of optical transitions strength and polarization. Therefore, investigation of their electronic structure is a basis for further understanding and optimization with respect to carrier scattering processes and its lifetime, optical transitions strength and polarization type. Quantum-dash structures, fabricated by self assembly growth techniques, were initially used to realize long wavelength lasers and amplifiers for telecom applications. This new type of laser material allows to achieve new device features not possible by conventional device designs. For example, broadband laser material with a gain bandwidth of more than 300 nm could be obtained to cover the extended telecommunication wavelength range between 1.4 and 1.65 μm. In spite of intensive experimental investigation directed toward employing interband transitions in quantum dashes, intersubband transitions suitable for mid- and far-infrared emission, until recently, did not attract attention at all. This thesis delivers an overall picture of the mid-infrared spectral properties of quantum dash nanostructures, while focusing in particular on its dependence on electron concentration, the size fluctuation of QDHs in ensemble, and thicknesses of the wetting layer. Before going into all the details of the mathematical equations underlying the physical model of intersuband absorption in quantum dashes, the thesis set the stage by providing an overview of some general answers on questions regarding the quantum dashes: What is a quantum dash, how quantum dashes form an ensemble, what are their basic geometrical and optical properties, and finally, how we can use them in semiconductor photonic components. After these introductory chapters, details of theoretical model of band structure and intersubband absorption, as well as method for their numerical calculations, are presented. The electronic band structure of a single QDHs is modeled by using the single-band Schrödinger equation in the envelope function approximation. The band structure calculation is performed self-consistently in the parabolic approximation by using the finite elements method. Additionally, a newly developed, efficient and simple method for the band structure and wavefunction calculation of quantum dashes is presented. The method is a combination of the coordinate transformation of the QDH geometry and its vicinity into rectangular computational domain and the single-band Schrödinger equation solved in this domain by the finite differences method. The method versatility comes from a broad set of function families which can fit the QDH heterointerface and generate the computational domain of the proper form and size. For the proposed coordinate transformations, applicability, robustness and convergence of the method are investigated and analyzed by varying the FDM grid density and size of the computational domain. The method provides sufficient accuracy, stability and flexibility with respect to the size and shape of the quantum dash and above all, extreme simplicity, which is promising and essential for an extension of the method to the multiband Schrödinger equation case. By choosing appropriate periodic functions this method can be used to investigate the band structure of a quantum dash array and its dependence on geometrical parameters. The miniband structure of the InAs/GaAs quantum dash array is calculated without taking into account the influence of the barrier doping and confined electrons on the conduction subband structure. The calculation shows that miniband energy significantly depends on the dash height and width, while the miniband width depends on the array period and the dash width. For large periods and some critical dimensions of QDHs, adjacent excited minibands may exhibit the effect of zero miniband gap and the multiple anticrossings, which are followed by the swapping of the wavefunction character of adjacent minibands top and bottom. Due to the wetting layer, higher minibands are clustered in the vicinity of the well top, providing a kind of the “bound continuum,” which might be relevant for control of capture and relaxation of excited carriers. However, it is assumed that lower minibands, responsible for intersubband transitions when ionized donors and corresponding electrons are taken into account, can be approximated by single energy levels, while the calculated Hartree potential remains unperturbed by their presence. Starting from the band structure, averaged intersubband absorption of a QDH ensemble is calculated in the dipole approximation, for incident light polarization in directions of stronger and weaker confinement in the QDHs, assuming that stochastic size distribution of QDHs in the ensemble can be modeled by two-dimensional Gaussian distribution. The absorption spectrum is pronounced for incident light polarization in the direction of weak confinement in QDHs and, depending on the material system and average dimensions of the QDHs in ensemble, dominantly occupies the mid-infrared wavelength range from 4 to 16 μm. It is observed that, with an increase in the QDH size fluctuations, the absorption spectra become broader, more asymmetric and flattened. As far as the wetting layer is taken into account, the electron concentration significantly affects the absorption, while the spectrum can be tailored by adjusting the spacer layers. The increase in the wetting layer thickness considerably reduces the absorption and leads to the redshift in the spectra, especially for smaller QDH size fluctuations. Finally, careful design of the spacer layers thickness can be used for optimization of the absorption spectrum regarding its maximization. Depending on the material system, the nominal thickness of the InAs layer and growth parameters, the self-assembled QDHs may have various cross-section profiles. In order to make a fair comparison and investigate the influence of QDH cross-section profile on the absorption spectrum, the criteria of equivalence for commonly used shape approximations (lens-like, trapezoidal, triangular and rectangular) are defined. It is observed that, regardless of the size fluctuation, absorption spectra for lens-like and equivalent trapezoidal cross-section profiles are similar. Using a rectangular approximation of the lens-like profile is justified for small-size fluctuations. The absorption spectra for ensembles with equivalent triangular QDHs exhibit significant red-shift and somewhat larger peak absorption in comparison to the ensembles with lens-like QDHs. In conclusion, the influences of the wetting layer thickness, the size distribution, and the remote doping by the spacer layer thickness on the mid-infrared absorption spectra, show interesting trends which are not fully intuitive. Therefore, the results presented in this thesis could help experimentalists to optimize specific intersubband transition properties in quantum dash nanostructures for future mid-infrared applications.
false
Određivanje karakteristika optičkih filtara sa mikro-prstenastim rezonatorom metodom konačnih elemenata
Characterization of optical filters based on micro-ring resonator by finite element method
U doktoratu je dat prikaz modelovanja optičkih filtara sa mikro-prstenastim rezonatorom primenom metode konačnih elemenata. Predloženi su algoritmi za određivanje parametara mikro-prstenastog rezonatora koji bi rezultirali željenom prenosnom karakteristikom. Poluprečnik prstena je određen iz dvodimenzionalnog modela na osnovu re- zonatne frekvencije prstena i slobodnog spektralnog opsega mikro-prstenastog rezonatora (FSR - ”Free Spectral Range”). Rezonatna frekvencija mikro-prstenastog rezonatora, dobijenog iz postupka modelovanja, upoređena je sa sopstvenim vrednostima (”Eigenvalue”) Helmholco- ve jednačine osno-simetričnog dvodimenzionalnog modela. Kako osno-simetri- čni model ima eksperimentalnu verifikaciju, iskorišćen je za verifikaciju predloženih modela. Red konvergencije rešenja je određen iz dvodimenzional- nog osnosimetričnog modela. Rastojanje između prstena i pravolinijskog svetlovoda je određeno iz trodi- menzionalnog modela. Trodimenzionalni model je takođe verifikovan pomoću osno-simetričnog modela. Kombinacijom metode konačnih elemenata i metode spregnutih modova zavisnih od vremena, određene su prenosne karakteristike mikro-prstenastih rezonatora. Za modelovanje optičnih filtara na bazi mikro-prstenastih rezonatora, pre- dložen je 2,5 dimenzioni model i metoda sa određivanjem parametara rasejanja na diskretnom skupu frekvencija. iv
Dissertation presents the modeling of optical filters with micro-ring resonators using finite element method. Algorithms for determining the parameters of the micro-ring resonator that would meet the desired transmission characteristics are proposed. e radius of the ring is determined from the two-dimensional model based on the resonant frequency of the ring and the free spectral range of micro-ring resonator (FSR - ”Free Spectral Range”). Resonant frequency of micro-ring resonators, resulting from the process of mode- ling, was compared to Eigenvalues of Helmholtc’s equation of axial-symmetric two- dimensional model. Since, axysymmetric model is experimentally verified, it was used to verify the proposed model. Convergence order of solutions is determined from the two-dimensional axisymmetrical model. e distance between the ring and the BUS waveguide is determined from a three- dimensional model. A three-dimensional model is also verified by using the axially- symmetrical model. By combining the finite element method and the method of temporal coupled modes, transmission characteristics of micro-ring resonators was determined. For the modeling of optical filters, based on micro-ring resonators, 2.5 dimensional model is proposed and method of determining the scaering parameters by excitation of ring resonator at discrete set of frequencies.
true
Metod za računarski efikasnu simulaciju energetskih pretvarača zasnovan na modelu stanja i superpoziciju prekidačkih efekata
A method for computationally efficient simulation of power electronic converters based on the state space model and superposition of switching effects
Upotreba prekidačkih konvertora je danas široko rasprostranjena, kako u smislu snage, naponskih nivoa, struje, tako i u smislu različitih aplikacija...
Today, usage of switching power converters is wide spread, as in the sense of power, voltage levels, current, also in the sense of different applications...
false
Detekcija i izolacija otkaza u separatoru pare termoenergetskih postrojenja
Fault detection and isolation in thermal power plant steam separator
Nedozvoljena devijacija najmanje jednog karakterističnog parametra ili osobine nekog sistema od njegovog uobičajnog ponašanja se može proglasiti greškom, odnosno otkazom u sistemu. Otkazi smanjuju efikasnost sistema, kvalitet proizvoda i ponekad mogu dovesti do potpunog zaustavljanja procesa, odnosno pada sistema. Ovakve pojave ne uzrokuju samo ekonomske gubitke već u nekim slučajevima mogu dovesti i do ljudskih žrtava. Rana detekcija otkaza može biti način sprečavanja ili smanjenja ranije pomenutih gubitaka. Stoga su monitoring sistema i detekcija otkaza postali esencijalni deo modernih sistema upravljanja. Kontrolni uređaji koji se u novije vreme eksploatišu da bi se unapredile performanse industrijskih procesa uključuju sofisticirane tehnike projektovanja digitalnih sistema i kompleksni hardver (ulazno-izlazni senzori, aktuatori, komponente i procesorske jedinice). Da bi se smanjila verovatnoća pojavljivanja otkaza na ovoj opremi potrebno je projektovati sistem za automatsko nadgledanje procesa koji bi se koristio za što je moguće raniju detekciju i izolaciju otkaza. U poslednje tri decenije, problem detekcije i izolacije otkaza u dinamičkim procesima je privukao veliku pažnju stručne javnosti i razvijen je čitav spektar pristupa koji se baziraju na modelima sistema. Za projektovanje sistema za robusnu detekciju i dijagnozu otkaza je značajno poraslo interesovanje jer se korišćenjem istih mogu smanjiti gubici u proizvodnji, sprečiti kvarovi na opremi i povećati sigurnost osoblja. Pouzdan sistem je sistem koji ima sposobnost za: 1. Što raniju detekciju otkaza 2. Tačnu dijagnozu istih 3. Što brži povratak sistema u nominalni režim rada U ovoj disertaciji je predložen novi pristup detekciji i identifikaciji otkaza u generatorima pare u okviru termo-energetskih postrojenja. Realni sistem, na kome su vršena merenja i predložen sistem za detekciju i izolaciju otkaza, je TEKO B1 blok termoelektrane Kostolac u Srbiji, čija je nominalna snaga 330 MW. Za takvu prirodu procesa i dostupne podatke, implementirani sistem za detekciju i izolaciju otkaza predstavlja kombinaciju dva poznata pristupa, jedan koji se bazira na modelima procesa i drugi koji se bazira na merenjima. Analizirana su tri moguća tipa otkaza, u senzorima nivoa vode, protoka vode i protoka pare. Prvi korak u predloženom algoritmu je identifikacija procesa. Merenje nivoa vode u separatoru su izvršena pod ekstremno visokim pritiscima pare uz nestacionarni dotok vode i ispuštanje pare. Stoga su dostupni podaci o nivou vode veoma nepouzdani, a prisutan je i sporadični, ali veoma jak, šum merenja. Kao posledica toga, standardne procedure za identifikaciju sistema nisu mogle da daju zadovoljavajuće rezultate i u ovoj tezi je predložena alternativa koja predstavlja robusnu estimaciju parametara sistema. U sledećem koraku je izvršeno statističko testiranje hipoteza pri čemu se ne koriste vektori merenja već se vektori parametara sistema izvode na osnovu procedure robusne identifikacije. Rezultati pokazuju izuzetnu efikasnost u detekciji i izolaciji otkaza jednog od moguća tri tipa otkaza koji su analizirani. Pristup detekciji otkaza koji je predložen u ovoj tezi, u poređenju sa rešenjima dostupnim u literaturi, je specifičan u sledećem: 1. Zasnovan je na linearnom modelu procesa sa odgovarajućim brojem ulaza i izlaza koji se identifikuje u zatvorenoj sprezi. 2. Predložena tehnika identifikacije je robusna po svojoj prirodi što je veoma važno u slučajevima kada je prisutan sporadičan jako izražen šum merenja. 3. Otkazi nisu detektovani na osnovu reziduala u merenjima, što je uobičajeni pristup u literaturi, već na osnovu parametara identifikovanog modela. 4. Predloženi metod za detekciju i izolaciju otkaza je kombinacija tehnika na bazi modela i na bazi merenja tako da je moguće napraviti kompromis između verovatnoće pojavljivanja lažnog alarma i vremenskog kašnjenja u detekciji otkaza. 5. Predloženi metod je primenjen na realnom sistemu separatora u termoelektrani i pokazao se kao veoma efikasan.
An un-permitted deviation of at least one characteristic property or parameter of a system from standard condition is referred as a fault. Faults result in reduced efficiency of the system, reduced quality of the product, and sometimes complete breakdown of the process. This not only causes economic losses but may also result in fatalities. An early detection of faults can assist to avert these losses. Therefore, fault detection and process monitoring is becoming an essential part of modern control systems. The control devices which are nowadays exploited to improve the overall performance of industrial processes involve both sophisticated digital system design techniques and complex hardware (input-output sensors, actuators, components and processing units). In such a way, the probability of failure occurrence on such equipment may result significant and an automatic supervision control should be used to detect and isolate anomalous working conditions as early as possible. Since the last three decades, the problem of fault detection and isolation in dynamic processes has received great attention and a wide variety of model-based approaches has been proposed and developed. A robust fault detection and diagnosis (FDD) system design has attained increased attention for reducing production loss, avoiding equipment damage, and increasing human safety. A more dependable system is a system that has the ability to: 1. Detect faults as fast as possible. 2. Diagnose them accurately. 3. Recover the system to the nominal performance as much as possible. This thesis presents a fault detection and identification approach for steam generators at thermal power plants. The real system in which the measurements were made and the FDI algorithm implemented is located at the TEKO B1 Unit of the Kostolac Thermal Power Plant in Serbia, whose nominal power output is 330MW. Given the nature of the process and available data, the implemented FDI algorithm is a trade-off of sorts between the model-based and the data-driven approach. Three possible types of faults are analyzed, of water level, water flow and steam flow sensors. The first step of the proposed algorithm is to identify the process. water level measurements in a separator are conducted under extremely high steam pressures, accompanied by constant unsteady water inflow and steam drain. therefore, available water level data are high unreliable and there is a sporadic of high-intensity measurement noise. As a consequence, standard process identification procedures have been shown not to yield satisfactory results and the thesis proposes a robust alternative to parameter estimation. The next step included statistical testing of the hypotheses, not using the measured data vector but the parameter vector derived from a robust identification procedure. The results demonstrated exceptional detection and isolation efficiency of one of the three possible and most frequent faults which were analyzed. The approach to sensor fault detection proposed in this work, compared to the solutions reported in the literature, is specific in the following respects: 1. It is based on a linear model of the process, with the corresponding number of inputs and outputs, identified in a closed loop. 2. The proposed identification technique is robust by its very nature, which is very important in the case of systems where sporadic high-intensity measurement noise is present. 3. Faults are not detected based on measurement residuals, which is the usual approach in the literature, but based on the parameter vectors of the identified model. 4. The proposed fault detection and isolation method is a combination of model- based techniques and the data-driven approach, such that a simple trade-off is possible between the probability of false alarm and the fault detection time delay; 5. The proposed method was applied in a real steam separator at a thermal power plant and demonstrated as highly efficient.
true
Psihografska analiza u segmentaciji turističkog tržišta
Psychographic analysis in the segmentation of the tourist market
Za razliku od robnih proizvoda, turističke usluge su „neopipljive“ i „nemerljive“, pa se teško mogu definisati u robno-novčanom smislu. Potrošač u turizmu kupuje u skladu sa svojim sistemom vrednosti, osobinama ličnosti i najviše prema stilu života. Stil života se manifestuje kroz modele ponašanja potrošača, koji su u fokusu marketinškog istraživanja. Karakteristike savremenih potrošača u kupovini proizvoda i usluga najviše dolaze do izražaja kroz analizu stila života, pa je važno kako i čime ga meriti. Psihografska analiza je instrument za merenje stila života koji čine aktivnosti, interesi, mišljenja, osobine ličnosti i druge interne dimenzije potrošača. Predmet ove disertacije je primena psihografske analize u svrhu segmentacije turističkog tržišta. Između različitih psihografskih tehnika merenja, autor se opredelio za „vrednosti i stil života“ (Values and Lifestyles) ili VALS 2 model. Prema autorovim saznanjima, ovo je prva primena VALS 2 tehnike za potrebe segmentacije turističkog tržišta. Empirijsko istraživanje obavljeno je u Republici Hrvatskoj na uzorku inostranih posetilaca, koji su anketirani putem VALS 2 upitnika. Primenjena je modifikovana verzija standardnog VALS 2 upitnika, prilagođen za korisnike turističkih usluga. Cilj disertacije je da se kroz psihografsku analizu, primenom VALS 2 tehnike identifikuje što više segmenata, u sklopu postojećih osam, prema VALS 2 tipologiji. Naučno-istraživačka metodologija sastoji se od teoretskog istraživanja relevantne literature iz navedenih oblasti, kao i empirijskog (kvalitativnog i kvantitativnog) istraživanja. Korišćenjem klaster analize, metodom aglomeracije, identifikuje se osam segmenata potrošača. Rezultati istraživanja služe kao osnova za razvoj marketing strategija. Metodologija i rezultati istraživanja su originalni i korisni, kako za naučno-istraživačku, tako i za poslovnu praksu. Očekuje se njihova primena, ne samo u regiji istraživanja, već i u zemljama okruženja, uključujući i Republiku Srbiju.
Compared to the market goods, the tourist services are „intangible“ and „immeasurable“, what makes them difficult to define in the monetary terms. At the tourist market, the customer is purchasing „anticipated benefits“complying with his value system, personality traits and lifestyle. Characteristics of the contemporary customers in the purchase of goods and services are mostly outspoken through their lifestyle, so it is importatant how to measure it. Psychographic analysis is an instrument for measuring the lifestyle, composed of attitudes, interests, opinions, personality traits and other internal dimensions of the customer. The objective of this dissertation is the application of psychographic analysis in the segmentation of the tourist market. Among a few measuring techniques, the author opted for the „Values and Lifestyles“ (VALS 2) measurement model. According to the author's findings, it is the first application of VALS 2 models for the segmentation of the tourist market. An empiric study is conducted on the sample of foreign visitors in Croatia by distributing to them a VALS 2 questionnaire. A modified version of the standard VALS 2 questionnaire, adopted for the tourist market was used. The aim of the study is to obtain as many as obtainable segments, up to eight, according to VALS 2 tipology by psychographic analysis. The scientific analysis is composed of the literature search from the above area and from the field study (qualitative and quantitative). By using the cluster analysis with agglomeration method, the eight consumer segments are identified. The research data is used as a basis for development of marketing strategies. The methodology and the results are authentic and useful for the scientific, as well as for the practical application. They are expected to be implemented not only in the study region, but wider, including the Republic of Serbia.
true
Računarski sistem za akviziciju, arhiviranje, pregledanje i obradu slika dobijenih gama kamerom
Computer system for acquiring, storing, retrieving and processing images obtained by gamma camera
Gama scintilaciona kamera (GSK) je medicinski uređaj koji prikazuje raspodelu γ-fotona emitovanih iz radiofarmaka unetog u organizam intravenski, oralno ili inhalacijom. Razvoj tehnologije integrisanih kola je imao uticaja na konstrukciju glave GSK i na fleksibilnost, brzinu, cenu i povećanu funkcionalnost kompjuterskih sistema za akviziciju, arhiviranje i analizu slike sa GSK. Iako je era digitalnih kamera počela još 80-tih godina 20. veka, mnogi nuklearno-medicinski centri u svetu (posebno u zemljama u razvoju) i dalje imaju analogne ili polu-analogne GSK u kojima i mehanika i scintilacioni kristal, tj. “srce” glave GSK, pravilno funkcionišu. Optimalno rešenje je digitalizacija GSK i integracija sa modernim računarskim sistemima. Digitalizacija analognih i polu-analognih GSK obezbeđuje: 1) produžetak korišćenja starih GSK i odlaganje finansijskih ulaganja u novu instrumentaciju; 2) primenu najboljih postojećih metoda analize nuklearno-medicinskih studija (u skladu sa aktuelnim principima medicine zasnovane na dokazima, MZD); 3) ispitivanje i razvoj novih metoda analize radi poboljšanja tretmana pacijenata, pod uslovom da je arhitektura sistema “otvorena”. Cilj istraživanja koje je prikazano u ovoj doktorskoj tezi je razvoj računarskog sistema (GammaKey sistem) koji daje funkcionalnost (akvizicija, arhiviranje i analiza slika) polu-analognim GSK, kompatibilan je i može se koristiti i sa savremenim digitalnim GSK, a koji svojom otvorenom arhitekturom omogućava razvoj i kliničku primenu novih dijagnostičkih tehnika. GammaKey sistem zadovoljava sledeće zahteve: 1) jednostavnu zamenu originalnog računarskog sistema (npr. MicroDELTA, Siemens Gammasonics Inc., Ilinois, SAD) novim sistemom - jednostavno priključenje na GSK pri čemu obuka korisnika minimalno traje zahvaljujući intuitivnom Windows grafičkom interfejsu; 2) pouzdanost sistema - zadovoljava standarde kvaliteta nuklearno medicinske instrumentacije definisanih od strane udruženja National Electrical Manufacturers Association (NEMA) i testiran je u paraleli sa pređašnjim računarskim sistemom (MicroDELTA, Siemens Gammasonics Inc., Ilinois, SAD); 3) kompatibilnost vi sa drugim (ne)medicinskim sistemima korišćenjem standardnih formata podataka (JPEG, BITMAP, ASCII, DICOM), i sinhronizacija podataka između dve radne stanice (jedna radna stanica je za akviziciju, arhiviranje i obradu studija dobijenih sa GSK, a druga je samo za obradu podataka); 4) brza, jednostavna i jeftina nadgradnja za nove generacije računara i analogno-digitalnih konvertora (ADK) – omogućena izborom LabVIEW softverskog okruženja (National Instruments, Teksas, SAD) i komercijalnog National Instruments ADK; 5) razvoj i dodavanje novih nestandardnih akvizicionih protokola i alatki za obradu studija. U osnovnoj konfiguraciji GammaKey sistema uključene su: 1) standardne funkcije za obradu slika (resemplovanje na dimenziju 1024x1024 piksela bilinearnom interpolacijom, nine-point smoothing, Batervortov i Vinerov filter); 2) alatke za manipulaciju slikom (zumiranje, podešavanje kontrasta, brisanje dela slike koji nije od interesa); 3) standardne funkcije za obradu scintigrafskih studija (crtanje do 10 regiona od interesa (ROI), generisanje dinamskih krivih nad ROI, aritmetičke operacije nad slikama i dinamskim krivama, N-point median i cubic spline filtriranje dinamskih krivih, automatska segmentacija ROI na statičkim studijama, korekcija radioaktivnosti ROI). U proširenoj varijanti, GammaKey sistem uključuje i tri nestandardne analize: 1) EffSpleen akvizicija i analiza za predviđanje efikasnosti splenektomije u bolesnika sa hroničnom trombocitopenijskom purpurom; 2) Submarine analiza za lokalizaciju paratiroidnih tumora; 3) SalivaryScan analiza za kompletnu kvantifikaciju funkcije pljuvačnih žlezda. EffSpleen rutina omogućava kompletnu akviziciju i analizu in vivo ispitivanja kinetike trombocita radi procene efikasnosti splenektomije u bolesnika sa hroničnom imunom trombocitopenijskom purpurom prema metodi razvijenoj u Centru za nuklearnu medicinu Kliničkog centra u Beogradu. Razvijena rutina je omogućila automatizaciju u procesu zahtevne obrade serije statičkih studija i proračuna indeksa sekvestracije trombocita. Submarine rutina predstavlja implementaciju originalne metode za lokalizaciju paratiroidnih tumora na osnovu dinamskih paratiroidnih scintigrama i razvijena je u saradnji sa lekarima iz Instituta za onkologiju i radiologiju Srbije u Beogradu. U tezi je prikazana evaluacija Submarine metode na grupi pacijenata sa histopatološki dokazanim paratiroidnim tumorima i ukazano je na njenu efikasnost pri lokalizaciji hiperplazija i paratiroidnih lezija zaklonjenih tiroidnim nodusima, vii rešavajući time tipične promašaje standardnog subtrakcionog protokola. Predloženi su i algoritmi za vizuelizaciju i automatsko izdvajanje regije paratiroidnih lezija. SalivaryScan rutina omogućava automatski proračun 11 kvantitativnih indeksa koji opisuju funkciju pljuvačnih žlezda, za razliku od komercijalnih softvera koji obično daju samo jedan parameter (ejekcionu frakciju). Uvedena je i značajna modifikacija indeksa koji opisuje brzinu sekrecije. Razvijena alatka je bazični instrument za dalja multidisciplinarna istraživanja u domenu scintigrafije pljuvačnih žlezda. U tezi je prikazan i primer jedne nefrološke kliničke studije na podacima snimljenim pomoću realizovanog sistema, pri čemu je celopkupna analiza urađena u posebnoj third-party aplikaciji razvijenoj od strane International Atomic Energy Agency (IAEA). GammaKey sistem osim što je omogućio nastavak osnovne kliničke delatnosti, obezbedio je i praćenje savremenih trendova, kontinuitet u istraživanju, definisanje validnih referentnih vrednosti kliničkih parametara i mogućnost validacije rezultata drugih aplikacija, čime je u potpunosti podržao MZD koncept. GammaKey sistem se koristi u svakodnevnoj kliničkoj praksi u dva velika medicinska centra u Srbiji: Kliničkom centru Vojvodine u Novom Sadu (od 2005. godine, Centar za Laboratorijsku medicinu, Odeljenje za nuklearnu medicinu, http://www.kcv.rs) i Kliničkom centru Srbije u Beogradu (od 2007. godine, Centar za nuklearnu medicinu, Odeljenje za neurologiju, urologiju i endokrinologiju, http://www.kcs.ac.rs). EffSpleen rutina se koristi u Kliničkom centru Srbije u Beogradu (Odsek za nuklearnu hematologiju) od 2008. (preko 250 pacijenata). Submarine rutina se koristi u Institutu za onkologiju i radiologiju Srbije u Beogradu za dijagnostiku paratiroidnih tumora i u istraživačke svrhe od 2012. godine (preko 100 pacijenata, uključujući i retroaktivnu preglede za pacijente od 2008. godine). Više od 30 000 pacijenata je bilo ispitano korišćenjem GammaKey sistema (više od 10 000 pacijenata u Kliničkom centru Vojvodine i više od 20 000 pacijenata u Kliničkom centru Srbije).
Gamma scintillation cameras (GSC) are medical devices for the analysis of the distribution intravenously injected, inhaled or ingested radiopharmaceuticals. The development of integrated circuit technology had influence on construction of GSC’s head and on flexibility, speed, cost and expandability of GSC’s supported systems for image acquisition, storage and analysis. Although the era of digital GSCs began long ago (1980s), many nuclear medical centres still have analogue or semi-analogue GSCs with fully operational mechanics and scintillation crystals, which are the core of nuclear imaging. Optimal solution is digitizing GSC and integration with modern computer systems. Digitizing of analogue and semi-analogue GSCs ensures: 1) continuity of using old GSCs and the postponement of buying new instrumentation; 2) application of best available analysis of nuclear-medical studies (in accordance with the principles of Evidence Based Medicine, EBM); 3) testing and development of new analysis techniques to improve the treatment of patients, if the system architecture is “opened”. The aim of the research presented in this doctoral thesis is the development of a computer system (GammaKey system) that gives the functionality (acquisition, storage and image processing) to semi-analogue GSCs, compatible and also can be used with modern digital systems, has an open architecture enabling the development and clinical application of new diagnostic techniques. GammaKey system meets the following requirements: 1) simple replacement of the original computer system (e.g. MicroDELTA, Siemens Gammasonics Inc., Ilinois, USA) to the new configuration – a simple connection to GSC and minimal training thanks to the intuitive Windows graphical interface; 2) reliability – meets the quality standards of nuclear medical instrumentation defined by National Electrical Manufacturers Association (NEMA), no loss of information about patients and studies, accurate and precise production of results; 3) compatibility with other (non)medical systems by using standard data formats (JPEG, BITMAP, ASCII, DICOM) and data synchronization between two workstations ix (the first one for acquisition, storage and processing, the second one only for image processing); 4) fast, simple and cheap upgrade to new generations of computers and analog-to-digital convertors (ADC) – enabled by using LabVIEW software environment (National Instruments, Teksas, SAD) and commercial National Instruments ADC; 5) development and implementing novel non-standard acquisition protocols and tools for study processing. Basic configuration of GammaKey system includes: 1) standard functions for image processing (resampling to the matrix 1024x1024 pixels by bilinear interpolation, nine-point smoothing, Butterworth and Winer filter); 2) tools for image manipulation (zoom, contrast adjustment, deleting parts of the image that are not of interest); 3) standard functions for scintigraphic study processing (up to 10 regions of interest (ROI), calculating time-activity-curves (TAC), frame and TAC arithmetic opearations, N-point median and cubic spline TAC filtering, automatic ROI segmentation of static studies, ROI radioactivity correction). Advanced configuration of GammaKey system includes three non-standard features: 1) EffSpleen acquisition and analysis for efficacy of the splenectomy prediction in patients with chronic immune thrombocytopenic purpura; 2) Submarine analysis for localization of parathyroid tumors; 3) SalivaryScan analysis for complete quantitative investigation of salivary glands function. EffSpleen routine enables complete acquisition and analysis of in vivo examination of platelet kinetics in order to estimate splenectomy efficacy according to the method developed in the Center for Nuclear Medicine, Clinical Center of Serbia, Belgrade. Developed routine allows the automation of time- consuming processing of sequential static studies and calculation of platelet sequestration index. Submarine routine is the implementation of the original method for the localization of parathyroid tumors based on dynamic parathyroid scintigrams, developed in collaboration with physicians from the National Cancer Research Center of Serbia, Belgrade. Evaluation of Submarine method is presented in the group of patients with histopathologically proven parathyroid tumors and its efficacy has been demonstrated for localization of hyperplasia and parathyroid lesions covered by thyroid nodules, solving typical pitfalls of standard subtraction protocol. Algorithms for visualization and automatic extraction of regions with parathyroid lesions are suggested. SalivaryScan routine enables automatic calculation of 11 quantitative indices that x describe the function of salivary glands, unlike commercial software that usually gives only one parameter (ejection fraction). A significant modification of secretion velocity index is introduced. Developed tool is a basic instrument for further multidisciplinary research in the field of salivary scintigraphy. An example of nephrology clinical study based on studies acquired by the realized system is also presented, where the analysis is done in a third-party application developed by the International Atomic Energy Agency (IAEA). GammaKey system enables the continuity of basic clinical practice, as well as following modern trends, continuity in research, defining valid reference values and the possibility of validation of other applications. All these facts describe the system that fully supports EBM concept. GammaKey system has been used in daily clinical practice in two leading medical centers in Serbia: Clinical Center of Vojvodina in Novi Sad (since 2005, Center for Clinical Laboratory Medicine, Nuclear Medical Department, http://www.kcv.rs) and Clinical Center of Serbia in Belgrade (since 2007, Center for Nuclear Medicine, Neurology, Urology and Endocrinology Departments, http://www.kcs.ac.rs). The EffSpleen protocol has been used in the Clinical Center of Serbia (Nuclear Haematology Department) since 2008 (more than 250 patients). The Submarine routine has been used in the National Cancer Research Center of Serbia, Belgrade, Serbia, for diagnostic and research purposes since 2012 (more than 100 patients, including retroactive analysis since 2008.). More than 30 000 patients have been examined using the GammaKey system (more than 10 000 patients at the Clinical Center of Vojvodina and more than 20 000 patients at the Clinical Center of Serbia).
true
Improving the performance of LoRa based wireless transceivers
Poboljšanje performansi bežičnih primopredajnika zasnovanih na LoRa modulaciji
Savremene bežične tehnologije značajno utiču na svakodnevni život modernog društva. Internet stvari, nova paradigma interakcije čoveka sa njegovom okolinom, omogućava brzo prikupljanje i selektovanje informacija iz realnog sveta koristeći različite vrste senzora. Jedna od novih tehnologija, koja je potencijalni kandidat da postane najmasovnije upotrebljavana tehnologija u takvim sistemima jeste upravo LoRa. Ova tehnologija je, do sada, ispitivana kroz brojne teoretske i eskperimentalne studije. Da bi se obezbedio dublji uvid u mogućnosti tehnologije, razvijen je novi LoRa primopredajnik sa poboljšanim karakteristikama u smislu osetljivosti i imunosti na smetnje. U tezi je prikazana metodologija razvoja novog primopredajnika i dati su rezultati laboratorijskih ispitivanja koji verikuju poboljšane performanse. Takodje su uradjena uporedna ispitivanja novorazvijenog primopredajnika sa dva veoma rasprostranjena komercijalna primopredajnika. Rezultati ispitivanja su pokazali, da novi LoRa primopredajnik ima poboljsanu osetljivost u opsegu 15 {25 dB, dok su rezultati ispitivanja u realnim uslovima dodatno utvrdili i veću imunost na spoljnje smetnje. U drugom delu teze su dati rezultati eksperimenta izvršenih u realnim uslovima. Novi LoRa primopredjanik je ispitivan na vrlo dugačkim i izuzetno dugačkim trasama u dva karakteristična propagaciona scenarija: kada je zadovoljen uslov optičke vidljivosti i kada je optička vidljivost blokirana izmedju prijemnika i predajnika. U ovim eksperimentima linkovi su formirani iznad kopna i iznad mora. Rezultati eksperimenata su pokazali da, ukoliko je obezbedjena optička vidljivost izmedju prijemnika i predajnika, linkovi u dužini od 112 km iznad kopna u potpunosti su izvodljivi, koristeći omnidirekcione antene i standardnu izlaznu snagu u oba ISM opsega, 868MHz i 434 MHz. Takodje, pokazano je da takvi linkovi mogu biti uspostavljeni za bilo koju kombinaciju LoRa parametara, ukljućujući i komercijalne primopredajnike, s obzirom na to da su eksperimenti izvršeni za tri karakteristična seta LoRa parametara, naime: koristeći faktore širenja SF 2 f7; 10; 12g unutar ksnog LoRa opsega od BW = 125 kHz i upotrebom zaštitnog kodovanja iii CR = 4=6. Analiza parametara linka u vremenskom domenu je pokazala da je glavni uzrok gubitka paketa mulitpath propagacija signala, dok je radio-interferencija imala zanemariv doprinos gubitku paketa. Dalje je pokazano da su izuzetno dugački linkovi, dužine 316 km, izvodljivi koristeći novorazvijeni primopredajnik u oba ISM opsega od interesa, upotrebljavajuci omnidirekcione antene i standardne izlazne snage. Konačno, eksperimenti iznad morske površine su pokazali da, ukoliko postoji optička vidljivost izmedju prijemnika i predajnika, linkovi u dužini od 22 km su mogući i to u oba ISM opsega od interesa, koristeći omnidirekcione antene i standardne izlazne snage; margina linka od 15 dB u oba frekvencijska opsega implicira da su moguća i veća rastojanja. Eksperimentalno je utvrdjeno i da, ukoliko je optička vidljivost preko morske površine izmedju prijemnika i predajnika blokirana, moguće je uspostaviti link u dužini od 28 km na frekvenciji od 434 MHz, upotrebom antena sa većim pojačanjem, ali kvalitet prenosa u ovom slučaju ne može biti garantovan.
The use of wireless technologies dominates the everyday life of a modern society. Internet of Things, the new paradigm of interaction between humans and their environment, unlocks advanced scenarios of collecting information from the real world using various sensors. One of the recently developed technologies, that is a potential candidate to become a prevalent technology exploited in the Internet of Things arena, is LoRa. The technology has been extensively explored in numerous theoretical and experimental studies, and a signicant insight has been gained. However, to provide a deeper insight into LoRa technology capabilities, the novel LoRa transceiver with improved characteristics in terms of sensitivity and immunity to external interferences has been developed. This work presents the development methodology, as well as the laboratory experiments that veried the transceiver's superior characteristics. In addition, it provides performance comparison with two widely adopted commercial solutions. The results have shown that the novel LoRa transceiver has an improved sensitivity of 15{25 dB, while the eld experiments revealed higher immunity to external interferences. The second part of the thesis explores LoRa capabilities and limitations by deploying the novel transceiver along with the two commercial transceivers in various experimental scenarios. Namely, the transceivers were deployed in very long and ultra-long distance links providing two propagation conditions: line-of-sight over land and line-of-sight over seawater. In addition, obstructed line-of-sight scenario over sea water has been investigated. The results indicate that 112km long links over land are fully feasible using low-cost, rubber-duck, omnidirectional antennas and standard output powers in two ISM frequency bands, 868MHz and 434 MHz, provided that line-of-sight is guaranteed. Furthermore, it has been shown that such links can be established for virtually any combination of LoRa parameters, since the experiments were carried out deploying three dierent, characteristic, set of LoRa parameters, namely: for spreading factors of SF 2 f7; 10; 12g, within a xed bandwidth of BW = 125 kHz, and deploying xed coding rate of CR = 4=6. The short and long-term temporal analysis of the links revealed that the main cause of packet reception rate deterioration are multipath propagation eects, while the i radio-interference has negligible eects on packet reception rate. Furthermore, it has been demonstrated that ultra-long distance links of 316 km are also possible, in both ISM bands of interest, but using only novel LoRa transceiver. The signicance of the results is amplied by the fact that omnidirectional antennas were used in the experiments with standard output powers. Finally, the experiments over seawater revealed that links up to 22 km in line-ofsight mode are fully feasible using low-cost, rubber-duck, omnidirectional antennas and standard output powers with link margin yielding more than 15 dB in both ISM frequency bands. Moreover, the links in obstructed line-of-sight mode over 28 km are also possible in ISM 434MHz band; however, there are certain limitations when it comes to their practical implementations since costly and bulky antennas are needed on both sides and transmission reliability is not guaranteed.
false
Primena T2 kontrolnih dijagrama i skrivenih Markovljevih modela na prediktivno održavanje tehničkih sistema
Application of T2 control charts and hidden Markov models in predictive maintenance of technical systems
U današnjoj industriji primena najboljih strategija odrţavanja je veoma vaţan zadatak, kako bi se obezbedila stabilnost i pouzdanost tehničkih sistema. U literaturi se moţe naći veliki broj radova i knjiga o različitim strategijama odrţavanja i uglavnom se svuda ističe prednost prediktivnog odrţavanja (eng. predictive maintenance) u odnosu na planirano odrţavanje (eng. time-based maintenance). Prediktivno odrţavanje produţava vreme u kojem sistem radi dobro, smanjuje nepotrebno zaustavljanje sistema, redukuje materijalne gubitke i sprečava nastajanje katastrofalnih otkaza. Iako je sa razvojem visoko sofisticiranih tehnologija ova oblast istraţivanja veoma uznapredovala, i dalje ima puno prostora za poboljšanje postojećih tehnika, kao i za razvoj novih. U ovoj tezi predloţena je inovativna tehnika prediktivnog odrţavanja sa ciljem isticanja prednosti prediktivnog odrţavanja u odnosu na planirano odrţavanje. Predloţeni algoritam primenjen je na konkretan problem koji se javlja planiranim odrţavanjem udarnih ploča mlinova koje melju ugalj u podsistemu za mlevenje uglja u termoelektrani “TEKO”, Kostolac, Srbija. Planiranim odrţavanjem predviĎa se zamena udarnih ploča nakon odreĎenog broja radnih sati, meĎutim u zavisnosti od kvaliteta uglja i same udarne ploče ova zamena je nekad potrebna ranije ili kasnije. Posledice ovakvog odrţavanja su veliki materijalni gubici koji nastaju zbog čestog prevremenog zaustavljanja celog podsistema za mlevenje uglja, kao i mogućnost da otkaz nastupi pre zamene. Prvi korak u primeni predloţenog algoritma bio je sakupljanje podataka, odnosno snimanje akustičkih signala. Akustički signali su snimljeni pomoću usmerenog mikrofona, dok je podsistem za mlevenje uglja u funkciji, bez ometanja rada mlina. Sakupljeni podaci zatim su obraĎeni izdvajanjem korisnih parametara akustičkih signala koji su predstavljeni u frekvencijskom domenu pomoću spektrograma (eng. spectrogram). Inovativnost predloţenog algoritma zasniva se na izboru tehnike prognoze otkaza (eng. failure prognostics). Izabrana je tehnika koja je zasnovana na pokretnim podacima (eng. data driven), zato što se ne traţi poznavanje modela, a dostupni su nam podaci nadgledanja stanja (snimljeni akustički signali). Najpre su kontrolni dijagrami (eng. control chart) primenjeni na izdvojene parametre akustičkih signala u frekvencijskom domenu, a zatim je formiran skriveni Markovljev model (eng. Hidden Markov Model) čije su opservacije parametrizovane statistike sa kontrolnih dijagrama. Softverska realizacija algoritma izvršena je u programu Matlab. Naučni doprinosi ove teze su dobijeni rezultati koji su potvrdili hipoteze koje su postavljene na početku ovog istraţivanja. PotvrĎeno je da je predloţena kombinacija T2 kontrolnih dijagrama i skrivenih Markovljevih modela (SMM) efikasna u prediktivnom odrţavanju. Drugim rečima, kontrolni dijagrami se mogu efikasno primeniti na izdvojene parametre akustičkih signala u frekvencijskom domenu, a nestacionarnost procesa moţe se pokriti uvoĎenjem SMM-a koji definisanjem različitih stanja podrazumeva dinamički model i nestacionarnost opservacija. TakoĎe, potvrĎeno je da su akustički signali informativni kada je u pitanju informacija o stanju udarnih ploča mlina, što je značajan doprinos, s obzirom da se u literaturi uglavnom daje prednost signalima vibracije. Predloţena metoda predstavlja kompromis izmeĎu već ponuĎenih rešenja koja se mogu naći u literaturi, a zbog svoje inovativne strukture svakako predstavlja naučni doprinos oblasti prediktivnog odrţavanja. Predloţena metoda ima nekoliko prednosti. Kao prvo, metoda nije invazivna i ne zahteva zaustavljanje celog podsistema za mlevenje uglja za snimanje akustičkih signala. Softverska realizacija predloţenog algoritma nije mnogo sloţena i nije vremenski zahtevna kada se SMM jednom adekvatno obuči, što bi bilo od značaja ako bi se ovakav način prediktivnog odrţavanja primenio u realnom vremenu. Dodatna prednost predloţene metode jeste što se zasniva na obradi akustičkih signala koji su jeftiniji za snimanje i jednostavniji za obradu od signala vibracije.
In today’s industry, application of the best maintenance strategies is very important task in ensuring stability and reliability of technical systems. Numerous papers and books about different maintenance strategies can be found in literature, and almost everywhere are emphasized merits of predictive maintenance in regard to time- based maintenance. Predictive maintenance extends the period of time during which the system functions well, decreases unnecessary shutdowns, reduces material losses and prevents catastrophic failures. Although this field of research is very much advanced with the development of high sophisticated technologies, there is still a lot of room for improvement of existing techniques and the development of new ones. In this thesis innovative technique of predictive maintenance is proposed with the aim of highlighting the benefits of predictive maintenance compared to time-based maintenance. The proposed algorithm is applied to a specific problem that occurs when time-based maintenance is applied on grinding tables of the coal mill, in coal grinding subsystem at the Thermoelectric Power Plant “TEKO”, Kostolac, Serbia. Time-based maintenance provides replacement of grinding tables after certain number of working hours, but depending on the quality of the coal and grinding table itself, this replacement sometimes needs to be made before or after planned replacement. The consequences of such maintenance are great material losses incurred because of frequent shutdowns of the entire coal grinding subsystem, as well as the possibility that the failure occurs before replacement. First step in application of proposed algorithm was data acquisition, i.e. recording of acoustic signals. Acoustic signals were recorded by means of directional microphone, while the coal grinding subsystem was operating, without interfering with mill operation. Collected data were processed by the feature extraction of acoustic signals that were presented in frequency domain with spectrogram. Innovativeness of the proposed algorithm is based on the selection of failure prognostic technique. Data- driven technique is chosen because model knowledge is not acquired, and condition monitoring data (recorded acoustic signals) are available. Control charts were first applied on extracted parameters of acoustic signals in frequency domain, and then Hidden Markov Model (HMM) is formed with observations that are parameterized statistics from control charts. Software realization of the algorithm is realized in program Matlab. The main scientific contributions of this thesis are obtained results that confirmed hypotheses from the beginning of this research. It is confirmed that proposed combination of T2 control charts and HMM is effective in predictive maintenance. In other words, control charts can be effectively applied on extracted parameters of acoustic signals in frequency domain, and HMM is introduced to account for the fact that the process is non-stationary, where different states are defined to reflect a dynamic system and non-stationary observations. It is also confirmed that acoustic signals are informative when information about grinding table condition is in question. This is significant contribution, given that in the literature vibration signals are mostly preferred. Proposed method is trade-off between solutions already offered in the literature, and because of its innovative structure it certainly represents scientific contribution to the field of predictive maintenance. Proposed method has several advantages. The method is non-invasive and does not require stopping of the coal grinding subsystem for acoustic signals recording. Software realization of the proposed algorithm is not too much complex and it is not time consuming when HMM is once trained. This fact is very important if this method would be applied in real time. Proposed method is based on processing of acoustic signals which are cheaper and much simpler for processing than vibration signals. This is additional advantage of the proposed method.
true
Modeliranje i definisanje dimenzija logističke usluge u elektronskoj trgovini
Modelling and defining logistics service dimensions in electronic commerce
U XXI veku došlo je do značajnih promena u B2C odnosu usled razvoja različitih digitalnih tehnologija. MeĎutim, taj odnos se nije promenio u smislu načina, mesta i vremena u kome se poslovanje/kupovina odvija, već se promenila sama priroda odnosa i širi kontekst ponašanja kupaca. Izvršene su funkcionalne, suštinske i kontekstualne promene u B2C odnosu, pri čemu su izmenjeni i poslovni i socijalni okviri takvog odnosa. Sam koncept B2C interakcije stoga je redefinisan u smislu motivatora i vrednosti, pri čemu su kupci mnogo više aktivniji i participativniji, što B2C odnos čini simbiotičnim. Za prodavce je vaţno da identifikuju najbolje prakse i afirmativne aspekte odgovarajućih kanala prodaje. Čvrsto poslovno razmišljanje, vizionarsko voĎstvo, temeljna analiza konkurencije, finansijska analiza i artikulacija dobro osmišljene strategije e-trgovine od suštinskog su značaja za uspešnu e-prodaju. Prepoznavanje pokretača promena u offline i online prodaji od vitalnog značaja je za kreiranje boljeg kupovnog iskustva. Primenom omni-channel strategije prodavci nastoje da povećaju zadovoljstvo svojih kupaca čije potrebe iz dana u dan postaju sve veće i raznovrsnije. Kupci danas rado prihvataju prednosti koje donosi digitalna tehnologija, npr. uvid u cenu proizvoda jednim klikom miša, iščitavanje komentara ostalih kupaca i eksperata itd. MeĎutim, takoĎe ţele da iskoriste i prednosti koje nude tradicionalne prodavnice, kao što su npr. interakcija licem u lice sa osobljem prodavnice, isprobavanje proizvoda na licu mesta itd. Svaki prodavac treba da ponudi kupcima ono što kupci od njega ţele, a jedan od načina da u tome uspe je da im planirano i precizno pribliţi različite kanale prodaje. Generalno, proces kupovine započinje odabirom proizvoda, zatim sledi isprobavanje ili testiranje, potom se proizvod isporučuje, i na kraju se prihvata/preuzima ili vraća u nekim slučajevima. Poznavanje kretanja kupaca tokom procesa kupovine je od presudne vaţnosti za omni-channel prodavca. S druge strane, nemoguće je kreirati besprekorno kupovno iskustvo ako ne postoji koordinacija izmeĎu svih logističkih procesa. Da bi se zadovoljila potraţnja kupaca sa bilo kog mesta (do bilo kog mesta), odluke o dizajnu mreţe, upravljanju zalihama, skladištenju i transportu trebaju biti usklaĎene sa opcijama realizacije porudţbina i obećanim vremenima isporuke. U dizajniranju mreţe vaţno je uzeti u obzir da je osim za ostvarivanje ekonomije obima i troškovnih benefita, neophodno obezbediti i lokalno prisustvo prodavca. Bez lokalnog prisustva prodavca gotovo je nemoguće kupcima ponuditi kratka vremena isporuke proizvoda. Pored toga, zalihe trebaju biti vidljive u realnom vremenu na svim lokacijama, kako bi prodavci mogli ponuditi različite opcije realizacije porudţbina. Sistemi upravljanja skladištem i transportom moraju se isprojektovati na način da se operacije iskladištenja, pakovanja i transporta obavljaju u okviru obećanog vremenskog roka. Koji kanal prodaje/distribucije je najbolji, odnosno kada se očekuje da će omni-channel strategija zaista dodati vrednost poslovnim ciljevima prodavca, pitanja su koja treba paţljivo obraditi, i koja su upravo obraĎena u ovoj disertaciji. Predmet istraţivanja ove disertacije su dimenzije logističke usluge koje vode ka stvaranju pozitivnog kupovnog iskustva i različiti kanali distribucije u e- trgovini. Na bazi sistematizacije dosadašnjih teorijskih i praktičnih doprinosa u istraţivanjima sistema distribucije proizvoda u e-trgovini, posebno sa aspekta dimenzija logističke usluge, u radu su razvijena dva modela: jedan koji obezbeĎuje podršku odlučivanju o dimenzijama logističke usluge u e-trgovini, i drugi za vrednovanje najpogodnijeg kanala distribucije u e-trgovini sa aspekta e-kupca, e-prodavca i logističkog provajdera. TakoĎe, u disertaciji je predstavljen i alat za merenje i praćenje zadovoljstva korisnika u odnosu na odreĎene dimenzije logističke usluge u e-trgovini. Pomenuti modeli, kao i merni instrument (skala) omogućiće e-kupcima, e-prodavcima i logističkim provajderima da bolje sagledaju i vrednuju različite kanale distribucije u e-trgovini, ali i da definišu ključne dimenzije logističke usluge u e-trgovini koje će dovesti do zadovoljstva krajnjih potrošača.
In the 21st century, there have been significant changes in the B2C relationship, due to the development of diverse digital technologies. Nevertheless, this relationship has not changed in terms of the manner, place and time in which a business/purchase takes place; rather, the very nature of the relationship and the broader context of customer behaviour have altered. Functional, essential and contextual changes have been made in the B2C relationship, and both business and social frameworks of such a relationship have been transformed. The very concept of B2C interaction has therefore been redefined in the sense of motivators and values, with customers being much more active and participatory, making the B2C relationship symbiotic. For salespeople, it is important to identify best practices and affirmative aspects of appropriate sales channels. Strong business thinking, visionary leadership, thorough competition analysis, financial analysis, and the articulation of a well-designed e-commerce strategy are essential for successful e-commerce. Identifying the initiators of change in offline and online sales is vital to creating a better shopping experience. By applying the omni-channel strategy, retailers strive to increase the satisfaction of their customers, whose needs are becoming greater and more diverse day by day. Customers today are eager to accept the benefits of digital technology, e.g. insight into the price of the product by a click of the mouse, reading the comments of other customers and experts, etc. However, they also want to take advantage of the benefits offered by traditional stores, such as face-to-face interaction with store staff, on-site product testing, etc. Every salesperson should offer customers what customers want from them, and one of the ways to succeed in that is to bring them closer to different sales channels in a planned and precise manner. Generally, the purchase process begins with product selection, followed by trial or testing, after which the product is delivered, and finally accepted/taken or returned in some cases. Knowing customer movements during the purchase process is crucial for an omni-channel retailer. On the other hand, it is impossible to create a flawless shopping experience if there is no coordination between all logistics processes. In order to meet customer demand from any place (to any place), decisions about network design, inventory management, warehousing and transportation need to be aligned with order options and promised delivery times. When designing a network, it is important to consider that, in addition to achieving economies of scale and cost benefits, it is necessary to ensure the local salesperson presence. Without the local presence of the salesperson, it is almost impossible to offer customers short product delivery times. Additionally, inventory should be visible in real time at all locations, so that retailers can offer different order options. Warehouse and transport management systems have to be designed in such a way that warehousing, packaging and transport operations are completed within the promised timeframe. Thinking which sales/distribution channel is the best, i.e. when the omni-channel strategy is expected to really add value to the retailer’s business goals, implies questions that need to be carefully addressed, and that have been addressed in this dissertation. The research topic of this dissertation are the dimensions of the logistics services that lead towards the creation of a positive shopping experience and different distribution channels in e-commerce. Based on the systematization of previous theoretical and practical contributions in the research of product distribution systems in e-commerce, especially from the aspect of logistics service dimensions, two models have been developed in the thesis: one that provides support to decision-making on logistics service dimensions in e-commerce, and the other for the evaluation of the most suitable distribution channel in e-commerce from the aspect of the e-customer, e-retailer and logistics provider. In addition, the dissertation also presents a tool for measuring and monitoring customer satisfaction in relation to certain dimensions of logistics services in e-commerce. The mentioned models, as well as the measuring instrument (scale), will enable e-customers, e-retailers and logistics providers to understand and evaluate different distribution channels in e-commerce better, but also to define key dimensions of logistics services in e-commerce that will lead to the end-customer satisfaction.
true
Pogodnosti ambijenta, preduzetništvo i ruralni razvoj Srbije
Amenities, entrepreneurship and rural development of Serbia
Uticaj koncepta pogodnosti na rast ruralne ekonomije je relativno novija tema u međunarodnoj akademskoj javnosti koja sve više zaokuplja pažnju. Sa ekonomskim rastom i razvojem kvalitet ruralnog ambijenta postaje sve značajniji faktor oblikovanja migracionih kretanja stanovništva na relaciji urbane – ruralne oblasti. Koncept ruralnih pogodnosti čine prirodne, izgrađene i kulturne pogodnosti koje imaju komplementarnu ulogu u ruralnom rastu i razvoju. Ruralni poljoprivredni pejzaži se vrednuju ne samo s obzirom na njihove mogućnosti za proizvodnju različitih poljoprivrednih proizvoda, već i sa stanovišta mogućnosti obezbeđenja ruralnih ekoloških javnih dobara. Prirodne pogodnosti i drugi kvaliteti lokalnog ambijenta koji unapređuju uslove za život identifikovani su kao ključni faktori rasta stanovništva, dohotka i zaposlenosti u ruralnim područjima. Predmet naučno-istraživačkog rada u okviru doktorske disertacije je analiza uticaja pogodnosti ambijenta i preduzetništva na ruralni razvoj, sa posebnim osvrtom na Srbiju. Unapređenje kvaliteta ljudskog kapitala i rast preduzetničkih aktivnosti u modelima rasta ruralne privrede može se posmatrati kao jedan od mogućih mehanizama transformacije, rasta i razvoja ruralne ekonomije. Osnovni ciljevi naučno-istraživačkog rada u okviru doktorske disertacije su vezani za klasifikaciju ruralnih oblasti u Srbiji sa stanovišta raspoloživih pogodnosti ambijenta i analizu uticaja koncepta ljudskog kapitala i preduzetništva na ekonomski rast ruralnih područja u našoj zemlji. Pogodnosti ambijenta, preduzetništvo i ruralni razvoj Srbije Doktorska disertacija Sonja Josipović ii U razvijenim ekonomijama različite pogodnosti ambijenta imaju snagu privlačenja značajnog obima ljudskog i preduzetničkog kapitala neophodnih za razvoj ruralne privrede. Ruralne oblasti sa većim učešćem ljudskog kapitala imaće bolje ekonomske performanse zahvaljujući inovacijama, visokom nivou preduzetništva i kreativnim oblicima poslovanja. Ruralna područja visokih pogodnosti ambijenta privlače nove ruralne rezidente koji su okrenuti ka preduzetništvu.
The impact of the concept of rural amenities on the growth and development of rural economy is a relatively recent topic which has been gaining more and more attention. Along with the economic growth and development the quality of the rural environment becomes an increasingly important factor in shaping the migration trends of the population between the urban and rural areas. The concept of rural amenities comprises the natural, built and cultural amenities that have a complementary role in rural growth and development. Rural agricultural landscapes are valued not only because of their ability to produce different agricultural outputs, but also because they can become rural ecological public goods. Natural amenities and other qualities of a local environment that improve living conditions have been identified as key factors for the growth of rural population, income and employment. The subject of research in the current doctoral dissertation is an analysis of the impact of the rural amenities and entrepreneurship on rural development, with a special emphasis on Serbia. Improving the quality of human capital and the growth of entrepreneurial activities in the models of rural economy growth can be seen as one of the possible mechanisms of transformation, growth and development of rural economy. The basic goals of research in the current doctoral dissertation are connected with the classification of rural areas in Serbia with respect to outdoor amenities and the analysis of the influence of the concept of human capital and entrepreneurship on the economic growth of rural areas in our country. In the developed economies, rural amenities have the power to attract significant human and entrepreneurial capital necessary for the development of rural economy. Rural areas with a higher share of human capital will have better economic performance due to innovations, high level of entrepreneurship and creative forms of Погодности амбијента, предузетништво и рурални развој Србије Докторска дисертација Соња Јосиповић iv businesses. Rural areas with high outdoor amenities attract new rural residents who are focused on entrepreneurship activities.
true
Semantička integracija senzorskih mreža
Semantic-based integration of sensor networks
azvoj CMOS tehnologije visokog stepena integracije prouzrokovao je intenzivan razvoj senzorskih tehnologija, bežičnih komunikacija i energetski efikasnih procesora, koji zajedno čine senzorske čvorove, sposobne da opažaju našu okolinu, vrše obradu podataka i da razmenjuju podatke sa drugim uređajima i korisnicima na Internetu. Takav tehnološki razvoj je doveo do pojave vizije Interneta Stvari (eng. Internet of Things - IoT), sa ciljem da se korisnicima na Internetu pruže informacije iz realnog sveta koji nas okružuje. Kao jedan od preduslova za realizaciju zamišljenih IoT servisa i proizvoda, neophodna je horizontalna integracija raspoređenih senzorskih mreža, implementacijom platformi koje bi omogućile integraciju heterogenih senzorskih mreža, sastavljenih od različitih senzorskih uređaja, koji koriste različite komunikacione protokole i formate poruka i u mogućnosti su da opslužuju veći broj korisnika. Predmet ove disertacije su arhitekture za integraciju senzorskih mreža, koje u opštem slučaju pružaju podršku radu IoT aplikacijama sa zahtevima za visokim performansama. Od osnovnih generičkih tipova, izdvojena je arhitektura zasnovana na semantici podataka i cilj ove disertacije je analiza arhitektura koje omogućavaju semantičku integraciju senzorskih mreža korišćenjem semantičkih veb tehnologija radi omogućavanja interoperabilnosti senzorskih podataka i mreža. U radu su najpre identifikovani osnovni tipovi generičkih arhitektura i date su njihove ključne karakteristike i način realizacije, a za arhitekture zasnovane na prolaznom uređaju i brokeru poruka urađena je evaluacija performansi u radu sa podacima u realnom vremenu kreiranjem simulacionog okruženja. Zatim su arhitekture sa semantičkom integracijom senzorskih mreža klasifikovane identifikovanjem tipičnih projektanskih pristupa. Izdvojene su dve grupe pristupa, i to pristupi orijentisani ka senzorskim mrežama i aplikativno orijentisani pristupi, a svaka grupa sadrži dalje po četiri tipa arhitekturalnih pristupa. Za svaki arhitekturalni tip data je analiza prednosti i nedostataka korišćenog pristupa i kratak opis konkretnih predstavnika. Na osnovu vi uočenih nedostataka analiziranih arhitektura, predložena je nova arhitektura koja koristi pristup zasnovan na distribuiranim bazama podataka. U srži te arhitekture je distribuirano RDF skladište visokih performansi, specijalizovano za rad sa semantičkim senzorskim podacima u operativnoj memoriji, u cilju podrške efikasnog skladištenja i pretraživanja semantičkih senzorskih podataka u uslovima intenzivnog generisanja novih senzorskih opažanja sa većeg broja senzora. U predloženom RDF skladištu dat je niz projektantskih rešenja kako bi se ostvarile željene performanse, uključujući način kreiranja indeksnih struktura za čuvanje i pretragu RDF trojki, predikatski zasnovanu strategiju distribucije RDF trojki po računarskim čvorovima, algoritam kreiranja plana izvršavanja SPARQL upita i način održavanja višedimenzionalnih indeksnih struktura za čuvanje rezultata merenja, vremenskih i prostornih podataka senzorskih opažanja. Evaluacija performansi je izvedena sa skupom podataka senzorskih merenja sa nekoliko hiljada meteoroloških stanica u SAD za vreme uragana Čarli, tokom 2004. godine, simuliranjem situacije generisanja senzorskih opažanja u realnom vremenu. Izmerena vremena izvršavanja vremensko-prostornih-vrednosnih semantičkih upita paralelno sa intenzivnim dodavanjem novih senzorskih opažanja pokazuju da je realizovano RDF skladište veoma efikasno i da postiže znantno bolje performanse od javnog dostupnog RDF skladišta Virtuoso, u rasponu od 3 do 5 puta kraće vreme izvršavanja upita i to samo pri najnižem opterećenju. Sa većom količinom senzorskih podataka vreme izvršavanja upita kod Virtuoso skladišta se nekontrolisano povećava, dok realizovano RDF skladište efikasno skalira i blago povećava vreme izvršavanja upita.
The progress of CMOS technology very large scale integration has resulted in the intensive development of sensor technologies, wireless communications, and energy efficient processors, which together make sensor nodes capable of perceiving our environment, processing data and exchanging data with other devices and users on the Internet. Such technological development has led to the emergence of the Internet of Things (IoT) vision, with the goal of providing information of the real world that surrounds us to the Internet users. As one of the prerequisites for the realization of imaginary IoT services and products, the horizontal integration of deployed sensor networks is required, by implementing platforms that are able to serve a large number of users and would allow the integration of heterogeneous sensor networks consisted of different sensor devices, using different communication protocols and message formats. The subject of this dissertation is the architectures for the integration of sensor networks, which generally support IoT applications with high performance requirements. Among the available generic types, an architecture based on the data semantics is selected and the aim of this dissertation is the analysis of the architectures that enable semantic integration of sensor networks using Semantic Web technologies in order to achieve the interoperability of sensor data and networks. The basic generic architecture types were first identified and their key characteristics and method of implementation were given. For architectures based on the gateway and the message broker, the performance evaluation was performed in the scenario of real-time sensor messages delivery by creating a simulation environment. Then the architectures with semantic-based integration of sensor networks are classified by identifying typical design approaches. Two groups of approaches are identified, approaches oriented to sensor networks and application- oriented approaches, whereas each group contains four architectural types. For each architectural type, an analysis of the advantages and disadvantages of the used approach is given, as well as a brief description of the concrete representatives. As a result of observed shortcomings of the analyzed architectures, a new architecture is proposed which uses an approach based on distributed databases. At the core of this architecture is a distributed high-performance RDF triple store specialized for viii working with semantic sensor data in RAM, in order to efficiently support storage and search of semantic sensor data in the conditions of intensive generation of new sensory observations from a large number of sensors. In the proposed RDF triple store, a number of innovative techniques are implemented in order to achieve the desired performance. Such techniques include the method of creating index structures for storing and searching RDF triples, the predicate-based distribution strategy of RDF triples to computer nodes, the algorithm of the SPARQL query execution plan, and the multidimensional indexing structures for keeping measurements, temporal, and spatial data of sensor observations. Performance evaluation was performed with a dataset of sensor data measurements from several thousands of meteorological stations in the United States during hurricane Charlie in 2004, simulating the situation of real-time sensor sensing. The measured execution time of semantic spatio-temporal-thematic queries over a different volume of sensor data shows that the implemented RDF triple store is very efficient and achieves significantly better performance than the publicly available RDF triple store Virtuoso, in the range from 3 to 5 times shorter query execution time, but only at the lowest workload. By increasing sensor data volume, the query execution time at the Virtuoso triple store increases uncontrollably, whereas the realized RDF triple store effectively scales and slightly increases the query execution time.
true
Metodologija projektovanja sistema za inteligentno prosleđivanje pitanja napisanih na prirodnom jeziku
A methodolology for the intelligent question routing systems design
Uprkos velikom razvoju veštačke inteligencije, ljudski mozak je još uvek superiorniji uzimajući u obzir razumevanje i manipulaciju delimično poznatim činjenicama. Jedna od oblasti u kojoj je ovo naročito istaknuto odnosi se na problem odgovaranja na pitanja. Kada je potrebno dati odgovor, posebno onaj koji se eksplicitno ne nalazi u tekstu korpusa koji se pretražuje, prednosti eksperta – čoveka predstavljaju različite sposobnosti kao što je sposobnost objašnjavanja, kombinovanja složenih odgovora i apstraktnog rezonovanja. Sistemi za Inteligentno Prosleđivanje Pitanja (SIPP) imaju za svrhu razmenu znanja na nivou proizvoljne oblasti ekspertize i od značaja su za veliki broj aplikacija u kojima se zahteva intenzivna komunikacija između korisnika. Korist od primene ovakvih sistema uključuje: (a) smanjenje nepotrebnog opterećenja eksperata koji predstavljaju vredan resurs i (b) povećanje kvaliteta usluga institucije (univerzitet, vlada, preduzeće), imajući u vidu zadovoljstvo korisnika s obzirom da su njihova pitanja prosleđena relevantnim osobama. U ovom radu predstavljena je metodologija projektovanja sistema za inteligentno prosleđivanje pitanja napisanih na prirodnom jeziku – SIPP. Na početku je dat detaljan pregled ove oblasti gde je poseban akcenat stavljen na realizaciju faza SIPP procesa. Takođe, u ovom poglavlju predstavljena je originalna prezentaciona paradigma koja generalizuje suštinu svih raspoloživih SIPP rešenja iz otvorene literature. Prezentaciona paradigma sadrži tri osnovne faze izvršavanja koje se odnose na tri glavna problema prilikom realizacije sistema: analizu pitanja, prosleđivanje pitanja i profilisanje korisničkog znanja. Na osnovu ove paradigme izvršena je detaljna 7 analiza i evaluacija ovakvih sistema, a kao zaključak naveden je predlog rešenja uočenih problema. U nastavku disertacije opisana je realizacija predloženih rešenja u vidu prototipa SIPP sistema. U okviru modula za obradu pitanja realizovan je pristup koji omogućava vizualizaciju pitanja, što obezbeđuje intuitivnu predstavu specifičnih odnosa između koncepata, kao i njihovog značaja u pitanju. Takođe, ovaj pristup kombinuje potpuno automatsku obradu teksta i ručnu korekciju rezultata, pružajući korisniku mogućnost povećanja tačnosti izlaza. Istovremeno, realizovani modul za obradu teksta upotrebljen je i za analizu odgovora. Nakon toga analizirani su i diskutovani postojeći pristupi za određivanje semantičke sličnosti dva kratka teksta, pogodni za jezike sa vrlo ograničenim elektronskim lingvističkim resursima, gde je poseban akcenat stavljen na srpski jezik. Na osnovu donetih zaključaka predložen je novi algoritam, nazvan LInSTSS, koji prilikom određivanja semantičke sličnosti dva kratka teksta uzima u obzir i specifičnosti reči koje ovi tekstovi sadrže. Takođe, realizovan je korpus parafraza za srpski jezik nad kojim je izvršena evaluaciju. Rezultati dobijeni nad ovim korpusom pokazali su da predloženi algoritam pruža bolje rezultate u odnosu na postojeća rešenja. Konačno, na osnovu evaluacije nad korpusima parafraza za srpski i engleski izvršeno je fino podešavanje parametara, a takođe stečena iskustva upotrebljena su za realizaciju modula za određivanje semantičke sličnosti između pitanja i korisničkog profila. U okviru realizacije faze prosleđivanja pitanja, diskutovane su specifičnosti problema poređenja pitanja i korisničkih profila, i predložen je novi algoritam nazvan P2Q. Ovaj algoritam određuje najveću (maksimalnu) sličnost između koncepata identifikovanih u pitanju i onih koji se nalaze u korisničkom profilu. Konačno, analizirani su dostupni veb portali i odabran je jedan čiji podaci se upotrebljeni za formiranje korpusa pitanja i odgovora. U 8 korpusu su izdvojene tri različite vrste korisnika, koje modeluju: (1) interesovanje, (2) znanje i (3) istovremeno i znanje i interesovanje. Formirani korpus je zatim upotrebljen za evaluaciju celokupnog sistema i testiranje polaznih hipoteza. Dobijeni rezultati su pokazali da P2Q pristup pruža znatno bolje rezultate u odnosu na ostale evaluirane pristupe. Takođe, uočeno je da upotreba semantičke ekstrakcije informacija iz teksta može poboljšati rezultate. Doprinos izložene doktorske disertacije je u domenu analize i sinteze jednog ovakvog softverskog sistema, koji treba da omogući inteligentno prosleđivanje pitanja napisanih na prirodnom jeziku. Takođe, na osnovu rezultata evaluacije zaključeno je da pravilno dodeljene težine mogu poboljšati performanse celokupnog sistema, ali takođe u slučaju da nisu pravilno postavljene mogu ih znatno pogoršati. Konačno, utvrđeno je da pri profilisanju kompetentnosti korisnika da pruži odgovor na postavljeno pitanje, nije važno samo razmotriti njegove najbolje odgovore, odnosno profilisati njegovo znanje, već je takođe potrebno uzeti u obzir i pitanja koja je postavio, s obzirom da ona mogu izraziti interesovanje.
In spite of great developments in artificial intelligence, human brain is still more powerful, concerning the comprehension and manipulation with partially known facts. One domain where this is prominent is related to a problem of question answering. When it comes to giving the answers, especially those that do not explicitly exist in the text corpus, the advantages of a human expert are abilities like explaining, combining complex answers, and abstract reasoning. Intelligent Question Routing Systems (IQRS) serve as a knowledge exchange medium in an arbitrary field of expertise, where intensive communication between users is required. The benefit coming from deployment of such systems includes: (a) reducing unnecessary “pinging” of experts, which are a valuable resource and (b) increasing the system owners’ (e.g. enterprise, government, university) quality of service, since users are more satisfied with answers, because their questions are answered by the right persons. This dissertation represents a methodology for IQRS systems design. It starts with a survey of the existing research in this domain, where the emphasis was put on the implementation of phases of the IQRS process. The survey also introduces an original presentation paradigm that generalizes the essence of approaches found in the open literature. The presentation paradigm includes three basic processing stages related to the three major problems of system implementation: question analysis, question forwarding, and users’ knowledge profiling. The outcome of this analysis is a proposal for a new approach that tackles identified problems. The rest of the dissertation describes an IQRS prototype which implements the proposed ideas. The question analysis module implements an approach that enables question visualization, thus it provides an intuitive representation of specific relations 10 between concepts and their importance in the question. Also, this approach combines a fully automatic text processing and manual correction of the results, giving a users ability to increase the accuracy of the output. Simultaneously, the implemented text processing module is used for the analysis of answers. Next, an analysis and discussion of existing approaches for determining the semantic similarity of two short texts is given, where particularly the focus was put on those suitable for languages with very limited electronic linguistic resources, like the Serbian language. Based on the conclusions a new algorithm is proposed, named LInSTSS, which for calculating the semantic similarity between two short texts includes the specificity of words that these texts contain. Additionally, a Serbian paraphrase corpus is constructed and the results obtained using this corpus showed that the proposed algorithm provides better results when compared to existing solutions. Finally, evaluation on paraphrase corpora both for English and Serbian is used to fine-tune algorithm parameters. Also, the lessons learned are applied to design the module for calculating the semantic similarity between a question and a user profile. In the question forwarding phase issues related to the problem of comparing questions and user profiles are discussed and a new algorithm called P2Q is proposed. This algorithm determines the highest (maximum) similarity between the concepts identified in question and those from the user profile. Finally, the analysis of available web portals is carried in order to find the one suitable for the creation of a questions and answers corpus. In the created corpus three different types of users are extracted, which model: (1) interests, (2) knowledge, and (3) both knowledge and interests. Corpus is then used to test the initial hypothesis and to evaluate the overall system performances. The results showed that P2Q approach provides significantly better results than other evaluated approaches. It was also noted that the use of semantic information extraction from text can improve results. Scientific contribution of the dissertation is in the field of analysis and synthesis of a software system, which should enable an intelligent questions routing. Also, based on the evaluation results it was found that properly assigned weights can 11 improve the overall performances of the system, but also if not assigned correctly performances can be significantly decreased. Finally, it was concluded that for profiling the user competence to give an answer for the provided question it is important not only to consider answers which the user best answered, i.e. to profile knowledge, but also it is important to consider questions which can express interests.
true
Geometrijska generalizacija Ajnštajnove teorije gravitacije
Geometric generalization of einstein theory of gravity.
Ajnxtajnova teorija gravitacije uspexno opisuje pojave u Sunqevom sistemu. Ona takodje predvia postojanje crnih rupa, gravitacionih sočiva i gravitacionih talasa, xto je uspešno opservirano. Medjutim Ajnštajnova teorija nije dovoljno proverena na velikim kosmičkim rastojanjima. Zbog toga posmatramo nelokalnu modifikaciju gravitacije i dobijamo nova rešenja za skalirajunje faktor a(t). Takodje, posmatramo i prostorno-vremenske pertur- bacije de Siterovog prostora.
Einstein theory of gravity successfully describes the Solar system. It also predicts the existence of the black holes, gravitational lenses and gravitational waves, which have been observed successfully. On the other hand Einstein theory of gravity is not tested on the large cosmic sccale. Therefore, we consider the nonlocal modied gravity and get new solutions for the cosmic scale factor a(t). Moreover we consider space-time perturbations of the de Sitter space.
false
Analiza kardiosignala pomoću druge generacije talasić
Analysis of cardiosignals using second generation wavelets.
Današnja istraţivanja vezana za kardiovaskularne bolesti usmerena su ka raĉunarskoj analizi kardiosignala, digitalnih zapisa kardiovaskularne aktivnosti i funkcionalnosti, imajući u vidu dva aspekta: ekonomski i telemedicinski. U cilju izbegavanja komplikovanih i skupih sistema za analizu kardiološkog stanja, efikasne tehnike automatske (mašinske) obrade signala imaju veliki znaĉaj za potrebe primarne nege. ObezbeĊivanje visokokvalitetnog signala koji opisuje stanje organizma, radi dobijanja informacije o patofiziologiji, predstavlja prioritet. U ovoj disertaciji se pod kardiosignalima podrazumevaju elektrokardiogram (ECG) i fonokardiogram (PCG), kao tzv. primarni kardiosignali, koji predstavljaju reprezentativne primere elektriĉnog i vibroakustiĉnog zapisa rada kardiološkog sistema. Talasna (vejvlet) transformacija se ĉesto koristi za potrebe reprezentacije sloţenih signala, pored ostalih i kardiosignala, i izdvajanja njihovih vaţnih karakteristika. Ova transformacija je otvorila vrata razvoju razliĉitih tehnika koje mogu biti od znaĉaja za konkretne primene, kao što je kliniĉka praksa u medicini. Nakon uvoĊenja talasne transformacije, javila su se i poboljšanja u vidu tzv. druge generacije talasića. Druga generacija predstavlja samo alternativan pristup za implementaciju diskretne talasne transformacije, dok je lifting šema efikasan algoritam za konstrukciju FIR filtar banki zasnovanih na talasićima (vejvletima). U ovoj disertaciji je pretpostavljeno da se upotrebom talasne transformacije i lifting šeme moţe efikasno analizirati kardiosignal. Naime, pretpostavljena je veza izmeĊu kardiosignala i primene diskretne talasne transformacije u cilju bolje vizuelizacije njegovog sadrţaja. Za takvu namenu znaĉajni su spektralni i spektrogramski prikazi signala. Zdruţena vremensko-frekvencijska transformacija, kao što je kontinualna talasna transformacija, multifraktalan spektar, ali i empirijski pristup, moţe biti primenjen u analizi kardiosignala. Shodno potrebi analize, neophodna su rešenja koja omogućavaju izdvajanje i/ili detektovanje srĉanih dogaĊaja od kliniĉke vaţnosti. Neadekvatna primena diskretne talasne transformacije u redukciji šuma utiĉe na visokofrekvencijske komponente kardiosignala. Ovakve komponente ne moraju biti amplitudski dominantne i jasno uoĉljive, a mogu biti kliniĉki relevantne u dijagnostici. Manifestacija klik-sindroma u fonokardiografiji je dobar primer. Za potrebe istraţivanja obavljena je akvizicija fonokardiograma pedijatrijskih pacijenata, zdravih i pacijenata sa prolapsom mitralne valvule (PMV). Njihovom upotrebom i analizom, predloţen je model za vizuelnu verifikaciju nalaza primenom diskretne talasne transformacije. Upotrebljena je interpretacija detalja pomoću greške predikcije i njihova efikasna lokalizacija u vremenskom domenu. Izbor reda prediktora kontrolisan je detekcijom dominantnih srĉanih zvukova i indikacijom potencijalnih abnormalnosti. Ovakva kontrolisana analiza smanjuje rizik zanemarivanja kliniĉki vaţnih informacija, kao što je klik-sindrom. Predloţeni model i eksperimentalni rezultati ukazuju na nekoliko mogućnosti za dalja istraţivanja i razvoj samopodešavajućih pristupa za analizu kardiosignala.
Current research of cardiovascular diseases is oriented towards the computer analysis of cardiosignals, digital records of cardiovascular activity and functionality, bearing in mind two aspects: economic and telemedical. In order to avoid complex and expensive systems for cardiovascular state analysis, efficient automatic (machine) signal processing techniques are of great importance in primary care. Providing high-quality signal which decribes the state of the organism, in order to obtain information about the pathophysiology, is a priority. In this thesis, electrocardiogram (ECG) and phonocardiogram (PCG) are refered to as cardiosignals, so called primary cardiosignals, which are representative examples of electrical and vibro-acoustical records of cardiac system functionality. Wavelet transform is often used in representation of complex signals, among others cardiosignals, and their valuable feature extraction. This transform opened the door for development of different techniques that can be of importance in particular applications, such as in clinical practice in medicine. Following the wavelet transform, novel enhancement approches are introduced, like so called second generation wavelets. The second generation represents an alternative approach for discrete wavelet transform implementation, where a lifting scheme is an efficient algorithm for FIR wavelet filterbank construction. In the thesis it is assumed that wavelet transform and lifting scheme can be used for efficient cardiosignal analysis. Namely, a relationship between cardiosignal and application of discrete wavelet transform is hypothesized in order to improve the visualization of its content. For such purpose, spectral and spectrogram based representation of a signal are significant. Joint time-frequency representation, such as continious wavelet transform, multifractal spectrum, as well as empirical approach, can be applied in cardiosignal analysis. In accordance with the analysis, it is inevitable to develop approaches that enable extraction and/or detection of cardiac events of clinical importance. Inadequate application of discrete wavelet transform in noise reduction may affect high frequency cardiosignal components. Such components are not necessarily either characterized by dominant amplitude or even clearly noticeable, but they are usually clinically relevant in diagnostics. Click-syndrome manifestation in phonocardiorgaphy is such an example. For the purpose of research, phonocardiograms are recorded from pediatric patients: healthy patients and patients with prolapsed mitral valve (PMV). Their use and analysis have led to the proposition of a model for visual click finding verification via discrete wavelet transform. Details are interpreted through prediction error and efficiently localized in time domain. The choice of prediction order is carried by detection of dominant heart sounds and potential abnormality indication. Such controlled analysis decreases the risk of neglecting clinically important information, such as the click-syndrome. Proposed model and experimental results show several possibilities for further investigations and development of self-adjusting approaches for cardiosignal analysis.
false
Numerička simulacija krize ključanja u isparivačkim cevima
Numerical simulation of boiling crises mechanism in evaporation pipes
Predmet ovog rada je numerička simulacija i analiza trodimenzionalnog dvofaznog strujanja i istraživanje mehanizama krize ključanja u isparivačkim cevima. Kriza ključanja je nepovoljna pojava. Nagli porast temperature zida cevi izaziva termomehanička naprezanja, koja mogu dovesti do pojave prskotina i pucanja cevi pod pritiskom. Razvijen je matematički model dvofaznog strujanja u isparivačkoj cevi,koji se sastoji od jednačina održanja mase, količine kretanja i energije za obe faze. Sastoji se od sistema parcijalnih diferencijalnih jednačinakoje su rečavane za definisane granične uslove.Bilo je neophodno uvesti i modeliranje i numeričku simulaciju na mikro nivou, jer je bilo potrebno ispitati i proces pojave i rasta mehura, kao i ponašanje dvofazne mešavine na zagrejačkoj površini na makroskopskom nivou. Rezultati numeričkih simulacija su upoređeni sa raspoloživim eksperimentalnim rezultatima.
The subject of this thesis is numerical simulation and analyses of three dimensional two-phase flow and mechanism of boiling crises in evaporation pipes. Boiling crises is characterized by a dried out heat surface and can bring physical destruction of the heater. A developed mathematical model of two phase flow in evaporation pipe consists of prescribed mass, momentum and energy conservation equations for liquid and vapor two-phase flow. It consists of set of partial differential equations which were solved for specific boundary conditions. Modeling of the micro scale level was necessary to take into account processes of the bubble rise and behavior of the two-phase mixture at the heating surface. Results of the numerical simulation are compared with similar available results of the experiments from the literature.
false
Metode i instrumentacija za procenu aktivnosti motornog sistema na osnovu elektromiografskih signala
Methods and instrumentation for assessment of motor function by the means of electromyography
Metoda merenja i analize električnih potencijala snimljenih površinskim elektrodama sa skeletnih mišića naziva se elektromiografija. Primenom ove metode omogućeno je praćenje mišićne i nervne aktivnosti, a samim tim omogućena je i direktna ocena motornog i indirektna ocena senzornog sistema. Cilj istraživanja je usavršavanje tehnologije merenja i analize elektromiografskih signala koje bi omogućile razvoj novih modela u senzorno-motornom sistemu čoveka. Novorazvijeni sistem, koji je prikazan u ovoj disertaciji, ima dve osnovne namene: podrška bazičnim istraživanjima u domenu neuronauka i podrška kliničkom radu. Prikazana je efikasnost sistema u oceni senzorno-motornog statusa kod pacijenata koji su preživeli moždani udar. Kroz 2 prilaza u merenju i analizi elektromiografskih signala prikazano je kako se može oceniti status senzorno-motornog sistema kod zdravih i kod pacijenata. Prvi prikaz se sastoji u proučavanju spontane mišićne aktivnosti tokom kontrolisanog voljnog pokreta članka i tokom hoda. Ovaj prikaz je iskorišćen za proučavanje kratkoročnih i dugoročnih efekata terapije na senzorno-motorni oporavak pacijenata. Drugi prikaz se sastoji iz proučavanja evociranih motornih potencijala. Na primeru zdravih ispitanika pokazano je kako se na pouzdan i ponovljiv način može iskoristiti projektovana instrumentacija za ocenu senzorno- motornog sistema. U analizi mišićnih signala predstavljene su dve nove metode proučavanja spontanih i evociranih potencijala mišića. Za potrebe ocene spontane mišićne aktivnosti uvedene su nove mere kokontrakcije koje omogućavaju ocenu mišićnih aktivacija i koje omogućavaju istovremenu ocenu vremenskih i amplitudskih osobina signala. U oceni evociranih potencijala uvedeni su novi algoritmi za automatsko upravljanje električnom stimulacijom i akvizicijom signala, kao i novi vid vizuelizacije i ocene prostorne raspodela evociranih potencijala. Klinička primena projektovanog mernog sistema i V procedura merenja i analize elektromiografskih signala ima poseban značaj jer omogućuje praćenje oporavka u procesu rehabilitacije i može da pomogne u izboru terapijskih modaliteta. Funkcionalna mera oporavka koja se zasniva na elektromiografiji ima zadatak da bude dopuna konvencionalnim kliničkim skalama koje u većini slučajeva nisu dovoljno osetljive da bi ukazale na izvore patologija i na promene stanja pacijenta tokom lečenja.
Electromyography is a method for measurement and analysis of electrical potentials of muscles in human. Nervous and muscular systems in humans can be assessed by the means of electromyography. This assessment enables direct evaluation of motor status, and indirect evaluation of sensory status. The aim of this research is the development of new technologies for measurement and analysis of electromyography signals, which would enable design of new sensory-motor models. The novel system, presented in this dissertation, has two main purposes: to improve research methodology in basic neurosciences and to support clinical practice. The functioning of the system is presented for the assessment of sensory-motor recovery in stroke patients. Two main approaches are used in measurement and analysis of electromyography signals in healthy subjects and in patients. The first approach consists of studying voluntary contractions by the means of electromyography during ankle movement and during gait. This approach is used for the studying long-term and short-term therapy effects on sensory-motor system during rehabilitation process in stroke patients. The second approach enables measurement and analysis of evoked motor potentials. Reliability and reproducibility of a designed instrumentation for assessment of motor status by studying evoked motor potentials is presented in healthy subjects. Two new methods for the analysis of voluntary muscle contraction and evoked potentilas are presented. For studying voluntary muscle contraction, the cocontraction measurement was introduced. These parameters reflected amplitude and temporal characteristics of recorded signals. For the analysis of evoked motor potentials new algoritms for the control of electrical stimulation and acquisition are presented. The novel method for vizualisation and assessment of spatial distribution of evoked potential is designed for the purposes of sensory-motor assessment. Clinical applications of the designed instrumentation, measurement procedures and analysis of electromyography VII signals have an impact in modern clinical practice. It enables assessment of recovery progress and rehabilitation procedures. Developed functional measurement of recovery by the means of electromyography should be used as an additional assessment method to stadard clinical scales, since in most cases those scales have less sensitivity in studying senzory-motor status than functional measurements.
true
Uticaj dimenzija ličnosti na kardiovaskularnu reaktivnost kod pacijenata sa dijagnozom koronarne bolesti
The effects of personality dimensions on cardiovascular reactivity in patients with coronary artery disease
UVOD: Jedna od mera stresne reakcije zdravih i obolelih osoba od koronarne bolesti je kardiovaskularna reaktivnost koja se definiše kao promena veličine kardiovaskularnih parametara (npr. puls, sistolni pritisak) tokom izlaganja fizičkom ili psihološkom stresoru, u odnosu na njihove vrednosti u stanju mirovanja. Na kardiovaskularna reaktivnost utiču mnogobrojni činioci kao što su genetska predispozicija, pol, životno doba, vrste stresora, strategije prevladavanja stresa, aktuelno raspoloženje, itd. Primenom različitih inventara ličnosti potvrđeno je da i pojedine dimenzije ličnosti značajno dorinose individualnim razlikama kardiovaskularne reaktivnosti na stres. CILJ: Glavni cilj našeg istraživanja bio je da se ispita uticaj opštih dimenzija ličnosti i Tipa D ličnosti na kardiovaskularnu reaktivnost kod pacijenata sa dijagnozom koronarne bolesti, uzimajući u obzir potvrđene moderatore (pol, depresivnost, stilovi prevladavanja stresa). METOD: Uzorak je činilo je 100 ispitanika izabranih metodom prigodnog uzorkovanja, podeljenih u dve grupe: a) studijska grupa (n=79) kojom su obuhvaćeni pacijenati sa dijagnozom stabilne koronarne bolesti i b) kontrolna grupa zdravih ispitanika (n=21) odgovarajućih sociodemografskih karakteristika. Istraživanje je sprovedeno u periodu od 2001. do 2007. godine u Klinici za kardiovaskularne bolesti KCS. Kod svih ispitanika izveden je mentalni stres test kombinovan sa testom fizičkim opterećenjem, u saradnji sa kardiologom, što je praćeno i ehokardiografski. Eksperimentalna faza istraživanja. Mentalni stres test kombinovan sa testom fizičkim opterećenjem izvođen je kod svih ispitanika u podne (1200h), u kontrolisanim klimatskim uslovima. Test se sastojao iz tri faze: a) psihijatrijski interviju, neposredno pre eksperimenta; b) faza odmora: u tihoj i delimično zamračenoj prostoriji ergo-kabineta ispitanik se 10-15 minuta odmarao u ležećem položaju; c) test mentalnog stresa koji obuhvata dva zadatka: usmeno računanje i zadatak koji reaktivira ljutnju. Instrumenti procene. Prvo je sproveden polustrukturisani intervju u cilju prikupljanja podataka relevantnih za istraživanje (sociodemografske karakteristike, faktori rizika za koronarnu bolest). Nakon laboratorijskog testa i psihijatrijskog intervjua, svi ispitanici popunili su bateriju testova koja je uključivala sledeće instrumente: a) Revidiran nov upitnik ličnosti, forma S (NEO Personality Inventory - Revised, NEO-PI-R) za pocenu dimenzija ličnosti; b) Skala D14 (The D-Scale 14, DS- 14) za procenu komponenti distres ličnosti - Negativni afektivitet i Socijalna inhibicija; c) Upitnik o stilovima prevladavanja stresa (Ways of Coping Questionnaire - WCQ-R); d) Bekova skala za procenu depresije (Beck Depression Inventory - BDI). vii REZULTATI: Rezultati ovog istraživanja pokazali su da ispitanike sa dijagnozom koronarne bolesti karakteriše viši nivo Neuroticizma i niži nivo Ekstraverzije i Otvorenosti za nova iskustva u odnosu na grupu ispitanika zdrave kontrolne grupe. Nisu nadjene statistički značajne razlike na preostalim dimenzijama NEO-PI-R-a. Daljom analizom, utvrdjeno je da je učestalost Tipa D ličnosti statistički značajno viša u grupi obolelih od koronarne bolesti u odnosu na zdrave ispitanike. Na dimenzijama Negativni afektivitet (NA) i Socijalna inhibicija (SI), prosečne vrednosti bile su statistički značajno više kod ispitanika sa dijagnozom koronarne bolesti u odnosu na zdrave ispitanike. Kod ispitanika sa koronarnom bolešću, dimenzija Negativni afektivitet bila je statistički značajno pozitivno udružena sa Neuroticizmom, dok je udruženost sa Ekstraverzijom, Savesnošću i Saradljivošću bila, takodje, značajna ali u ovom slučaju odnos je inverzan. Neuroticizam i Ekstraverzija objašnjavaju 48% varijanse dimenzije Negativni afektivitet. Pored toga, kod ispitanika sa koronarnom bolešću, dimenzija Socijalna inhibicija visoko statistički značajno pozitivno korelira sa Neuroticizmom, a negativno sa Ekstraverzijom. Registrovana je statistički značajna negativna udruženost izmedju socijalne inhibicije i Otvorenosti za nova iskustva i Savesnosti. Daljom analizom utvrdjeno je da Neuroticizam i Ekstraverzija objašnjavaju 38% varijanse dimenzije socijalne inhibicije. Prosečna vrednost depresivnosti (procenjena BDI upitnikom) bila je viša kod ispitanika sa dijagnozom koronarne bolesti, u odnosu na zdrave kontrolne ispitanike, ali bez statističke značajnosti. Kod ispitanika sa dijagnozom koronarne bolesti koji su zadovoljili kriterijum za Tip D ličnosti prosečna vrednost depresivnosti bila je statistički značajno viša u odnosu na ispitanike iste grupe bez Tipa D ličnosti. Sličan trend zapažen je i kod zdravih ispitanika, ali dobijene razlike nisu bile statistički značajne. Depresivnost je visoko značajno pozitivno korelirala sa Neuroticizmom, dok je i korelacija izmedju depresivnosti i Ekstraverzije bila visoko statistički značajna, ali u negativnom smeru. Ispitanici sa dijagnozom koronarne bolesti češće su angažovali strategije orijentisane na emocije (npr. distanciranje, samokontrolu i izbegavanje), dok su zdravi ispitanici predominantno koristili aktivne strategije (npr. plansko rešavanje problema i pozitivnu reevaluaciju). Statistički značajna razlika između ispitivanih grupa bila je prisutna na skalama izbegavanje, socijalna podrška i plansko rešavanje problema. Osobe sa koronarnom bolešću manifestovale su izbegavajući stil ponašanja, ređe su tražili socijalnu podršku a manje koristili analitički pristup rešavanju problema. Kod ispitanika sa dijagnozom koronarne bolesti utvrdjeno je da su dimenzije Savesnost i Saradljivost bili prediktori kardiovaskularne reaktivnosti tokom izvođenja zadatka mentalne aritmetike. Tokom izvođenja zadatka koji izaziva ljutnju, prediktori kardiovaskularne reaktivnosti bili su, takođe, Savesnost i Saradljivost, ali i Neuroticizam. Dimenzija Otvorenost za nova iskustva izdvojila se kao prediktor povišenja dijastolnog krvnog pritiska tokom izvođenja oba zadatka. Uticaj Tipa D ličnosti na promenu vrednosti srčane frekvencije tokom eksperimenta nije bio statistički značajan. Dalja, intraindividualna analiza ukazala je da je između promene vrednosti srčane frekvencije i Tipa D ličnosti izmedju faze 1 i faze 2 tokom eksperimenta, registrovana statistički visoko značajna interakcija. Dakle, uočeno je sniženje vrednosti srčane frekvencije nakon faze odmora kod ispitanika sa Tipom D ličnosti značajno manje u odnosu na ispitanike bez Tipa D ličnosti (produžen period oporavka, „viši tonus“). Nasuprot tome, uticaj Tipa D ličnosti na promenu vrednosti sistolnog krvnog pritiska viii tokom eksperimenta bio je statistički značajan. Početne prosečne vrednosti sistolnog pritiska bile su veće kod ispitanika sa Tipom D ličnosti, da bi tokom izvodjenja testa mentalnog stresa (pre svega zadatka koji aktivira ljutnju), prosečne vrednosti sistolnog pritiska bile gotovo identične kod obe grupe ispitanika. Kod ispitanika sa Tipom D ličnosti zapažena je slabija reaktivnost sistolnog pritiska u odnosu na ispitanike bez Tipa D ličnosti. Tip D ličnosti nije uticao na promenu vrednosti dijastolnog pritiska tokom eksperimenta. Dalja intraindividualna analiza pokazala je da je uticaj Tipa D ličnosti na vrednost dijastolnog pritiska statistički značajan između faze 1 i faze 2, kao i izmedju faze 2 i faze 3 eksperimenta. Dakle, sniženje vrednosti dijastolnog pritiska nakon faze odmora kod ispitanika sa Tipom D ličnosti bilo je značajno manje u odnosu na ispitanike bez Tipa D ličnosti (produžen period oporavka). Ujedno, kod ispitanika sa Tipom D ličnosti zapažena je slabija kardiovaskularna reaktivnost u odnosu na ispitanike bez Tipa D ličnosti (manja promena vrednosti dijastolnog pritiska-početna i dijastolnog pritiska - tokom oba zadatka). ZAKLJUČAK: Na osnovu raspoložive literature, rezultati našeg istraživanja su prvi koji ukazuju na sniženu kardiovaskularnu reaktivnost kod pacijenata sa koronarnom bolešću koje su sklone distresu (Tip D ličnosti). Pored toga, pokazali smo da specifične dimenzije ličnosti (npr. Savesnost i Saradljivost) značajno doprinose povišenju kardiovaskularne reaktivnosti. Stoga je u cilju jasnijeg rasvetljavanja faktora rizika za razvoj, tok i prognozu koronarne bolesti, neophodna je saradnja između kardiologa i psihijatara u oblasti bihejvioralne kardiologije kao i pristup orijentisan ka osobi.
Introduction: Over the past few decades, cardiovascular reactivity (CVR) has come to be viewed as a meassure of the stress reaction in patients with coronary heart disease as well as in healthy subjects. CVR is defined as a change of the measures of cardiovascular activity (i.e. heart rate, blood pressure) during exposure to physical or psychological stress, comparing to their values in the rest state. Many factors might contribute to the individual differences in CVR, such as genetic predisposition, gender, age, types of stressors, way of coping, affectivity, etc. Furthermore, many studies confirmed that specific personality dimensions play a significant role in the indivdual differences in CVR. Purpose: The main purpose of our investigation was to examine the effects of personality dimensions on cardiovascular reactivity (CVR) to experimentally induced stress, taking into account the confounding variables (i.e. gender, depression, and coping with stress). Methodology: The sample consisted of non-consecutive case series of seventy nine patients with clinically stable and angiographically confirmed coronary artery disease (CAD) who had been admitted to the Clinic of Cardiology, University Clinical Centre, and twenty one healthy persons of the control group. The participants were included in the study during the period of six years (from 2001 to 2007). The test was conducted in three phases: a) Initial psychiatric assessment; b) Rest phase: participants rested in a lying position for 10-15 minutes in a partially darkened room; c) Application of the Mental stress test (MST) that consists of two tasks: mental arithmetic and an anger recall task. Continuous 12-lead ECG and blood pressure were monitored and recorded every 5 minutes throughout the rest and the MST procedure. Afterwards, all participants were administered the following battery of psychological tests: a) NEO Personality Inventory - Revised, NEO- PI-R); b) The D-Scale 14, DS-14; c) Ways of Coping Questionnaire - WCQ-R; and d) Beck Depression Inventory - BDI. Results: The results of the study showed that patients with coronary heart disease scored higher than healthy controls on Neuroticism and lower on Extraversion and Openness to Experience. There were no statistically significant differences on Agreeableness and Conscientiousness. The prevalence of Type D personality was higher in patients with stable coronary heart disease than in the healthy control group. The mean NA/SI scores, also, were higher in patients with stable coronary heart disease than in the healthy control group. Neuroticism and Extraversion shared about 48% and 38% variance with Negative affectivity and Social inhibition, respectively. Correlation analyses indicated that both Type D dimensions (NA, SI) were positively associated with Neuroticism, while negatively associated with Extraversion and Conscientiousness. In addition, we found negative x correlation between Negative affectivity and Agreeableness, as well as, negative correlation between Social inhibition and Openness to Experience. Significant difference was found in mean BDI scores for patients with coronary heart disease and healthy control participants. Additionally, the cross-tabulation was used to examine the overlap between depression/Type D subgroup of patients - depressed Type D, depressed non-Type D and non-depressed Type D. These findings have shown that Type D personality and depression most probably are two distinctly different categories of psychological distress. In addition, the mean value of BDI was found to be significantly higher in Type D personality patients, than in non-Type D personality patients. Depression was positively correlated with Neuroticism and negatively with Extraversion. Patients with the diagnosis of CVD more often use emotion-focused coping strategies (e.g. distancing, self-control, avoidance), while healthy controls predominantly used active coping strategies (e.g. planful problem solving, positive reappraisal). The statistically significant difference between tested groups was found on the scales avoidance, social support and planful problem solving. Individuals with CVD were characterized by avoidant behavioural style, and were less likely to use social support, and analytical approach to problem solving. Patients with coronary heart disease with higher scores on Agreeableness and Conscientiousness showed higher heart rate reactivity, as well as diastolic blood pressure reactivity during the mental arithmetic task. Agreeableness and Conscientiousness, as well as Neuroticism were associated with higher cardiovascular reactivity (especially systolic blood pressure reactivity) in patients with coronary heart disease during the anger recall task. During the both tasks, Openness to Experience was associated with lower cardiovascular reactivity. The impact of Type D personality on heart rate score changes (Tip D x HR) during the both tasks proved to be statistically irrelevant. Additional test of within-subjects effects indicated that the interaction between the change in the heart rate and type D personality between the phase 1 and phase 2 was statistically highly significant. Therefore, the non-Type D personality subjects have shown statistically significantly higher drop in HR after the rest phase, compared to the subjects with Type D personality. In contrast, the tests of between- subjects effects further showed that the impact of personality to the change of systolic blood pressure (Type D x SBP) during the four phases of experiment was statistically significant The contrast analysis pointed to the statistically significant interaction between Type D personality x SBP between the phases 3 and 4 of the experiment and/or reduction of SBP during the performance of anger recall task with subjects with Type D. Finally, the between-personality comparison showed that the Type D personality had no impact in terms of changing the diastolic blood pressure during all four phases of the experiment. The within-subjects analysis showed that the interaction Type D x DBP was statistically significant between the phases 1 and 2 and that this interaction was highly statistically significant between the phases 2 and 3 of the experiment. Therefore, the subjects of non- Type D personality had statistically significantly higher drop in DBP following the phase of break, and the higher increase in DBP during the exposure to acute stressor (mental arithmetic task). xi Conclusion: The review of relevant literature suggests that the results of our study are the first to indicate reduced CVR in patients with CVD who are prone to distress (Type D personality). Moreover, our study revealed that specific personality dimensions such as Conscientiousness and Agreeableness strongly contribute to the increase of CVR. According to the findings of our study, aimed to elucidate the risk factors for onset, course and outcome of CVD, we emphasize the need of collaboration between cardiologists and pszchiatrists in the field of behavioral cardiology as well as necessity of the person oriented approach.
true
Izbor tipa, lokacije i parametara fleksibilnih regulacionih uređaja u elektroenergetskim sistemima primenom višekriterijumske optimizacije
Selection of the type, location and parameters of facts devices in the electric power systems using multicriteria optimization
Tradicionalni načini upravaljanja elektroenergetskim sistemima kao što su redispečing aktivnih i reaktivnih snaga proizvodnje kao i promena topologije mreţe u današnje vreme nisu dovoljni za kvalitetno i pouzdano upravljanje sistemom. Konstantan rast potrošnje, razvoj trţišta električne energije, priključenje distribuiranih izvora električne energije, ekološki zahtevi i sl., nametnuli su potrebu za novim načinima upravljanja. Jedan od savremenih načina za upravljanje sistemom je upotreba novih kontrolnih ureĎaja baziranih na primeni elemenata energetske elektronike velikih snaga. Ovi ureĎaji poznati su pod jedinstvenim nazivom FACTS (Flexible Alternate Current Transmission Systems) ureĎaji. Osnovna namena FACTS ureĎaja je povećanje kontrolabilnosti i prenosne sposobnosti sistema, odnosno direktna regulacija tokova snage ţeljenim prenosnim koridorima, kao i regulacija napona odabranih čvorova u mreţi. Ovi ureĎaji omogućavaju bolje iskorišćenje postojećih prenosnih vodova i odlaganje potrebe za izgradnjom novih. Pored toga što FACTS ureĎaji mogu da upravljaju tokovima aktivnih i reaktivnih snaga i naponskim prilikama u elektroenergetskoj mreţi, oni imaju i pozitivan uticaj i na druge pokazatelje koje karakterišu rad prenosnog sistema. Imajući sve ovo u vidu i sa ţeljom da se FACTS ureĎaji iskoriste na najefikasniji način, od posebne je vaţnosti izbor tipa, lokacije i parametara FACTS ureĎaja. Cilj ove disertacije je uvoĎenje nove procedure za nalaţenje optimalnog tipa, lokacije i parametara FACTS ureĎaja primenom višekriterijumske optimizacije. Osnovna prednost razvijene procedure je što omogućuje uključivanje proizvoljnog broja kriterijumskih funkcija. Na taj način mogu se obuhvatiti svi relevantni kriterijumi koji mogu uticati na izbor FACTS ureĎaja. Kao rezultat višekriterijumske optimizacije dobija se set optimalnih rešenja, za razliku od jednokriterijumske optimizacije gde se kao rezultat dobija jedno jedinstveno rešenje. Set optimalnih rešenja, generalno gledano, daje kompletniji opis optimizacionog problema, a konkretno kod izbora FACTS ureĎaja daje uvid u kontrolne mogućnosti ovih ureĎaja. Pored toga zbog postojanja više kompromisnih optimalnih rešenja krajnjem korisniku je data veća sloboda kod izbora finalnog rešenja pri čemu on iz seta ponuĎenih rešenja moţe da izabere ono koje njemu najviše odgovara. MeĎutim, zbog sloţenosti optimizacionog problema izbor konačnog rešenja iz velikog broja ponuĎenih optimalnih rešenja moţe da bude veoma teţak i nepraktičan. Ovo je posebno izraţeno kada se razmatra veći broj kriterijumskih funkcija. Već kod tri kriterijumske funkcije, grafički prikaz dobijenih rešenja u 3D prostoru nije praktičan za odabir finalnog rešenja, a za slučaj više od tri kriterijumske funkcije ne postoji način da se dobijena rešenja adekvatno grafički prikaţu. U okviru nove procedure razvijena je metoda za izbor finalnog kompromisnog rešenja koja ne favorizuje ni jednu od analiziranih kriterijumskih funkcija već na ravnopravan način uvaţava uticaj svih kriterijumskih funkcija. Kao matematički alat, za rešavanje višekriterijumskog optimizacionog problema, odabrana je optimizaciona tehnika evolutivnog tipa u literturi poznata kao genetički algoritam sa sortiranjem prema dominaciji (algoritam NSGA-II). Algoritam NSGA-II je izabran jer predstavlja jednu od standarnih procedura za rešavanje problema višekriterijumske optimizacije. U ovoj disertaciji NSGA-II algoritam je dodatno unapreĎen tako što je uvedeno lokalno pretraţivanje za članove populacije koji se nalaze na krajevima tekućeg Pareto fronta. Na taj način je povećana brzina i efikasnost pri nalaţenju optimalnih rešenja. Nova procedura za izbor FACTS ureĎaja ilustrovana je na dva realna inţenjerska problema. Prvi problem se odnosi na odreĎivanje optimalne lokacije i parametara FACTS ureĎaja u prenosnoj mreţi. Proračun je izvršen na realnom sistemu elektroprivrede Republike Srbije sa ciljem da se dobije optimalan plan razvoja prenosne mreţe uvoĎenjem samo regulacionih ureĎaja bez izgradnje novih vodova i transformatorskih stanica. Proračun je uraĎen za različite godine razvoja ovog sistema imajući u vidu povećanje nivoa opterećenja i planiranu izgradnju novih proizvodnih kapaciteta. Kao kriterijumi za izbor razmatrani su kako globalni pokazatelji rada elektroenergetske mreţe, kao što su gubici aktivne snage u sistemu, tako i lokalni pokazatelji koji se odnose na nivoe opterećenja prenosnih vodova i napone čvorova, pokazatelji sigurnosti, pokazatelji vezani za rad sistema u trţišnom okruţenju i ekonomski pokazatelji. Nova procedura kroz pretraţivanje velikog broja potencijanih varijanti i kroz uporednu analizu uticaja pojedinih tipova FACTS ureĎaja na analiziranu mreţu daje konkretne predloge za FACTS ureĎaje čijom bi se ugradnjom na predloţene lokacije ostvario plan razvoja prenosne mreţe bez potrebe za izgradnjom novih prenosnih kapaciteta. Pored toga ugradnjom FACTS ureĎaja unapredile bi se performanse prenosne mreţe imajući u vidu analizirane kriterijumske funkcije. Drugi problem koji je obraĎen u ovoj disertaciji odnosi se na nalaţenje optimalne lokacije i parametara FACTS ureĎaja u distributivnoj mreţi uz prisustvo distribuiranih izvora električne energije. Kod ovog proračuna osnovni kriterijum za izbor bilo je obezbeĎivanje napajanje krajnjih potrošača kvalitetnom električnom energijom za širok dijapazon mogućih radnih reţima uz uvaţavanje svih pogonskih ograničenja. Proračun je uraĎen na realnoj srednjenaponskoj distributivnoj mreţi koja napaja nekoliko sela u regionu Banata u Srbiji. Proračunom su dobijene konkretne loakcije za ugradnju FACTS ureĎaja u analiziranoj mreţi. Razvijena procedura za izbor FACTS ureĎaja u primeni na realne prenosne i distributivne mreţe daje veoma dobre rezultate. Procedura pretraţuje veliki broj potencijalnih rešenja i u procesu optimizacije daje rešenja koja su za analiziranu mreţu najefikasnija, imajući u vidu kriterijumske funkcije kojima je modelovan optimizacioni problem.
In present days, traditional ways of controlling power systems such as re- dispatching of active and reactive power and changing the network topology, are not sufficient for quality and reliable system management. The constant growth of consumption, development of electricity market, connection of distributed generation, environmental requirements etc, imposed the need for new ways of system management. One of the modern ways to control power systems is to use new control devices based on application of power electronics. These devices are known as Flexible Alternate Current Transmission Systems (FACTS) devices. The main purpose of FACTS devices is to increase controllability and capacity of the transmission system by controlling active and reactive line flow in desired transmission corridors and voltage control of the selected buses in the network. The use of FACTS devices in transmission systems enabled a better utilization of existing transmission lines, instead of constructing the new ones. Those devices have a positive influence on many parameters characterizing the operation of transmission networks. Bearing all this in mind and with the desire to take advantage of FACTS devices in the most efficient manner, proper choice of type, location and parameters of FACTS devices are particularly important. The aim of this dissertation is to introduce a new procedure for finding the optimal type, location and parameters of FACTS devices by using multi-criteria optimization. The main advantage of this procedure is that it allows inclusion of any number of criteria functions. This way, we can cover all the relevant criteria that may influence the choice of FACTS devices. Result of multi-criteria optimization is a set of optimal solutions, as opposed to a single criterion optimization where we obtain only one solution. Set of optimal solutions, in general, provides a more complete description of the optimization problem. In particular, in the choice of FACTS devices, provides insight into the control features of the FACTS devices. In addition, set of optimal solutions enables the decision maker to choose a solution that seems fit. However, large number of solutions and large number of possible choices can be confusing, so selecting the final choice may be quite difficult. This is particularly evident when large number of criterion functions is considered. For three criterion functions, it is not practical to select the final solution by graphical representation of the obtained solutions in a 3D space. In case of four and more criterion functions, adequate graphical representation of obtained solutions is not possible. With the new procedure, a method for the choice of a unique optimal solution from the set of solutions was developed, that does not favor any of the analyzed criteria functions, but equally respects the influence of all criteria functions. Evolutionary optimization algorithm known as Nondominated Sorted Genetic Algorithm (NSGA-II) is selected as a mathematical tool for solving multi-criteria optimization problem. NSGA-II algorithm was chosen because it represents one of the standard procedures for solving multi-criteria optimization problems. In this dissertation, NSGA-II algorithm is further improved by introducing a local search for members of the population at the edges of the current Pareto front. This way, speed and efficiency in finding the optimal solutions are increased. New procedure for selection of FACTS devices is illustrated on two realistic engineering problems. The first issue relates to the determination of the optimal location and parameters of FACTS devices in a transmission network. The calculation is done on the Republic of Serbia power system in order to obtain an optimal plan for transmission network by introducing only FACTS devices without building new transmission lines and substations. The calculation is done for different years of development of this system taking into account increase in the consumer load and planned construction of new production capacities. The following selection criteria were considered: system active power losses, transmission line load levels, bus voltages, security indicators, indicators related to the operation of the system in the market environment and economic indicators. New procedure through a searching of a large number of potential variants and comparative analysis of the impact of certain types of FACTS devices on the observed network provides specific proposals for installation of FACTS devices. Installation of FACTS devices at proposed locations would fulfill the development plan of the transmission network, without the need to build new transmission capacities. In addition, installation of FACTS devices would improve the performance of the transmission system by taking into account the analyzed criterion functions. Second problem addressed in this dissertation refers to finding the optimal location and parameters of FACTS devices in a distribution network in presence of distributed power sources. In this calculation, the main selection criterion was to provide the quality electrical energy to the end users for a wide range of possible operating modes with respect to all operating limits. The calculation is performed on a real medium voltage distribution network which supplies several villages in the Banat region in Serbia. The calculation proposed the specific locations to install the FACTS devices in the analyzed network. Implementation of developed procedures for selection of FACTS devices to the real transmission and distribution network gives very good results. The procedure searches a large number of potential solutions and provides the most efficient solutions for the analyzed network in optimization process, keeping in mind the criteria functions used to model the optimization problem.
true
Razvoj mašinskog učenja inteligentnog mobilnog robota baziran na sistemu veštačkih neuronskih mreža
Machine learning of inteligent mobile robot based on arti ficial neural networks
Unutrašnji transport sirovina, materijala i gotovih delova podrazumeva brzo, efikasno i ekonomično obavljanje postavljenog transportnog zadatka. Unutrašnji transport u okviru tehnološkog okruženja moguće je unaprediti uvođenjem naprednih metoda i tehnologija. Razvoj novih sistema unutrašnjeg transporta treba da rezultira inteligentnim sistemima unutrašnjeg transporta koji su u stanju da sprovedu transportni zadatak bez direktnog nadzora od strane operatera. Primena mobilnih robota u okviru inteligentnog tehnološkog sistema za potrebe unutrašnjeg transporta omogućila bi efikasnije i ekonomičnije obavljanje postavljenog transportnog zadatka. Stepen razvoja i primene mobilnih robota dostigao je nivo neophodan za ispunjavanje zahteva koji su određeni tehnološkim i proizvodnim procesima. U istraživanjima sprovedenim u okviru ove doktorske disertacije razvijene su nove metode neophodne za primenu mobilnih robota u inteligentnom tehnološkom sistemu. Fokus istraživačkih napora predstavlja sposobnost mobilnog robota da uz primenu odgovarajuće matematičko-softverske podrške i akvizicijom informacije od kamere, samostalno odredi svoj položaj i položaj karakterističnih objekata u okruženju. Problem simultanog ocenjivanja položaja mobilnog robota i karakterističnih objekata u okruženju rešen je primenom linearizovanog Kalmanovog filtra i veštačkih neuronskih mreža. Analizirana je i mogućnost ostvarivanja inteligentnog ponašanja mobilnog robota u formi razvoja hibridne upravljačke arhitekture, koja obezbeđuje robustnost zahvaljujući primeni koncepta mašinskog učenja na bazi veštačkih neuronskih mreža. VIII Poseban segment istraživanja posvećen je razvoju hibridnog upravljačkog algoritma namenjenog za eksploataciju mobilnog robota u okviru inteligentnog tehnološkog sistema bez dodatne transportne infrastrukture. Predložene metode su verifikovane kroz razvoj simulacije u MATLAB® programskom okruženju i putem eksperimentalnog metoda. Eksperimentalni postupak je sproveden u laboratorijskom modelu tehnološkog okruženja korišćenjem LEGO Mindstorms NXT mobilnog robota i Khepera II mobilnog robota uz primenu sistema prepoznavanja na bazi kamere. Simulacioni i eksperimentalni rezultati ukazuju da razvijene metode podržavaju predloženi koncept primene mobilnih robota za potrebe unutrašnjeg transporta u okviru inteligentnog tehnološkog sistema.
Material Handling Systems in manufacturing environment imply efficient and economical transport solutions. Automated Guided Vehicles (AGVs) are a common choice made by many companies for Material Handling in manufacturing systems. Nowadays, AGV based internal transport of raw materials, goods and parts is becoming improved with advances in technology. Demands for fast, efficient and reliable transport imply the usage of the flexible AGVs with onboard sensing and special kinds of algorithms needed for daily operation. These transport solutions can be modified and enhanced by applying advanced methods and technologies. New generation of internal transport systems should operate autonomously, without direct human control. Level of development of mobile robots insures reliability and efficiency needed for dayily operations within manufacturing environemnt. In this thesis, the implementation of mobile robots for internal transport within Material Handling System is analyzed and new solutions are proposed. Focus of research efforts is devoted to the ability to estimate position and orientation of mobile robot within manufacturing environment using newly developed algorithms and sensory information. Simultaneous localization (of the mobile robot) and mapping (of the working environment) is one of the most important problems in mobile robotics community. The soultion to this problem insures autonomous navigation and henceforth autonomous operation for transport purposes within manufacturing/industrial facility without direct human control. In this thesis, new algorithm for state estimation is proposed and analyzed; the algorithm is based on integration of Extended Kalman Filter and feedforward neural networks (Neural Extended Kalman Filter) and camera is used as exteroceptive sensor. Furhermore, to achieve intelligent behavior, the X new robotic hybrid control architecture is developed and analyzed. Finally, the new hybrid control algorithm for guidance of mobile robot is proposed. Two building blocks form the hybrid algorithm: visual servoing and position based control. Neural Extended Kalman Filter is used for state estimation of the mobile robots, and at each time instant the robot knows its position and orientation. The proposed solutions are developed in MATLAB® environment by developing a specific software code and tested using Khepera II mobile robot, WEB camera and LEGO Mindstorms NXT mobile robot. Simulation and experimental results show usability of the proposed solutions for material handling within an Intelligent Manufacturing System.
true
Štete kod potrošača usled prekida isporuke električne energije
Customers costs incurred due to interruptions of the supply with the electrical energy
U doktorskoj disertaciji je predstavljen nov probabilistički model za realne i efektivne procene materijalnih i nematerijalnih šteta koje kod različitih potrošača prouzrokuju prekidi isporuke električne energije. Osnovni doprinos ovog rada je unapređenje metodologije proučavanja posledica prekida napajanja na različite potrošače kroz detaljnije modelovanje aktivnosti domaćinstava i komercijalnih potrošača, koje u kombinaciji sa odgovarajućim modelima kvarova daju podatke o verovatnoći i očekivanim trajanjima narušavanja tih aktivnosti i time obezbeđuju realnu podlogu za procenu šteta. Pored navedenog, u okviru modela su predloženi i pokazatelji za kvantifikovanje nematerijalnih posledica koji mogu da posluže kao dopunska merila za ocenu kvaliteta isporuke električne energije. Za potrebe doktorske disertacije i u cilju unapređenja modela, temeljno su pripremljene i sprovedene odgovarajuće ankete. Na osnovu odgovarajućih matematičkih modela, razvijeni su softverski programi u Matlabu pomoću kojih su sprovedeni proračuni šteta. Detaljno je analiziran uticaj pojedinih aktivnosti na štete kod potrošača, a za proračune u slučaju prekida koji su se dogodili, analiziran je uticaj sezona, tipa dana (radnog i neradnog) i načina grejanja na štete kod potrošača usled prekida napajanja. Sa gledišta veličine šteta, kod domaćinstava najznačajnija aktivnost je frižider sa oko 50% ukupnih troškova, a zatim daljinsko grejanje sa oko 15% ukupnih troškova. Na ukupne troškove komercijalnih potrošača značajno utiču troškovi vezani za aktivnosti rashlađivanja kvarljivih artikala, kao i troškovi aktivnosti zaštite objekata sa velikom vrednošću imovine (robe). Analogno domaćinstvima, kod komercijalnih potrošača koji u osnovnoj delatnosti koriste rashladne uređaje, troškovi u vezi ovih aktivnosti su najznačajniji i iznose oko trećine ukupnih troškova. U slučaju proračuna šteta za prekide koji su se dogodili, utvrđeno je da kod domaćinstava ukupne štete ne zavise značajno od sezone, tipa dana i načina grejanja, dok se kod komercijalnih potrošača javljaju značajne razlike. Za komercijalne potrošače koji u osnovnoj delatnosti ne koriste rashladne uređaje niti uređaje za zaštitu objekata (npr. knjižare), uticaj sezone i tipa dana je značajan. Za komercijalne potrošače koji koriste uređaje za zaštitu objekata, ali ne i rashladne uređaje (npr. pošte), uticaj sezone i tipa dana je izražen ali je manje značajan. Komercijalni potrošači koji radi obavljanja osnovne delatnosti koriste rashladne uređaje za kvarljive artikle, npr. restorani, obično nemaju neradni dan, a zbog relativno velikih ukupnih troškova izazvanih troškovima aktivnosti rashladnih uređaja, uticaj sezonskih aktivnosti grejanja prostorija se kreće oko 10% ukupnih troškova. Doprinosi doktorske disertacije se ogledaju u razvoju efikasnih i jednostavnih za primenu matematičkih modela za realne procene nematerijalnih i materijalnih šteta kod potrošača usled prekida isporuke električne energije. Na bazi bolje zasnovanih tehničko-ekonomskih analiza iz doktorske disertacije, dobijaju se korisne informacije za eksploataciju i planiranje poboljšanja postojećih ili izgradnju novih distributivnih sistema, što omogućava poboljšanje kvaliteta isporuke električne energije i unapređenje i razvoj distributivnih elektroenergetskih sistema. Unapređenjem modela i određivanjem realnih šteta kod potrošača usled prekida isporuke električne energije postignuto je ne samo zadovoljenje zakonskih i regulatornih obaveza već je i povećan stepen zadovoljstva potrošača, što predstavlja dobrobit za celo društvo.
This doctoral dissertation presents a new probabilistic mathematical model for real and effective assessment of material and non-material supply interruption costs for various customers. The main contribution of this paper is enhancing the methodology of the evaluation of the consequences of supply interruptions for various customers by modelling more detailed the household and commercial activities, which in combination with the appropriate fault models provide information about the probability and expected duration of the interruption of those activities, that provide a realistic basis for costs assessment. In addition, under the proposed model and data for quantifying non material cost consequences that can provide an additional criteria to assess the quality of electricity supply. For the purposes of doctoral dissertation and in order to enchase model, appropriate surveys are thoroughly prepared and implemented. Based on appropriate mathematical models, software programs are developed in Matlab by which calculations are performed. The detailed analysis of influences of certain activities on the customer’s costs is made. In the calculation case for system faults have occurred, further analyses of the influence of seasons, type of day (weekdays and holidays) and kind of heating on customers costs incurred due to interruptions of the supply, are performed. From the perspective of outage costs, for residential customers fridge is the most important activity with about 50% of total costs follows by district heating to about 15% of total costs. Refrigeration and safety activities for objects with a large value of the property (goods) significantly affect to total costs of commercial customers. Analogously residential customers, for commercial customers that apply fridge devices for main activity, the costs for these activities are the most significant, amounting to about a third of the total cost. In the calculation case for system faults have occurred, it was found that residential customers total costs do not depend significantly on the seasons, type of day and kind of heating, while for commercial consumers significant differences occur. For commercial customers that do not apply fridge and safety devices for main activity, for example bookstores, influence of season and type of day is important. For commercial customers that use the safety devices but not fridge, for example post offices, the influence of season and type of day are expressed but are less significant. For commercial customers that apply fridge devices for main activity, for example restaurants, usually have a holiday, and because of the relatively large total costs caused by the cost of the activities of refrigeration equipment, the influence of seasonal activity space heating is about 10 % of total costs. Contributions of dissertation reflect the development of efficient and simple for application of mathematical models for a realistic assessment of non-material and material customers costs incurred due to interruptions of the supply with the electrical energy. Based on the better based technical- economic analysis of the dissertation, we get useful information for planning and exploitation of existing or construction of new distribution system, which allows the improvement of the quality of electricity supply and the promotion and development of electric power distribution systems. Improvement of the model and determining the actual cost due to interruption of electrical energy supply is achieved not only to meet the legal and regulatory obligations but further to raise the level of customer satisfaction, which is a benefit for the whole society.
true
Prostiranje elektromagnetnih talasa kroz hiralne metamaterijale u terahercnom frekventnom opsegu
Propagation of electromagnetic waves through chiral metamaterials in terahertz frequency range.
Sve ve´ca primena terahercnog zraˇcenja u oblasti komunikacija, spektroskopije i imidˇzinga, posledica je njegovih jedinstvenih osobina, a posebno ˇcinjenice da prodire u mnoge izolatorske materijale. Pomenute primene omogu´cene su realizacijom uredaja koji su operativni na terahercnim frekvencijama. Njihova konstrukcija moˇze biti zasnovana na metamaterijalima, veˇstaˇcki napravljenim strukturama koje se sastoje od periodiˇcno rasporedenih rezonantnih elemenata. Ovim materijalima je mogu´ce posti´ci jaku interakciju sa terahercnim elektromagnetnim talasima, usled ˇcega su u fokusu aktuelnih istraˇzivanja. Hiralni metamaterijali, kao podvrsta metamaterijala karakteristiˇcne geometrije, su od posebnog interesa zbog drugaˇcijeg odziva na upadne levo i desno cirkularno polarizovane talase. Ova osobina omogu´cava brojne primene od kojih su najznaˇcajnije enantiomerski senzori i komponente potrebne za manipulaciju cirkularno polarizovanih terahercnih talasa...
The unique features and especially the property of terahertz radiation that it can penetrate into many insulting materials make them ideal for diverse applications in the field of communications, spectroscopy and imaging. These applications are enabled by the realization of devices which operate at terahertz frequencies. Their construction can be based on metamaterials, artificially made structures which consist of periodically distributed resonant elements. Metamaterials provide a strong interaction with terahertz electromagnetic waves, which places them in the focus of current investigations. Chiral metamaterials, as a class of metamaterials with characteristic geometry, are of the special interest due to the different response to left and right circularly polarized waves. This feature allows numerous applications among which the most significant are enantiomeric sensors and components necessary for manipulation of circularly polarized waves...
false
Stabilno prekidačko upravljanje DC-DC konvertorima
Stable switching control of DC-DC converters.
Kola energetske elektronike se koriste za efikasnu konverziju i kontrolu elektri ˇcne snage. Konvertori predstavljaju kola energetske elektronike koja vrše konverziju jednosmernog ili naizmeniˇcnog ulaznog napona ili struje u jednosmerni ili naizmeniˇcni izlazni napon ili struju. U ovoj disertaciji se analiziraju DC-DC konvertori koji vrše konverziju jednosmernog ulaznog napona u jednosmerni izlazni napon...
Power electronics circuits are used for controlling and providing ecient power conversion. Furthermore, these converters are used for conversion of the input DC or AC voltage (or current) to output DC or AC voltage (or current). In this dissertation, DC-DC converters that are used for converting input DC voltage to output DC voltage will be considered...
false
Uticaj pretretmana vitaminom D na strukturne i biohemijske promene uzrokovane prolaznom globalnom ishemijom mozga džerbila
The effects of vitamin D pretreatment on structural and biochemical changes in the brain of gerbils exposed to transient global cerebral ischemia
Cerebralna ishemija nastupa kada protok krvi opadne ispod 20% normalnog protoka, dovodeći do smanjenog dotoka kiseonika i glukoze u mozak, i posledično poremećaja jonske homeostaze, smanjenje sinteze ATP-a, povećano oslobađanje glutamata i ekscitotoksičnost. U patogenezi moždane ishemije važno mesto ima povećano oslobađanje slobodnih radikala (ROS) odnosno nastanak oksidativnog stresa. Važno je što ranije ponovo uspostaviti cirkulaciju, međutim reperfuzija uzrokuje još veću produkciju slobodnih radikala i iscrpljivanje antioksidativnih kapaciteta ćelije. Potencijalni izvor ROS u ishemiji i reperfuziji mozga predstavljaju enzimi NADPH oksidaze (nikotinamid adenin dinukleotid fosfat oksidaze), čija je fiziološka funkcija stvaranje superoksidnog anjona. U ćelijama CNS-a (neuronima, astrocitima, mikrogliji), detektovana je ekspresija NOX2 (gp91phox) proteina, koji u asocijaciji sa p22phox, GTP-Rac, p47phox, p67phox i p40phox produkuje superoksidni anjon, dok je NOX4 je najzastupljenija izoforma na nivou vaskularne mreže. Kako NOX2 i NOX4 mogu predstavljati značajan izvor oksidativnog stresa nakon ishemije mozga ove izoforme predstavljaju potencijalne terapijske mete. Ishemijsko/reperfuziono (I/R) oštećenje pokreće i proces apoptoze i izaziva poremećaj procesa autofagije pri čemu uloga koju autofagija ima u ishemiji mozga još uvek nije razjašnjena. Sve se više proučava uloga vitamina D (vitD) u prevenciji kardiovaskularnih i cerebrovaskularnih bolesti. Dosadašnje studije su pokazale da je snižena koncentracija vitamina D u serumu povezana sa povećanim rizikom od ishemijskog moždanog udara, kao i sa pogoršanjem ishemijskog oštećenja, težim funkcionalnim oporavkom i kognitivnim disbalansom. Uloga i mehanizmi delovanja vitamina D u cerebralnoj ishemiji još uvek nisu u potpunosti razjašnjeni. Cilj: Ispitati efekte pretretmana vitaminom D na parametre oksidativnog stresa, apoptoze, autofagije i zastupljenost pojedinih ćelija centralnog nervnog sitstema upotrebom odgovarajućih markera u prefrontalnom korteksu i hipokampusu mongolskih džerbila izloženih desetominutnoj prolaznoj globalnoj ishemiji mozga i reperfziji u trajanju od jednog, tri ili sedam dana...
Cerebral ischemia occurs when blood flow drops below 20% of normal flow, leading to a reduced supply of oxygen and glucose to the brain, which further affects the disorder of ionic homeostasis, ATP synthesis, glutamate release, and excitotoxicity, which ultimately results in ischemic neuronal damage. Another mechanism thats plays an important role in the pathogenesis of cerebral ischemia is the disturbance of oxidative balance, which manifests itself through increased release of free radicals (FR) and oxidative stress. In ischemic tissue, it is of utmost importance to establish circulation, but reperfusion leads to even greater free radical production, exhaustion of antioxidative capacity and further oxidative damage. Potential FR source in ischemia and brain reperfusion is NADPH oxidase family (nicotinamide adenine dinucleotide phosphate oxidase), who’s main physiological function is superoxide anion production. Expression of NOX2 (gp91phox) protein was detected in CNS cells (neurons, astrocytes, microglia), which in association with p22phox, GTP-Rac, p47phox, p67phox and p40phox, produces a superoxide anion. NOX4 is the most common isoform at the level of the vascular network. As NOX2 and NOX4 could be a significant source of FR after brain ischemia, these isoforms represent potential therapeutic targets. In addition to oxidative stress, ischemic/reperfusion (I/R) injury also triggers the apoptosis and distorts the autophagy. The role of autophagy in brain ischemia has not yet been clarified. Recent studies focused on the prevention of cardiovascular and cerebrovascular diseases have been researching the impact of vitamin D. It has been shown that the reduced serum concentration of vitamin D is associated with an increased risk of ischemic stroke, as well as with exacerbation of ischemic damage, more severe functional recovery, and cognitive impairment. Calcitriol exhibits effects in the body through the vitamin D (VDR) nuclear receptor, which is present in the brain in both neurons and glial cells. The role and mechanism of vitamin D in cerebral ischemia have not yet been fully clarified...
false
Dekodovanje kodova sa malom gustinom provera parnosti u prisustvu grešaka u logičkim kolima
Sve ve´ca integracija poluprovodniˇckih tehnologija, varijacije nastale usled nesavršenosti procesa proizvodnje, kao zahtevi za smanjenjem napona napajanja cˇine elektronske ured¯aje inherentno nepouzdanim. Agresivno skaliranje napona smanjuje otpornost na šum i dovodi do nepouzdanog rada ured¯aja. Široko je prihvac´ena paradigma prema kojoj se naredne generacije digitalnih elektronskih ured¯aja moraju opremiti logikom za korekciju hardverskih grešaka...
Due to huge density integration increase, lower supply voltages, and variations in technological process, complementary metal-oxide-semiconductor (CMOS) and emerging nanoelectronic devices are inherently unreliable. Moreover, the demands for energy efficiency require reduction of energy consumption by several orders of magnitude, which can be done only by aggressive supply voltage scaling. Consequently, the signal levels are much lower and closer to the noise level, which reduces the component noise immunity and leads to unreliable behavior. It is widely accepted that future generations of circuits and systems must be designed to deal with unreliable components...
false
Distribuirani sistem za funkcionalnu električnu stimulaciju
Distributed functional electrical stimulation system
Funkcionalna električna stimulacija se koristi za vraćanje funkcije nervno- mišićnom sistemu čoveka oštećenom fizičkom povredom, bolešću ili nekim drugim uzrokom. Kontrolisanim strujnim ili naponskim impulsima se vrši pobuñivanje nervnih motornih ili senzornih vlakana što se odražava kroz aktivaciju mišića. Posledica toga je restauracija funkcionalnog kretanja pojedinačnih ili grupe mišića. Počeci primene funkcionalne električne stimulacije sežu u šezdesete godine dvadesetog veka, kada je projektovan i proizveden prvi upotrebljiv električni stimulator, namenjen korekciji poremećaja padajućeg stopala. Nada da će tehnologija ubrzo omogućiti paralizovanim osobama da se ponovo pokreću podstakla je u narednim godinama razvoj mnogih rešenja za sisteme električne stimulacije. Meñutim, današnje stanje, pola veka kasnije, pokazuje da i pored jasne motivacije i ogromnog napretka na polju tehnologije još uvek nema značajnog pomaka na tom putu. Konzervativne strategije upravljanja električnom stimulacijom se zasnivaju na centralizovanom upravljanju. Raspoloživi istraživački sistemi se odlikuju manjim ili većim brojem senzorskih ulaza i nezavisnih stimulacionih kanala. Generalne mane su mnoštvo provodnika, komplikovane procedure postavljanja i podešavanja i vezanost za laboratoriju. Komercijalni sistemi su jednostavni za korišćenje ali imaju slabe funkcionalne performanse. U ovoj disertaciji je predstavljena nova arhitektura distribuiranog sistema za električnu stimulaciju inspirisana prirodnim biološkim sistemom. Nova filozofija u razumevanju mehanizma kretanja počiva na ideji hijerarhijski kontrolisane distribuiranosti u nervnom sistemu. Nekoliko kontrolnih sprega na različitim nivoima nervnog sistema doprinose ovoj kontroli: silazna motorna kontrola iz motornog korteksa (u mozgu), mreža generatora ritma u kičmenoj moždini (spinalni stepping generator) i senzorna povratna sprega za kontrolu hoda (sa zglobova u pokretu, mišića i kože). vii Biomimetičkim preslikavanjem hipotetičke strukture biološkog sistema dobijena je opšta arhitektura distribuiranog sistema za električnu stimulaciju koja je predmet ove disertacije. Mozak je preslikan na centralni kontroler, hipotetički autonomni centri u kičmenoj moždini na koordinator, a lokalne mišićno-čulne sprege na set periferijskih stimulacionih jedinica. Po ugledu na prirodni sistem definisane su dva režima rada sistema, režim motornog učenja i režim utvrñene primene. Za svaki režim rada analizirani su aktivnost i protok informacija izmeñu komponenti svakog hijerarhijskog nivoa. U režimu motornog učenja sistem karakteriše veliki protok informacija na svim nivoima i visoka aktivnost centralnog kontrolera. Nasuprot tome, u režimu utvrñene primene protok informacija je drastično smanjen a aktivnost je prebačena na periferijske jedinice. Rad sistema u režimu učenja daje mogućnost razvoja kompleksnih algoritama u kojima su upravljanje i nadzor pod kontrolom moćnog centralnog kontrolera, tj. računara, kojim upravljaju terapeuti, naučnici. Smanjena aktivnost i protok informacija u režimu utvrñene primene daje mogućnost rada sistema sa smanjenom potrošnjom i protokom podataka, što ga čini dugotrajnim i pouzdanim pomagalom. Uzimajući u obzir sadašnje stanje tehnike predložena je jedna implementacija opšte arhitekture u kojoj se komunikacija izmeñu komponenti sistema odvija bežičnim radio putem. Predstavljene su opšte karakteristike sistema sa stanovišta hardvera i softvera. Bežična komunikacija je razvijena na bazi jednog poznatog standarda bežične komunikacije i prilagoñena je ovoj nameni. Projektovan je novi protokol rutiranja u dva skoka koji uzima u obzir fizičke karakteristike sistema i obezbeñuje brz i siguran protok podataka. Za potrebe verifikacije i testiranja razvijen je prototip sistema korišćenjem komercijalno dostupnih komponenti. Opisane su značajne pojedinosti realizacije koje odslikavaju sadašnje stanje tehnike. Verifikacija sistema je izvedena kroz aplikaciju kontrole hvata kod osobe sa delimično oštećenom funkcijom usled moždanog udara. Zadatak je bilo rešavanje poznatog problema kompenzacije tenodeze tj. neželjenog savijanja ruke u zglobu usled stimulacije mišića koji savija prste. U režimu učenja upravljanje u zatvorenoj sprezi je iskorišćeno da se proceni i optimizuje optimalni stimulacioni profil koji obezbeñuje željeno kretanje. U režimu utvrñene primene pokazalo se da sistem obezbeñuje željenu funkciju sa minimalnim protokom podataka i niskom aktivnošću centralnog kontrolera. viii Samim tim ova aplikacija je potvrdila osnovne hipoteze ove disertacije. Pored primene u stimulaciji, senzorski podsistem je verifikovan kroz nekoliko aplikacija za objektivnu procenu parametara hoda.
Functional electrical stimulation (FES) is used for restoration of neuro-muscular system of patients with motor impairment provoked by physical injury, disease of some other cause. Stimulation of neural motor or sensory fibres is performed with controlled voltage or current pulses which perform muscle activation resulting with restoration of movement provoked provoked by individual muscle or muscle group. Application of FES begins in the 1960s, with design and production of the first usable electrical stimulator that aimed to correct drop foot disturbance. In the following years, there was a hope that new technology will soon allow restoration of movement for paralyzed persons which encouraged development of numerous systems for electrical stimulation. However, fifty years later, after immense technological progress and clear motivations, still there is no significant breakthrough in this area. Conservative strategies for control of electrical stimulation are based on centralized controls. Available research systems feature smaller or higher number of sensor input and independent stimulation channels. General fault of these systems are numerous conductors, complicated mounting procedures and adjustments and dependency on laboratory environment. Commercial systems are user-friendly but with weaker functional performances. In this dissertation, new architecture of a distributed system for functional electrical stimulation is presented, inspired by natural biological system. New philosophy in understanding movement mechanism relies on hierarchically controlled distribution in neural system. It is believed that motor control is performed on three levels: descending control from the motor cortex (brain), rhythm-generating network in the spinal cord (spinal stepping generator) and sensory feedback from joints, muscle and skin of extremities. Using biomimetic mapping of hypothetical structure of biological system, general architecture of distributed system for electrical stimulation has been made which is the x main subject of this dissertation. Brain has been mapped to central computer, hypothetical autonomous centers in spinal cord are mapped as network coordinator, and local sensory-motor feedbacks are mapped as peripheral stimulation units. Looking up to biological system, we defined two working modes: motor learning and application of mastered movement. For each mode, activity and information flow between components of all hierarchical levels have been analyzed. Motor learning mode is characterized by large data flow on all levels and high activity of central controller. On the opposite, application of mastered movement has drastically lower data flow and the activity is transferred to peripheral units. Motor learning mode allows possibility of development of complex algorithms where control and supervision are supervised by powerful central controller (computer) operated by therapist and/or researchers. Decreased activity and data flow during application of learnt movement allow low-consumption working regime and low data flow which makes this system long-lasting and reliable assistance. Considering today’s state-of-the-art in this area, one implementation of general architecture with radio-frequency wireless communication between system components has been proposed. General characteristics regarding hardware and software have been proposed. Wireless communication was developed based on one known standard of wireless communications which has been adapted for this application. New two-hop routing protocol has been created which considers physical characteristics of the system and provides fast and secure data flow. System prototype has been created for verification and testing, using commercially available components. Significant features related to realization of the system which describe current technical situation have been described. System verification is performed through application for assisting patient with partially impaired motor function has been selected. The goal was to resolve one familiar problem – compensation of involuntary wrist flexion due to muscle stimulation which provides finger flexions. For application of learned movement, the system provides desired function with minimal data flow and central controller activity. Hereby, this application has confirmed hypothesis of this dissertation. Besides stimulation, the sensor subsystem is verified through several applications for objective gait assessment. xi
true
Određivanje karakteristika elektromagnetske emisije uređaja merenih u Faradejevom kavezu pomoću tehnike obrade retkih signala
Evaluation of radiating-source parameters by measurements in faraday cages and sparse processing.
U ovoj doktorskoj disertaciji razvijena je nova metoda koja omogućava određivanje nivoa električnog polja (u slobodnom prostoru) nepoznatog izvora korišćenjem skalarnih merenja u Faradejevom kavezu. Nova metoda omogućava upotrebu Faradejevog kaveza kao jeftine alternative standardizovanim mernim mestima za ispitivanje elektromagnetske kompatibilnosti (EMK), kao što su (polu)anehoične sobe, otvoreni merni prostor i reverberacione sobe (kavezi). Nova metoda, u odnosu na standardizovani merni postupak u reverberacionim sobama, omogućava uspešno merenje uređaja na niskim frekvencijama, koje su ispod najniže rezonantne frekvencije kaveza. Metoda se zasniva na aproksimaciji električnog polja zračenja ispitivanog uređaja nizom ekvivalentnih dipola. Parametri ekvivalentnih dipola su određeni primenom tehnike obrade retkih signala, odnosno primenom l1 regularizacije kojom se umanjuje neodređenost rešenja inverznih elektromagnetskih problema. Složenost prostiranja signala u sredinama kao što je Faradejev kavez povoljno utiče na tačnost rešenja. Predloženi metod je verifikovan skalarnim merenjima izvora u Faradejevom kavezu. Rezultati obrade su upoređeni sa onima dobijenim u anehoičnoj sobi i rezultatima procene parametara izvora korišćenjem analizatora spektra i numeričke obrade. Tom prilikom korišćen je namenski projektovan i napravljen izvor elektromagnetskog zračenja. U disertaciji je, osim toga, prikazano precizno modelovanje Faradejevog kaveza (uključujući razne nesavršenosti), a dat je i opis merenja odziva ekvivalentnih dipola u Faradejevom kavezu. Pri tome, bilo je neophodno odrediti mernu nesigurnost ispitivanja u Faradejevom kavezu i mernu nesigurnost modela Faradejevog kaveza. Takođe, precizno je definisana potrebna merna postavka pri merenjima izvora u Faradejevom kavezu. Na kraju disertacije predložene su smernice za budući rad.
In this dissertation a new algorithm has been developed that enables determination of free- space field levels of radiating sources using amplitude measurements in Faraday cages. The new approach enables using Faraday cages as a low-cost alternative to the standard EMC test sites, such as (semi)anechoic chambers, open-area test sites, and reverberation chambers. In comparison to the standardized EMC tests in reverberation chambers, the new method enables measurements at low frequencies, even below the lowest resonant frequency of the cage. The algorithm computes the radiation parameters of a DUT in a free space, by means of equivalent dipoles and sparse processing. In order to extract the parameters of the equivalent dipoles, we use sparse processing and l1 regularization for solving ill-posedness of this inverse electromagnetic problem. By exploiting the multipath from the Faraday cage walls, floor, and ceiling, we improve the estimation accuracy. The algorithm has been verified experimentally by taking phaseless measurements of a known device in the Faraday cage. The results of the reconstruction have been compared with those obtained in an anechoic chamber and with directly assessed parameters of the source used a spectrum analyzer and a full-wave numerical model. The source of electromagnetic radiation was designed and developed specifically for this purpose. In this dissertation the model of the Faraday cage and numerical evaluation of the responses of the equivalent dipoles in the Faraday cage are described. Consequently, it has been necessary to determine the measurement uncertainty of testing in the Faraday cage and the measurement uncertainty of the model Faraday cage. Also, the measurements setup for measurements of the radiating sources in Faraday cages is defined. At the end of this dissertation guidelines for the future work are proposed.
true
Eksperimentalno fotometrijsko ispitivanje strukture laminarnog plamena nadstehiometrijske smeše propan - butan - vazduh
Experimental photometric investigation of lean premixed laminar propane-butane-air flame structure
Razvijena je eksperimentalna optička metoda za identifikaciju i merenje relativne koncetracije CH* radikala u cilju identifikacije karakterističnih zona fronta laminarnog predmešanog plamena gasovitih goriva. Izvršena su eksperimentalna istraživanja karakterističnih lokalnih zona prisustva radikala, raspodele koeficijenta viška vazduha i zona nastajanja NO u laminarnom plamenu savremenog atmosferskog gorionika. Sistematski su varirani koeficijent viška vazduha od 1,0 do 1,4 i toplotne snage gorionika u opsegu 2:1, pri laminarnim režimima strujanja (Re brojevi plamenih otvora u opsegu od 100 do 250). Istražena je korelacija relativne koncentracije CH* radikala, zona formiranja CH i OH radikala, raspodele koeficijenta viška vazduha, temperature i emisije NOx. Pokazano je da je ovom fotometrijskom metodom moguće kvalitativno pratiti front plamena i optimirati proces sagorevanja u cilju ostvarivanja stabilnog plamena i smanjene emisije NOx u realnim sistemima sagorevanja.
An experimental optical method has been developed in order to identify and measure relative concentration of CH* radicals in order toidentifycharacteristic zones of laminar premixed laminar gaseous fuels-air flames. Experimental investigation of characteristic local zones in terms of presence of radicals, distribution of air coefficient and production of NO in laminar flames of an up to date atmospheric burner was conducted. Systematic variation of air coefficient (from 1.0 to 1.4) and heating power (2:1 span) was performed. Re number of flame ports varied between 100 and 250. Correlation between CH* radical and CH and OH radicals zones, air coefficient, temperature distributions and NOx emission was found. It was shown that, based on this photometric method, it was possible to monitor flame front qualitatively and to optimize combustion in order to establish stable flame and lower NOx emission in real combustion systems.
true
Javne agencije i zaštita životne sredine u pravu Evropske unije i u pravu Srbije
Javna uprava, kao i način na koji ona obavlja svoje poslove, doživela je veliku transformaciju u poslednjih trideset godina usled značajnih globalnih političkih, pravnih i ekonomskih promena. Važan aspekt upravnih reformi predstavlja široko rasprostranjeno korišćenje javnih agencija kao oblika organizacije uprave, u cilju obavljanja rastućeg broja poslova koji zahtevaju visok nivo specijalizovanih znanja. Predmet ove disertacije su različiti aspekti agencifikacije uprave, pre svega na evropskom nivou, ali i u Srbiji u kontekstu evropskih integracija, uz ispitivanje specifičnosti koje ovaj oblik organizovanja dobija u oblasti zaštite životne sredine. Težište ovog istraživanja se postavlja na pravna pitanja, u prvom redu na regulatornu funkciju agencija na evropskom nivou, njihovu interakciju s drugim relevantnim telima i akterima u regulatornom procesu, obavljanje različitih upravnih poslova, ali najviše na učešće u donošenju akata (odluka) , kako opštih tako i pojedinačnih, kao i na poslove inspekcijskog nadzora, koji je od ključnog značaja za valjanu implementaciju Prava životne sredine. U studiji se ispituju osobenosti procesa agencifikacije i transformacije javne uprave i uticaj evropskih modela na organizaciju obavljanja poslova u oblasti zaštite životne sredine u Republici Srbiji. Cilj istraživanja bio je da se sagledaju evolucija i trenutni položaj i uloga agencija u oblasti zaštite životne sredine u EU i Srbiji, njihova interakcija sa složenim institucionalnim aranžmanima na evropskom nivou i preispita svrsishodnost njihovog daljeg osamostaljivanja uz širenje nadležnosti, pre svega regulatornih. Cilj istraživanja bio je i da pruži doprinos teorijskom ispitivanju koncepta agencije kao oblika organizacije uprave na nacionalnom i supranacionalnom nivou, kao i njegovim funkcionalnim mogućnostima na nivou Evropske unije i u Republici Srbiji, a posebno osobenosti u oblasti zaštite životne sredine. Disertacija „Javne agencije i zaštita životne sredine u pravu Evropske unije i u pravu Srbije“ sastoji se od Uvoda, tri dela i Zaključka. Nakon uvodnog razmatranja predmeta, ciljeva, društvene i naučne relevantnosti teme, kao i primenjene metodologije, u prvom delu predmet istraživanja predstavljaju fenomen agencifikacije na nacionalnom i supranacionalnom nivou, položaj i vrste agencija u pravu EU, kao i učešće agencija u donošenju odluka u Evropskoj uniji, što prepoznajemo kao najizazovnije pravno pitanje u vezi sa funkcionisanjem evropskih agencija danas. U prvom delu disertacije je, takođe, izvršen uvod u problematiku agencifikacije i oslikan značaj ovog fenomena za upravnopravnu analizu i razumevanje upravnih sistema Evropske unije i Srbije – oba u kontinuiranoj transformaciji. U drugom delu, nakon analize podele nadležnosti između država članica i Unije u poslovima zaštite životne sredine, ispituju se položaj i uloga dve postojeće agencije u ovoj oblasti – Evropske agencije za životnu sredinu i Evropske agencije za hemikalije. Analizom u drugom i trećem delu doktorata, uočeno je da je oblast zaštite životne sredine veoma pogodna za sagledavanje (interdisciplinarnog) fenomena agencifikacije, kako u Evropskoj uniji tako i u Srbiji. Ovo je, pre svega, posledica prirode pravnih izazova sa kojima se savremena javna uprava, bilo nacionalna bilo nadnacionalna, bilo kao zajednička/integrisana/jedinstvena suočava u ovoj oblasti. U trećem delu, nakon uvodnog razmatranja agencifikacije u Srbiji i pravnog okvira za delovanje javnih agencija u našem pravu prema Zakonu o javnim agencijama, ispituje se položaj i uloga Agencije za zaštitu životne sredine (jedinstvene agencije u našem pravu, koja je po svom položaju organ u sastavu nadležnog ministarstva) kao i uticaj koji je model Evropske agencije za zaštitu životne sredine imao na oblikovanje poslova naše agencije. Konačno, u trećem poglavlju se posmatraju dodatni aspekti agencifikacije i deagencifikacije naše javne uprave u oblasti zaštite životne sredine, kroz analizu položaja i uloge (ukinute) Agencije za hemikalije i organizacije koja je, uprkos težnjama da se ukine, nastavila sa radom – Agencije za zaštitu od jonizujućih zračenja i nuklearnu sigurnost Srbije. U toku rada na istraživanju u Republici Srbiji su se odigrali paralelni procesi agencifikacije i deagencifikacije, a domen zaštite životne sredine prepoznat je kao lokus ovih promena, uzbudljivih u istraživačkom i teorijskom, ali i u praktičnom smislu. Ovi nalazi poslužili su za formulisanje zaključaka u vezi sa daljim tokom procesa agencifikacije i neizvesnim mogućnostima za formiranje agencije sa širim upravnim ovlašćenjima u oblasti zaštite životne sredine u našem upravnom sistemu.
Public administration organisation and the performance of public tasks have significantly altered over the past 30 years through global political, legal and economic change. An important aspect of the administrative reforms is the widespread use of public agencies as a model of administrative organisation to perform an increasing number of specific tasks requiring high levels of expertise. This PhD thesis analyses the agencification of public administration and the specificities of the process in the field of environmental protection at the European level and in Serbia. The main focus of the research is on the legal aspects of the phenomenon, namely the regulatory functions of agencies at the European level: their interaction with other relevant bodies and actors in the regulatory process; their performance of different administrative tasks (including inspections); and their decision-making roles. The thesis examines specific aspects of the agencification process, the transformation of Serbian public administration in the context of European integration, and the influence this transformation has had on environmental governance. The goal of the research is to examine the evolution and current position of environmental agencies in the European Union’s and Serbia’s legal orders, to analyse the agencies’ interaction with complex administrative arrangements developed at the supranational level, and to argue for enhancing their independence and (regulatory) competencies. The further goal is to contribute to legal scholarship on agencification at national and supranational levels, and especially in the field of environmental protection. The thesis consists of the introduction, three parts and the conclusion. The introduction frames and contextualises the research topic, goals and scientific relevance, and it provides a methodological framework. Part one examines the agencification process primarily at the European level but also nationally. The thesis draws attention to EU-level agency position, typology and administrative decision-making, which are highlighted as challenging issues for legal analysis. The second part examines environmental agencies in the EU. This is done first by studying the division of competencies between the EU and member states for the implementation of environmental acquis. The thesis then turns to the development of the European Environmental Agency and European Chemicals Agency, the tension over their functions, and conflicting perspectives on the organisation of community administration and the implementation of relevant regulatory regimes. The third part examines agencification in the Serbian context by studying the transformation of Serbian public administration through the process of European integration and harmonisation with environmental acquis. We study the influence of the process on the creation of three agencies in the area of environmental protection; however, this is contrasted with the de- agencification process that occurred during this research project. It enabled to draw valuable conclusions and sketch the chaotic legislative and political context in which protection of the environment and the expertise necessary for the required highly complex tasks is being sidetracked and discredited. All findings from the analysis are used to gather conclusions on the legal transformation of agencies both at the European level and in Serbian law, as well as these agencies’ position and prospects in a changing legal landscape. The findings are also used to suggest possibilities for further development of the agency model in Serbian law, but also to contextualise Serbian environmental agencies in an ever-changing European governance process.
true
Modifikacije metode promenljivih okolina i njihove primene za rešavanje problema raspoređivanja prenosa datoteka
Modifications of the variable neighborhood search method and their applications to solving the file transfer scheduling problem
Metoda promenljivih okolina se u praksi pokazala vrlo uspesnom za resavanje pro- blema diskretne i kontinualne optimizacije. Glavna ideja ove metode je sistematska promena okolina unutar prostora resenja u potrazi za boljim resenjem. Za opti- mizaciju funkcija vise promenljivih koriste se metode koje nalaze lokalni minimum polazeci od zadate pocetne tacke. U slucaju kada kontinualna funkcija ima mnostvo lokalnih minimuma, nalazenje globalnog minimuma obicno nije lak zadatak jer najcesce dostignuti lokalni minimumi nisu optimalni. Kod uobicajenih implementa- cija sa ogranicenim okolinama razlicitih dijametara iz proizvoljne tacke nije moguce dostici sve tacke prostora resenja. Zbog toga je strategija koriscenja konacnog broja ogranicenih okolina primenjiva na probleme kod kojih optimalno resenje pripada nekom unapred poznatom ogranicenom podskupu skupa IRn. U cilju prevazilazenja pomenutog ogranicenja predlozena je nova varijanta meto- de, Gausovska metoda promenljivih okolina. Umesto denisanja niza razlicitih okolina iz kojih ce se birati slucajna tacka, u ovoj metodi se sve okoline pokla- paju sa celim prostorom resenja, a slucajne tacke se generisu koriscenjem razlicitih slucajnih raspodela Gausovog tipa. Na ovaj nacin se i tacke na vecem rastojanju od tekuce tacke mogu teorijski dostici mada sa manjom verovatnocom. U osnovnoj verziji metode promenljivih okolina neophodno je unapred denisati sistem okolina, njihov ukupan broj i velicinu, kao i tip raspodele koja ce se koristiti za odabir slucajne tacke unutar tih okolina. Gausovska metoda promenljivih okolina za razliku od osnovne verzije ima manje parametara jer su sve okoline teorijski iste velicine (jednake celom prostoru pretrage) i imaju jedinstvenu jednoparametarsku familiju raspodela Gausovu raspodelu slucajnih brojeva sa promenljivom dispe- rzijom. Problem raspored-ivanja prenosa datoteka (File transfer scheduling problem - FTSP) je optimizacioni problem koji svoju primenu pronalazi u mnogim oblastima poput telekomunikacijama, LAN i WAN mrezama, raspored-ivanju u okviru MIMD (multiple instruction multiple data) racunarskih sistema i dr. Spada u klasu NP teskih problema za cije resavanje se uobicajeno koriste heuristicke metode. Za- datak optimizacije FTSP sastoji se u trazenju odgovarajuceg rasporeda pojedinacnih prenosa datoteka, tj. vremenskih trenutaka kada ce svaka datoteka zapoceti svoj prenos tako da duzina vremenskog intervala od trenutka kada prva datoteka zapocne prenos do trenutka u kom poslednja zavrsi bude sto manja...
The Variable neighborhood search method proved to be very successful for solving discrete and continuous optimization problems. The basic idea is a systematic change of neighborhood structures in search for the better solution. For optimiza- tion of multiple variable functions, methods for obtaining the local minimum starting from certain initial point are used. In case when the continuous function has many local minima, nding the global minimum is usually not an easy task since the obta- ined local minima in most cases are not optimal. In typical implementations with bounded neighborhoods of various diameters it is not possible, from arbitrary point, to reach all points in solution space. Consequently, the strategy of using the nite number of neighborhoods is suitable for problems with solutions belonging to some known bounded subset of IRn. In order to overcome the previously mentioned limitation the new variant of the method is proposed, Gaussian Variable neighborhood search method. Instead of dening the sequence of dierent neighborhoods from which the random point will be chosen, all neighborhoods coincide with the whole solution space, but with die- rent probability distributions of Gaussian type. With this approach, from arbitrary point another more distant point is theoretically reachable, although with smaller probability. In basic version of Variable neighborhood search method one must dene in advance the neighborhood structure system, their number and size, as well as the type of random distribution to be used for obtaining the random point from it. Gaussian Variable neighborhood search method has less parameters since all the neighborhoods are theoretically the same (equal to the solution space), and uses only one distribution family - Gaussian multivariate distribution with variable dispersion. File transfer scheduling problem (FTSP) is an optimization problem widely appli- cable to many areas such as Wide Area computer Networks (WAN), Local Area Ne- tworks (LAN), telecommunications, multiprocessor scheduling in a MIMD machines, task assignments in companies, etc. As it belongs to the NP-hard class of problems, heuristic methods are usually used for solving this kind of problems. The problem is to minimize the overall time needed to transfer all les to their destinations for a given collection of various sized les in a computer network, i.e. to nd the le transfer schedule with minimal length...
false
Redukcija talasnosti kod direktne kontrole momenta asinhronog motora primenom komparatora sa više nivoa
Ripple reduction in direct torque control of induction motor by using multilevel comparators
Savremeni elektromotorni pogoni zahtevaju od upravljačkih algoritamskih struktura širok propusni opseg i malu grešku u regulisanoj veličini. Takođe, kao bitan cilj se postavlja i eliminisanje davača na vratilu motora sa namerom povećanja robustnosti pogona. Direktna kontrola momenta (DTC – Direct Torque Control) kao najmlađa tehnika kontrole mašina naizmenične struje koja u svojoj osnovi pripada tehnikama koje ne zahtevaju davač na vratilu, odlikuje se jednostavnošću i širokim propusnim opsegom po momentu kao upravljačkoj veličini. Direktna kontrola momenta, od svog nastanka do danas, pretrpela je veliki broj modifikacija u cilju eliminisanja njenih nedostataka od kojih se kao najveći ističe talasnost (ripl) momenta. Ove modifikacije uspešno smanjuju vrednost intenziteta ripla momenta na jednoj strani dok na drugoj strani neretko podrazumevaju smanjenje brzine odziva momenta motora, zahtevaju složene estimatore fluksa i proračune koji produžavaju vreme izvršavanja algoritma čime se narušavaju karakteristike pogona i jednostavnost originalne DTC. U ovoj tezi predložen je algoritam direktne kontrole momenta motora koji uvodi višestruke naponske vektore i odgovarajuće komparatore sa više nivoa u cilju redukcije ripla momenta. Predloženi algoritam zadržava proračunsku jednostavnost konvencionalne DTC i brz odziv momenta. Umesto šest osnovnih nenultih naponskih vektora predloženi algoritam koristi veći broj diskretizovanih naponskih intenziteta (Discretized Voltage Intensities - DVI-DTC) u cilju smanjenja inkremenata momenta u okviru prekidačkog ciklusa i ukupnog ripla momenta. Za razliku od do sada publikovanih rešenja gde je povećanje broja korišćenih naponskih intenziteta skopčano sa problemima vezanim za proširenje prekidačke tabele, predloženi algoritam nema potrebu za izmenom konvencionalne prekidačke tabele sa promenom broja definisanih naponskih intenziteta. Predloženi algoritam uvodi raspregnut izbor smera i intenziteta odgovarajućeg naponskog vektora. Rasprezanje izbora smera i intenziteta dalje omogućava jednostavno prilagođenje algoritma u zavisnosti od željenog broja definisanih naponskih intenziteta bez izmene prekidačke tabele. Da bi ripl momenta u zavisnosti od broja definisanih naponskih intenziteta bio tačnije određen, izvršena je analiza uticaja indukovane elektromotorne sile na inkremente momenta i statičku grešku momenta u odnosu na zadatu vrednost. U radu je predložen način za eliminisanje uticaja indukovane elektromotorne sile čime je postignuto da inkrement momenta u okviru prekidačkog ciklusa zavisi samo od intenziteta primenjenog naponskog vektora. Dobijeni rezultati su najpre potvrđeni simulacijom, a potom i eksperimentalnim putem uz detaljan opis eksperimentalne postavke i postupka izvođenja eksperimenata. U radu je izvršeno višestruko uzorkovanje terminalnih veličina motora u cilju paralelne estimacije fluksa i momenta mašine. Paralelna estimacija sa višestruko uzorkovanim veličinama omogućila je uvid u ponašanje momenta u okviru prekidačkog ciklusa i tačnije određivanje njegovog ripla. Takođe, izvršena je analiza stepena iskorišćenosti napona jednosmernog međukola i brzine odziva momenta kod predloženog DVI-DTC algoritma uz selektivnu kompenzaciju indukovane elektromotorne sile. Najznačajniji rezultati eksperimentalne analize dati su na kraju rada gde je prikazan stepen redukcije ripla momenta u zavisnosti od broja definisanih naponskih intenziteta. U toku izrade teze analizirana je brojna naučna literatura kroz hronološki pregled stanja i istoriju razvoja oblasti direktne kontrole momenta. Referentna literatura je obrađena u prvom poglavlju i predstavlja dobru podlogu za razradu teme ove doktorske disertacije.
Modern electric drives demand high system bandwidth and small regulation error from control algorithm structures. Also, elimination of shaft sensors within drives in order to intensify robustness of the drives is very important aim. Direct Torque Control (DTC) as the youngest technique for control of AC machines, is in its basis shaft-sensorless technique and stands out for its simplicity and high torque bandwidth. Direct torque control has gone through various modifications from its first appearance till today in order to eliminate its disadvantages the biggest of which is torque ripple. These modifications have successfully suppressed the torque ripple, but on the other hand they have often reduced the bandwidth of motor torque response, and usually demand complex flux estimators and calculations that prolong algorithm execution which deteriorate drives characteristics and simplicity of the original DTC. This thesis proposes DTC algorithm that implements multiple voltage vectors and adequate multilevel comparators with the goal to reduce torque ripple. Suggested algorithm keeps calculation simplicity of the conventional DTC and quick torque response. Instead of six basic active voltage vectors, proposed algorithm uses larger number of Discretized Voltage Intensities (DVI-DTC) aiming to reduce torque increment within switching cycle and overall torque ripple. Contrary to solutions published by now, where larger number of used voltage intensities are linked to problems concerning switching table modification, proposed algorithm doesn’t require changes in the conventional switching table as a consequence of changing the number of defined voltage intensities. Suggested algorithm brings in a possibility to separately choose direction and intensity of the appropriate voltage vector. Separating the choice of direction and intensity makes further possible simple adjustments of the algorithm depending on the number of defined voltage intensities without switching table modification. Analysis of the induced electromotive force’s impact on torque increments and steady- state torque error has been conducted in order to determine torque ripple more accurately in relation to the number of defined voltage intensities. This thesis proposes a way to eliminate the influence of the induced electromotive force which further provides that torque increment within a switching cycle depends solely on the intensity of the applied voltage vector. Obtained results were first confirmed by means of simulation, followed by experimental verification. Detailed description of the experimental setup and procedure of the experiment performance are presented. Within this thesis oversampling the motor’s terminal electrical signals has been done with a purpose of parallel estimation of the flux and torque of the machine. Parallel estimation with oversampled quantities provided the insight into the torque behaviour within a switching cycle and estimating the ripple more accurately. Also, the degree of utilization of the DC link voltage and torque rise time with proposed DVI-DTC algorithm was analysed with selective compensation of the induced electromotive force. The most important results of the experimental analysis are given in the later part of the thesis where the degree of the torque ripple reduction is shown in correlation to the number of defined voltage intensities. In the process of writing this thesis, numerous scientific literature was analysed through chronological overview and history of the development of the direct torque control. Significant literature references were processed in the first chapter and represent a good foundation to further develop the essence of this doctoral dissertation.
true
Detekcija epileptiformnih aktivnosti u EEG signalima primenom statističkog prepoznavanja oblika
Detection of epileptiform activities in EEG signals based on statistical pattern recognition
Epilepsija predstavlja najučestaliji poremećaj moždane aktivnosti i ispoljava se iznenadnim i ponavljajućim napadima koji su posledica prekomernog i sinhronog električnog pražnjenja velikog broja neurona u mozgu koje se može u većini slučajeva vizuelno uočiti na elektroencefalogramu (EEG). Obzirom da standardni farmakološki način lečenja ovog poremećaja vrlo često ne daje očekivane rezutate najnovija istraživanja su usmerena u pravcu razvoja sistema za direktnu električnu stimulaciju mozga koji bi se u bliskoj budućnosti mogli implantirati u lobanju čoveka i na osnovu snimljenih EEG signala bili u stanju da najpre predvide ili što je pre moguće detektuju epileptični napad a zatim ga električnom stimulacijom izbegnu ili stave pod kontrolu odnosno skrate njegovo trajanje. Jedan od najkritičnijih delova jednog takvog sistema jeste pouzdana tehnika za detekciju epileptiformnih aktivnosti u EEG signalima čiji rezultat zapravo dalje inicira električnu stimulaciju. Pored toga pouzdana automatska tehnika za detekciju bi omogućila i dosta bržu i objektivniju analizu već snimljenih EEG signala što zna da bude veoma naporan posao i zahteva angažovanje eksperta čime bi takođe i eventualne greške u analizi bile značajno umanjene. Glavni cilj istraživanja u okviru ove doktorske disertacije jeste razvoj i testiranje jedne nove tehnike za detekciju epileptiformnih aktivnosti u EEG signalima. Nova tehnika je zasnovana isključivo na informacijama sadržanim u EEG signalima i pokušava prevazići određene nedostatke već postojećih tehnika i na taj način dati svoj dodatni doprinos u ovoj i dalje veoma atraktivnoj istraživačkoj oblasti. Pored pažljivog izdvajanja i optimalne selekcije obeležja iz nekoliko domena i kliničkih podopsega EEG signala kao ključnih koraka nova tehnika obuhvata i projektovanje optimalnog deo po deo kvadratnog klasifikatora čija je tačnost testirana na segmentima EEG signala snimljenih u Centru za epilepsiju Univerziteta u Bonu i koji je u poređenju sa već postojećim vii tehnikama pokazao prilično visoku tačnost i time potencijal za dalje testiranje u realnim kliničkim uslovima.
Epilepsy is the most common brain activity disorder characterized by sudden and recurrent seizures. Epileptic seizures are a consequence of excessive and synchronous electrical discharges of a large number of neurons that is in most cases evident in the electroencephalogram (EEG). Since the most common treatment of epilepsy is based on regularly administered medication quite often does not give the expected results, a lot of recent studies have focused on the development of a system for direct electrical stimulation of the brain. In the near future, such a system would be implanted in the head of a patient, and based on the recorded EEG signals, we would at first be able to predict or as soon as possible detect seizures and later on using electrical stimulation avoid or control them. One of the most critical points of such a system is a reliable technique for the detection of epileptiform activities in the EEG signals, whose output in fact triggers the electrical stimulation. In addition, reliable and automatic detection would ensure an objective and facilitating analysis of long-term EEG recordings, which is a very tedious and time-consuming process for an expert. The possibility that the expert could misread the data and fail to make a proper decision would be decreased as well. The main goal of the research within this doctoral disertation is the development and testing of a new technique for the detection of epileptiform activities in the EEG signals. The new technique is based only on information contained in the EEG signals. It aims to overcome some disadvantages of already existing techniques available in the literature, and in that way give its own contribution to the research field that is still attractive. Apart from a very careful and optimal selection of features from a few domains and sub-bands of the EEG signals, as its key steps the new technique also includes the design of an optimal piecewise quadratic classifier whose accuracy is tested on the EEG signals recorded at the Center for Epilepsy, University of Bonn, Germany. The results obtained demonstrated a quite high accuracy in comparison to the existing tecniques, and thus a potential for further testing under real clinical conditions. ix
true
Značaj ekspresije faktora proliferacije i proteina regulacije ćelijskog ciklusa u gigantocelularnom tumoru kosti
The significance of expression proliferation factors and cell cycle regulation proteins in giant cell tumor of bone
Gigantocelularni tumor kosti (GCTB) je primarni koštani tumor koga karakteriše prisustvo dve ćelijske populacije: džinovske ćelije tipa osteoklasta i monojedarne stromalne ćelije koje predstavljaju pravu neoplastičnu tumorsku komponentu. Biološko ponašanje GCTB je nepredvidivo. Terapija izbora za gigantocelularni tumor kosti je hirurška. Međutim, pojava recidiva nakon hirurške intervencije je opisana kod značajnog broja pacijenata i kreće se od 10 do 25%. Postoji veliki broj studija koje se bave analizom biološkog ponašanja GCTB, pre svega sa ciljem predviđanja nastanka recidiva. One su veoma značajne jer ukazuju da GCTB sa agresivnijim biološkim ponašanjem zahteva i agresivniji terapijski pristup. Naše istraživanje se odnosi na određivanje različitih kliničko-radiološko-patohistoloških parametara koji bi mogli uticati na predikciju biološkog ponašanja GCTB, odnosno predviđanje pojave recidiva. Izdvajanjem i jasnim definisanjem karakteristika tumora koje bi mogle ukazivati na povećan rizik za ponovnu pojavu GCTB, nakon hirurške terapije, moglo bi eventualno opravdati i ekstenzivnost/agresivnost terapijskog pristupa. Ciljevi: Gigantocelularni tumor kosti je benigni tumor, koji i danas ima nepredvidljivo ponašanje, morfološko i kliničko. U cilju preciznijeg utvrđivanja ponašanja tumora i utvđivanja novih prognostičkih faktora, postavili smo sledeće ciljeve: 1. Utvrđivanje korelacije ekspresije Ki-67, p53 i Ciklin D1 u tkivu gigantocelularnog tumora kosti sa biološkim ponašanjem ovog tumora (pojavom rest-recidiva ili pojavom plućnih metastaza). 2. Utvrđivanje korelacije ekspresije Ki-67, p53 i Ciklin D1 u tkivu gigantocelularnog tumora kosti sa kliničko-patološkim prognostičkim parametrima (klinička GTM, Enneking-ova klasifikacija). Metodologija: Ova teza je izvedena kao studija preseka, u koju je uključeno 164 pacijenata kojima je u periodu od 1964-2008. god. na Institutu za patologiju Medicinskog fakulteta u Beogradu dijagnostikovan gigantocelularni tumor kosti, a koji su operisani na Institutu za Ortopedsko- hirurške bolesti „Banjica“. Kriterijumi za uključenje u studiju bili su: patohistološki verifikovan gigantocelularni tumor kosti. Kriterijumi za isključenje iz studije bili su: nekompletnost podataka preuzetih iz Istorije Bolesti IOHB "Banjica”, period praćenja kraći od 3 godine. Rezultati: U cilju određivanja potencijalnih prediktivnih parametara za nastanak recidiva analizirani su kliničko-radiološko-patohistološke karakteristike tumora. Univarijantnom logističkom regresijom(ULR) je od 16 analiziranih varijabli, koje predstavljaju potencijalne prediktorne faktore za nastanak recidiva GCTB, izdvojno 4 visoko-statistički značajna parametra: 1) lokalizacija lezije, 2) broj hirurških intervencija,3) nuklearna ekspresija p53 u monojedarnim ćelijama, 4) nuklearna ekspresija CyclinaD1u džinovskim, multijedarnim ćelijama. Utvrđeno je češće recidiviranje GCTB u humerusu (HR=3,641, p<0,001) i radijusu (HR=4,798, p=0,027), nego u femuru. Zapaženo je da sa povećanjem broja hirurških reintervencija raste i verovatnoća za nastanak recidiva. Nakon treće intervencije rizik za ponovno pojavu tumora je 11 puta veći, u odnosu na primarnu biopsiju (HR=11,026, p<0,001). Od imunohistohemijskih parametara koji se mogu koristiti na biopsijskom uzorku, prilikom inicijalne dijagnostike pozitivnost monojedarnih ćelija na p53 je gotovo 7 puta bila češća kod tumora koji su dali recidive (HR=6,677, p<0,001). Tumori koji su imali Cyclin D1 pozitivna jedra u dživskim ćelijama su 2 puta češće recidivirali od lezija koje nisu pokazivale ovu pozitivnost (HR=2,193, p=0,005). Univarijantna regresija je ove varijable izdvojila kao nezavisne prognostičke parametre. Takođe su stitstičkim alatima (projekcija ULR na veći uzorak) u potencijalno značajne prediktore uključena i sledeća dva kriterijuma: 1) jasna destrukcija korteksa sa prisustvom ekstrakomapartmentalne lezije i 2) histološki kriterijumi za malignost. U našem uzorku je 2 puta češća pojava recidiva kod destrukcije korteksa (HR=2,346, p=0,075), dok maligni izgled tumora 3 puta češće ukazivao na pojavu recidiva (HR=3,418, p<0,043). Sve izdvojene varijable su uključene u multivarijantnu logističku regresiju (MLR), čime je utvrđeno da je ekspresija p53 u monojedarnim ćelijama najznačajniji prediktorni faktor (HR=6,181 p<0,001), čija pozitivnost ukazuje Na 6 puta veću verovatnoću za pojavu recidiva u GCTB. Ostali parametri nisu multivarijantnom regresijom potvrđeni kao prediktori nastanka recidiva. Međutim, uočeno je da je ekspresija Ciclina D1 u džinovskim ćelijama, koje sadrže manje od 15 jedara, takođe stistički značajna (HR=8,398, p=0,038) za predikciju recidiviranja i ukazuje na 8 puta češći nastanak recidiv kod pozitivnih tumora. Zaključci: Analiza ekspresije faktora proliferacije, kao i regulatora ćelijskog ciklusa, ipak ima značajanu ulogu u proceni i predviđanju ponašanja GCTB. Najznačajnija uloga se ogleda u imunohistohemijskoj nuklearnoj pozitivnosti stromalnih ćelije na p53, i on je najznačajniji parametar u predikciji recidiva. Drugo po značajnosti je nuklearna pozitivnost Cyclina D1, i to u ćelijama sa manje od 15 jedara i takođe se može reći da predstavlja predikcioni faktor. I poslednja ali ne manje važna je ekspresija β-katenina u citoplazmi monojedarnih, kao i u jedrima džinovskih ćelija, koja visoko korelira sa nastankom patološke fracture kod GCTB. Pozitivnost β-katenina u ćelijama GCTB ukazuje i na aktivaciju Wnt puta, a to predstavlja doprinoseći faktor za lokalno destruktivno ponašanje ovog tumora. Kada ove parametre povežemo sa kliničkim pokazateljima agresivnog ponašanja GCTB zaključujemo, da se postavkom ovih parametara može, ipak sa većom sigurnošću postaviti preciznija indikacija, za agresivniji terapijski – hirurški pristup u terapiji GCTB.
Giant cell tumor of bone (GCTB) is a primary bone tumor, characterized by the presence of two cell populations: osteoclast-like giant cells and mononuclear stromal cells, which are, in fact, the real neoplastic tumor component. Biological behavior of the GCTB is unpredictable. Apart from the locally aggressive growth followed by the destruction of bone tissue and propagation in the surrounding soft tissue, as well as frequent incidence of fractures and relapses, the occurrences of pulmonary metastases are also reported. The therapy of choice for giant cell tumor of bone is surgical. However, the recurrence after surgical intervention is found in significant number of patients, i.e. at a rate of 10 to 25%. A large number of studies deal with analyzing the biological behavior of GCTB, primarily in order to predict the recurrence. Those studies are of great importance, since they imply that GCTB with a more aggressive biological behavior also requires a more aggressive therapeutic approach. Our research refers to determining various clinical, radiological and pathohistological parameters which may have an influence on predicting the biological behavior of GCTB, i.e. on predicting the incidence of relapses. Detecting and clearly defining the tumor characteristics which could indicate an increased risk of GCTB recurrence after surgical therapy might account for the extensiveness/aggressiveness of therapeutic approach. Objective: Giant Cell Tumor of Bone is a benign tumor, which still has an unpredictable behavior, morphology and clinical. In order to precisely determine the behavior of the tumor and the determining of new prognostic factors, we set the following objectives: 1. Establishing correlations expression Ki-67, p53 and Cyclin D1 in the tissues of giant cell tumor of bone with the biological behavior of this tumor (the appearance of rest-relapse or emergence of pulmonary metastases). 2. Establishing correlations expression of Ki-67, p53 and Cyclin D1 in the tissues of giant cell tumor of bone with clinicopathological prognostic parameters (clinical GTM, Enneking's classification). Methodology: This thesis was carried out as a cross-sectional study, which included 164 patients in the period from 1964 to 2008.god. at the Institute of Pathology, School of Medicine in Belgrade diagnosed with Giant Cell Tumor of Bone, which were operated at the Institute for Orthopedic Surgery "Banjica". Criteria for inclusion in the study were: histologically verified Giant Cell Tumor of Bone. Criteria for exclusion from the study were: the completeness of the data taken from the History of Diseases IOHB „Banjica”, a shorter follow-up period of 3 years. Results: In order to determine possible predictive parameters for recurrence, the clinical, radiological and pathohistological characteristics of tumors were analyzed. Out of 16 analyzed variables, which presented possible predictive factors for the incidence of relapse of GCTB, univariate logistic regression (ULR) was used to extract 4 highly statistically significant parameters: 1) lesion localization, 2) number of surgical interventions, 3) nuclear p53 expression in mononuclear cells, 4) nuclear Cyclin D1 expression in giant multinuclear cells. A more frequent recurrence of GCTB was detected in humerus (HR=3,641, p<0,001) and radius (HR=4,798, p=0,027) than in femur. It was observed that, with the increase of surgical reinterventions, the probability of relapse also increased. After the third intervention, the risk of tumor recurrence was 11 times greater in comparison to the primary biopsy (HR=11,026, p<0,001). Out of immunohistochemical parameters which could be used on biopsy samples, upon initial diagnostics, the positivity of mononuclear cells on p53 was almost 7 times more frequent in recurrent tumors (HR=6,677, p<0,001). Tumors displaying Cyclin D1 positive nuclei in giant cells were 2 times more recurrent than lesions lacking in this positivity (HR=2,193, p=0,005). Univariate regression detected these variables as independent prognostic parameters. Additionally, statistical tools (ULR projection on a larger sample) were used to include the following two criteria into potentially significant predictors: 1) clear destruction of cortex with the presence of extracompartmental lesion and 2) histological criteria for malignancy. Our sample revealed recurrence which was 2 times more frequent when cortex was destructed (HR=2,346, p=0,075), whereas malignant tumor appearance indicated 3 times more frequent incidence of relapse (HR=3,418, p<0,043). All detected variables were included in the multivariate logistic regression (MLR), revealing that p53 expression in mononuclear cells was the most significant predictive factor (HR=6,181 p<0,001), the positivity of which indicated 6 times higher probability for recurrence in GCTB. There were no other parameters confirmed as recurrence predictors using the multivariate regression technique. However, it was identified that the expression of Cyclin D1 in giant cells, containing less than 15 nuclei, was also statistically significant (HR=8,398, p=0,038) for predicting the recurrence, and that it demonstrated 8 times more frequent recurrence in positive tumors. Conclusions: Analysis of the expression of the factors of proliferation and cell cycle regulators has a significant role in assessing and predicting the behavior GCTB. The most important role is reflected in the immunohistochemical nuclear positivity in stromal cells on p53, and he is the most important parameter in predicting recurrence. The second significance of nuclear cyclin D1 positivity, in cells with less than 15 sails and can also be said to represent a predictive factor. And last, but not least, is the expression of β-catenin in the cytoplasm mononuclear, as well as giant cell nuclei, which is highly correlated with the development of pathologic fracture at GCTB. β- catenin positivity in the cells indicates the activation of Wnt putway, and this represents a contributing factor for the local destructive behavior of this tumor. When these parameters, connected with clinical indicators of aggressive behavior GCTB, we can conclude that the setting of these parameters can, however, a more reliable set of indications, for aggressive surgical approach.
true
Termomehanički procesi pri sušenju prehrambenih materijala u sistemima sa recirkulacijom vazduha
Thermomechanical processes in drying of food materials in systems with air recirculation
Predmet rada u ovoj disertaciji je istraživanje termomehančkih procesa pri konvektivnom sušenju prehrambenih materijala vlažnim vazduhom kao agensom sušenja u sistemima sa recirkulacijom. Matemtičko-numeričkim metodama i regresionom analizom definisan je fizičko-matematički model koji omogućava simulaciju promene vlažnosti materijala u bezdimenzionom obliku tokom procesa sušenja, a koji uzima u obzir uticaj termomehaničkih osobina vlažnog vazduha (temperature, relativne vlažnosti i brzine strujanja) na proces sušenja. U cilju verifikacije postavljenog modela, osmišljena je i izrađena laboratorijska instalacija za konvektivno sušenje prehrambenih materijala u tankom sloju, koja radi sa potpunom recirkulacijom vazduha. Odabrani materijali (jabuka, krompir i banana) su sušeni pod istim režimima, sa temperaturom sušenja u opsegu 35 ÷ 55°C , relativnom vlažnošću u opsegu 10 ÷ 30% i brzinom strujaja vlažnog vazduha u opsegu 1÷ 2 m s . Jabuke su sušene u komadićima oblika kocke dimenzija 10 mm i 13 mm, krompir u listovima debljina 2 mm i 3 mm, a banana u kolutovima debljina 6 mm i 7 mm. Analiziran je uticaj termomehaničkih osobina vlažnog vazduha na kinetiku procesa sušenja, na osnovu dobijenih vrednosti efektivnog koeficijenta difuzije. Rezultati istraživanja su potvrdili da termomehaničke osobine vlažnog vazduha značajno utiču na proces sušenja, pri čemu se u slučajevima slabijeg uticaja na proces sušenja pojedine veličine mogu zanemariti. Analiziran je rad sistema sušenja koji je podržan radom toplotne pumpe, na osnovu izračunatih vrednosti osnovnih pokazatelja efikasnosti rada sistema. Ustanovljeno je da se pravilno utvrđenim parametrima rada instalacije može uticati na povećanje energetske efikasnosti sistema i smanjenje specifične potrošnje energije po količini izdvojene vlage. Verifikacija dobijenih rezultata izvršena je poređenjem rezultata sopstvenog modela sa rezultatima sprovedenog eksperimentalnog istraživanja, a potom i sa modelima i eksperimentalnim istraživanjima drugih autora.
The motivation of the research presented in this thesis was to investigate thermomechanical processes that appears during the convective drying of food materials in systems with air recirculation. The mathematical model was formed to describe the change of moisture ratio of the material during the time, by taking into account the influence of all relevant thermomechanical properties of drying air (temperature, relative humidity and velocity) on the drying process. Convective laboratory scale dryer with possibility of full drying air recirculation was designed and built in order to achieve and measure all relevant drying process parameters. Apples, potatoes and bananas were used in the experiments. The chosen materials were dried under the same drying conditions. Drying air temperature was in range 35-55oC, relative humidity in range 10-30% and velocity in range 1-2 m/s. The apples were cut into cubes with 10 mm and 13 mm side dimension, potatoes were sliced into chips of 2 mm and 3 mm thickness and bananas were sliced into the cylinders with 6 mm and 7 mm thickness. Analysis of error appearance probability was performed. The effective moisture diffusivity was used to describe drying process efficiency. The experimental results confirmed that thermomechanical properties of drying air have significant influence on drying process. It was discovered that drying air, relative humidity could be ignored in some occasions. The performance of heat pump drying system was analyzed in order to discover system efficiency. The system energy and mass flows, air circulation behavior, drying air properties and system components behavior were monitored in order to calculate the main parameters of system performance. It was discovered that appropriate system performance optimization could increase system energy efficiency and reduce specific energy consumption. The results of mathematical model equation were compared with the experimental results and with the models and with the experimental results of the other authors, in order to be verified.
true
Adaptivna ekvalizacija linearnih vremenski - disperzivnih radio kanala pri vrlo brzom prenosu podataka u zatvorenim prostorima
Equalization of Linear Time-Dispersive Intra- and Interbuilding Radio Data Channels
Za telekomunikacione sisteme unutar zgrada, koncept “bežičnog je uveo “novi” prenosni medijum, odnosno kanal za prostiranje radio talasa u zatvorenim prostorima, koji se koristi za povezivanje terminala i ostale opreme u lokalnu računarsku mrežu (LAN). Eksperimentalne studije, urađene u cilju opisivanja kanala za prostiranje talasa u zatvorenim prostorima, su pokazale da su karakteristike prostiranja talasa u zatvorenim prostorima: velika slabljenja pri prostiranju, prostiranje po više puteva između predaje i prijema, spore vremenske promene kanala I feding. Neke primene bežičnih kanala u zatvorenim prostorima zahtevaju veoma brz prenos podataka koji je ograničen prostiranjem po više puteva I fedingom. Kad se poveća brzina prenosa podataka, prostiranje po više puteva prouzrokuje interferenciju među simbolima u toku procesa detekcije simbola. Uobičajeni način borbe protib fedinga koji nastaje pri prostiranju po više puteva, a kao posledica toga i interferencije među simbolima, je upotreba adaptivne ekvalizacije na prijemu. Problem osetljivosti digitalnog radio telekomunikacionog sistema u zatvorenim prostorima na prostiranje po više puteva pri promenjivim parametrima kanala i sitema, je obrađen u ovoj tezi, na primeru 4-QAM telekomunikacionog sistema, sa LMS linearnim ekvalizatorom sa kašnjenjem od jednog simbola između susednih koeficijenata kao i sa LMS nelienarnim (DF) ekvalizatorom sa kašnjenjem manjim od jednog simbola između susednih koeficijenata na prijemu. Rešavanju problema je pristupljeno iz 3 pravca. Prvo je data teorijska analiza problema. Metodom simulacija na računaru je izvršena detaljna procena performansi sistema za veoma brzi prenos podataka po disperzivnim kanalima sa sporim fedingom, koriteći statistički eksponencijalni model kanala sa Relejevim fedingom koji je predložen u jednoj studiji, kao i Rajsov model kanala sa fedingom koji se bazira na merenjima izvršenim u radnim prostorijama i fabričkom okruženju. Konačno. Impulsni odzivi kanala izmereni u naučno-istraživačkoj laboratoriji pri različitim uslovima prostiranja, topografijama (konfiguracijama) i frekvencijama, su takođe korišćeni za procenu performansi sistema. Prosečna verovatnoća greške po bitu je korišćena kao performansni kriterijum. Krive prosečne verovatnoće greške po bitu ukazuju da, bilo kod kanala generisanih uz pomoć računara ili kod izmerenih kanala, linearni ekvalizator smanjuje grešku približno 10 puta (odnos signal-šum od 8dB) u poređenju sa neekvalizovanim kanalom, a da DF ekvalizator nadmašuje linearni ekvalizator dalje smanjujući grešku 2 puta (SNR=8dB). Zapravo, počevši od slučaja frekvencijski-ravnog fedinga, ustanovljeno je da ekvalizator može da smanji grešku za kanale generisane uz pomoć računara skoro do normalizovane vrednosti brzine prenosa podataka od 4 (Relejev slučaj), I, 2.2, 1 I 0.5 (Rajsov slučaj) I vrednost Rajsovog faktora od 2, 6.8 I 11dB, respektivno, dok su u slučaju izmerenih kanala vrednosti minimuma krivih verovatnoće greške između 2 i 3, a pozicije minimuma zavise od veličine prostorije i topografije. Posle minimuma, performanse sistema degradiraju kad se bilo brzina prenosa ili disperzija kanala povećaju, zbog negativnog efekta interferencije među sibolima koji potire poboljšanje usled implicitne (vremenske) raznovrsnosti. Kad je odnos signal-šum nizak (do 12dB zakjlučno), primećeno je da se javlja efekat prostiranja greške. Kod DF prijemnika. Ipak, pokazano je da umetanjem nove, “modifikovane” automatske funkcije čišćenja u filtar povratne sprege kod DF ekvalizatora, se značajno može smanjiti neželjen uticaj efekta prostiranja greške, a samim tim i prosečna verovatnoća greške po bitu i do 2 reda veličine. Na kraju, performanse ekvalizatora su upoređene i sa teorijskom performansom prilagođenog filtra za kanale sa Relejevim fedingom, koji predstavljaju performansnu granicu za sve tehnike eliminacije interferencije među simbolima. Ustanovljeno je da DF ekvalizator sa tačnim povratnim simbolima, koji ima bolje performanse od bilo linearnog ili DF ekvalizatora sa detektovanim povratnim simbolima, daje prosečnu verovatnoću greške po bitu samo oko dva puta veću od “optimalnog” prijemnika čije performanse predstavljaju teorijsku granicu najvećeg mogućeg poboljšanja koje se može dobiti korišćenjem mnogo komplikovanijih prijemnika sa detektorima koji odlučuju po principu funkcije najvećeg izgleda.
The concept of wireless in-building systems has introduced a “new” medium the indoor radio propagation channel, which can be used for interconnecting terminals and computers in LAN-s. Experimental studies have shown have shown that indoor radiowave propagation can be characterized by large propagation losses, multipath, slow time variations and fading. Some computer applications require very high signaling rates. When the data rate is increased, themultipath propagation in the indoor channel causes intersymbol interference in the detection process, thus adversely affecting the system performance. A common approach to combat multipath fading and hence intersymbol intereference, is to use an adaptive equalizer. This thesis considers a 4-QAM communication system with LMS symbol-spaced linear and LMS fractionally-spaced decifion feedback adaptive equalizer at the receiver. The problem is approached in three ways. The system performance evaluation for high-speed signaling over slowly fading dispersive diversity channels is considered first, using the statistical exponential Rayleigh fading channel model, and the Rician fading channel model. These models are based on propagation measurements performed in office and factory environments. Next, the channel impulse responses measured in the research laboratory building for different propagation environments, topographies and frequencies are used for system performance predictions based on computer simulation. Finally, a theoretical performance analysis of the problem is given. The average bit error rate (BER) is used as the system performance quality measure. The average BER curves indicate, for both computer generated and measured channels, that a linear equalizer reduces the BER by approximately a factor of ten at SNR=8dB compared with unequalized channel case. The decision-feedback equalizer surpasses the linear equalizer by further reducing the BER by approximately a factor of two at SNR=8dB. Using the frequency-flat fading case as a reference the decision-feedback equalizer is shown to be able to reduce the BER almost up to the normalized data rate of 4 for the case of Rayleigh fading, and up to 2.2, 1 and 0.5 for Rician fading with Rice factor values of 2, 6.8 and 11dB, respectively. The error performance curves for the measured channels have their best values between the normalized data rates of 2 and 3, with the actual minimum positions depending on the room size and lay-out. Beyond the BER minima, system performance degrades when either the data rate or the multipath spread is increased. This is ue to the negative effect of intersymbol interference, which offsets the time divesity gains. Low signal-to-noise ratio (SNR ≤ 12dB) are observed to cause error propagation effects in the decision-feedback receiver. Insertion of a novel modified automatic “clear” function in the backward filter of the decision-feedback equalizer significantly reduces the error propagation effect, decreasing the BER by up to two orders of magnitude. Finally, equalizer performance is compared with the theoretical matched filter performance for Rayleign fading diversity channels; this represents a lower bound for all intersymbol interference elimination techniques. The decision-feedback equalizer with correct symbols fed back, being better in performance than either a linear equalizer or the decision-feedback equalizer with detected symbols fed back, is shown to perform worse than the ideal diversity combining receiver by a factor of approximately two, which represents the theoretical limit for possible performance improvement, using much more complex receivers with maximum likelihood detectors.
false
Unapređenje metoda za proračun zvučne izolacije u zgradama
Improvement of methods for calculation of sound insulation in buildings
Predmet istraživanja u ovom radu su metode proračuna zvučne izolacije prema klasičnom modelu prenosa zvuka u zgradama i izražavanje izolacije jednobrojnom vrednošću. Cilj rada je ispitivanje mogućnosti unapređenja standardnih metoda kvantifikovanja i proračuna zvučne izolacije, radi postizanja veće tačnosti dobijenih brojnih vrednosti i njihove korelacije sa subjektivnim osećajem akustičkog komfora u zgradama. Postupci proračuna zvučne izolacije i njenog predstavljanja jednobrojnim vrednostima koji se koriste u inženjerskoj praksi opisani su međunarodnim standardima i normama iz serija EN 12354, ISO 717, kao i predlogom novog standarda ISO 16717. Norma EN 12354 je prvi i jedini važeći međunarodni standard koji propisuje metode za proračun izolacije u zgradama, od vazdušnog i udarnog zvuka unutar zgrada i vazdušnog zvuka iz spoljašnje sredine. Metode su zasnovane na klasičnom teorijskom modelu prenosa zvuka u zgradama i u opštem slučaju daju vrednosti frekvencijski zavisnih deskriptora zvučne izolacije. Izražavanje izolacije jednobrojnom vrednošću uvodi se radi pojednostavljenja ocene izolacionih svojstava pregrada i efikasnije standardizacije kriterijuma zvučne izolacije u zgradama. Procedure za izračunavanje jednobrojnih vrednosti definisane su u aktuelnom standardu ISO 717 i novom predlogu istog standarda ISO 16717 i zasnovane na primeni referentnih krivih izolacionih moći i spektara buke. Tako dobijena jednobrojna vrednost zvučne izolacije dodatno treba da bude korelisana sa subjektivnim doživljajem akustičkog komfora, a objektivni deskriptori zvučne izolacije što približniji njenoj subjektivnoj oceni. U radu su najpre detaljno opisani relevantni algoritmi proračuna frekvencijski zavisnih, a zatim i širokopojasnih, jednobrojnih vrednosti zvučne izolacije. Zbog brojnosti i složenosti modelovanih akustičkih fenomena vezanih za prostiranje zvuka u zgradama, postupak proračuna zvučne izolacije iziskuje kompleksne matematčke procedure i povezan je sa mnogim teškoćama, uključujući i ograničenu tačnost (do nekoliko decibela) u poređenju sa realnim, izmerenim vrednostima. Dominantni uzrok iv smanjenja tačnosti je u ograničenoj dostupnosti tačnih vrednosti ulaznih parametara neophodnih za proračun. Stoga je u ovom radu analizirano nekoliko aspekata metodologije proračuna izolacije i mogućnosti za uvođenje izvesnih poboljšanja kojim bi se popravila objektivna tačnost, subjektivna relevantnost, kao i razumno pojednostavnjenje matematičkih procedura. Poseban akcenat je na izolaciji od vazdušnog zvuka između susednih prostorija i homogenim i masivnim pregradama, koje su većinski zastupljene u zgradama u Republici Srbiji. Prvi analizirani aspekt je izražavanje izolacije od vazdušnog zvuka jednobrojnim vrednostima korišćenjem referentnih spektara buke. Monitoringom buke generisane u zgradama i spoljašnje, saobraćajne buke i njihovom statističkom analizom utvrđena je veza između pojedinih vremenskih (impulsivnost) i spektralnih karakteristika buke i jednobrojnih vrednosti izolacije. Sa stanovišta spektra posebno je analiziran uticaj buke na niskim frekvencijama u sveukupnoj oceni izolacije, dat je osvrt na upotrebu referentnih spektara buke, a u slučaju unutrašnje buke u zgradama i njihovu korekciju u skladu sa najčešćim izvorima buke u domaćinstvima. Dodatno, izražavanje impulsivnosti buke u zgradama primenom standardnih i nestandardnih (emergence) deskriptora predloženo je radi potpunije karakterizacije ovog tipa buke. Kao rezultat toga, očekuje se poboljšana ocena izolacije jednobrojnim vrednostima, u smislu njihovog približavanja subjektivnom doživljaju akustičkog komfora uz uključivanje relevantnih vremenskih i spektralnih karakteristika najčešćih izvora buke. Drugi analizirani aspekt je unapređenje metoda za proračun frekvencijski zavisnih deskriptora zvučne izolacije u zgradama. Najpre je dat osvrt na tehničke teškoće u sprovođenju standardnih procedura proračuna, moguće dileme pri njihovoj primeni u realnim okolnostima i usklađivanju sa nacionalnom regulativom u oblasti. To je demonstrirano kroz izradu softverske implementacije matematičkih algoritama. Zatim je brojnim numeričkim analizama različitih scenarija zvučne izolacije kvantifikovan uticaj pojedinih ulaznih parametara (osnovnih fizičkih parametara pregrada i njihovih izolacionih svojstava, kao i deskriptora spojeva među njima) na tačnost rezultata proračuna. To omogućava pažljiviji pristup izradi proračuna uz povećanu kontrolu i pouzdanost rezultata, ali i otvara mogućnosti za dalja pojednostavljenja proračuna, usmeravanjem samo na najuticajnije parametre, uz očuvanje približno identičnih rezultata proračuna. Ovo je posebno važno imajući u vidu da i pored složenosti v algoritama za proračun, njihova tačnost u praksi je ograničena, sa procenjenom standardnom devijacijom odstupanja od stvarnih vrednosti oko 2 dB. Analizom rezultata proračuna za tipične scenarije zvučne izolacije u zgradama ustanovljeno je da nekoliko parametara, poput izolacionih moći pregrada i indeksa prenosa vibracija spojeva ima najznačajniji uticaj na proračunate vrednosti, dok je doprinos ostalih ulaznih parametara daleko manji. Na kraju, izvšeno je i poređenje rezultata detaljnog proračuna frekvencijski zavisnih parametara zvučne izolacije i takođe standardizovanog pojednostavljenog modela, zasnovanog na jednobrojnim parametrima i svedenijim matematičkim algoritmima, na velikom uzorku realnih građevinskih konstrukcija. Utvrđeno je da u mnogim realnim postavkama proračuna mogu postojati značajna odstupanja između rezultata dva tipa proračuna i do nekoliko decibela. Stoga se predlaže dodatna korekcija rezultata pojednostavljenog proračuna, kako bi se rezultati približili rezultatima sveobuhvatnijeg i preciznijeg detaljnog modela. U skladu sa tim nalazima, razmotrena je i mogućnost daljeg pojednostavljenja proračuna izolacije u najčešćim, tipičnim postavkama, uz prihvatljivo smanjenje tačnosti u odnosu na kompletan detaljni model.
The main object of this work are the methods for calculation of sound insulation based on the classical model of sound propagation in buildings and single-number rating of sound insulation. The aim of the work is inspection of the possibilities for improvement of standard methods for quantification and calculation of sound insulation, in order to achieve higher accuracy of the obtained numerical values and their correlation with subjective impression of the acoustic comfort in buildings. Procedures for calculation of sound insulation and its representation with single-number values, which are used in engineering practice, are described in the international standards and norms EN 12354, ISO 717, as well as the proposed ISO 16717. The norm EN 12354 is the first and the only existing international standard which describes methods for calculation of sound insulation – from airborne and structure born sound inside the buildings as well as airborne outdoor sound. The methods are based on the classical theoretical model of sound propagation in buildings and in general give values of frequency dependent descriptors of sound insulation. Rating sound insulation with single-number values has been introduced to simplify assessment of the insulation of building elements and make more efficient standardization of sound insulation criteria. Procedures for calculation of single-number values are defined in the current standard ISO 717 and the new standard proposal ISO 16717, and are based on the implementation of the reference curves of sound insulation and the reference noise spectra. The obtained single-number values of sound insulation are expected to correlate well with subjective impression of the acoustic comfort, making the objective sound insulation descriptors closer to the subjective criteria. The first part of the work gives a detailed description of the relevant algorithms for calculation of frequency dependent, as well as wideband single-number values of sound insulation. Due to the large number and complexity of the modelled acoustic phenomena associated with the propagation of sound in buildings, calculation of sound vii insulation assumes complex mathematical procedures, still being related to many difficulties, such as the limited accuracy (up to several decibels) compared to the real, measured values. The dominant cause of the lower accuracy is limited availability of accurate values of the input parameters, which are needed for the calculation. Therefore, this work analyses several aspects of the methodology and the possibilities for introducing certain improvements, in order to increase the objective accuracy, improve subjective relevancy, and introduce reasonable simplification of the mathematical procedures. Special emphasis is on insulation of airborne sound between adjacent rooms and homogeneous and heavyweight building elements, being the most frequent in buildings in the Republic of Serbia. The first analysed aspect is rating airborne sound insulation with single-number values, using the reference noise spectra. By monitoring the noise generated inside buildings and outdoor traffic noise and using statistical analysis, the relation between certain temporal (impulsiveness) and spectral characteristics of noise and single-number values of sound insulation has been determined. Influence of noise at low frequencies in the overall rating of sound insulation is particularly analysed, use of the reference noise spectra is discussed, as well as their correction in the case of indoor noise in buildings and the most frequent sources of noise in dwellings. In addition to this, a quantification of impulsiveness of noise in buildings is suggested, using standard and non-standard (emergence) descriptors, for more comprehensive characterisation of this type of noise. As a result, improved assessment of sound insulation using single-number values is expected, which agrees better with subjective impression of acoustic comfort, due to the inclusion of relevant temporal and spectral characteristics of the typical noise sources. The second analysed aspect is the improvement of the methods for calculation of frequency dependent descriptors of sound insulation in buildings. Firstly, the overview of the technical difficulties in the implementation of standard calculation procedures is given. The work considers potential uncertainties for the implementation of the procedures in real circumstances and harmonization with the national noise protection regulations. They are demonstrated through the development of software implementation of the mathematical algorithms. Secondly, the influence of certain input parameters (basic physical parameters of the building elements, descriptors of their insulation properties and the junctions between them) on the accuracy of the results is viii quantified by extensive numerical analyses of different sound insulation scenarios. This allows more thorough approach to the calculation, with higher control and reliability of the results, but also opens the possibilities for further simplification of the algorithms, while keeping the satisfactory accuracy by putting in focus only the most influential parameters. This is especially important having in mind that in spite of the complexity of the algorithms, their accuracy in practice is rather limited, with assessed standard deviation from the real values around 2 dB. After the analysis of calculation results for typical sound insulation scenarios in buildings, it is concluded that several parameters, such as sound reduction index of the building elements and vibration reduction index of the junctions, have the greatest influence on the resulting values, while the contribution of other input parameters is much smaller. Finally, the results of the detailed model for calculation of frequency dependent sound insulation descriptors and (also standardized) simplified model, which is based on single-number valued parameters and simplified mathematical algorithms, are compared over a large sample of real building elements. It is shown that with many realistic calculation setups, significant deviations between the results of the calculation using the two models may occur, up to several decibels. For that reason, additional correction of the results obtained using the simplified model is suggested, which should improve its agreement with the more comprehensive and accurate detailed model. Using these findings, the possibilities for further simplification of the calculations in typical scenarios are considered, with acceptable drop of accuracy, compared to the full detailed model.
true
Primena soft computing tehnika za predviđanje nivoa buke drumskog saobraćaja
Aplication of soft computing techniques in traffic noise prediction
Intenzivan tehnolški i industrijski razvoj, iako doprinosi napretku civilizacije, ostavlja negativne posledice na čovekovu životnu i radnu sredinu. Pored zagađ enja vazduha, zemlji²ta i vode, razvoj industrijskih i saobraćajnih kapaciteta prouzrokuje pove¢anje komunalne buke koja može ugroziti psihozičko zdravlje, a posledično i kvalitet života, kao i produktivnost stanovništva. Rezultati strateškog mapiranja buke na teritoriji Evropske unije nedvosmisleno pokazuju da drumski saobra¢aj predstavlja dominantni izvor buke u urbanim sredinama. U literaturi su denisani različiti modeli za predvi đanje nivoa buke drumskog saobra¢aja čija primena omogućava procenu ugroženosti stanovništva saobraćajnom bukom i planiranje odgovaraju¢ih mera za zaštitu životne sredine od povišenih nivoa zvuka. Primenom postojećih modela dobijaju se vrednosti koje značajno odstupaju od eksperimentalnih rezultata merenja nivoa buke na teritoriji Republike Srbije, što ukazuje na potrebu za razvojem novog modela za procenu ekvivalentnog nivoa buke drumskog saobraćaja. U okviru ove disertacije, na osnovu analize sastava saobraćaja, definisane su značajne kategorije motornih vozila koje karakteriše različiti uticaj na ekvivalentni nivo buke. Na osnovu eksperimentalnih podataka za svaku od denisanih kategorija vozila, primenom soft computing tehnika, odre đen je prosečan nivo buke, čime je omogu¢eno predviđ anje ekvivalentnog nivoa saobra¢ajne buke u okruženju sa zanemarljivom reeksijom zvuka, a na osnovu informacija o protoku i strukturi saobra¢aja. Razmatrana je primena optimizacionih metoda zasnovanih na inteligenciji roja i evolucionim algoritmima u uspostavljanju analitičke veze izme u nivoa saobra¢ajne buke i parametara saobra¢ajnog toka. Kako bi se omogućila prognoza buke u okruženju sa izraženim uticajem reeksije, denisani su i odgovaraju¢i korekcioni faktori kojima se uzima u obzir uticaj okruženja, kao i određ enih karakteristika saobra¢ajnice, na nivo buke na mestu prijema. Pored razvijenog matematičkog modela, kreirana je i vešta£ka neuralna mreža za predvi anje ekvivalentnog A-ponderisanog nivoa buke oko drumskih saobraćajnica. Validacija razvijenog matematičkog modela i kreirane neuralne mreže izvržena je statističkom analizom odstupanja izračunatih od izmerenih nivoa buke, kao i korelacionom analizom ovih nivoa. Postupcima statističke i korelacione analize utvr eno je dobro slaganje merenih i proračunatih vrednosti. Uporedna analiza rezultata dobijenih primenom predloženih modela, kao i nekih od najčešće koriš¢ćenih modela za prognozu buke drumskog saobraćaja, pokazala je da primena novoformiranih modela omogu¢ava tačnija predvi đanja ekvivalentnog nivoa saobraćajne buke...
Although technological and industrial development contributes to the progress of civilization, it has negative inuences on human's living and working environment. In addition to air, soil and water pollution, the development of industrial and transport capacities causes increase in levels of communal noise which has negative impact on the psycho-physical health and productivity of the population. The results of strategic noise mapping in the European Union clearly indicate that road trac represents dominant noise source in urban areas. Many authors in available literature dened dierent models for road trac noise prediction, whose application enables noise mapping and noise protection planning. However, noise levels predicted by existing models deviate signicantly from the experimental results of noise level measuring in the territory of Republic of Serbia, which indicates the need for development of a new model for equivalent noise level estimation. Within the framework of this dissertation, signicant categories of motor vehicles are dened based on the analysis of trac structure and its inuence on the equivalent noise level. On the basis of experimental data, for each of dened categories, the average noise level is estimated by application of soft computing techniques. Optimization methods based on swarm intelligence and evolutionary algorithms were used for establishing an analytic relationship between trac noise level and trac ow parameters. In order to enable prediction of noise level at an arbitrary distance from the road in an environment with signicant sound reection, correction due to sound reection and distance correction are dened. In addition to the developed mathematical model, an arti- cial neural network for prediction of equivalent A-weighted level of road trac noise has been designed. The validation of developed mathematical model and created neural network was performed by statistical analysis of the deviations between predicted and measured noise levels, as well as the correlation analysis of these levels. Results of statistical and correlation analysis show good agreement between measured and calculated values. A comparative analysis of the results obtained by proposed models and some of frequently used models for road trac noise prediction has shown that the application of proposed models enables more precise prediction of trac noise levels...
false
Razvoj modela za optimizaciju gotovosti motornih vozila sa stanovišta troškova
Development of model for availability optimization of motor vehicles based on maintenance costs
Disertacija je nastala sa ciljem praktičnog rešavanja problema optimizacije intervala periodičnog održavanja, koji se pojavljuje u oblasti održavanja motornih voyila i tehničkih sistema uopšte...
The investigation of this disertation was accomplished with a goal of solving practical problems arising in the field of vehicle maitenance, related to optimization of preventive maitenance intervals...
false
Measuring black hole masses in active galactic nuclei using the polarization of broad emission lines
Одређивање масе црних рупа код активних галактичких jезгара помоћу поларизациjе у широким емисионим линиjама
Supermasivne crne rupe (SMCR-e) se nalaze u centru gotovo svake masivne galaksije u svemiru. Većina su neaktivne, ali kada u okolini ima do- sta gasa, započeće aktivna faza u kojoj se formira aktivno galaktičko jezgro (AGJ). U aktivnoj fazi, SMCR imaju veliki uticaj na evoluciju galaksije koja sadrži AGJ i njenog okruženja. Pouzdano merenje mase SMCR je zato važan zadatak savremene astronomije. U tu svrhu, Afanasiev i Popović (2015) su ne- davno predložili metodu koja koristi rotaciju pozicionog ugla polarizacione ravni duž profila široke emisione linije kako bi detektovali keplerovsko kretanje i odredili masu SMCR-e. Ova metoda pretpostavlja da su široke li- nije emitovane iz pljosnatog diskolikog regiona sa keplerovskim kretanjem, dok je glavni mehanizam polarizacije rasejanje na unutrašnjem zidu koplanarnog to- rusa prašine – tzv. ekvatorijalno rasejanje. Cilj teze je teorijski ispitati mo- gućnosti ove metode. Izvršili smo brojne Monte Karlo simulacije za modeli- ranje ekvatorijalnog rasejanja kod AGJ-a koristeći kˆod za prenos zračenja stokes (Goosmann & Gaskell 2007). Razmatrali smo i uticaj kompleksnog kretanja emi- sionog regiona koje može biti u formi radijalnog priliva i vertikalnog od- liva gasa, ili usled prisustva dvojne supermasivne crne rupe (DSMCR-e). Ta- kođe, izabrali smo četiri dobro poznata AGJ-a koja smo posmatrali: NGC 4051, NGC 4151, 3C 273 i PG0844+349. Spektropolarimetrija je urađena koristeći SCORPIO spektrograf montiranim na 6-metarskom BTA teleskopu Specijalne astrofizičke opservatorije Ruske akademije nauka (SAO RAN-a). Modelirali smo svaki od ovih AGJ-a koristeći podatke iz literature i uporedili smo re- zultate naših modela sa posmatračkim podacima. Dobili smo da se ova metoda može koristiti za nezavisno merenje masa SMCR-a kod AGJ-a. Uticaj priliva i oticanja gasa se može ignorisati ukoliko su njihove brzine manje od brzina keple- rovskog kretanja. Dodatno, kada je rasejavajući region blizu širokolinijskog re- giona, ova metoda ne zavisi od inklinacije posmatranja. Za sistem DSMCR-e ova metoda se ne može koristiti, međutim, dobili smo jedinstvene profile koje nisu uobičajeni za jednu SMCR-u, što može biti korišćeno za određivanje mogućih kandidata za DSMCR-e. Procenjene mase SMCR-a za četiri posmatrana AGJ-a slažu se dobro sa masama koje su procenjene drugim metodama, pre svega metodom reverberacionog mapiranja. Metoda za nezavisno merenje masa SMCR-a je u ovom radu teorijski i eksperimentalno proverena, što je od izuzetnog značaja za buduća izstraživanja koja se bave uticajem SMCR-a na svoju neposrednu okolinu.
Supermassive black holes (SMBHs) reside in the heart of nearly every massive galaxy in the Universe. Most of them lie dormant, but when the nearby gas is abundant, it will enter an active phase and form an active galactic nucleus (AGN). When in their active phase, SMBHs have a profound effect on the host galaxy evolution and play an important role in shaping their environment. Reliable SMBH mass measuring is therefore an important task in modern astronomy. For that purpose, a method has been recently proposed by Afanasiev & Popovic (2015) that uses the rotation of the polarization plane position angle across the broad emission line profile in order to trace the Keplerian motion and determine the SMBH mass. The goal of the thesis is to theoretically explore the possibilities of this method. In order to do that, we performed numerous radiative transfer simulations for modeling equatorial scattering in AGNs using the code STOKES. We included complex motions present in the system in the form of inflows and outflows, and we also compared our results with observations. We have demonstrated under which circumstances this method can be used to measure the SMBH mass in a new independent way. The second part of this work involves predicting the broad lines polarization when AGNs are powered by supermassive binary black holes (SMBBHs). We treated four different cases with sub-pc SMBBHs, and again modeled equatorial scattering. We obtained a unique polarization signature across the broad lines. We have shown that spectropolarimetry could be a powerful tool and a first step for searching SMBBHs in the future systematic spectropolarimetric surveys.
true
Hardverska realizacija jednoprolaznog brzog kodeka sa visokim stepenom kompresije i minimalnim zahtevanim resursima
Hardware realization of one pass fast codec with a high compression ratio and minimal required resources
Kompresija i kodovanje digitalne slike bave se minimizacijom memorijskih resursa potrebnih za predstavljanje, pohranjivanje i prenos digitalne slike. Korišćenje tehnika za kompresiju dovodi do značajnih ušteda. Brzi algoritmi i arhitekture za kompresiju i dekompresiju su od primarnog značaja za primene u kojima su potrebne performanse u stvarnom vremenu. Postojeći postupci za kompresiju digitalnih slika sa visokim stepenom kompresije su dugotrajni i zahtevaju složene procesore i veliku memoriju velike propusne moći, što takođe povećava potrošnju. I vreme procesiranja i kapacitet memorije postojećih postupaka za kompresiju slike sa gubicima rastu sa stepenom kompresije. U ovoj doktorskoj disertaciji pristupilo se rešavanju problema hardverske realizacije kodeka koji bi minimizirao potrebne memorijske resurse, bio brz i imao visok stepen kompresije. Korišćenjem kombinovanih naučnih metoda u tome se i uspelo. Minimizacija potrebnih memorijskih resursa postignuta je korišćenjem optimizovanih nestacionarnih filtara i novog algoritma dvodimenzionalne podopsežne transformacije, kao i jednoprolaznom realizacijom procenjivača verovatnoća. Brzina rada postignuta je korišćenjem jednoprolaznog algoritma za procenu verovatnoća i izbegavanjem deljenja ulazne slike na blokove. Visok stepen kompresije postignut je sjedinjenim delovanjem sastavnih delova rešenja: podopsežnog transformatora, kvantizera, jednoprolaznog procenjivača verovatnoća i entropijskog enkodera. Hardverska realizacija iz ove disertacije verifikovana je u primeni sistema za kompresiju slike sa mrežnim interfejsom, gde je pokazano da je moguće u jedno programabilno logičko kolo nižeg cenovnog razreda smestiti ceo enkoder slike u boji, bez korišćenja spoljašnje memorije.
Digital image compression and coding deal with minimization of the memory resources needed for presentation, storage and transmission of digital image. Use of compression techniques leads to substantial savings. Fast algorithms and architectures for compression and decompression are of a primary importance in the applications where real time performance is required. Existing techniques for digital image compression with high compression ratio are slow and require complex processors and large amounts of high bandwidth memory, at the same time increasing power consumption. Both processing time and memory capacity increase as the compression ratio increases. In this doctoral dissertation, we approached the problem of hardware realization of a codec which minimizes the required memory resources, is fast and has a high compression ratio. The task has been accomplished by means of combined scientific methods. The minimization of the required memory resources is achieved by means of optimized nonstationary filters and a new algorithm of two-dimensional subband transformation, as well as a one pass realization of probability estimators. The speed of operation is achieved by means of one pass algorithm for probabilities estimation and the fact that the input image is not processed partially. The high compression ratio is achieved by joint effects of all the subsystems in the realization: subband transformer, quantizer, one pass probabilities estimator and entropy encoder. The hardware realization from this dissertation has been verified by means of image compression system with network interface. It is shown that it is possible to implement entire color image encoder within a low cost programmable logic device, with no external memory components.
true
Sinteza, karakterizacija i primena aktivnih ugljeničnih materijala dobijenih od ploda platana
Synthesis, characterization and application of activated carbon materials obtained from the fruit of plane tree
Predmet ove teze, u širem smislu, je bila sinteza, karakterizacija i primena aktivnih ugljeničnih materijala dobijenih od ploda platana. Poznato je da postupci sinteze utiču na mikrostrukturu i svojstva materijala. Ugljenični materijali mogu biti napravljeni sa širokim spektrom struktura, kompozicija i svojstava, u zavisnosti od prirode organskih prekursora i procesnih parametra. Bilo koji materijal koji sadrži ugljenik može se koristiti za proizvodnju ugljeničnih materijala. Zbog toga, biomasa kao prirodna mešavina ugljovodonika, dostupna u velikom obimu odličan je preduslov. Isto tako, to je ekološki obnovljivi resurs. Postoje značajne ekonomske i tehničke prednosti u korišćenju biomase. Za prevođenje biomase (ploda platana) u ugljenični materijal korišćena je hidrotermalna sinteza na niskim temperaturama kao i piroliza na višim temperaturama. Ugljenični materijal je izložen metodama aktiviranja kao što su fizička (CO2) i hemijska (H3PO4) aktivacija u pokušaju da se dobije aktivni ugljenični materijal sa visokim stepenom adsorpcije za određenu primenu. Adsorpcija na aktivnom uglju ima primenu u prehrambenoj, farmaceutskoj, hemijskoj, automobilskoj, nuklearnoj industriji, itd. Primeri za to su razni filteri, adsorbenti, nosači katalizatora, baterije, kao elektrodni materijal u elektrohemiji i raznim drugim različitim granama industrije. 6 Ugljenični materijali se uveliko intenzivno proučavaju kao elektrodni materijali za elektrohemijske kondenzatore usled njihove visoke specifične površine, dobre električne provodljivosti i niske cene. Istraživači, različitim mnogobrojnim istraživanjima pokušavaju da sintetišu različite forme ugljeničnih materijala sa razvijenom specifičnom površinom i razvijenom poroznošću u cilju što boljeg skladištenja naeleketrisanja. Karakterizacijom datih materijala u pogledu vrednosti specifične površine, veličine i raspodele pora, prisustva i raspodele površinskih grupa, stepena uređenosti i čistoće, pokušava da se ustanovi koji su ključni parametari za najefikasnije skladištenje energije ugljeničnog kondenzatora. Sintetisani su mikroporozni i mezoporozni ugljenični materijali dobijeni od ploda platana sa različitim, visokim specifičnim površinama na različitim temperaturama. Uticaj temperature i vremena aktivacije na fizičke, strukturne i površinske osobine dobijenog mikroporoznog aktivnog ugljenika su karakterisane Rendgenostrukturnom analizom, elementarnom analizom, adsorpcijom azota i FTIR spektroskopijom. Cilj je da se parametrima aktivacije kao što su vreme i temperatura kontoliše specifična površina ugljeničnog materijala kao i raspodela mikro/mezopora. Takođe, prateći promene teksturalnih parametara (specifične površine, zapremine i površine mikro i mezopora) sa parametrima aktivacije pokušaće da se da uvid u mehanizam CO2 aktivacije ispitivanog materijala. Priroda dvojnog električnog sloja ispitivanih aktivnih ugljenika, sa različitom raspodelom mikro/mezopora, biće ispitivana Cikličnom Voltametrijom u različitim vodenim elektrolitima (KOH, H2SO4 i Na2SO4), sa akcentom na kulonski kapacitet materijala. Plan rada je da se dođe do zaključka da li materijal ima potencijala da se koristi u elektrohemijskim kondenzatorima i ako ima koji je vodeni rastvor kao elektorlit najpogodniji. Iako su razni ugljenični materijali (ugljenične nanotube, ugljenična nanovlakna, aktivni ugljenici, grafeni), široko ispitivani u polju elektrohemijskih kondenzatora, naročito poslednjih godina, i dalje postoje brojni nejasni odgovori u pogledu mehanizma skladištenja naelektrisanja na granici ugljenična elektroda/elektrolit. Zapravo, iako se teži da se dobiju materijali što veće specifične površine, u literaturi ne postoji generalna korelacija između specifične površine i kapaciteta. Naime, pojedini autori dobijaju rast kapaciteta sa specifičnom površinom dok drugi dobijaju suprotno. Specifični kapaciteti 7 aktivnih ugljenika će biti korelisani sa njihovom specifičnom površinom kao i raspodelom mikro/mezopora, a sa ciljem da se dođe do zaključka koji su ključni parametri za brzu kinetiku formiranja dvojnog sloja ispitivanih aktivnih ugljenika. Naime, dobiće se odgovor na pitanje da li stvarno materijal sa najrazvijenijom površinom isporučuje najveći kapacitet i da li su mikro ili mezopore ili dobro balansiran odnos mikro/mezopora odgovorni za najbolje skladištenje naelektrisanja. Izgled cikličnih voltamograma će takođe biti korelisan sa raspodelom mikro i mezopora, u sva tri vodena elektrolitička rastvora, sa ciljem da se izvedu dodatni zaključci koji bi pomogli rasvetljavanju mehanizma skladištenja naelektrisanja na faznoj granici elektroda/elektrolit. Pored kapaciteta, biće pokazano kako izbor elektrolita utiče na skladištenje vodonika u aktivnom ugljeniku i na jačinu C-H veze prilikom skladištenja. Široku primenu u prečišćavanju zagađenih površinskih, podzemnih i industrijskih voda imaju aktivni ugljenični materijali kao adsorbent. Karakteristiku dobrog adsorbenta ima zbog velike specifične površine, raspodele pora kao i sposobnosti da neselektivno u značajnoj meri adsorbuje mnoge štetne sastojke iz zagađene vode. Zbog velikog zagađenja javila se potreba za uklanjanjem i smanjivanjem zagađenja. Dobijeni aktivni ugljenični materijal hidrotermalnom karbonizacijom sa H3PO4 biće korišćen kao adsorbent za uklanjanje katjonske boje metilen plavo (MP) iz vodenih rastvora.
In general terms, the subject of this thesis, was the synthesis, characterization and application of active carbon materials obtained from the plane tree fruit. It is known that synthetic methods influence the microstructure and properties of the material. Carbon materials can be obtained with a wide range of structures, compositions and properties, depending on the nature of organic precursors and process parameters. Any carbon-containing material can be used for the production of carbon materials. Therefore, biomass as a natural mixture of hydrocarbons, available on a large scale, is a great prerequisite. Likewise, it is an ecological renewable resource. There are significant economic and technical advantages in the use of biomass. Hydrothermal synthesis at low temperatures, as well as pyrolysis at higher temperatures, was used to convert biomass (plane tree fruit) into carbon material. Carbon material is exposed to activation methods such as physical (CO2) and chemical (H3PO4) activation in an attempt to obtain an active carbon material with a high degree of adsorption for a particular application. Adsorption on activated carbon has application in food, pharmaceutical, chemical, automotive, nuclear industry, etc. Examples are various filters, adsorbents, catalyst carriers, batteries, electrode materials in electrochemistry and various other industries. 9 Carbon materials are extensively studied as electrode materials for electrochemical condensers due to their high specific surface area, good electrical conductivity and low price. Researchers, by various numerous studies, are trying to synthesize different forms of carbon materials with a developed specific surface and developed porosity in order to improve storage of electrical charge. By specifying the given materials in terms of the value of the specific surface area, size and distribution of pores, the presence and distribution of surface groups, the degree of orderliness and purity, an attempt is made to determine the key parameters for the most efficient energy storage of the carbon capacitor. Microporous and mesoporous carbon materials derived from the plane tree fruit with different, high specific surfaces, were synthesized at different temperatures. The influence of the temperature and activation time on the physical, structural and surface properties of the obtained microporous active carbon is characterized by X-ray diffraction analysis, elemental analysis, nitrogen adsorption and FTIR spectroscopy. The aim is to activate the parameters of activation and to control the specific surface of the carbon material as well as the distribution of the micro / mesopores. Also, following the changes in the textural parameters (specific surfaces, volumes, and surfaces of the micro and mesopores) with the activation parameters, an attempt has been made to inspect the CO2 activation mechanism of the tested material. The nature of the dual electric layer of the tested active carbon, with a different micro / mesopores distribution, was tested with cyclic voltammetry in various aqueous electrolytes (KOH, H2SO4 and Na2SO4), with an emphasis on the culonic capacity of the material. The idea of this investigation was to find out whether the material has the potential to be used in electrochemical capacitors and, if so, which of aqueous solution as was the most suitable as an electrolyte. Although various carbon materials (carbon nanotubes, carbon nanoparticles, active carbon, graphene) have been widely studied in the field of electrochemical capacitors, especially in recent years, there are still many unclear responses regarding the storage mechanism of the charge at the carbon electrode / electrolyte boundary. Actually, in spite of intention to produce materials of as large specific surface as possible, there is no general correlation in the literature between the specific surface area and the capacity. Namely, some authors gained the capacity with specific surface, while others obtained 10 the opposite results. Bearing this in mind, the specific capacities of the active carbon will be correlated with their specific surface as well as the distribution of the micro / mesopores in order to come to a conclusion on the key parameters for the rapid kinetics of the formation of the double layer of studied active carbons. Namely, the objective of this investigation will be to get to the answer to the question is whether the material with the most developed surface delivers the greatest capacity, and whether the micro or mesopores or a well-balanced micro-mesoporous relationship are responsible for the best storage of the charge. The appearance of cyclic voltamograms will also be correlated with the distribution of micro and mesopores, in all three aqueous electrolytic solutions. These results will draw additional conclusions that would help to clarify the electrode / electrolyte phase electrode storage mechanism. In addition to the capacities, it will be shown that electrolyte selection will influence the storage of hydrogen in active carbon and the C-H bond strength during storage. Active carbon materials have a wide application as an adsorbent in purification of contaminated surface, underground and industrial waters. Good adsorbent capacity is due to the large specific surface area, the distribution of pores and non-selective and significant ability to adsorbe numerous harmful substances from the contaminated water. Due to the large pollution there was a need for removal and reduction of these contaminants. The resulting activated carbon material obtained by hydrothermal carbonization with H3PO4 will be used as an adsorbent for the removal of cationic color methylene blue (MP) from aqueous solutions.
true
Lokalni operator za reprodukciju slika širokog dinamičkog opsega uz očuvanje detalja
Local tone mapping operator for detail preserving reproduction of high dynamic range images.
Opseg osvetljaja koji se javlja u prirodnim scenama uveliko prevazilazi mogućnosti standardnih uređaja za snimanje i reprodukciju slike. Ljudski vizuelni sistem je evoluirao, tako da omogući efikasno funkcionisanje i percepciju detalja u uslovima velike promene osvetljaja. Kako bi se omogućila što realnija reprodukcija slika i video sadržaja, potrebno je obezbediti mogućnost snimanja i reprodukcije što šireg dinamičkog opsega osvetljaja. Razvoj tehnika za snimanje je napredovao i danas postoji mogućnost snimanja celokupnog dinamičkog opsega osvetljaja scene korišćenjem standardnih senzora. Razvoj displeja je međutim napredovao sporije i većina displeja koji su danas u upotrebi ima skroman dinamički opseg osvetljaja. Operator za redukciju dinamičkog opsega predstavlja ključnu komponentu sistema za reprodukciju scena širokog dinamičkog opsega (HDR), na standardnim displejima nižeg dinamičkog opsega (LDR). Operatori za redukciju dinamičkog opsega HDR slika se mogu podeliti na globalne i lokalne operatore. Globalni operatori su jednostavniji, robusniji i mogu se lakše kontrolisati, ali imaju tendenciju potiskivanja detalja. Lokalni operatori omogućavaju superiorniji kvalitet slike u pogledu reprodukcije detalja, po cenu povećane kompleksnosti i teže kontrole. Takođe, nijedan od analiziranih lokalnih operatora nema mogućnost nezavisnog podešavanja različitih aspekata izlazne slike. U okviru ove disertacije je predloženo rešenje lokalnog operatora za reprodukciju HDR slika, koje kombinuje dobre osobine lokalnih i globalnih operatora. Operator je projektovan sa ciljem da se predloži robusno i fleksibilno rešenje, koje svoju primenu može naći u različitim uređajima široke namene. Zbog toga je predloženi operator baziran na dekompoziciji na bazni sloj i sloj detalja, pri čemu se redukcija dinamičkog opsega postiže globalnim mapiranjem baznog sloja. Predložen je fleksibilan postupak ove globalne redukcije dinamičkog opsega, kojim se postiže robusnost za najrazličitije tipove ulaznih slika. Takođe, omogućeno je nezavisno podešavanje različitih aspekata izlazne scene poput osvetljaja, oštrine, lokalnog kontrasta, odnosa lokalnog i globalnog kontrasta i zasićenosti boja. v Performanse predloženog operatora su testirane na velikoj bazi HDR slika i upoređene sa 7 različitih operatora za redukciju dinamičkog opsega, korišćenjem 9 objektivnih metrika kvaliteta. Testovi su sprovedeni tako, što je za svaku sekvencu, svakom od operatora dodeljivan rang, u zavisnosti od rezultata odgovarajuće metrike. Operatoru koji je postigao najbolje rezultate, dodeljivan je rang 1, dok je operatoru koji je postigao najlošije rezultate, dodeljivan rang 8. Testiranja su pokazala da predloženi operator ima najbolje karakteristike po pitanju kvaliteta izlazne slike. Prosečan rang koji je ostvaren na celoj bazi i po svim metrikama je 2.74, dok prvi naredni operator ima rang 3.96. Kako bi se poboljšale performanse operatora, za scene sa ekstremnim vrednostima osvetljaja, predložena je automatska korekcija parametra osvetljaja. Uvođenjem ove korekcije, robusnost predloženog operatora se značajno poboljšava, tako da je sada ostvareni rang na celoj bazi HDR slika 1.84, dok prvi sledeći operator ima rang 4.15. Nakon toga je predložena temporalna kontrola, kojom se omogućava primena operatora za reprodukciju HDR videa. Operatori koji se koriste za mapiranje HDR videa, obično se projektuju ili sa ciljem temporalne adaptivnosti ili sa ciljem temporalne koherentnosti. Temporalna adaptivnost podrazumeva da se operator prilagođava trenutnom osvetljaju scene, tako da se u svakom trenutku postigne što bolji prostorni kontrast. Temporalna koherentnost podrazumeva očuvanje poretka osvetljaja između frejmova video sekvence, odnosno očuvanje temporalnog kontrasta, u procesu redukcije dinamičkog opsega. Prostorni i temporalni kontrast se ne mogu postići istovremeno. Predložena kontrola omogućava da se pomoću istog operatora, promenom parametra adaptacije, postigne maksimalni prostorni kontrast, maksimalni temporalni kontrast ili neki odnos između ova dva. Pokazano je da se korišćenjem ovakve temporalne kontrole, predloženi operator može prilagoditi, tako da daje rezultate slične temporalno adaptivinim ili temporalno koherentnim operatorima. Na kraju je analizirana mogućnost efikasne implementacije velikih filtara osetljivih na ivice, koji predstavljaju najzahtevniji deo predloženog operatora. Analizirane su performanse algoritma brzog vođenog filtra, koji se bazira na izračunavanju većine operacija u decimiranom domenu. Predloženo je unapređenje osnovne formulacije brzog vođenog filtra, pri čemu se kvalitet aproksimacije povećava za 21dB u proseku. Predložena je hardverska realizacija ovog algoritma, kojom se omogućava procesiranje vi videa rezolucije 1920 1080  , brzinom od 60 frejmova po sekundi, na FPGA čipu nižeg cenovnog ranga. Postojeće hardverske arhitekture vođenog filtra, koje omogućavaju ovakve performanse, su ASIC tipa. Takođe je pokazano da predloženo rešenje ima manje zauzeće memorijskih resursa i značajno manju potrošnju u odnosu na postojeća hardverska rešenja. Fleksibilnost operatora koja se ogleda u mogućnosti nezavisne kontrole različitih aspekata izlazne slike i robusnost koja je demonstrirana na velikoj bazi HDR slika, kvalifikuje predloženi operator za upotrebu u različitim proizvodima široke namene. Fleksibilna temporalna kontrola omogućava upotrebu ovog operatora u reprodukciji HDR videa. Pri tom je moguće postići, temporalnu koherentnost ili temporalnu adaptivnost, različitim podešavanjima kontrolnog bloka. Efikasna hardverska arhitektura unapređenog brzog vođenog filtra otvara mogućnost za primenu ovog operatora u namenskim sistemima male potrošnje.
Light intensity variations in natural scenes greatly exceed the capabillities of standard imaging and display devices. The human visual system has evolved to deal with these lightning conditions and enable efficient perception of details. In order to enable realistic reproduction of natural images and video, it is necessary to develop techniques and devices for capturing and reproduction of the high dynamic range content. Capturing techniques have evolved and now it is possible to capture entire dynamic range of the scene using standard sensors. The development of displays, however, has progressed more slowly and most of the displays that are used today exhibits modest dynamic range capabilities. Tone mapping operator is a key component that enables reproduction of the high dynamic range (HDR) images on the low dynamic range (LDR) displays. Tone mapping operators can be divided in two main categories: global and local operators. Global operators are simple, robust, and easier to control, but they tend to suppress details. Local operators, provide superior image quality in terms of detail reproduction, at the cost of increased complexity and more difficult control. Also, none of the analyzed local operators has the ability to independently adjust different aspects of the output image appearance. In this thesis, we present local tone mapping operator, which combines the good features from both categories. The operator is designed to offer a robust and flexible solution that can be used in wide variety of applications. To enable this, we use decomposition of the input image to the base and detail layers, and reduce dynamic range by applying global tone mapping on the base layer. We propose a flexible method for the dynamic range reduction of the base layer, which achieves robust operation for wide variety of input images. Also, it is possible to independently adjust different aspects of the output image appearance such as brightness, sharpness, local contrast, trade-off between local and global contrast, and saturation of colors. viii The performance of the proposed operator has been tested on the large HDR image database and results are compared with 7 different state-of-the-art tone mapping operators using 9 objective image quality metrics. The tests were conducted in such a way that each tone mapping operator was assigned a rank, ranging from 1 to 8, where 1 corresponds to the best performing operator regarding objective quality metric. The tests showed that the proposed tone mapping operator has the best performance regarding image quality. The average ranking achieved across all input images, and all objective metrics of the proposed operator was 2.74 while the second best operator achieved ranking of 3.96. In order to improve operator’s performance for the scenes with extreme illumination conditions, automatic correction of the brightness parameter value is proposed. With this correction, the robustness of the proposed operator is significantly improved, and achieved ranking rises to 1.84 while the second best operator achieved ranking of 4.15. Next, temporal control of the local tone mapping operator is proposed, in order to enable its usage for the HDR video reproduction. Video HDR tone mapping operators are usually designed with the goal of achieving either temporal adaptivity or temporal coherency. Temporal adaptivity means that operator adjusts its parameters in order to achieve maximal spatial contrast. Temporal coherency means that operator tries to preserve brightness relations between different frames in the video sequence, which represents temporal contrast, during the process of tone mapping. This is due to inability to simultaneously achieve both spatial and temporal contrast. Proposed temporal control was designed in such way to enable different tradeoffs between temporal and spatial contrast using the same tone mapping operator. That means that, just by different tuning of control block, proposed operator can achieve temporally adaptive or temporally coherent behavior. This was verified using available HDR video database and several operators that are designed for different target applications. Finally, different possibilities for the efficient implementations of edge aware filters are analyzed, and the algorithm of fast guided filter was choosen as the best solution. Efficient implementation of the edge-aware filters is essential since it represents the most demanding part of the proposed tone mapping operator. Approximation quality of the fast guided filter is analyzed and improvement of the original algorithm is proposed. This improvement increase approximation quality by 21dB on average. ix Improved algorithm was implemented on the low cost FPGA chip and it achieves real time performance of 60fps when processing 1080p video. From the currently available hardware solutions, only those implemented in ASIC were able to provide such high throughput. Proposed hardware architecture also uses less memory and logic resources, and have significantly lower power consumption when compared to the existing solutions. Flexibility and ease of tuning along with the demonstrated robustness qualifies the proposed operator for the usage in wide variety of applications. Flexible temporal control enables its usage in HDR video applications. Both temporal coherency and temporal adaptivity can be achieved, just by different tuning of the proposed control block. Efficient hardware architecture of the fast guided filter opens the ability to use this operator in the low power embedded solutions.
true
Procjena preostalog radnog vijeka sistema za prenos snage teretnih vozila
Estimation of the residual lifetime of truck power transmission system
Predmet rada je procjena radnog vijeka elemenata i sistema za prenos snage terertnih vozila od datog početnog stanja pa do dostizanja definisanog graničnog stanja. Granična stanja se određuju na osnovu tehničkih i ekonomskih kriterijuma. Granična stanja po tehničkim kriterijumima nastaju kada dođe do pada kvaliteta rada i funkcionalne sposobnosti sistema, što se dešava pri pojavi odgovarajućeg nivoa pohabanosti, deformacije ili loma elementa. Po ekonomskim kriterijumima granično stanje nastupa kada dođe do pada ekonomske efektivnosti sistema ispod granice isplativosti dalje eksploatacije. Uslovi rada i vrijednosti opterećenja kao i konstrukcione karakteristike primjenjenih materijala elemenata sistema za prenos snage teretnih vozila su dominantno stohastičkog karaktera, pa se zbog nemogućnosti tačnog proračuna vrši procjena njihovog radnog vijeka. U radu su prikazane sistemske metode određivanja opterećenja i proračuna prenosnika snage vozila. Procjena radnog vijeka elementa do loma usled zamora materijala bazirana je na primjeni linearnih hipoteza o akumulaciji oštećenja materijala. Prikazane su metode determinističkog i vjerovatnosnog proračuna, odnosno procjene radnog vijeka elemenata i sistema. Radni vijek elementa izražen u km pređenog puta dat je kao odnos ukupne raspoložive radne sposobnosti elementa i njegove potrebne radne sposobnosti po jednom km puta. Pored prethodno prikazane procjene radnog vijeka elementa i sistema, koji polazi od projektovanog stanja kao početnog, prikazana je i procjena preostalog radnog vijeka elementa i sistema u odnosu na dato izmijenjeno početno stanje. U tom cilju razmatra se promjena stanja elemenata, prvenstveno kroz proces habanja elementa i formiranje zazora u njihovim vezama. Prikazan je uticaj zazora na tehničke i ekonomske pokazatelje od značaja za preostali radni vijek elemenata i sistema za prenos snage vozila. Uticaj zazora na opterećenje elemenata i sistema za prenos snage vozila razmatran je na uprošćenom dvomasenom modelu i detaljnom modelu sistema: pogonski motor-sistem za prenos snage-pogonski točak-vozilo-put (PM-SPS-PT-V-P). Odgovarajući matematički i simulacioni modeli verifikovani su u značajnoj mjeri eksperimentalnim ispitivanjem. Eksperimentalnim putem ispitan je uticaj zazora na poluvratilu jednog putničkog vozila na vrijednost momenta u prelaznom procesu i ustaljenom režimu kretanja. U radu je prikazana i procjena preostalog radnog vijeka sistema za prenos snage do dostizanja graničnog stanja po ekonomskim kriterijumima. Prikazana je procjena optimalnog radnog vijeka po kriterijumima minimuma specifičnih troškova i maksimalne dobiti u eksploataciji vozila. Na osnovu izloženih sadržaja izvedeni su odgovarajući zaključci koji pokazuju naučni doprinos i upotrebni značaj rezultata. Na kraju rada ukazano je na potrebu i pravce daljih istraživanja problematike koja je predmet rada
Aim of this paper is to estimate the residual lifetime of truck transmission system and elements starting from initial conditions up to a defined limit state. Limit state is defined on the basis of technical and economic criteria. Limit state, according to technical criteria, occurs when there is a reduction in quality of work and functional capabilities of a system, which happens due to an increased wear and tear, deformation and element fracture. According to economic criteria, limit state occurs when there is a decrease in economic effectiveness of the system below the level of profitable exploitation. Exploitation conditions and load values together with the construction properties of the truck power transmission elements are dominantly of stochastic nature, so it is impossible to accurately calculate their lifetime. This paper shows methodology of how to determine load and calculate power transmission system residual lifetime. Estimation of the residual lifetime of the transmission elements up to the moment of fracture due to material fatigue is based on the application of linear damage accumulation hypothesis. We have shown the methods of deterministic and probabilistic calculation, i.e. estimation of the residual lifetime of the elements and the system. Lifetime of the elements, indicated in km of the travelled distance, is given as the relation between the total work capability of the elements and their required work capability per one km. Apart from the abovementioned estimation of the elements and the system lifetime, where a projected status is an initial status, we have also shown the residual lifetime of the elements and the system in relation to the given alternated initial state. In this light we have considered changes of the elements states, primarily through wear and tear process and formation of clearance in their connexions. We have also shown the impact of clearance on the technical and economic indicators which are significant for the residual lifetime of a vehicle power transmission elements and systems. The impact of clearance on the load of the vehicle power transmission is considered on a simplified model with two masses and a detailed system model: drive engine - power transmission system - drive wheel – vehicle - road (PM-SPS-PT-V-P). Appropriate mathematical and simulation models were largely verified by experimental research. We have experimentally tested impact of clearance on a passenger vehicle half-shaft regarding momentum values during the transitional process and regular motion rates. This paper also shows the residual lifetime of power transmission up to the moment of reaching the limit state, according to economic criteria. Furthermore, we have shown an estimation of optimum lifetime, according to the criteria of minimum costs and maximum exploitation of a vehicle. On the basis of these results, we have drawn some conclusions which show scientific contribution and practical significance of our results. In the end we have indicated further possibilities of how to investigate this topic that our paper deals with
false
Eksperimentalna karakterizacija i Monte Karlo simulacija dozimetrijskih parametara MOSFET strukture u poljima jonizujućeg zračenja
Experimental characterization and monte carlo simulation of the dosimetric parameters of the mosfet structure in the fields of ionizing radiation.
Rezultati naučnih istraživanja pokazuju da jonizujuće zračenje može dovesti do funkcionalnog ili trajnog oštećenja poluprovodničkih naprava različitog tipa u ekstremnim radnim uslovima (u okruženju nuklearnog reaktora, akceleratorske instalacije, nuklearne eksplozije ili prilikom testiranja naprava za eksploataciju u kosmosu). Cilj ovog rada je primena specifične metodologije kojom se preciznije prati uticaj jonizujućeg zračenja na karakteristike MOSFET komponente. Specifičnost metodologije se ogleda u tome da pored toga što se određuje zavisnost promene napona praga od eksperimentalno određene apsorbovane doze u MOSFET-u koji se nalazi u polju jonizujućeg zračenja, uvode se i Monte Karlo proračuni veličina od kojih zavise dozimetrijski parametri MOSFET strukture i sagledava se njihova veza sa rezultatima eksperimenta. Jedan od predmeta istraživanja u disertaciji obuhvata eksperimentalni rad na proučavanju izabranih parametara MOSFET komponente kojima se može sprovoditi dozimetrija u poljima jonizujućih zračenja. Eksperimenti su obuhvatili određivanje zavisnosti napona praga VT i promene napona praga ΔVT od ukupne apsorbovane doze zračenja za slučajeve različitih poluprovodničkih, strukturnih i konstrukcionih karakteristika izabranih komercijalnih MOSFET komponenti, što se postiže promenama u debljini oksida gejta, u dopiranosti poluprovodničke osnove, u različitoj vrsti poluprovodničkog materijala, u prisustvu slojeva za pasivizaciju, u dužini i širini kanala između sorsa i drejna, u naponu napajanja gejta u toku zračenja. Eksperimentalna istraživanja su sprovedena sa PMOS tranzistorima u polju sa gama, X ili elektronskim zračenjem, pri čemu su korišćene različite energije zračenja, jačine doza zračenja, tako da su komponente bile izložene različitim ukupnim dozama zračenja. U radu je prikazana i primena numeričkih simulacija na osnovu metoda Monte Karlo pri transportu fotona i elektrona kao čestica jonizujućeg zračenja kroz geometrijsku konfiguraciju MOSFET strukture. Na kraju ovog proračuna jedan od rezultata su vrednosti deponovane energije u zapremini svake materijalne zone strukture MOSFET komponente. Odatle sledi da se polazeći od proračunatih vrednosti deponovane energije, uz korišćenje fundamentalne definicije u dozimetriji zračenja, određuju vrednosti apsorbovane doze u oksidu gejta (SiO2) i supstratu (Si) kao veličine koje su preko odgovarajućih relacija povezane sa dozimetrijskim parametrima MOSFET komponente. Sprovedeno je i istraživanje mogućnosti da se utvrdi relacija između eksperimentalno određenog dozimetrijskog parametra, kao što je promena napona praga MOSFET komponente i vrednosti apsorbovane doze u dozimetrijski osetljivoj zapremini oksida gejta, koje su dobijene kao rezultati numeričkih simulacija za transport fotona i elektrona jonizujućeg zračenja kroz MOSFET strukturu. U postupku dizajniranja MOSFET komponente važnu ulogu igra odluka o izboru materijala za oklapanje strukture. Ovaj inženjerski problem je ovde takođe bio predmet posebnog istraživanja u kojem se sprovode numerički proračuni na bazi Monte Karlo simulacija za MOSFET strukture sa i bez materijalne zone kojom se vrši oklapanje i na taj način se definiše faktor fizičke zaštite elektronske komponente.
The results of scientific research show that ionizing radiation can lead to functional or permanent damage of semiconductive devices of various types in extreme operating conditions (in the environment of a nuclear reactor, accelerator installation, nuclear explosion, or when testing the devices for exploitation in the cosmos). The aim of this study is to establish a specific methodology for more accurate monitoring of the impact of ionizing radiation on the characteristics of the MOSFET component. The specificity of this methodology is reflected in the fact that besides the determination of the dependence of the change in the threshold voltage of the experimentally determined absorbed dose in the MOSFET, which is located in the field of ionizing radiation, there is also the introduction of the Monte Carlo size calculations which the dosimetric parameters of MOSFET structure depend on and the examination of their connection with the results of the experiment. One of the objects of research in the thesis includes experimental work on studying the selected parameters of the MOSFET component in the fields of ionizing radiation. The experiments included the determination of the dependence of the threshold voltage VT and the changes in the threshold voltage ΔVT on the total absorbed dose of radiation for various semiconductor, structural and constructional characteristics, of selected MOSFET components which is achieved by changing the thickness of the gate oxide, the level of doping the semiconductor base, by using a different type of semiconductor material, the presence of passivation layers, the length and width of the channel between the source and the drain, the voltage of the power supply for the gate during the radiation. Experimental studies have been conducted with PMOS transistors in the field with gamma, X or electron radiation, with different radiation energy and the strength of the radiation dose used, so that the components were exposed to different total doses of radiation. This paper also presents the application of numerical simulations based on the Monte Carlo method in transporting the photons and electrons as particles of ionizing radiation through the geometric configuration of the MOSFET structure. One of the results of this calculation are the values of deposited energy in the volume of each material zone of the structure of the MOSFET component. It follows that, starting from the calculated values of deposited energy, with the use of fundamental definition in the dosimetry of radiation, the values of the absorbed dose in the gate oxide (SiO2) and the substrate (Si) can be determined as the sizes which are through appropriate relations connected with the dosimetric parameters of the MOSFET component. What was also conducted was the exploration of the possibilities to establish a relation between the experimentally determined dosimetric parameter, such as the change in threshold voltage of the MOSFET component and the values of the absorbed dose in the sensitive dosimetric volume of the gate oxide, obtained as the results of numerical simulations for the transportation of photons and electrons of the ionizing radiation through the MOSFET structure. In the process of designing a MOSFET component the choice of materials for structural shielding performs an important role. This engineering problem was also the subject of a separate study in which numerical calculations are carried out based on Monte Carlo simulations for MOSFET structures with and without a material zone used for shielding, thus defining the factor of the physical protection of the electronic component.
true
Kooperativna automatska klasifikacija signala po tipu modulacije korišćenjem mreže senzora
Cooperative Automatic Modulation Classification by Using Multiple Sensors
U okviru disertacije razmatran je problem kooperativne automatske klasifikacije signala po tipu primenjene modulacije (Automatic Modulation Classification, AMC) koja se ostvaruje korišćenjem mreže prostorno distribuiranih senzora. Pri tome, izvršena je detaljna analiza postojećih rešenja sa stanovišta mogućnosti njihove primene u realnim uslovima rada koje definišu: nekooperativna priroda izvođenja AMC postupka, različiti prostorni raspored senzora, i realno propagaciono okruženje koje karakterše pojava fedinga usled višestruke propagacije. Sprovedenom numeričkom analizom utvrđeno je da se primenom ranije predloženih (postojećih) rešenja za kooperativnu AMC sa centralizovanom fuzijom odluka ili distribuiranim procesiranjem ne može ostvariti uspešna klasifikacija za najveći broj analiziranih realnih scenarija primene. Na osnovu uočenih nedostataka postojećih rešenja, koja ispoljavaju izuzetno veliku osetljivost na problem neusklađenosti referentnih vrednosti koje se koriste u procesu fuzije, predložen je alternativni koncept fuzije podataka. U ovom novom konceptu, umesto donošenja nezavisnih lokalnih odluka u senzorima mreže, i njihovog naknadnog kombinovanja, za potrebe fuzije koriste se vrednosti onih veličina na osnovu kojih se obavlja AMC, odnosno donose odluke u procesu klasifikacije. Na osnovu predloženog koncepta fuzije podataka definisan je određen broj novih metoda fuzije, pri čemu je kao osnov za razvoj ovih rešenja izabran AMC postupak na osnovu kumulanata četvrtog reda. Sprovođenjem numeričke analize, putem sveobuhvatnog procesa Monte-Carlo simulacija, pokazano je da se putem primene ovde predloženih metoda fuzije ostvaruju znatno bolje AMC performanse u poređenju sa postojećim rešenjima, i to kako za idealizovane scenarije primene tako i za ovde posebno razmatrane realne scenarije primene kooperativne AMC. Pri tome, najveći dobici ostvareni su u slučaju radio kanala u kojima se javlja frekvencijski-selektivan feding, dok su najlošiji rezultati dobijeni u slučaju radio kanala sa ravnim fedingom. Iz tog razloga, predložen je postupak združene korekcije kumulanata na nivou mreže, na osnovu koga su značajno poboljšanje performanse za realne scenarije primene kooperativne AMC. Osim toga, u cilju smanjivanja problema neusklađenosti referentnih veličina predložen je originalni postupak dvostepene hibridne fuzije, u kome se u prvom Kooperativna automatska klasifikacija signala po tipu modulacije korišćenjem mreže senzora mr Goran B. Marković koraku obavlja fuzija podataka, a i drugom fuzija odluka. Odgovarajućom numeričkom analizom pokazano je da se primenom ovog rešenja značajno popravljaju performanse predloženih rešenja za kooperativnu AMC, i to upravo za scenarije u kojima se primenom centralizovane fuzije odluka i/ili podataka postižu najlošiji rezultati. Na kraju, predložena je modifikacija postupka klasifikacije u okviru samih senzora mreže. U slučaju kanala sa ravnim ili frekvencijski-selektivnom fedingom povećanje dužine opservacije ne omogućava poboljšanje kvaliteta procene kumulanta, pa samim tim ni uspešnosti klasifikacije. Stoga je predložen postupak segmentacije uzoraka signala, uz sprovođenje fuzije odluka i/ili podataka dobijenih za posmatrane segmente. Pokazano je da se primenom ovog postupka značajno povećava uspešnost klasifikacije u senzorima mreže, što dalje uslovljava povećanje ukupnih AMC performansi na nivou celokupne kooperativne mreže. Na osnovu dobijenih i prikazanih rezultata, može se zaključiti da su sprovedenom detaljnom analizom, kao i kroz same predloge velikog broja novih rešenja u smislu definisanja novih metoda fuzije kao i samog načina izvođenja postupka klasifikacije u okviru kooperativne AMC mreže, u potpunosti ispunjeni osnovni zahtevi i ciljevi postavljeni na početku procesa istraživanja. Ponuđena su odgovarajuća rešenja za sve posmatrane tipove kanala sa višestrukom propagacijom, za koje je putem numeričke analize utvrđeno da ostvaruju znatno bolje performanse od ranije predloženih rešenja, kako za idealizovane tako i za posmatrane realne uslove primene kooperativne AMC. Ovde predložena rešenja omogućavaju zanačajno poboljšanje uspešnosti klasifikacije u odnosu na klasično rešenje za AMC na osnovu prijema signala sa jednim senzorom.
The main subject of the study portrayed in this thesis, was the proper formulation of the cooperative automatic modulation classification (AMC) solutions by using the spatially dispersed and networked sensors. A detailed analysis of the so far proposed solutions was conducted in order to evaluate their performances under the realistic application conditions defined with: the inherent uncooperativeness of the AMC process, different spatial distributions of sensors, and realistic propagation conditions characterized by the multipath fading channels. Through the comprehensive numerical analysis, it is shown that by applying the previously proposed centralized decision fusion and distributed processing under the conditions of here analyzed realistic application scenarios only a very low AMC performance can be achieved. Based on the detected deficiencies of the existing solutions for the cooperative AMC, which exhibit extremely high sensitivity to the problem of mismatched reference values that are used in decision fusion process, we have formulated a new alternative data fusion concept. In this new concept, instead of making the individual decisions at the each sensor and performing their combination, we apply the data fusion on the very features (cumulants) that are used in the classification process. In line with this new proposed data fusion concept, several new fusion methods are proposed by using the AMC algorithm based on the fourth-order cumulant. By further numerical analysis through the comprehensive Monte-Carlo trials, it was proven that here proposed cooperative AMC solutions greatly outperform the existing ones for the idealized and also for here considered realistic application scenarios. Thereby, the highest AMC performance gains were achieved for the frequency-selective fading channels, while the worst result were obtained for the flat fading channels. Therefore, additional joint cumulant estimation correction method was proposed, and it is proven that significant AMC performance enhancements are achieved by applying this solution in the realistic application scenarios. Also, in order to suppress the negative effects of using mismatched references, that is found in practice, a novel hybrid two-stage fusion (HyTSF) was proposed, with the data fusion applied in the first and the decision fusion in the second step of the proposed procedure. Through the Kooperativna automatska klasifikacija signala po tipu modulacije korišćenjem mreže senzora mr Goran B. Marković appropriate numerical analysis, it is shown that the application of here proposed HySTF scheme enables substantial performance improvements compared to the existing solutions. Moreover the largest gains are obtained in the case of the realistic application scenarios, in which the worst AMC performances have been observed when centralized decision fusion or centralized data fusion were used. Finally, the modified AMC procedure for the individual sensors is proposed. In the flat fading or frequency-selective fading channels cumulant estimates, and consequently the AMC performances, are not further impoved when the signal observation period is prolonged. Therefore, the segmentation of the input sample sequence is proposed, followed with the data or decision fusion applyed on the cumulant estimates obtained for thus formed separate segments. It is also verified that by applying here proposed segmentation and fusion procedure large AMC perfomance gains are achieved at individual sensors, thus allowing further gains on the cooperative network level. Based on here obtained and presented results, it can be concluded that the basic requrement and objectives set at the beginning of the research proces are completely fullfiled by the means of comprehensive numerical analysis and a number of originaly proposed solutions (in the form of concrete fusion methods, but also in the way that the classification process is performed within the cooperative network). The appropriate AMC solutions are offered for all considered multipath propagation channels. These solutions are verified to achieve significantly better AMC performance than the existing (referent) solutions, for the both types of application scenarios, i.e. idealized and realistic. Also, it is shown that proposed solutions enable large AMC performance gains compared to the classic AMC solution that uses only one sensor for the signal reception.
true
Unapređivanje SMT rešavače korišćenjem CSP tehnika i tehnika paralelizacije
Improving SMT solvers using CSP techniques and parallelization techniques
SMT rešavači implementiraju procedure odlučivanja za ispiti- vanje zadovoljivosti logičkih formula prvog reda u odnosu na neku unapred zadatu odlučivu teoriju. Oslanjaju se na moćne tehnike za rešavanje problema iskazne zadovoljivosti (SAT) u kombinaciji sa specifičnim procedurama odlučivanja koje omogućavaju rezonovanje u konkretnoj teoriji. Tipična primena SMT rešavača je u verifikaciji softvera, pa su zato teorije koje se standardno koriste u SMT rešavačima prilagođene toj vrsti primene. Proširivanje primenjivosti na druge oblasti može zahtevati definisanje novih SMT teorija, kao i razvoj odgovarajućih procedura odlučivanja. U tom smislu, u ovoj tezi razmatramo unapređivanje SMT rešavača u pravcu pro- širivanja njihove primenjivosti na rešavanje CSP problema. Ovi problemi podrazumevaju konačne domene za zadate promenljive, a cilj je pridružiti vrednosti promenljivama tako da budu zadovoljena sva zadata ograničenja. Kada je u pitanju rešavanje CSP problema, glavni nedostatak postojećih SMT teorija je to što nisu prilagođene problemima sa konačnim domenima, a nemaju ni mogućnost rezonovanja o globalnim ograničenjima koja igraju značajnu ulogu u većini CSP problema. Zbog toga je u ovoj tezi definisana nova SMT teorija koja omogućava prirodno izražavanje CSP problema, jer uključuje podršku za neka često korišćena globalna ograničenja. Za ovu teoriju razvijena je procedura odlučivanja zasnovana na postojećim tehnikama pozajmljenim iz tradicionalnih CSP rešavača, ali adaptiranim tako da se mogu koristiti u okviru SMT rešavača. Procedura odlučivanja u ovom trenutku podržava dva tipa globalnih ograničenja: alldifferent ograničenje i linearno ograničenje. U slučaju alldifferent ograničenja, najznačajniji doprinos je novi algoritam za generisanje objašnjenja propa- gacija i konflikata, što je neophodno u procesu analize konflikata. U slučaju linearnog ograničenja, razvijen je novi algoritam filtriranja koji uzima u obzir postojanje alldifferent ograničenja u datom problemu, što može pojačati moć propagacije. Drugi pravac unapređivanja SMT rešavača koji se razmatra u tezi je korišćenje tehnika paralelizacije. Razmatra se paralelizacija simpleks algoritma koji se u SMT rešavačima koristi kao procedura odlučivanja za linearnu aritmetiku. Takođe se razmatra i paralelni portfolio, kao i hibridizacija ova dva pristupa. Daje se opsežna eksperimentalna evaluacija na velikom broju, kako industrijskih, tako i slučajno generisanih instanci. Za potrebe svih navedenih istraživanja, u toku rada na tezi razvijen je i novi SMT rešavač argosmt otvorenog koda, zasnovan na DPLL(T ) arhitekturi.
SMT solvers implement decision procedures that check for satisfiabil- ity of first-order logical formulae with respect to some decidable theory. They rely on the powerful techniques that are used for solving boolean satisfiability prob- lem (SAT), combined with specific procedures that permit reasoning in a particular theory. Typical applications of SMT solvers are mostly found in software verifica- tion, and for this reason the theories frequently considered in SMT solvers are best suited for such kind of applications. Extending the applicability of SMT solvers to other areas may require defining new SMT theories, as well as developing the corresponding decision procedures for such theories. In that sense, in this thesis we consider improving of SMT solvers by extending their applicability to solving CSP problems. Such problems consider variables with finite domains, and the goal is to find the assignment of values to the variables of the problem satisfying all the imposed constraints. When CSP solving is considered, the main drawback of the existing SMT theories is that they are not adapted to the problems with finite domains, and are also incapable of reasoning about global constraints, which play the significant role in most CSP problems. For this reason, in this thesis a novel SMT theory is defined that enables natural representation of CSP problems, since it includes support for some frequently used global constraints. For such theory, we developed a decision procedure based on the existing techniques borrowed from the traditional CSP solvers, but these techniques are adapted in the way that permits their usage within SMT solvers. The decision procedure currently supports two types of global constraints: the alldifferent constraint and the linear constraint. In case of the alldifferent constraint, the main contribution is a novel algorithm for generating explanations of conflicts and propagations, which is required for the conflict analysis. In case of the linear constraint, a novel filtering algorithm is de- veloped that takes into consideration the existence of the alldifferent constraints in the given problem, which can lead to a stronger propagation. Another direction of improving SMT solvers that is considered in the thesis is the usage of the par- allelization techniques. We consider the parallelization of the simplex algorithm, which is used in SMT solvers as the decision procedure for the linear arithmetic. We also consider the parallel portfolio approach, as well as the hybridization of the two approaches. A comprehensive experimental evaluation is provided on a large number of instances, both industrial and randomly generated. For the purpose of all the mentioned research, during the work on the thesis a new open-source SMT solver called argosmt is developed, based on DPLL(T ) architecture.
true
Karakterizacija zvučnog polja u prostorijama primenom multifraktalne analize
Room sound field characterization using multifractal analysis
Iz ugla akustike, svaka prostorija se može oceniti na osnovu postojanja reflektovane zvučne energije u njoj, gde se iz vremenske strukture reflektovane zvučne energije (impulsni odziv) mogu izvući parametri za opisivanje zvučne slike. Opisivanje impulsnog odziva predstavlja izazov koji još uvek nije odakvatno razrađen u literaturi. Tema ove disertacije jeste razvoj grupe alata za nelinearnu obradu signala impulsnog odziva koji omogućava automatizovanu analizu strukture impulsnog odziva prostorije u domenu Hölderovih eksponenata, singulariteta dobijenih primenom multifraktalne analize. Predložena je metodologija koja obuhvata metode za detekciju granice završetka oblasti ranih refleksija u impulsnom odzivu prostorija, detekciju ranih refleksija u toj oblasti, kvantifikovanje svojstava impulsnog odziva prostorije i mogućnost akustičke klasifikacije impulsnih odziva. Primenom multifraktalne analize izvršena je analiza globalnih karakteristika signala impulsnog odziva prostorije, a sama analiza realizovana je primenom metode velikih devijacija (Large Deviation) za izračunavanje multifraktalnog spektra. Iterativnom analizom segmenata signala impulsnog odziva izračunati su njihovi multifraktalni spektri, pri čemu su posmatrane promene globalnih parametara multifraktalnog spektra u zavisnosti od pozicije analiziranih uzoraka u okviru impulsnog odziva. Utvrđena je veza između promene globalnih parametara multifraktalnog spektra i promena globalnih akustičkih karakteristika impulsnih odziva prostorije. Ovakav iterativni pristup omogućio je uvid u lokalnu raspodelu promena globalnih parametara multifraktalnog spektra koje su povezane sa lokalnim promenama u impulsnom odzivu. Na osnovu ovih zapažanja kreiran je algoritam za izdvajanje oblasti ranih refleksija. Predložena metodologije je testirana na različitim impulsnim odzivima prostorija, različitih akustičkih karakteristika. Tokom istraživanja posebna je pažnja posvećena analizi ranih refleksija u impulsnom odzivu prostorije, što predstavlja značajnu temu u akustičkom dizajnu. U okviru ovog istraživanja, analizom signala impulsnog odziva u domenu singulariteta multifraktalne analize obezbeđen je precizniji uvid u lokalnu strukturu impulsnog odziva. Takav pristup je omogućio realizaciju algoritma za detekciju ranih refleksija. Izračunavanje Hölderovih eksponenata i njihova selekcija u postupku multifraktalne analize izvršeni su u 1D i 2D domenu, pri čemu je primenjena metoda „box-counting“ uz korišćenje nekoliko različitih mera. Za potrebe 2D multifraktalne analize, impulsni odziv prostorije je predstavljen slikom spektrograma primenom kratkovremene Furijeove transformacije (STFT). U poslednjoj fazi istraživanja izvršena je analiza mogućnosti kvantifikovanja akustičkih svojstava impulsnih odziva na osnovu globalnih multifraktalnih parametara iz multifraktalnog spektra dobijenog metodom Velikih devijacija. Utvrđena je zavisnost vrednosti globalnih multifraktalnih parametara dobijenih iz multifraktalne analize impulsnih odziva i vrednosti difuznosti u fizičkom modelu difuznih prostorija u kojima su analizirani impulsni odzivi snimljeni. Klasterizacija impulsnih odziva iz različitih prostorija realizovana je analizom vrednosti globalnih multifraktalnih parametara što ukazuje na potencijalnu primenu pomenutih globalnih parametara za klasifikaciju impulsnih odziva. Analiza predloženih metoda izvršena je na realnim impulsnim odzivima, impulsnim odzivima smimljenim u fizičkim modelima i impulsnim odzivima generisanim u softverskim modelima. Eksperimentalni rezultati predloženog algoritma pokazali su visoku tačnost u detekciji refleksija bez obzira na geometriju prostorije, vrednosti difuznosti, vrednost apsorpcije i položaj snimanja impulsnih odziva u samoj prostoriji.
From the acoustic point of view, each room can be estimated based on the existence of its reflected sound energy and the parameters extracted from the temporal structure of the reflected sound energy named the room impulse response. The impulse response description is a challenge that has not yet been sufficiently elaborated in the literature. The subject of this dissertation is the development of a group of tools for impulse response nonlinear signal processing, which enable an automated analysis of the structure of the room impulse response in the domain of Hölder exponents, i.e. singularities obtained using multifactal analysis. The proposed methodology includes methods of detecting the time boundary of the end of the area of early reflections in the impulse response, detection of early reflections in that area, the quantification of the impulse response properties and finally the possibility of acoustic classification of impulse responses. Using the multifractal analysis, the global characteristics of the impulse response signal were analyzed by using the Large Deviation multifractal method of calculating the multifractal spectrum. During this analysis, the segments of the impulse response signal were analyzed iteratively by calculating their multifunctional spectrum, whereby the changes in the calculated values of global parameters of the multifunctional spectrum were monitored, depending on the analyzed segments of signal. The relationship between the changes of the global parameters of the multifractal spectrum and the changes of the global acoustic characteristics of the room impulse response was established. This approach allowed an insight into the local distribution of global parameter changes, which are associated with local changes in impulse response. Based on these observations, an algorithm for extracting the area of early reflection was created. The proposed methodology was tested on different room impulse responses, with different acoustic characteristics. During the research, special attention was devoted to the analysis of early reflections in the impulse response, which is an important topic in the room acoustic design. Within this research, the signal analysis of the impulse response in the domain of singularity of multifractal analysis provided a more precise insight into the local structure of the impulse response. This approach enabled the development of an algorithm for the detection of early reflections. The calculation of Hölder exponents and their selection in the multifractal analysis was carried out in 1D and 2D domains, using the "box-counting" method and several different measures. For the purposes of 2D multifractal analysis, the room impulse response was represented by the spectrogram image using the short-time Fourier transform (STFT). The last phase of this research included an analysis of the possibilities of quantification of acoustic impulse response properties based on global multifractal parameters from the multifractal spectrum obtained by the Large Deviations method. The global multifractal parameters obtained from the impulse response multifractal analysis are correlated with the diffusivity indexes in the physical model of а room in which the analyzed impulse responses were recorded. The clusterization of different room impulse responses was realized by analyzing the values of global multifractal parameters, which indicates the potential application of using the global parameters for the classification of room impulse responses. The analysis of the proposed methods was carried out on realistic impulse responses, impulse responses recorded in physical models and impulse responses generated in software models. Experimental results of the proposed algorithm showed high accuracy in detection of the reflections, regardless of room geometry, diffusiness rate, absorption, and the position in the room where the impulse responses were recorded.
true
Optimizacija kognitivnih sistema sa kontrolisanim nivoom interferencije primenom višeantenskih tehnika sa adaptivnom kontrolom snage
Optimization of cognitive systems with controlled interference level by using multi-antenna techniques with adaptive power control
Zbog sve rasprostranjenijeg korišćenja bežičnih komunikacionih sistema koji omogućavaju prenos podataka velikim brzinama, slobodni frekvencijski opsezi postali su veoma ograničen i tražen resurs. Regulatorna tela koja dodeljuju licence za korišćenje određenih opsega su već izvršila raspodelu velikog dela spektra, zbog čega razvoj novih ili proširenje postojećih servisa predstavlja poseban izazov. Brojne analize ukazuju da nedovoljna fleksibilnost u načinu pristupa spektru predstavlja veći problem od samog fizičkog nedostatka slobodnih frekvencijskih opsega. Koncept kognitivnog radio sistema predstavlja potencijalno rešenje za efikasnije korišćenje spektra. U ovoj disertaciji analiziran je kognitivni radio sistem sa kontrolisanim nivoom interferencije, koji podrazumeva da primarni (licencirani) i sekundarni (kognitivni) korisnici spektra istovremeno pristupaju istom opsegu učestanosti. Sekundarnom korisniku je dozvoljeno emitovanje signala pod uslovom da na ulazu primarnog prijemnika generiše interferenciju koja je manja od unapred definisanog praga. Ovaj zahtev se može ispuniti korišćenjem adaptivne kontrole snage na sekundarnom predajniku, čime se direktno ograničava kapacitet sekundarnog linka. U disertaciji je analizana primena višeantenskih tehnika na sekundarnom linku, kojima se mogu poboljšati performanse sekundarnog korisnika. Razmatrana je primena diverzitija zasnovanog na kombinovanju sa maksimalnim odnosom (Maximal Ratio Combining, MRC), kao i primena tehnike optimalne selekcije predajne antene (Transmit Antenna Selection, TAS). Analizirana je i kombinovana tehnika združene selekcije predajne antene i prijemnog diverzitija (Transmit Antenna Selection/ Maximal Ratio Combining, TAS/MRC), kao i primena ortogonalnih prostorno-vremenskih blok kodova (Orthogonal Space Time Block Codes, OSTBC). Izvršena je analiza performansi sekundarnog sistema sa primenom svake od navedenih tehnika i adaptivnom kontrolom snage. Analiza je data za slucaj kada se statističke karakteristike fedinga u propagacionom okruženju mogu opisati Nakagami-m raspodelom. Izvršena je detaljna analiza performansi sekundarnog sistema koji je ograničen maksimalnom dozvoljenom snagom interferencije na primarnom prijemniku i maksimalnom dostupnom snagom sekundarnog predajnika. Za slučaj sekundarnog linka bez primene diverzitija, kao i sekundarni link sa primenom MRC i OSTBC tehnika izvedeni su analitički izrazi u zatvorenom obliku za ergodični kapacitet sekundarnog linka u okruženju sa Nakagami-m fedingom. Takođe, izvedeni su i odgovarajući asimptotski izrazi koji su validni u slučaju kada važi jedan od uslova ograničenja rada sekundarnog korisnika. Izveden je analitički izraz za ergodični kapacitet sekundarnog linka sa primenom TAS/MRC u okruženju sa Rejlijevim fedingom, pod uslovom ograničene snage interferencije na primarnom prijemniku. Izvedeni izrazi su validni za proizvoljan broj predajnih i prijemnih antena sekundarnog korisnika i potvrđeni su nezavisnim simulacionim postupkom. Analizirana je prosečna verovatnoća greške po bitu pri prenosu signala na sekundarnom linku kognitivnog radio-sistema. Izveden je izraz u zatvorenom obliku za verovatnoću greške po bitu za više različitih modulacionih postupaka, u slučaju kada su sekundarni predajnik i prijemnik opremljeni sa po jednom antenom, kao i za slučajeve kada su primenjene MRC i OSTBC tehnike. Posebno su razmatrane performanse višeantenskog sekundarnog linka u slučaju kada je informacija o stanju u kanalu od sekundarnog predajnika do primarnog prijemnika zastarela. Rezultati sprovedene analize ukazuju da je sekundarni sistem sa OSTBC manje osetljiv na zastarelost ove informacije u odnosu na sistem sa primenom TAS/MRC. Pokazano je da je izbor optimalne tehnike složen i da zavisi od dozvoljene verovatnoće premašenja interferencije, karakteristika fedinga u okruženju, dimenzija sistema, kao i vrednosti praga interferencije na primarnom prijemniku i maksimalne emisione snage sekundarnog predajnika.
Due to massive use of wireless communication systems that are enabling data transfer with high data rates, available frequency bands became very limited and scarce resource. Regulatory bodies have already granted licences for exclusive spectral use of large part of available spectrum, thus making development and expansion of new services extremely challenging. Numerous analyses indicate that the lack of flexibility in the spectrum access represents more significant problem than physical lack of available frequency bands. The concept of cognitive radio represents the potential solution, leading to the more efficient use of spectrum. In this thesis cognitive radio system with controlled interference level is analyzed, where primary (licensed) and secondary (cognitive) spectrum users simultaneously access the same spectrum band. Secondary user is allowed to transmit signal under the constraint that generated interference at the primary user’s receiver is below the predetermined threshold. This requirement can be achieved by applying adaptive power control at the secondary transmitter, with the consequence of limiting the secondary link capacity. In this thesis the application of multi-antenna techniques is analyzed, with an aim of improvement of secondary user’s performances. The employment of maximal ratio combining (MRC) diversity at the secondary user’s receiver is considered, as well as the application of transmit antenna selection (TAS) technique. Combined use of both above mentioned techniques (TAS/MRC) is also analyzed, together with possibility of application of orthogonal space time block codes (OSTBC) at the secondary user’s link. The analysis of the performance of the secondary system with the application of each of these techniques and adaptive power control is provided. The analysis is given for the case when the statistical properties of the fading in the propagation environment can be described by Nakagami-m distribution. Comprehensive analysis of the secondary system performances is provided, for the case when both maximal permitted interference level at the primary receiver and maximal available power of secondary transmitter constraints are applied. Closed-form analytical expressions are derived for the ergodic capacity of secondary link in the Nakagami-m propagation environment, for the case when no diversity is applied, as well as for the cases of secondary link with MRC diversity and OSTBC. Furthermore, asymptotic expressions valid for the case when only one of the constraints of secondary user transmission is applied are developed. Closed-form expression for the ergodic capacity of secondary link with TAS/MRC in Rayleigh fading environment is derived, for the case when interference power constraint is applied. Expressions are valid for an arbitrary number of transmit and receive antennas and are confirmed by using an independent simulation method. The analysis of error probability for the secondary link cognitive radio system is provided. Closed-form expressions are derived for different modulation techniques, for the cases when secondary user is equipped with single transmit and single receive antennas, as well as for the cases when MRC and OSTBC techniques are applied. Finally, multi-antenna secondary user link performances are analyzed for the case when channel state information (CSI) for the link between secondary transmitter and primary receiver is outdated. Results of the analysis indicate that secondary system with OSTBC is more robust to the outdated CSI when compared to system with TAS/MRC. It has been shown that the selection of optimal technique is complex and depends on probability of interference outage, fading characteristics in propagation environment, system dimensions, as well as interference threshold at primary user receiver and maximal transmit power constraint of the secondary user.
true
Optimizacija i implementacija naprednih protokola za rutiranje
Optimization and implementation of the advanced routing protocols
Ova disertacija u prvom delu analizira protokole rutiranja u topologijama data centara. Posebna pažnja je posvećena dvofaznim protokolima rutiranja koji u pravilnim topologijama data centara sa velikim brojem alternativnih putanja omogućavaju izbegavanje zagušenja u mreži. U disertaciji je predložen novi algoritam za optimizovano dvofazno balansiranje, LB-ECR, koji omogućava bolje iskorišćenje komunikacionih mreža data centra. Korišćenjem metoda linearnog programiranja LBECR maksimizuje protok bez gubitaka za date saobraćajne zahteve svičeva. LB-ECR se oslanja na ECMP rutiranje koje je uobičajeno u data centrima jer koristi alternativne putanje iste cene. Dvofazno balansiranje omogućava pojednostavljenje linearnog modela eliminacijom saobraćajnih tokova i smanjuje mogućnost zagušenja raspoređivanjem saobraćaja po manje iskorišćenim linkovima mreže. Pojednostavljenje linearnog modela omogućava njegovo rešavanje za mreže velikih data centara. Disertacija sadrži pregled topologija i načina rutiranja u data centrima i rezultate poređenja različitih algoritama rutiranja u tipičnim topologijama data centara. Pored optimizacije rutiranja, disertacija razmatra algoritme ažuriranja lukap tabela rutera. Drugi deo disertacije sadrži pregled lukap algoritama i algoritama ažuriranja, i ispitivane su performanse ažuriranja dva napredna lukap algoritma. Izvedene su formule za proračun najgoreg slučaja zauzeća memorija lukap bloka, dok rezultati simulacije pokazuju zauzeće memorije za tipične tabele rutiranja. Prikazani su i rezultati broja pristupa memorijama lukap bloka u toku ažuriranja, kao i kompleksnost algoritama ažuriranja i vreme izvršavanja za tipične tabele rutiranja. Treći deo disertacije obuhvata opis implementacija algoritama rutiranja sa balansiranjem i algoritama ažuriranja lukap tabela, kao i načina integracije ovih implementacija u okviru rutera.
The first part of this dissertation presents analysis of the routing protocols in data center topologies. Emphasis is put on two-phase routing protocols which enable congestion avoidance in data center topologies, which are regular and have significant number of alternative paths. This dissertation proposes new two-phase routing algorithm, LB-ECR, which enables better utilization of data center networks. Using the method of linear programming, LB-ECR maximises the loss-free throughput for the given required traffic of the switches. LB-ECR is based on ECMP routing which is common in data centers due to its utilization of alternative equal-cost paths. Two-phase balancing enables simplification of the linear model by eliminating traffic flows and decreases the possibility of congestion by distributing traffic among less used links. Simplification of the linear model simplifies its solution for large data centers. The first part of this dissertation gives an overview of the network topologies and the routing in data centers, and performance comparison of different routing algorithms in typical data center topologies. In addition to the routing optimization, this dissertation examines the algorithms for updating the lookup tables of Internet routers. The second part of this dissertation provides an overview of the lookup and update algorithms. Updating performance of two advanced lookup algorithms is examined. We develop formulas for the worst-case memory requirements for two fast lookup algorithms, while we show through simulations the memory requirements for typical routing tables. We also evaluate the number of memory accesses to the lookup modules during updates, the complexity of the updating algorithms, as well as their execution time for typical routing tables. The third part of this dissertation encompasses description of the implementations of the two-phase routing algorithms and of the lookup update algorithms, as well as the integration of these components inside the router.
false
Procena potencijala remedijacije sedimenta primenom imobilizacionih agenasa
Assessment of potential remediation of metal contaminated sediment using imobilisation agents
Predmet izučavanja ove disertacije je ispitivanje mogućnosti imobilizacije toksičnih metala u sedimentu primenom imobilizacione tehnike solidifikacije/stabilizacije portland cementom, kalcijum-oksidom, prirodnim zeolitom, letećim pepelom, kaolinitom, montmorilonitom, i različitim smešama ovih agenasa kao i određivanje njihove efikasnosti u zavisnosti od brojnih faktora. Efikasnost imobilizacionih postupaka praćena je korišćenjem tzv. testova izluživanja, odnosno određivanjem koncentracije metala koji su u toku određenog vremena iz imobilisane faze prešli u mobilnu fazu gde se smatraju potencijalno opasnim i biodostupnim. Krajnji cilj svakog testa izluživanja je mogućnost procene adekvatnosti primene određene imobilizacione, odnosno remedijacione tehnike. Efikasnost imobilizacionih postupaka zavisi i od određenih osobina: pH rastvora, dodatka imobilizacionog agensa što je istraživanjem ispitano. Na osnovu ovog testa, odnosno odnosa kumulativne frakcije metala koja je oslobođena iz imobilizacione smeše u mobilnu fazu i vremena određen je tip mehanizma koji omogućava transport metala (spiranje, difuzija, rastvaranje). Takođe su određeni i i parametri (koeficijenti difuzije, indeksi izlužljivosti) koji će poslužiti za ocenu efikasnosti prethodno primenjenih imobilizacionih tehnika. Primenjeni su i modifikovani testovi izluživanja sa ciljem što bolje simulacije realnih uslova. U modifikovanim testovima izluživanja korišćen je rastvor sa pH 3,25 da bi se simulirali uslovi koji oponašaju kisele kiše, odnosno uslovi u realnom sistemu. Takođe je koršćen i rastvor huminskih materija sa ciljem simulacije uslova velikog organskog opterećenja do kojeg bi moglo doći u slučaju akcidentnih situacija (na primer poplava) i generalno uslova koji se stvaraju u prirodi pri raspadanju organskog materijala (lišće, trava, itd.). Rezultati dobijeni simulacijom ovih uslova će dalje omogućiti modelovanje ponašanja metala u smislu dugoročnog "izluživanja" iz tretiranog sedimenta kao i procenu najefikasnijih agenasa za imobilizaciju različitih metala u sedimentu. Na osnovu dobijenih rezultata, zaključeno je da su optimalni agensi za imobilizaciju metala u sedimentu smeša cementa i kreča (5% cementa i 10% kreča), smeša montmorilonita i kreča (30% montmorilonita i 10% kreča), leteći pepeo (30%) i zeolit (30%). U ovim smešama dominantan mehanizam izluživanja je difuzija, a ove smeše se mogu smatrati inertnim otpadom po svim ispitivanim kriterijumima. Ovi rezultati se mogu upotrebiti za projektovanje i izgradnju pilot postrojenja na kome bi se ispitala efikasnost ovih agenasa za remedijaciju sedimenta u realnim uslovima. Dobijeni podaci su neprocenjivi sa aspekta ekonomski i ekološki prihvatljivog upravljanja sedimentom.
This work is concerned with exploring the possibilities of immobilization of toxic metals in sediments using solidification/stabilization as imobilization technique and using Portland cement, calcium oxide, natural zeolite, flying ash, kaolinite, montmorilonite, and various mixtures of these agents and determine their effectiveness depending on many factors. Performance of imobilization procedures was followed by the use of so-called leaching tests and determination of the metals concentration that are within a specified interval of time leached from the immobile phase and as such can be considered potentially hazardous and bioavailable. The ultimate goal of every leaching test is to assess the adequacy of the possibility of applying certain imobilization or remediation technique. Performance of imobilizaction procedures depends on certain characteristics: pH of solution, additon of imobilization agent, etc. Based on this test, ie relation between the cumulative fraction of metal leached from imobilization mixture and time, the type of leaching mechanism that allows the transport of metals (wash-off, diffusion, dissolution) was determined. Other parameters which will serve for evaluating the efficiency of the previously applied imobilization techniques (diffusion coefficients, leaching indices) were also determined. Modified leaching tests were applied with the aim of better simulation of real conditions on the field. In the modified leaching tests the solution with pH 3.25 was used to simulate conditions that mimic acid rain and conditions in the real system. The humic acid solution was also used with the aim to simulate high organic loads which could occur in the case accident situation (eg floods) and the general conditions that are created in naturewhen decaying organic material (leaves, grass, etc.). The results obtained by simulating these conditions will enable the modeling of behavior of metals in terms of long-term leaching period from the treated sediment and assess the most effective agents for the immobilization of various metals in the sediments. Based on these results, it was concluded that the optimal agents for immobilization of metals in the sediment mixture of cement and lime (5% cement and 10% lime), a mixture montmorilonite and lime (30% montmorilonite and 10% lime), fly ash (30%) and zeolite (30%). In these mixtures dominant leaching mechanism is diffusion, and these mixtures can be considered as inert waste by all tested criteria. These results can be used to design and builda pilot plant in which order to evaluate the effectiveness of these agents for remediation of sediment in real terms. The obtained data are invaluable from the aspect of economic and ecologically acceptable management of sediment.
false
Uticaj izbora scintilacionog kristala na odziv detektora kod pet uređaja
Influence of the scintillation crystal option on the detector response of pet devices
Pozitronska emisiona tomografija (PET) je tehnologija koja pruža jedinstvene i izuzetne mogućnosti u funkcionalnoj dijagnostici u smislu da se pomoću do sada najefikasnije i najpouzdanije metode dolazi do informacije o biohemijskoj aktivnosti i ćelijskom metabolizmu u telu, određivanjem tačne lokalizacije i vršenjem semikvantitativne procene distribucije radioaktivne supstance. U ovom radu porede se karakteristike nedavno uvedenih kristala na bazi lutecijuma (LYSO) sa karakteristikama konvencionalnih bizmut-orto-germanat (BGO) scintilatora. Ispitane su sledeće scintilacione karakteristike PET kristala: osetljivost, energetska rezolucija, svetlosni prinos, svojstvena energetska rezolucija, atenuacija γ zraka kroz scintilacione kristale, radijaciona dužina, Molijerov radijus, vreme gašenja svetlosnog impulsa u scintilacionom kristalu, vreme preleta (TOF eng. „Time of Flight“) i refrakcioni indeks. Razmatran je uticaj fizičko-hemijskih i električnih osobina scintilatora na odziv detektora, uparenost svih parametara PET scintilatora, dimenzionisanje detektorskih kristala, „organizacija“ u PET blok detektore i povezivanje sa PET fotomultiplikatorima, sve u cilju dobijanja što kvalitetnije PET medicinske slike.
Positron emission tomography (PET) is a technology that provides unique and exquisite possibilities in functional diagnostics, in the sense that it is the most efficient and most reliable method for obtaining information about biochemical activity and cellular metabolism in the body, by determining exact localization and performing semi- quantitative assessment of the distribution of a radioactive tracer. This work compares the characteristics of recently introduced lutetium based crystals (LYSO) to those of conventionally used bismuth-ortho-germanate scintillators (BGO) The following characteristis of PET crystals have been investigated: energy resolution, scintillation decay time, sensitivity, light yield, intrinsic energy resolution, atenuation of gama radiation in scintilation crystal, radiation length, time of flight, Moliere radius, refraction index. The influence of scintilators’ physical, chemical and electric properties on the detector response are considered and analyzed, as well as the PET scintilation parameters pairing, dimensioning of detector crystals, their organizing in PET block detectors and linking to PET photomultipliers, all in order to obtain a higher quality of PET medical image.
true
Sinergije hoda osoba posle moždanog udara
Synergy patterns of stroke subjects while walking
Moždani udar je oboljenje koje izaziva promene koje utiču na praktično sve segmente životnih aktivnosti. Posledice moždanog udara su izražene u senzorno-motornim funkcijama, ali i u mentalnim i emocionalnim elementima. Gubitak selektivne motorne kontrole, posebno hoda i pokreta šake, su od centralnog značaja u razvoju onesposobljenosti pacijenata. Tome dodatno doprinose poremećaji govora i psihičke promene. Hodanje je važna komponenta samostalnosti i neophodan činilac normalnih životnih aktivnosti. Ponovno omogućenje hoda pacijentu posle moždanog udara zato zauzima važno mesto u rehabilitaciji i predmet je izučavanja mnogobrojnih kliničkih studija. Osnovna hipoteza uspostavljanja hoda je da je on moguć i sa smanjenim senzorno-motornim resursima. Novi hod se može aktivirati kroz motorno učenje, koristeći svojstvo centralnog nervnog sistema da može da se adaptira na promene. Ovo motorno učenje je u osnovi slično procesu učenja u periodu posle rođenja. Motorno učenje hoda se pojačava klasičnim rehabilitacionim tehnikama (princip neurofiziološkog i motornog učenja), a danas često i u kombinaciji sa: 1) robotskim napravama koje asistiraju balansu, delimičnom rasterećenju težine i u određenoj meri pokretanju pojedinih segmenata, 2) povećanjem motivacije i motornog učenja primenom povratnih sprega (biofeedback, brain-computer interface - BCI) i 3) funkcionalnom električnom stimulacijom (FES) koja aktivira pojedine senzorno-motorne puteve od značaja za hod. Ova doktorska disertacija je nastala kao rezultat razvoja unapređenih sistema za asistenciju hodanja primenom električne stimulacije. Poznato je da električna stimulacija omogućava aktivaciju paralizovanih mišića delujući direktno na motorne nerve ili delujući na senzorne nerve i generišući funkcionalne reflekse. Ovaj metod asistencije je često korišćen za terapiju osoba posle moždanog udara koji rezultuje hemiplegijom, tj. jednostranom V paralizom tela. Električna stimulacija zahteva upravljački sistem koji će sinhrono drugim aktivnostima aktivirati mišiće koji su paralizovani. U ranijim istraživanjima je model upravljanja bio baziran na kopiranju sinergija koje su karakteristične za hod ljudi koji nemaju motorne probleme. U tim istraživanjima su dominantna dva pristupa: upravljanje sa povratnom spregom na bazi detaljnog modela hoda, i senzorski upravljana stimulacija na bazi neanalitičkog modela hoda. U merenjima promena koje nastaju na novo uspostavljenom hodu je uočeno da postoji značajna akomodacija biološke kontrole neparalizovanih delova tela u cilju kompenzacije i generisanja stanja koja omogućavaju integraciju pokreta sa spolja kontrolisanim pokretima paralizovanih (paretičnih) ekstremiteta. Ova saznanja su doprinela ideji da je neophodno detaljno istražiti karakteristike hoda osoba posle moždanog udara sa ciljem formiranja modela koji se može implementirati u sistem koji koristi višekanalnu električnu stimulaciju, a koji rezultuje generisanjem aktivnosti mišića koje na optimalan način doprinose oporavku hoda. Shodno tome je cilj istraživanja bio razvoj metode za praćenje karakteristika hoda (kinematike i dinamike) i karakteristika mišićne aktivnosti pri hodu u procesu učenja hoda. Hod se najčešće posmatra kao ciklični proces u kome se koraci ponavljaju, ali je zapravo skup koraka koji se međusobno razlikuju i treba ih karakterisati kao slučajan proces. Ova činjenica je uključena u istraživanja i za analizu hoda su primenjeni statistički modeli. U radu je posvećena pažnja i dinamičkom modeliranju hoda s obzirom na specifičnosti sistema koji čine nervni, mišićni i skeletni sistem, koji su odgovorni za hod i stajanje. Istraživanju je pristupljeno nakon detaljne analize postojećih sistema upravljanja višekanalnom električnom stimulacijom za rehabilitaciju hoda. Posebno su razmotrene različite metode upravljanja i njihova ograničenja. Snimanje karakteristika hoda na zdravim ispitanicima obavljeno je radi formiranja modela zdravog obrasca hoda. Više-senzorskim sistemom koji se postavlja na telo i omogućava snimanje većeg broja koraka, snimljene su VI kinematičke karakteristike hoda pacijenata nakon moždanog udara. Za merenje sile reakcije podloge, u istraživanju je korišćen ulošak za cipelu koji smo razvili u Laboratoriji za biomedicinsko inženjerstvo, Elektrotehničkog fakulteta u Beogradu. Sistemom za snimanje elektromiografskih signala (EMG) sa većeg broja mišića su registrovane mišićne aktivnosti mišića koji su dominantni za pokrete od interesa. U ovom istraživanju smo se bavili sa dva modaliteta hoda: hodom asistiranim konvencionalnim metodama (štap, terapeut) i hodom sa motorizovanom hodalicom Walkaround. Sva snimanja obavljena su na Klinici za rehabilitaciju „Dr Miroslav Zotović“, Beograd, na osnovu unapred definisanih procedura odobrenih od strane etičkog komiteta Klinike. U studiji su učestvovali pacijenti sa dijagnozom akutne (manje od 6 nedelja nakon moždanog udara), sub-akutne (od 6 nedelja do 6 meseci nakon moždanog udara) i hronične faze hemiplegije (više od 6 meseci nakon moždanog udara). Za obradu i analizu podataka, realizaciju stohastičke metode i neuralnih mreža korišćen je programski paket MATLAB, matematičko i simulaciono okruženje za matematičke proračune, simuliranje i analizu procesa. Analiza hoda na osnovu uobičajeno korišćenih kinematičkih parametara hoda (brzine, simetrije, trajanja jednooslonačke i dvooslonačke faze), proširena je kompleksnijom analizom uz primenu stohastičke metode (Principal Component Analysis - PCA). Detaljnija analiza kinematičkih sinergija omogućila je bolji uvid u mehanizme nervnog sistema koji rezultuju devijacijama hoda. Analiza hoda na bazi elektromiografskih signala pokazala je da li se i na koji način sinergijske šeme mišićnih aktivacija menjaju u zavisnosti od korišćenog asistivnog sredstva. Tehnike mašinskog učenja upotrebljene su da bi se na osnovu postojećih snimaka pacijenata sa hemiplegijom izvršilo mapiranje senzorske reprezentacije i mišićnih aktivacija hoda, radi formiranja individualnog modela hoda (baze znanja) koji karakteriše potencijalnog korisnika sistema. VII
Stroke is a unique disease with consequences affecting all life aspects. As a result of stroke, patients face physical, mental and emotional changes that decrease life potentials. The loss of selective control of limb segments and hand movements is the main contributor to the patients’ disability. This state is additionally affected by communication problems and mental disorders. Walking is an essential component of independence and an important part of various life activities. This is why restoration of walking has a significant role in rehabilitation of stroke patients and is often a central task of many clinical trials. Restoration techniques are based on the hypothesis that neuroplasticity (motor learning) allows the development of a new walking even with the existing limitations of a sensory-motor system after stroke. At present, gait restoration is largely based on different methods, used either independently or combined: 1) classical rehabilitation (neurophysiological and motor learning techniques), 2) robotic assistance of posture, body weight, and body segment movements, 3) feedback and brain-computer interface (BCI) used for motor learning, and 4) functional electrical stimulation (FES). This doctoral dissertation was written as a part of a larger study, where the systems for walking assistance using functional electrical stimulation are being developed. It has been shown that electrical stimulation activates paralyzed muscles by direct stimulation of motor system or reflex stimulation of afferent nerves. This principle is often used in movement restoration in patients after stroke which results in hemiplegia – paralysis of one side of the body. The use of FES requires external control that would integrate the biological and artificial control. In previous studies, the control of FES was primarily based on using synergies that characterize the walking of subjects with no sensory-motor impairments. Here two approaches are used: feedback control systems based on a detailed gait model and a sensor-driven stimulation based on a non- IX analytical gait model. The gait assessment showed that there is a significant accommodation of biological control of non-paralyzed extremities to match the disability of the affected leg, and that the restoration of healthy-like walking is most likely not the achievable goal. These findings have contributed to the idea that the characteristics of walking in hemiplegic patients after stroke need to be investigated in more details. This should lead to a model that could be implemented in a system that uses a multi-channel electrical stimulation, and generates muscle activities for optimal walking recovery. The important goal of this research was to develop a method for assessment of walking characteristics (kinematics and dynamics) and the characteristics of muscle activities during walking in clinical conditions. The gait is considered a cyclic activity; yet in reality there are substantial differences among the consecutive strides even when walking over ground with no obstacles. This directly suggests that the strides being part of the walking cycle should be characterized as a random process. This modeling requires the use of statistical models of gait. Not all elements of the process can be analyzed solely by kinematic measures. Knowledge of sensory-motor system responsible for locomotion is necessary, so the analysis includes dynamics of the walking as well. The inital step in the research was a detailed literature review in order to get a better insight into the existing control systems. Different control methods as well as their limitations were studied in detail. Walking characteristics of healthy subjects were recorded in order to form a model of healthy walking patterns (benchmark). Portable multi-sensor system that included goniometers and force sensing resistors (FSR) was used to record kinematical and dynamic characteristics of walking in patients after stroke. A sensor insole with FSR sensors developed in the Laboratory for biomedical engineering, at the Faculty of Electrical Engineering, Belgrade, was used as a force sensor. A multy-channel system for electromyography (EMG) was used to assess activities of muscles dominant for the walking. In this case, two modalities of walking were analyzed: walking assisted with a cane and walking assisted by a powered walker, called Walkaround that provides postural assistance. All recordings were done at the Institute for Rehabilitation „Dr Miroslav Zotović“, Belgrade, based on pre-defined X protocols approved by the Institute’s ethics committee following the Helsinki declaration and current EC rules and recommendations. The study included patients in their acute (less than 6 weeks after stroke), sub-acute (6 weeks to 6 months post-stroke) and chronic phases of hemiplegia (more than 6 months after stroke). Data processing and analysis, implementation of stochastic methods and neural networks, were done in MatLab. Principal Component Analysis – PCA, allowed a statistical gait analysis in addition to commonly used data analysis methods which consider kinematic and dynamic parameters of walking (speed, symmetry, single and double stance duration). The method provided a better understanding of the underlying mechanisms of the nervous system that result in walking deviations. Gait analysis based on electromyographic signals showed whether synergistic muscle activation patterns change depending on the asistive device used. Based on existing recordings of patients with hemiplegia, machine learning techniques were implemented to performe mapping of sensory representations to muscle activations of walking, in order to form an individual model (knowledge base), which characterizes a potential user of the system.
true
Zaštita nadzemnih vodova od atmosferskih prenapona u ekstremnim uslovima
Lightning protection of overhead power lines in extreme conditions.
Osnovni zadatak svakog elektroenergetskog sistema (EES) jeste snadbijevanje potrošača kvalitetnom i jeftinom električnom energijom. Da bi se ovaj zadatak ispunio mnogi problemi koji se javljaju u radu EES-a moraju biti rješeni, a jedan od izraženijih je zaštita njegovih dijelova od atmosferskih prenapona...
The fundamental task of every electric power system (EPS) is to provide the consumers with a quality and cheap electrical energy. To accomplish this task, many problems occurring in the operation of an EPS must be solved, and one of the significant ones is the protection of its parts against lightning overvoltages...
false
Energetska efikasnost rečnih samohodnih teretnih brodova
Energy efficiency of inland waterway self-propelled cargo vessels
Disertacija je proistekla iz višegodišnjeg istraživanja problema pogona rečnih samohodnih teretnih brodova. Hidrodinamičke pojave koje su posledica složenih uslova u kojima plove rečni samohodni teretni brodovi čine izbor adekvatnog propulzivnog kompleksa pri projektovanju veoma složenim, ali i izazovnim zadatkom. Poseban problem predstavlja nedostatak metoda za određivanje uticaja ograničenja plovnog puta na hidrodinamičku efikasnost. Odatle proističu i poteškoće pri određivanju ukupne energetske efikasnosti brodova ovog tipa. Na osnovu svih podataka prikupljenih iz literature do kojih je bilo moguće doći, ali i dostupnih izveštaja, oformljena je baza sa osnovnim dimenzijama i rezultatima merenja apsorpcije snage velikog broja različitih tipova rečnih samohotki. Formirana baza je poslužila kao osnova za istraživanja hidrodinamičkih pojava do kojih dolazi pri plovidbi. U okviru istraživanja je sproveden niz neposrednih merenja apsorpcije snage na brodovima u realnim uslovima. Ovo je omogućilo proširenje formirane baze, ali i ukazalo na kompleksnost postupka merenja iz čega proističu nekada i velike razlike između raspoloživih rezultata merenja apsorpcije snage na istom brodu, ali i između modelskih ispitivanja i ispitivanja brodova u prirodnoj veličini. Na osnovu rezultata modelskih ispitivanja razvijen je matematički model pomoću koga se za poznate dimenzije i nosivost broda, kao i ograničenja plovnog iv puta može proceniti snaga koju je neophodno angažovati u cilju postizanja željene brzine. Analizom rezultata ispitivanja brodova u prirodnoj veličini, i poređenjima sa rezultatima modelskih ispitivanja, uz primenu postojećih metoda za određivanje efekata spoljašnjih uticaja, izvedene su preporuke za procenu dodatka za službu. Ukazano je na put kojim se na osnovu ispitivanja brodova u prirodnoj veličini mogu odrediti koeficijenti propulzije u zavisnosti od ograničenja plovnog puta. Rezultati do kojih se došlo nikako nisu konačni. Buduća istraživanja će potvrditi i/ili unaprediti neke od zaključaka i postavki ovog rada. Na osnovu izvedenih zaključaka je nagovešten put kojim se može odrediti indeks energetske efikasnosti i uspostaviti referentne krive za rečne samohodne teretne brodove.
This dissertation has originated from many years of research on the issue of propulsion of inland-waterway self-propelled cargo vessels (IWV). Hydrodynamic effects stemming from complex conditions related to IWV navigation make selecting the appropriate propulsion system during the project phase a very complex and challenging task. A particularly important issue is the scarcity of methods for determining the influence of waterway restrictions on hydrodynamic efficiency. This produces difficulties when determining the overall energy efficiency of these ship types. Based on all the data collected from the available references and reports, a database was created containing basic dimensions and results of power absorption measurements of many different types of IWV. The data served as a basis for research devoted to navigation-related hydrodynamic effects. During the course of the research, numerous direct measurements of ship power absorption in realistic conditions were performed. This enabled the expansion of the aforementioned base, but also indicated the complexity of measurement methods, which occasionally produce considerable differences between the available results of power absorption measurements of the same ship, but also between examinations of models and corresponding full-scale IWV. Based on the available results of model-scale examinations, a mathematical model was developed that can determine the power required for achieving the desired speed, on the basis of the known ship dimensions and displacement, as well as waterway restrictions. The analysis of full-scale ships examinations and comparisons of model and full-scale measurements, along with the application of existing methods of determining the effects of external influences, led to the formulation of recommendations for estimating power vi power margin. A method was indicated, based on the examination of full-scale ships, whereby the propulsion coefficients depending on waterway restrictions can be determined. Resulting from the derived conclusions, a method was suggested that can determine the energy efficiency index and establish the reference curve/curves for inland waterway self-propelled cargo vessels.
true