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
Modelovanje gasnih detektora čestica visokih energija primenom tehnike elektronskih rojeva
Applications of electron swarm theory in modeling of gaseous particle detectors
Zahvaljujući svojim dobrim performansama i niskoj ceni po jedinici zapremine, gasni detektori su najčešće korišćeni detektori u eksperimentalnoj fizici visokih energija. Pored fizike visokih energija, ovi detektori nalaze primene u mnogim drugim oblastima, poput dozimetrije i zaštite od zračenja, medicine, fizike kosmičkog zračenja i geofizike. Postoje brojni modeli ovih uređaja. Bez obzira da li su analitički ili numerički, stohastički ili deterministrički, svi ovi modeli koriste transportne i/ili sudarne podatke za elektrone u gasovima. Nažalost, uobičajeno je da se unutar detektorske zajednice, ovi podaci kao i tehnike za njihov proračun primenjuju nekritički. Ova praksa, kao što je diskutovano u ovoj disertaciji, veoma često vodi ka pojednostavljenim modelima detektora i neadekvatnoj metodologiji za pristup i analizu. Imajući ove činjenice u vidu, osnovni cilj ove disertacije je da na osnovama transportne teorije rojeva elektrona, ukaže na čitav spektar važnih aspekata u proračunu i implemetanciji transportnih podataka u modelovanju kao i da na konkretnim modelima detektora pokaže kako podaci i njihova implementacija utiču na izračunate signale i performanse detektora. Osim toga, u ovom radu su detaljno analizirane netipične pojave u transportu elektrona indukovane eksplicitnim efektima nekonzervativnih sudara i njihove potencijalne implikacije u modelima. Specifičnosti transporta elektrona u ukrštenom električnom i magnetskom polju, ilustrovane su na primeru detektora tipa Time Projection Chamber (TPC) koji se koriste za trodimenzionalnu rekonstrukciju putanje čestica. Kod ovih detektora magnetsko polje ima ulogu u redukciji bočne difuzije od koje direktno zavisi prostorna rezolucija dok npr. linearnost rekonstrukcije zavisi od osetljivosti brzine drifta na temperaturu i nehomogenost magnetskog polja. U ovom kontekstu potencijalne optimizacije radnih uslova, a takođe imajući u vidu i neželjene varijacije parametara tokom rada, sistematski je razmotren uticaj električnog i magnetskog polja, temperature i pritiska gasa kao i udela nečistoća u gasnoj smeši, na transportne osobine elektrona u TPC detektoru. Posebna pažnja posvećena je detektoru sa pločastim elektrodama visoke otpornosti (Resistive Plate Chamber, RPC) koji se koriste za timing i triggering u brojnim eksperimentima fizike visokih energija kao i u drugim oblastima. Rešavanjem Boltzmannove jednačine i primenom Monte Karlo tehnike, koju smo sa jedne strane koristili da proverimo rezultate dobijene Boltzmannovom jednačinom, a sa druge strane za dobijanje prostorno razloženih transportnih parametara elektrona, identifikovani su i objašnjeni transportni fenomeni elektrona poput negativne diferencijalne provodnosti i grejanja zahvatom elektrona koji su uočeni u gasnim smešama RPC detektora koji se primenjuju na eksperimentima u CERN-u. U ovoj disertaciji, razmotreni su teorijski principi klasičnog fluidnog modela i modela zasnovanog na difuzionoj jednačini i hidrodinamičkoj aproksimaciji na osnovu kojih je razvijen numerički 1.5-dimenizionalni fluidni model RPC detektora. Ovim modelom, ispitivan je razvoj elektronske lavine i strimera pod dejstvom efekata prostornog naelektrisanja i fotojonizacije u gasu. Pokazano je da nepravilna implementacija transportnih podataka, zanemarivanjem eksplicitnih efektata nekonzervativnih sudara, može dovesti do greške od nekoliko stotina procenata u proračunu indukovanog signala. Razmatrana je i osetljivost izračunatog signala detektora na promene setova preseka za rasejanje elektrona u individualnim gasovima gasnih smeša. Konačno, u ovoj disertaciji razvijen je i mikroskopski Monte Karlo model RPC detektora koji se zasniva na praćenju pojedinačnih trajektorija elektrona i njihovih sudara sa molekulima gasa. Pomoću ovog modela, ipitivana je stohastika elektronskog lavinskog procesa. Takođe, koristeći različite setove preseka za rasejanje elektrona u individualnim gasovima gasne smeše, izračunate su vremenska rezolucija i efikasnost RPC detektora koje se dobro slažu sa eksperimentalnim vrednostima.
Owing to their good performance characteristics and low price per unit volume, gaseous particle detectors remained the most commonly used detectors in high energy physics experiments. In addition to high energy physics, these detectors have also found applications in other fields such as radiation protection and dosimetry, medicine, cosmic ray physics and geophysics. A number of methods to model particle detectors have been developed. Being analytical or numerical, stochastic or deterministic, detailed knowledge of electron swarm transport properties as well as reliable cross sections for electron scattering are required as an input in modeling. The highly applied nature of the field, has inevitably driven the modeling of these systems more towards empiricism, and unfortunately often away from its roots in the fundamental transport theory of electron swarms. Thus, the main goal of this work is to bridge the gap between the fundamental transport theory of electron swarms and applications in the field of particle detectors. This goal is achieved by considering many elements of the theory which are important for accurate calculation and correct implementation of electron transport data in modeling. In addition, we provide the examples of specific detector models in which the incorrect implementation of data affects the calculated signal and detector performance characteristics. In this work, we discuss atypical manifestation of electron transport phenomena induced by the explicit effects of non-conservative collisions and potential implications arising from their inclusion in the models. The peculiarities of electron transport in crossed electric and magnetic fields are illustrated using the example of Time Projection Chamber (TPC), a detector employed for three-dimensional reconstruction of particle trajectories. In this detector, the magnetic field suppresses the transverse diffusion, which directly affects the spatial resolution. On the other hand, the reconstruction linearity depends on the sensitivity of drift velocity on the gas temperature and non-uniformity of the magnetic field. With this as one of the motivating factors, and also having in mind the unwanted variation of detector parameters with time, we systematically study the influence of the electric and magnetic fields, gas pressure and temperature, as well as the impact of impurities in the gaseous mixture, on the electron transport properties in TPC. Special attention is given to the Resistive Plate Chambers (RPCs) which are used for timing and triggering purposes in many high energy physics experiments and elsewhere. The Boltzmann equation is used for the determination of electron swarm transport properties under conditions when transport is greatly affected by non- conservative collisions. A Monte Carlo simulation technique has been used with the aim of verifying the results based on the Boltzmann equation as well as for the evaluation of spatially resolved electron transport data. This segment of data is used to explain the existence of certain kinetic phenomena, including negative differential conductivity and attachment heating, which are important for the detector behavior. Within the framework of the classical fluid model and using the diffusion equation in association with the hydrodynamic approximation, we have developed a numerical 1.5-dimension fluid model of an RPC. This model is used to study the electron avalanche and streamer development under the influence of space charge effects and photoionization. We have shown that improper use of the data, especially the lack of correct representation of the explicit effects of non-conservative collisions, can lead to errors of a several hundred percents for the calculated signals. In addition, we discuss the sensitivity of the output detector signals with respect to the sets of cross sections for electron scattering. Finally, in this thesis we present our microscopic Monte Carlo model of RPC based on the tracking of individual electron trajectories and their collisions with the gas molecules. Using this model, we study the electron avalanche fluctuations and the related processes. The detection efficiency and timing resolution are calculated using different sets of cross sections for electron scattering. We have found that our results agree very well with the measured data.
true
Modelovanje mehaničkog naprezanja u poluprovodničkim nanostrukturama
Mecahnical stress modelling in semiconductor nanostructures
Tokom poslednjih dvadeset godina mnogo puta je pokazano da je tehnološki moguće dobiti nanostrukture u obliku vertikalno i periodično poređanih kvantnih tačaka - KT stekova. Ovakve nanostrukture mogu naći potencijalnu primenu u poluprovodničkoj laserskoj tehnologiji, ili za izradu solarnih ćelija. Do danas nije objavljen nijedan naučni rad u kome su izvedeni analitički izrazi za izračunavanje mehaničkih deformacija unutar KT stekova. Izvestan broj radova bavio se izračunavanjem mehaničkih deformacija primenom numeričkih metoda, i to za KT stekove sastavljene od manjeg broja kvantnih tačaka. U ovoj doktorskoj disertaciji analizirani su vertikalni stekovi poluprovodničkih kvantnih tačaka bez dislokacija, ugrađenih u supstrat poluprovodnika druge vrste, za epitaksijalni rast u kristalnom pravcu (001). Izvedeni su analitički izrazi za deformaciju unutar oba poluprovodnika u slučaju kada su debljine kvantnih tačaka i supstratskih odstojnika dovoljno male u odnosu na poprečne dimenzije steka. U konkretnim slučajevima, kada se stek sastoji od InxGa1-xAs kvantnih tačaka unutar GaAs supstrata, gde x uzima vrednosti 0,25; 0,5; 0,75 i 1,0, izračunati rezultati prikazani su grafički. Ovi rezultati pokazali su da kontrolom odnosa broja monoslojeva u kvantnoj tački i broja monoslojeva u periodu steka (odnosa ) deformacija unutar kvantne tačke može biti precizno kontrolisana. Na taj način moguće je precizno kontrolisati i elektronska i optička svojstva kvantnih tačaka koje zavise od deformacije, što je i pokazano u slučaju svih ispitanih nanostruktura. vii Od elektronskih svojstava nanostrukture izračunate su i grafički predstavljene promene rastojanja dna provodne zone od vrha zona lakih i teških šupljina u slučaju svih ispitanih nanostruktura. Manja od tih dveju promena određuje i promenu efektivnog energetskog procepa. Promena rastojanja dna provodne zone od vrha zone teških šupljina se blago menja sa odnosom , dok se rastojanje od vrha zone lakih šupljina menja sa mnogo većim nagibom, i ova činjenica otvara mogućnosti za kontrolu energetskog procepa menjanjem odnosa . Promenu energetskog procepa sa promenom odnosa  diktira rastojanja dna provodne zone od vrha zona lakih šupljina, počevši od vrednosti odnosa  = 0,592, tj u intervalu 0,592    1, i ovo je upravo interval od interesa, u pomenutom smislu, i to za sve ispitane nanostrukture.
It was proven many times during the past twenty years that the vertical alignment of the semiconductor quantum dots into the vertical stacks - QD stacks, is technologically possible. This kind of nanostructures could be appropriate for the applications in the semiconductor laser technology, and for the solar cells fabrication. Until now, there is no published scientific paper about the analytical strain evaluation inside QD stacks. Some of the published papers have discussed results obtained by the numerical methods for the stacks which have consisted of several quantum dots. In this Doctoral Dissertation the vertical stacks of semiconductor quantum dots without dislocations, incorporated in the other semiconductor substrate, for the epitaxial growth in the crystal direction (001) were considered. Analytical expressions for the strain in the both semiconductors were derived for the case when the thicknesses of quantum dots and the substrate spacers are small enough in comparision with the stack lateral dimensions. In the particular cases when the stack consists of InxGa1-xAs quantum dots in the GaAs substrate, where x takes values 0.25; 0.5; 0.75 and 1.0, the obtained results were calculated and shown graphically. These results have shown that by controlling ratio between the number of monolayers in InAs quantum dots and number of monolayers in the stack period (ratio ), quantum dots strain could be precisely controlled. In this way, the electronic and optical properties of quantum dots, which depend on the strain, could be controlled precisely, and that was shown for all of the tested nanostructures. Among the nanostructure electronic properties, the changes in distances between the bottom of the conduction zone, and the tops of the heavy hole and light hole zones, for all of the tested nanostructures, were calculated and graphically presented. From these two changes, smaller one determines the change in the effective energy gap. ix Distance change from the bottom of conduction zone to the top of heavy hole zone slightly varies with the ratio , while the distance change from the bottom of conduction zone to the top of light hole zone varies with a pretty greater slope, and this fact opens the new possibilities for the energy gap control by changing ratio . Changing of energy gap with changing ratio  is dictated by the distance from the bottom of conduction zone to the top of light hole zone, starting with the ratio  = 0.592, i.e. in the interval 0.592    1, and this is the interval of interest, as interpreted above, for all of the tested nanostructures.
true
Optička detekcija ostataka supernovih i uticaj njihove emisije u liniji H[alfa] na određivanje stope formiranja zvezda
Optical supernova remnants and their influence on star formation rates derived from H[alpha]
U ovoj disertaciji je analiziran uticaj zracenja ostataka supernovih (OSN) u liniji H na stopu formiranja zvezda (SFZ) odredenu na osnovu emisije u ovoj liniji. Pri- kazana je detekcija 16 optickih OSN, od strane autora ove disertacije, u delu galaksije IC342, kao i dva potencijalna OSN u galaksiji NGC 185. Posmatranja su realizovana pomocu dvometarskog teleskopa na Nacionalnoj astronomskoj opservatoriji Rozen u Bugarskoj, koriscenjem uskopojasnih ltera: H, [S ii] i kontinuum ltera na tala- snoj duzini bliskoj talasnim duzinama ove dve linije. Takode, u sklopu rada na ovoj disertaciji, pretrazena je literatura kako bi se sastavila lista svih opticki detektovanih OSN van nase Galaksije. Prilikom odredivanja stope formiranja zvezda iz linije H, polazna pretpostavka je da zracenje potice iz jonizovanog gasa koji okruzuje mlade zvezde, tj iz tzv. Hii regiona. Zbog toga opticko zracenje ostataka supernovih u liniji H ,,kontaminira" uks na osnovu koga treba da se odredi SFZ. Nalazimo da je srednja vrednost udela H uksa OSN u ukupnom uksu galaksije, za uzorak od 18 galaksija sa optickim OSN, 5 5%. Zbog postojanja posmatrackih selekcio- nih efekata, ovaj procenat predstavlja tek donju granicu kontaminacije ukupnog H uksa galaksije zracenjem OSN. Za M83, galaksiju sa najvecim brojem detektova- nih optickih OSN, ovaj procenat je 9%, i ocekujemo da je ovaj broj blizu stvarnog procenta zagadenja u spiralnim galaksijama. Na primeru dva specicna slucaja, patuljastih galaksija Holmberg IX i NGC 185, pokazujemo da H uks proizveden od strane ostataka supernovih, ali i drugih objekata koji ne ukazuju na radanje zvezda, moze biti daleko veci, cak do deset puta, od onog uksa koji proizvode Hii regioni.
In this dissertation we discuss in uence of H emission from supernova remnants (SNRs) on H derived star formation rates (SFRs). We present the detection of 16 optical SNR candidates in a part of nearby spiral galaxy IC342, and two more poten- tial SNRs in NGC 185 galaxy. The candidates were detected by applying [S ii]/H ratio criterion on observations made with the 2 m telescope at Rozhen National Astronomical Observatory in Bulgaria. Also, in this dissertation we present the most up-to-date list of nearby galaxies with optically detected SNRs. When deri- ving H star formation rates, assumption is that the radiation is coming from the ionized gas surrounding hot and young stars { Hii regions. In this case, emission from SNRs contaminates H ux used to derive SFR. We found that the contribu- tion of SNRs' ux to the total H ux, for 18 galaxies in our sample of galaxies with optical SNRs, is 5 5%. Due to the observational selection effects, the SNR contamination of SFRs derived herein represents only a lower limit. M83 is the galaxy with the most detected optical SNRs and in this galaxy SNRs account for 9 per cent of the total H emission. We expect that fraction similar to this would be close to the real contribution of SNR emission to the total H emission in spiral galaxies. Using two dwarf galaxies as an example, Holmberg IX and NGC 185, we show that the contamination of H ux by SNRs, or other sources, can be much larger, up to ten times higher than total H ux coming from Hii regions of the observed galaxy.
false
Dobijanje lantan-galata za primenu u gorivnim ćelijama
Obtaining of lanthanum-gallate for fuel cells application
U ovom radu su sintetisani prahovi na bazi lantan-galata, citratnom sol-gel metodom i hemijskom sintezom u parnoj fazi (CVS), a njihovim presovanjem i sinterovanjem su dobijeni keramički uzorci kontrolisane mikrostrukture. Cilj disertacije je bio dobijanje guste keramike koja bi mogla da se primeni kao jonski provodni elektrolit u gorivnim ćelijama sa čvrstim elektrolitom (SOFC), a koje bi radile na srednjim temperaturama od 500–700 °C (IT-SOFC). Po prvi put su sintetisani nanoprahovi lantan-galata u gasnoj fazi i utvrđeno je da ovako dobijeni polazni prahovi, zbog svojih superiornih karakteristika omogućavaju snižavanje temperature sinterovanja za 150 °C, što je veoma povoljno kako sa stanovišta dizajniranja mikrostrukture, tako i zbog uštede energije u procesu proizvodnje gorivnih ćelija. Ipak, zbog nedovoljne kontrole hemijskog sastava i stehiometrije sintetisanih prahova, što je i najveći nedostatak CVS metode dobijanja perovskitnog LaGaO3, nije bilo moguće dobiti keramiku koja bi ispunjavala zahteve za primenu u SOFC. S druge strane, citratna sol-gel metoda pruža veliku kontrolu pomenutih parametara zbog čega je bila moguća sinteza čitavog niza čvrstih rastvora lantan-galata. Ovom metodom su tako sintetisani čist lantan-galat (LG) i dopirani prahovi kod kojih je deo lantana supstituisan stroncijumom, a deo galijuma magnezijumom: La0,85Sr0,15Ga0,85Mg0,15O3- La1-xSrxGa0,8Mg0,2O3-, gde je x = 0,10, 0,15 ili 0,20 (LSGM). Svi ovako sintetisani prahovi su zahtevali naknadnu kalcinaciju na 900 °C, ali su potrebne gustine (>95% teorijske gustine) postignute sinterovanjem na temperaturi od 1450 °C već nakon 2 h, što je izuzetno kratko imajući u vidu literaturne podatke. Takođe, sinterovani uzorci su po faznom sastavu bili čista LSGM keramika, a impedansna merenja su pokazala da je najveću provodljivost imao uzorak La0,85Sr0,15Ga0,8Mg0,2O3-. Ovaj sastav je dalje korišćen u cilju provere mogućnosti za dodatno povećanje provodljivosti te su pripremljeni i uzorci kod kojih je izvršena parcijalna supstitucija magnezijuma sa niklom ili kobaltom: La0,85Sr0,15Ga0,8Mg0,2-yMyO3-, gde je M = Ni ili Co, a y = 0,03 ili 0,05 (LSGMN i LSGMC). Pokazano je da dodatak male količine prelaznih metala značajno utiče na mehanizam provođenja, ali da je na višim temperaturama jonska provodljivost i dalje dominantna. Konstatovano je da se dodavanjem male količine Ni ili Co u LSGM mogu dobiti materijali koji bi služili kao elektroliti u IT-SOFC, pri čemu je potencijal nikla kao dopanta nešto veći nego kobalta.
Powders based on lanthanum-gallate have been synthesised in this work by using citrate sol-gel method in the liquid phase and by chemical vapour synthesis (CVS). As-synthesised powders were calcined, pressed and finally sintered in order to produce ceramic samples with controlled microstructure. The main goal of this dissertation has been obtaining of dense ceramics for application in ion conducting electrolyte for solid oxide fuel cells working at 500–700 °C (IT-SOFC). Lanthanum-gallate nanopowders have been synthesised in the gas phase for the first time and it has been determined that these starting powders posses superior properties which could lower down the sintering temperature for about 150 °C. This is very significant considering microstructure design, but also high energy consumption during the manufacturing process of fuel cells. However, it has not been possible to obtain ceramics with exact properties needed for SOFC application using CVS due to the lack of control of chemical composition and stoichiometry of the as-synthesised powders, which are the main drawbacks of this method. On the other hand, citrate sol-gel method offered a possibility to precisely control aforementioned parameters which enabled synthesis of a whole range of lanthanum-gallate sollid solutions. So, pure perovskite lanthanum-gallate (LG) has been synthesised by using this liquid phase method, but also doped powders where part of lanthanum and gallium was supstituted with strontium and magnesium, respectively: La0,85Sr0,15Ga0,85Mg0,15O3- and La1-xSrxGa0,8Mg0,2O3-, where x = 0.10, 0.15 or 0.20 (LSGM). Calcination at 900 °C was necessary step during the processing of these sol-gel powders but the sintering time at 1450 °C was only 2 h which is quite short in comparison with available literature data. Additionaly, the sintered samples were phase pure LSGM ceramics and impedance measurement showed that the highest conductivity had sample La0,85Sr0,15Ga0,8Mg0,2O3-. Accordingly, this composition has been used to check the possibility of further improvement of conductivity. A set of new samples has been prepared where one part of magnesium has been substituted with nickel or cobalt: La0,85Sr0,15Ga0,8Mg0,2-yMyO3-, where M = Ni or Co and y = 0.03 or 0.05 (LSGMN i LSGMC). It has been shown that addition of small amount of transition metals significantly influences conduction mechanism, but at higher temperatures the ionic conductivity is still dominant. It has been found that electrolyte materials for IT-SOFCs could be obtained by incorporation of small quantities of Ni or Co into LSGM and that nickel is more promissing for this purpose than cobalt.
false
Projektovanje i ispitivanje struktura baza podataka u upravljanju održavanjem vazduhoplovanih sistema
Visoki zahtevi koji se postavljaju sistemu za odrzavanje vazduhoplova izmedu ostalog zahtevaju temeljnu analizu postojece strategije odrzavanja i njeno znacajno unapredenje. Savremeni sistemi odrzavanja obuhvataju i upravljanje odrzavanjem, tj. zasnovani su na automatizovanoj podrsci odrzavanju primenom racunara (Computer Integrated Maintenance Management). U osnovi automatizovane podrske odrzavanju primenom racunara je model podataka. Veoma je vazno da je model podataka dovoljno bogat da bi mogao, na relativno lak nacin prihvatiti i u sebe ukljuciti razlicite koncepte realnog sistema zajedno sa svojim atribu- tima i medusobnim vezama. Hijerarhijske strukture podataka (strukture podataka tipa stablo) su posebno interesantne u sistemu odrzavanja vazduhoplova i one su predmet istrazivanja u ovom radu. Postavlja se sledeci zahtev. Obezbediti takvu implementaciju hijerarhijske struk- ture koja pruza mogucnost njene ekasne obrade. Rekurzija koja je u osnovi hijerarhijske strukture podataka odnosno stabla koji je reprezent ovakve strukture podataka, moze biti problem u ekasnom upravljanju sistemom odrzavanja vazduhoplova. U radu su date nove karakterizacije stabla, koje daju osnov za nerekurzivne modele hijerarhijske strukture poda- taka. U kontekstu ogranicenih resursa u sistemu za odrzavanje vazduhoplova posebno su intere- santne optimalne strategije odrzavanja. Bitan segment sistema za upravljanje odrzavanjem je planiranje odrzavanja. Jedan od osnovnih problema izmedu ostalog je neizvesnost dogadaja koji znacajno uticu na samo planiranje. Rec je o dogadajima koji su po svojoj prirodi slucajni. Posebno analizira se planiranje odrzavanja u slucaju kada skup aktivnosti ima hijerarhijsku strukturu i pri tome je trajanje pojedinih aktivnosti iz datog skupa aktivno- sti slucajna velicina. Pod pretpostavkom da aktivnost na visem nivou hijerarhije ne moze zapoceti pre nego sto su zavrsene sve aktivnosti na nizem nivou hijerarhije i da su poznati gubici koji nastaju u slucaju da se sa nekim aktivnostima kasni, odnosno da su poznati gu- bici u slucaju da su neke aktivnosti preuranjene, potrebno je formirati optimalnu strategiju aktiviranja aktivnosti iz datog skupa. U slucaju hijerarhije visine jedan problem se svodi na poznati newsboy problem, ali je u opstem slucaju isuvise komplikovan. U radu se predlaze heuristicki pristup kojim se do resenja dolazi iterativnom primenom jednostavnog resenja problema na dvononiovskoj hijerarhiji. Vazduhoplovi su savremena saobracajna sredstva vrhunske tehnologije sa velikim bro- jem komponenti i kompleksnom strukturom. Pojedine komponente vazduhoplova se koriste na razlicit nacin, trpe razlicita opterecenja i odrzavaju po programu koji je njima najbolje prilagoden. Jedan od nacina savladavanja slozenosti je takozvana hijerarhijska dekompozi- cija. Slozeni koncept se na jednom nivou apstrakcije posmatra kao jedinstvena celina. Na nizem nivou apstrakcije se posmatra kao koncept koji se sastoji od delova (komponenti). Uzastopnom primenom opisanog postupka dobija se takozvana hijerarhijska sastavnica. Da- kle jedan od bitnih koncepata u sistemu odrzavanja vazduhoplova jeste sastavnica. Poseban problem je obezbediti kvalitetan sadrzaj hijerarhijske strukture podataka, odnosno u sistemu za upravljanje odrzavanjem vazduhoplova treba obezbediti kvalitetnu sastavnicu. U radu je predlozena eksibilna struktura sastavnica koja u velikoj meri podrzava njeno sadrzajno rasirenje...
The high demands that are placed to aircraft maintenance systems among other things require thorough analysis of existing maintenance strategy and its signicant improvement. Modern maintenance systems include the management of maintenance, i.e. they are based on automated maintenance support by computer - Computer Integrated Maintenance Mana- gement. The basic component in automated maintenance support account is data model. It is important that the data model is rich enough to accept in a relatively easy way a variety of real system concepts together with their attributes and interrelations. Hierarchical data structure ( textit tree type data structure) are of special interest in the aircraft maintenance systems, and they are the subject of this study. This raise the following request. Implement hierarchical data structure which allows ecient processing. Recursion which is essential in implementation of hierarchical data structure, i.e., of the tree, which is typical representative of these data structures, can be a problem in ecient aircraft maintenance management. This study gives a new characterizations of tree, which provide a basis for nonrecursive models of hierarchical data structures. In the context of limited resources in the aircraft maintenance systems, optimal mainte- nance strategies are of special interest. An important segment of the maintenance manage- ment system is maintenance planning. One of the major problems is uncertainty of events which signicantly aect maintenance planning. These events are random by its nature. In particular, the planning of maintenance is analyzed in the case when the set of activities has a hierarchical structure and duration, of each activity from a given set of activities, is random value. Assuming that the activity at a higher level of hierarchy can not begin before the completion of all activities at a lower level of hierarchy, and that losses incurred in the case of delayed activities, i.e., losses incurred in the case of premature activity are known, it is necessary to establish optimal activating strategy for each activity from given set of activities. In the case of one level hierarchy, problem is reduced to the well known newsboy problem, but it is too complicated in general. In this study, a heuristic approach is proposed for the solution of general problem that is based on iterative application of the solution of a simple two levels hierarchy problem. Aircrafts are modern high-tech means of transport with a large number of components and complex structure. Components of aircraft are used in dierent ways, they suer dierent loads and are maintained by the programs that are best suited for them. One way to overcome this complexity is so-called hierarchical decomposition. Complex concept at one level of abstraction is seen as unied. At the lower level of abstraction it is seen as a concept that consists of parts (other components). Repeated application of described procedure gives the so-called hierarchical bill of material. So one of the important concepts in the aircraft maintenance systems is bill of material. A particular problem is to provide quality hierarchical data structures to the aircraft maintenance management system, i.e., to provide quality bill of material. This study proposes a exible structure of the bill of material which greatly supports disseminate its content...
false
Personalizovani programski vodiči za digitalnu televiziju
Personalized program guides for digital television.
Razvoj digitalne televizije je doveo do značajnog porasta broja TV sadržaja dostupnih korisnicima, ali je otežao izbor onog koji je od interesa. Sve do pojave personalizovanih programskih vodiča sposobnih da nauče korisnička interesovanja i preporuče odgovarajuće sadržaje nije postojalo rešenje koje je na adekvatan način razmatralo ovaj problem. Ranija rešenja, kao što su štampani i elektronski vodiči, su pretežno samo pretvarala problem viška informacija u drugi oblik. Napredak tehnologije i društva postavlja sve veće zahteve pred personalizovane programske vodiče za digitalnu televiziju, što zahteva njihovo pažljivo planiranje i projektovanje. Vodiči moraju da budu u mogućnosti da modeliraju različite načine donošenja odluka pojedinačnih korisnika, da rade u realnom vremenu na mobilnim uređajima s ograničenim hardverskim resursima, da vode računa o karakteristikama prikupljenih podataka, da uzimaju u obzir kontekst u kome se pristupa TV sadržaju i da štite privatnost svih korisnika, jer neki od njih nisu svesni mogućih opasnosti. Pažljivim izborom arhitekture i algoritma učenja, lokalno implementiran vodič baziran na neuralnim mrežama može da ispuni sve ove zahteve. S obzirom na to da korisnici znatno češće pružaju informacije o sadržajima koji im se dopadaju nego o onim koji im se ne dopadaju, u ekstremnim slučajevima se dešava to da su prikupljene samo pozitivne interakcije. Da bi se taj problem prevazišao, predložen je sistem s dva režima rada. U prvom režimu sistem uči i pruža preporuke samo na osnovu TV sadržaja koje korisnik voli, dok u drugom izjednačava uticaj sadržaja koje korisnik voli i onih koje ne voli na proces pružanja preporuka. Povećan uticaj pozitivnih interakcija dovodi do degradacije predikcije sadržaja koje posmatrač ne želi da gleda, te će se, usled greške u klasifikaciji, neželjeni sadržaji često pojavljivati u listi preporuka i na taj način smanjiti zadovoljstvo korisnika. Korišćenjem serije simulacija pokazali smo da je postignuto trajanje treniranja neuralne mreže kratko, čak i na uređajima s ograničenim hardverskim resursima. Zaključak je da je predloženi vodič veoma pogodan za implementaciju na mobilnim uređajima od kojih se očekuje da u budućnosti postanu dominantan način pristupa TV sadržajima.
The development of digital television significantly increased the quantity of media contents available to the users, but made it difficult to make the right choice. Before the invention of the personalized program guides capable of learning user preferences and recommending adequate contents, there were no means of properly addressing this problem. Former solutions, such as printed or electronic program guides, mostly converted the problem of having to deal with too much information into another form. The advancements in both technology and society put higher demands to the personalized program guides for digital TV, which require careful planning and design processes. Guides must be able to model various individual decision making approaches, work in real-time on mobile devices with limited hardware resources, take into account the characteristics of the collected data, take into consideration the program accessing context and protect the privacy of all users, since some of them are not aware of the possible risks. By carefully choosing the architecture and learning algorithms, a locally implemented guide based on neural networks can fulfil all the aforementioned requirements. Due to the fact that the users provide information about the content they like much more often than about the one they dislike, only positive interactions are collected in extreme cases. In order to overcome that situation, a system having two operating modes is proposed. The first mode enables the system to learn and give recommendations based on preferred TV contents, while the second equalizes the influence of the liked and disliked contents on the recommending process. The increased influence of positive interactions degrades the unwanted content prediction process, resulting in classification error, appearance of unwanted content in the recommendation list and user dissatisfaction. By applying a series of simulations, we showed the accomplished neural network training time to be short, even in cases of devices with limited hardware resources. It can be concluded that the proposed guide is very convenient for implementation on mobile devices which are expected to become a dominant way of accessing media contents in the future.
true
Višekriterijumska optimizacija sastava šarže za hidrometalurški proces dobijanja cinka
Multi-criteria optimization of batch composition for the hydrometallurgical for the hydrometallurgical process of zinc production
Ova doktorska disertacija ima za cilj definisanje originalnog algoritma za formiranje optimalne mešavine od raspoloživih koncentrata cinka, kroz sedam koraka, postupnim uvoĎenjem pojedinih metoda višekriterijumskog pristupa odlučivanju, u cilju definisanja integralnog modela za rangiranje pojedinih komponenti u mešavini i višekriterijumske optimizacije sastava šarže, kao inputa u tehnološki proces. Definisan je originalni AHP – OEW – PROMETHEE (GAIA) – TOPSIS model u fazi okruženju. Ovaj model zasnovan je na poreĎenju različitih koncentrata cinka prema kriterijumima koji su definisani u modelu (korisne komponente, štetne komponente, kao i nabavna cena koncentrata). Za rangiranje pojedinih vrsta koncentrata u sastavu šarže, korišćena je PROMETHEE (GAIA) metodologija, uz primenu AHP (Analytical Hierarchy Process) i OEW (Objective Entropy Weight) pristupa u dodeljivanju težinskih parametara kriterijumima. UvoĎenjem hibridnog modela F-PROMETHEE (Fazi PROMETHEE), vrednosti svake alternative u odnosu na kriterijume, predstavljene su kao fazi brojevi. U postupku definisanja Fazi TOPSIS metodologije, odreĎena su fazi pozitivna i fazi negativna idealna rešenja. Zatim, odreĎene su distance svih alternativa od idealno pozitivnog i idealno negativnog rešenja, što je omogućilo da se odrede relativne blizine idealnom rešenju za svaki od definisanih scenarija. Definisani hibridni model na primeru optimizacije sastava šarže za hidrometalurški postupak dobijanja cinka, predstavlja poboljšanje pristupa u rešavanju problema mešavine i kao takav može biti primenjen i na druge sisteme, s obzirom da u sebi sadrži elemente koji se mogu primeniti na bilo koji drugi sistem.
This doctoral dissertation is dealing with the development of original algorithm for optimal mixture formation from available zinc concentrates, through seven steps, by introduction of certain methods of multi-criteria decision making approach, in order to define an integrated model for ranking the individual components in the mixture and multi-criteria optimization of the batch composition, as input in the technological process. The original AHP – OEW – PROMETHEE (GAIA) – TOPSIS model in the fuzzy environment is defined. This model is based on a comparison of different zinc concentrates according to the criteria defined in the model (useful components, harmful components, as well as the purchase price of concentrate). PROMETHEE (GAIA) methodology was used for the ranking of certain types of concentrates in the batch composition, by application of the AHP (Analytical Hierarchy Process) and the OEW (Objective Entropy Weight) approach in the allocation of the weight parameters for the optimization criteria. With the introduction of the hybrid Fuzzy PROMETHEE model, the value of each alternative in relation to the criteria, are presented as fuzzy numbers. In the process of defining the Fuzzy TOPSIS methodology, fuzzy positive and fuzzy negative ideal solution were determined. Then, the distances of all the alternatives from ideal positive and ideal negative solutions were determined, which enabled us to determine the relative closeness to the ideal solution for each of the defined scenarios. Defined hybrid model in the case of the batch composition optimization for the hydrometallurgical process of zinc production, represents an improvement of the approach for blending problem solving and as such can be applied to other systems, since it contains elements that can be applied to any other system.
true
Plazmonske strukture za poboljšanje poluprovodničkih infracrvenih detektora
Plasmonic structures for enhancement of semiconductor infrared detectors.
Plazmonika je jedna od oblasti nauke koje se u današnje vreme eksplozivno razvijaju. Ona je posvećena elektromagnetici nanokompozitnih metamaterijala koji podržavaju rezonanciju površinskih plazmona polaritona (surface plasmons polaritons, SPP). SPP predstavljaju hibridne ekscitacije nastale sprezanjem elektromagnetnih talasa sa oscilacijama slobodnih nosilaca naelektrisanja na razdvojnim površima između dva materijala sa različitim znakovima relativne dielektrične permitivnosti, npr. provodnika i dielektrika. Posledica ovakvog sprezanja je između ostalog lokalizacija elektromagnetnog zračenja na podtalasnom nivou, osobina plazmonskih struktura koja je našla veliku primenu u spektroskopiji, integrisanoj optici, senzorici itd. Jedna od značajnih primena plazmonske lokalizacije je u oblasti u fotodetekcije, pre svega za poboljšanje performansi solarnih ćelija. Najveći problem proširenja primene plazmonike u fotodetekciji na drugu oblast od interesa, infracrvene (IC) detektore, predstavlja činjenica da je plazmonska učestanost većine provodnika (metala) u ultraljubičastom ili vidljivom delu spektra. Brojne tehnološki pogodne tehnike koje su dale izuzetne rezultate za poboljšanje solarnih ćelija ostale su zbog toga bez primene u IC oblasti. Ova disertacija se prevashodno bavi proširenjem primenljivosti plazmonike na srednjetalasnu i dugotalasnu infracrvenu oblast i metodama prevazilaženja ograničenja koje postavljaju sami materijali. U tu svrhu razmatrana su dva pristupa. Jedan od njih podrazumeva upotrebu submikrometarskih čestica od provodnog materijala. Funkcionalnost u IC oblasti postiže se kombinacijom izbora pogodnijeg materijala čestica (elektroprovodni optički providni oksid umesto metala) i imerzije čestica u dielektrik visokog indeksa prelamanja. Drugi pristup podrazumeva korišćenje tankih metalnih slojeva sa uređenom matricom apertura koji omogućuju pomeranje spektralne zavisnosti prema crvenom delu spektra menjanjem isključivo geometrijskih parametara matrice apertura. Oba pristupa nude mogućnost praktično proizvoljnog podešavanja frekvencije plazmonske rezonancije i time njenu upotrebu za IC detektore. Analiza ova dva pristupa rađena je numeričkim simulacijama, primenom metode konačnih III elemenata. Uticaj na performanse infracrvenih detektora određivan je kombinovanjem rezultata numeričkih modelovanja sa analitičkim modelom IC detektora. U okviru disertacije predložene su i analizirane prve konkretne strukture plazmonski poboljšanih IC detektora zasnovanih na provodnim podtalasnim česticama i urađena je njihova optimizacija. Predložena je primena antirefleksnih slojeva sa gradijentom indeksa prelamanja za maksimizaciju sprezanja prostirućeg talasa sa plazmonskim modovima čestica. Razmatrana je mogućnost povećanja funkcionalnosti ovakvih struktura uvođenjem podtalasnih čestica oblika izdužene ili spljoštene sfere, gde se frekvencija i intenzitet rezonancije odvojeno podešavaju veličinom poluosa sferoida. Kod struktura sa nizom podtalasnih apertura pokazano je kako se njihova geometrija može izmeniti radi dodatnog povećanja specifične detektivnosti IC detektora superponiranjem više pojedinačnih 2D matrica na jednoj istoj strukturi. Za ovaj slučaj prikazano je formiranje optičkih vrtloga u aktivnoj oblasti detektora sa mogućnošću podešavanja spektralne i prostorne pozicije optičkih vrtloga. Možda najvažniji rezultat je to što je pokazano kako se čisto optičkim putem može smanjiti nivo generaciono-rekombinacionog šuma u poluprovodničkim infracrvenim detektorima prostim smanjenjem ukupne zapremine aktivne oblasti bez gubitka performansi detektora zahvaljujući plazmonskoj lokalizaciji, iako se smatralo da je ovaj šum fundamentalan i da su neophodne aktivne metode da bi se prevazišao. Pokazano je da se na taj način mogu na sobnoj temperaturi postići performanse plazmonski poboljšanih IC detektora koje odgovaraju konvencionalnim napravama hlađenim tečnim azotom.
Plasmonics is one of the fastest growing fields in the contemporary science. Plasmonics studies properties of nanocomposite metamaterials which support surface plasmon polariton (SPP) resonance. SPP are formed by coupling electromagnetic waves with free charge carrier oscillations at an interface between materials with different signs of their relative permittivity i.e. conductor and dielectric. One of the results of this coupling is localization of electromagnetic radiation on subwavelength scale, property of plasmonic structures that has found practical use in the fields of spectroscopy, integrated optics, sensors, etc. One of the principal applications of light localization is in the field of photodetection, primarily for the enhancement of solar cells. The main problem with any attempt to apply plasmonics for photodetector enhancement at longer wavelengths, i.e. for infrared (IR) detectors, is that the plasmon resonance frequency of most conductive materials (metals) is in the ultraviolet or visible part of the spectrum. Because of this many convenient methods yielding excellent results for plasmonic enhancement of solar cells have not been utilized in the infrared. The main goal of this dissertation is bringing plasmonic enhancement of semiconductor photodetectors to medium and long wavelength infrared parts of the spectrum by overcoming limitations imposed by material properties. To achieve this two approaches are considered and analyzed. The first approach implies the use of submicrometer conductive particles. A sufficient shift of plasmonic resonance to the infrared is achieved by both a suitable choice of the particle material (transparent conductive oxides – TCO instead of metal) and by immersion of the particles in dielectric with a large index of refraction. The second approach is based on using thin metallic films with 2D array of holes drilled through them, where redshifting is achieved by tuning the geometrical properties of the hole array. It is shown that both approaches allow one to achieve practically arbitrary positioning of plasmonic resonance in the infrared. The finite element method was used for numerical simulations of the analyzed structures. A combination of the results of numerical modeling with the V analytical results for the IR detectors was used to determine the effects of the plasmonic enhancement. As a result, the first actual structures based on subwavelength plasmonic particles for the enhancement of detectors in medium wavelength IR are proposed and analyzed. The use of antireflective layers with gradient index of refraction is proposed to maximize coupling of the incident light with the plasmonic modes of the particles. Modifications of the particle geometry from spherical to spheroidal are utilized to achieve additional degrees of freedom in designing the structure by allowing to tune separately the spectral positioning and magnitude of the plasmonic resonance by changing the size of different spheroid axes. For the case of structures with subwavelength aperture arrays it is shown that an additional increase of specific detectivity can be achieved by modifying their geometry through a superposition of two or more 2D arrays of apertures in the same structure. It is shown that in this case optical vortices are formed in the photodetector active area with a possibility to tune the spectral and spatial positioning of the vortices. Perhaps the most important contribution of the dissertation is the proof that it is possible to reduce generation-recombination noise of IR photodetectors through plasmonic localization by reducing the active area volume without sacrificing detector performance. This means that it is possible to reduce g-r noise by a purely optical method, although it was maintained previously that, being fundamental, this noise could only be suppressed by active methods. Because of this it is possible to achieve plasmonicaly enhanced photodetectors at room temperature with their characteristics corresponding to those of conventional devices cooled with liquid nitrogen.
true
Verovatnosni računi sekvenata i klasifikacija neklasičnih logika zasnovana na entropiji
Probability sequent calculi and entropy based non-classical logics classification
Posle kratkog uvodnog pregleda, rad je podeljen na dva dela. Prvi deo se bavi prisustvom verovatnoće u logici (v. [16], [17], [18], [19], [22], [23] i [24]), a drugi je posvećen primeni entropije u klasifikaciji polivalentnih logika (v. [14], [15], [20], [21] i [25]). Osnovna ideja koja dominira prvim delom rada jeste obogaćivanje Gentzen- ovog računa sekvenata klasične logike iskaza jednim verovatnosnim operatorom definisanim nad sekventima Γ ⊢ Δ kako bi se izrazila činjenica da "verovatnoća istinitosti sekventa Γ ⊢ Δ pripada intervalu [a, b] ⊂ [0, 1]". Uvodimo sledeće sisteme: LKprob, LKprob(ε), NKprob i LKfuzz. Osnovna forma sekevnata u sistemu LKprob je Γ ⊢ b, a Δ sa gore datim značenjem. Sistem LKprob(ε) se fokusira na Suppes-ove forme Γ ⊢ n Δ koje omogućavaju formalizaciju rečenice "verovatnoća istinitosti sekventa Γ ⊢ Δ pripada intervalu [1 - nε,1] ⊆ [0, 1]", za neki n ∈ N. Sistem NKprob predstavlja prirodno-dedukcijski analogon računu sekvenata LKprob. Modeli zasnovani na Carnap–Popper–Leblance-ovoj semantici definisani su za svaki od ovih računa uz odgovarajuće rezultate saglasnosti i potpunosti. Konačno, račun LKfuzz je uveden sa opdžtijom formom sekvenata Γ ⊢ h Δ, gde je h element konačne mreže, sa ciljem da se opiše jedno rasplinuće računa LK. Značenje sekventa Γ ⊢ h Δ je "da je h mera rasplinuća sekventa Γ ⊢ Δ". Modeli za LKfuzz su dati sa rezultatima saglasnosti i potpunosti, a dokaz-teoretski tretman računa LKfuzz uključuje i teoremu eliminacije sečenja. Drugi deo rada istražuje činjenicu da svaki logički sistem povezan sa particijom indukovanom odgovarajućom Lindenbaum–Tarski–jevom algebrom omogućava definisanje njegove entropije. Definišemo entropiju logičkog sistema baziranoj na geometrijskoj raspodeli mera nad odgovarajućom particijom skupa formula. Ova definicija omogućava klasifikaciju polivalentnih iskaznih logika u odnosu na njihovu entropiju. Asimptotske aproksimacije entropije nekih beskonačnovalentnih logika su takođe date. Razmotreni primeri uključuju Lukasiewicz-evu, Kleene-jevu i Priest-ovu trovalentnu logiku, Belnap-ovu četvorovalentnu logiku, Gödel-ove i McKay-eve m-valentne logike, i Heyting-ovu i Dummett-ovu beskonačnovalentnu logiku.
After a brief introductory survey, this work is divided into two parts. The first part deals with presence of probability in logic (v. [16], [17], [18], [19], [22], [23] and [24]), and the second one is devoted to the application of entropy in classification of many–valued logics (v. [14], [15], [20], [21] and [25]). The basic idea, dominant in the first part of the work, is to enrich the Gentzen’s sequent calculus LK for propositional classical logic by a kind of probability operator defined over the sequents Γ ⊢ Δ in order to express the fact that ”the truthfulness probability of Γ ⊢ Δ belongs to the interval [a, b] ⊂ [0, 1]”. We introduce the following four systems: LKprob, LKprob(ε), NKprob and LKfuzz. The basic form of sequents in LKprob is Γ ⊢ b, a Δ with the above given intended meaning. The system LKprob(ε) is focused on the Suppes-forms Γ ⊢ n Δ enabling to formalize the sentence ”the truthfulness probability of Γ ⊢ Δ belongs to the interval [1 - nε,1] ⊆ [0, 1]”, for some n ∈ N. The system NKprob presents a natural deduction counterpart of the sequent calculus LKprob. The models founded on Carnap– Popper–Leblance probability semantics are defined for each of these calculi and accompanied by the corresponding soundness and completeness results. Finally, the calculus LKfuzz is introduced with a more general form of the sequents Γ ⊢ х Δ, where x is an element of a finite lattice, with the aim to describe a fuzzification of LK. The meaning of Γ ⊢ х Δ is that ”x is the fuzziness measure of Γ ⊢ Δ". Models for LKfuzz are given with soundness and completeness results, and a proof– theoretical treatment of LKfuzz includes the cut–elimination theorem. The second part of the work explores the fact that each logical system associated with the partition induced by the corresponding Lindenbaum–Tarski algebra makes it possible to define its entropy. We define the entropy of a logical system based on geometric distribution of measures over matching partition of set of formulae. This definition enables the classification of many–valued propositional logics according to their entropies. Asymptotic entropy approximations for some infinite–valued logics are proposed as well. The considered examples include Lukasiewicz’s, Kleene’s and Priest’s three–valued logics, Belnap’s four–valued logic, Gödel’s and McKay’s m– valued logics, and Heyting’s and Dummett’s infinite–valued logics.
false
Razvoj fotoakustičkog mernog sistema za termičku karakterizaciju tankih uzoraka
Development of photoacoustic measurement system for thermal characterisation of thin samples
Nastanak zvučnog talasa nakon osvetljavanja uzorka modulisanim svetlosnim snopom naziva se fotoakustički (FA) efekat. Teorijska istraživanja predviđaju da je uz pomoć mernih tehnika zasnovanih na FA efektu moguće izvršiti određivanje karakteristika tankih uzoraka. FA tehnike nisu invazivne i zbog toga izazivaju interesovanje u raznim oblastima nauke, medicine i tehnike. U zavisnosti od karakteristika koje se žele proučavati, primenjuju se različite tehnike zasnovane na FA efektu. U fokusu predloženog istraživanja je primena frekvencijske FA tehnike za termičku karakterizaciju, dok će sporedni pravac istraživanja razmatrati mogućnosti povezivanja frekvencijske FA tehnike sa FA spektroskopijom radi simultane optičke i termičke karakterizacije. Predmet istraživanja je razvoj FA mernog sistema (FAMS) za određivanje termičkih parametara tankih uzoraka, i to termičke difuzivnosti, termičke provodnosti i koeficijenta termičkog širenja. Sistem se sastoji od pobudnog dela, FA ćelije i detekcionog dela. Pobudni deo sistema čine izvor svetlosti i modulator svetlosnog snopa. Modulisani svetlosni snop pogađa uzorak u kome se apsorbuje deo svetlosne energije, što za posledicu ima formiranje akustičkog signala frekvencije jednake frekvenciji modulacije svetlosnog snopa. Generisani akustički signal se detektuje mikrofonom, dok se deo svetlosti koji se reflektuje od uzorka detektuje fotodetektorom, pa detekcioni deo sistema čine mikrofon sa pretpojačavačem, fotodetektor, pojačavači detektovanih signala i sistem za akviziciju. U okviru sistema za akviziciju se nalaze akviziciona kartica, lock-in detektor i računar. FA ćelijom se naziva zatvoreni prostor u kome se prostire akustički signal generisan FA efektom. Mada FA ćelija može biti izvedena kao komora u koju se postavljaju uzorak, mikrofon i fotodetektor, češće se uzorak pričvršćuje vii direktno za mikrofon, tako da prostor između uzorka i mikrofona predstavlja FA ćeliju. Za primenu frekvencijske FA tehnike (FFAT) opisani merni sistem se koristi tako što se računarskom obradom signala, koje su detektovali mikrofon i fotodetektor, određuje zavisnost odnosa amplituda i fazne razlike tih signala od modulacione frekvencije. Na osnovu poređenja tih frekvencijskih zavisnosti sa rezultatima novih teorijskih modela FA efekta mogu se odrediti termičke karakteristike uzorka metodama linearnog i nelinearnog fitovanja, određivanjem pozicija minimuma i maksimuma, kao i analizom rezonantnih frekvencija. Cilj istraživanja je određivanje prenosne funkcije FA u opsegu tri dekade audio-frekvencijskog spektra. S obzirom na veliku širinu prenosnog opsega fotodetektora, prenosna funkcija FAMS može da se predstavi proizvodom prenosnih funkcija FA ćelije, mikrofona, pojačavača i akvizicione kartice. Prema tome, frekvencijski opseg FFAT je ograničen frekvencijskim opsezima pobrojanih komponenata mernog sistema. Sporedni cilj istraživanja predstavlja istraživanje mogućnosti detektovanja rezonanci toplotne memorije merenjem FA efekta kod uzoraka sa niskim stepenom strukturnog uređenja (eng. low-order), kao što su porozni materijali i biološka tkiva. Generalizovani model FA efekta, koji uključuje toplotnu memoriju kao svojstvo materijala, predviđa postojanje toplotnih rezonanci koje su posledica konačne brzine prostiranja toplote kroz uzorke. Kod uzoraka sa niskim stepenom strukturnog uređenja teorija predviđa malu brzinu prostiranja toplote i rezonantne frekvencije unutar audio-frekvencijskog opsega (AFO) FAMS koji se razvija u okviru disertacije. Značaj istraživanja se ogleda u proširenju frekvencijskog spektra FA ispitivanja, koji će obezbediti mogućnost tačnijeg i preciznijeg određivanja termičkih parametara, kao i otvaranju mogućnosti otkrivanja rezonanci termičke memorije za materijale niskog stepena strukturnog uređenja. Pored toga, modularni karakter predloženog mernog sistema otvara mogućnost njegovog povezivanja sa drugim mernim tehnikama i istovremeno određivanje većeg broja parametara materijala, ne ograničavajući se samo na termičke. Značajan doprinos viii predstavlja i sam razvoj ovakvog sistema, posebno sa stanovišta modularnosti, jer se dalja istraživanja mogu sprovesti u različitim pravcima. Na kraju treba reći da je, zbog svoje prirode, ovakav sistem moguće koristiti u istraživanjima gde je dominantan efekat minijaturizacije, što je u poslednje vreme jedan od čestih zahteva.
The formation of sound waves after exposition of a sample to a modulated light beam is called the photoacoustic (PA) effect. Theoretical researches predict that it is possible to determine characteristics of thin samples with the help of measurement techiques based on the PA effect. PA techniques are not invasive and for this reason cause interest in many different areas of science, medicine and technique. Depending on the characteristics to be studied, different techniques are applied based on the PA effect. In the focus of the suggested research is the application of the PA frequency technique for thermal catheterization, while the secondary direction of the research will consider possibilities of connecting the frequency PA technique with the PA spectroscopy aimed at the simultaneous optical and thermal characterization. The subject of research is the development of the PA measurement system for determination of thermal parameters of thin samples, namely thermal diffusivity, thermal conductivity and linear expansion coefficient. The system consists of an excitation part, a PA cell and a detection part. The excitation part of the system consists of a light source and a modulator of the light beam. The modulated light beam affects the sample, which absorbs a part of the light energy, and due to the PA effect, the created acoustic signal has the same frequency as the light-beam modulations. Generated acoustic signal is detected by a microphone, whereas the part of light reflected from the sample is detected by a photodetector, so the detection part of the system consists of the microphone with a preamplifier, photodetector, amplifiers of the detected signals and an acquisition system. The acquisition system includes an acquisition card, a lock-in detector and a computer. The PA cell is a closed space in which an acoustic signal is generated by the PA effect. Although a PA cell can be a chamber in which the sample, microphone and photodetector x are placed, more often the sample is attached directly to the microphone, so that the space between the sample and the microphone represents a PA cell. For the application of PA frequency technique, the described measurement system is used by determining the dependence of the amplitude ratio and the phase difference of these signals on the modulation frequency by the computer processing of the signals detected by the microphone and photodetector. Based on the comparison of these frequency dependences with the new theoretical models of the PA effect, the thermal characteristics of the sample can be determined by the methods of linear and nonlinear fitting, the determination of the positions of the minimum and the maximum, as well as the analysis of resonant frequencies. The aim of the study is to determine the transfer function of the PA in a three - decade audio-frequency spectrum (AFS). Due to the wide bandwidth of the photodetector, the transfer function of the PA measurement system (PAMS) may be represented by the product of the transfer functions of a PA cell of a microphone, amplifier and acquisition card. Accordingly, the frequency band of the frequency PA technique is limited by frequency ranges of the listed components of the measurement system. The secondary objective of the study is to investigate a possibility of detecting thermal memory resonances by measurement of the PA effect in low-order samples, such as porous materials and biological tissues. The generalized model of the PA effect, which includes thermal memory as a property of the material, predicts the existence of thermal resonances as a consequence of the finite speed of heat transfer through the samples. In the case of samples with a low degree of structural arrangement, the theory predicts a low speed of heat transfer and resonant frequency within AFS of PAMS, which is elaborated in the dissertation. The significance of the research is reflected in the expansion of the frequency spectrum of PA testing, which will provide the possibility of more accurate and more precise determination of thermal parameters, as well as in opening of the possibility of detecting resonances of thermal memory in low-order materials. In addition, the modular character of the proposed measurement system opens the possibility of its connection with other measurement techniques and simultaneous determination of a larger number of material parameters, other than thermal ones. xi
true
Neke klase spektralno ograničenih grafova
Some classes of spectrally constrained graphs
Spektralna teorija grafova je grana matematike koja je nastala pedesetih godina pro²log veka i od tada se neprestano razvija. Njen zna£aj ogleda se u brojnim primenama, naro£ito u hemiji, zici, ra£unarstvu i drugim naukama. Grane matematike, kao ²to su linearna algebra i, posebno, teorija matrica imaju vaºnu ulogu u spektralnoj teoriji grafova. Postoje razli£ite matri£ne reprezentacije grafa. Najvi ²e su izu£avane matrica susedstva grafa i Laplasova (P.S. Laplace) matrica, a zatim i Zajdelova (J.J. Seidel) i takozvana nenegativna Laplasova matrica. Spektralna teorija grafova u su²tini uspostavlja vezu izme u strukturalnih osobina grafa i algebarskih osobina njegove matrice, odnosno razmatra o kojim se strukturalnim osobinama (kao ²to su povezanost, bipartitnost, regularnost i druge) mogu dobiti informacije na osnovu nekih svojstava sopstvenih vrednosti njegove matrice. Veliki broj dosada²njih rezultata iz ovog ²irokog polja istraºivanja moºe se na¢i u slede¢im monograjama: [20], [21], [23] i [58]. Disertacija sadrºi originalne rezultate dobijene u nekoliko podoblasti spektralne teorije grafova. Ti rezultati izloºeni su u tri celine glave, od kojih je svaka podeljena na poglavlja, a neka od njih na potpoglavlja. Na po£etku svake glave, u posebnom poglavlju, formulisan je problem koji se u toj glavi razmatra, kao i postoje¢i rezultati koji se odnose na zadati problem, a neophodni su za dalja razmatranja. U ostalim poglavljima predstavljeni su originalni rezultati, koji se nalaze i u radovima [3], [4], [47], [48], [49], [50], [51] i [52]. U prvoj glavi razmatra se druga sopstvena vrednost regularnih grafova. Postoji dosta rezultata o grafovima £ija je druga po veli£ini sopstvena vrednost ograni£ena odozgo nekom (relativno malom) konstantom. Posebno, druga sopstvena vrednost ima zna£ajnu ulogu u odre ivanju strukture regularnih grafova. Poznata je karakterizacija regularnih grafova koji imaju samo jednu pozitivnu sopstvenu vrednost (videti [20]), a razmatrani su i regularni grafovi sa osobinom 2 ≤ 1 (videti [64]). U okviru ove disertacije pro²iruju se rezultati koji se nalaze u radu [64], a predstavljaju se i neki op²ti rezultati koji se odnose na vezu odre enih spektralnih i strukturalnih osobina regularnih nebipartitnih grafova bez trouglova...
Spectral graph theory is a branch of mathematics that emerged more than sixty years ago, and since then has been continuously developing. Its importance is reected in many interesting and remarkable applications, esspecially in chemistry, physics, computer sciences and other. Other areas of mathematics, like linear algebra and matrix theory have an important role in spectral graph theory. There are many dierent matrix representations of a given graph. The ones that have been studied the most are the adjacency matrix and the Laplace matrix, but also the Seidel matrix and the so-called signless Laplace matrix. Basically, the spectral graph theory establishes the connection between some structrural properties of a graph and the algebraic properties of its matrix, and considers structural properties that can be described using the properties of the eigenvalues of its matrix. Systematized former results from this vast eld of algebraic graph theory can be found in the following monographs: [20], [21], [23] i [58]. This thesis contains original results obtained in several subelds of the spectral graph theory. Those results are presented within three chapters. Each chapter is divided into sections, and some sections into subsections. At the beginning of each chapter (in an appropriate sections), we formulate the problem considered within it, and present the existing results related to this problem, that are necessary for further considerations. All other sections contain only original results. Those results can also be found in the following papers: [3], [4], [47], [48], [49], [50], [51] and [52]. In the rst chapter we consider the second largest eigenvalue of a regular graph. There are many results concerning graphs whose second largest eigenvalue is upper bounded by some (relatively small) constant. The second largest eigenvalue plays an important role in determining the structure of regular graphs. There is a known characterization of regular graphs with only one positive eigenvalue (see [20]), and regular graphs with the property 2 ≤ 1 have also been considered (see [64]). Within this thesis we extend the results given in [64], and we also present some general results concerning the relations between some structural and spectral properties of regular triangle-free graphs...
false
Diferencijske sheme za rešavanje jednačine subdifuzije
The difference schemes for solving subdiffusion equation
U poslednje vreme poraslo je interesovanje za modeliranjem fizičkih i hemijskih procesa jednačinama u kojima se pojavljuju izvodi i integrali razlomljenog reda. Jedna takva jednačina je jednačina subdifuzije koja se dobija iz difuzione jednačine zamenom klasičnog izvoda po vremenskoj promenljivoj izvodom reda a; gde je 0 < a < 1: Predmet ove disertacije jeste početno-granični problem za jednačinu subdifuzije i njegova aproksimacija konačnim razlikama. Najpre je razmatran jednodimenzioni problem. Pokazana je egzistencija i jedinstvenost slabog rexenja. Dokazana je stabilnost i izvedena ocena brzina konvergencije implicitne i sheme sa težinom. Posebna pažnja je posvećena dvodimenzionom problemu subdifuzije, kako sa Laplasovim, tako i opštim diferencijalnim operatorom drugog reda. Pretpostavka je da njegovi koeficijenti zadovoljavaju standardne uslove eliptičnosti što garantuje postojanje rexeša u odgovarajućim prostorima tipa Soboljeva. U tom slučaju su, pored gore pomenutih, postavljene još i aditivna i faktorizovana shema. Ispitana je njihova stabilnost kao i brzina konvergencije u zavisnosti od glatkosti ulaznih podataka i generalisanog rešenja.
In recent years there has been increasing interest in modeling the physical and chemical processes with equations involving fractional derivatives and integrals. One of such equations is the subdiffusion equation which is obtained from the diffusion equation by replacing the classical first order time derivative by a fractional derivative of order a with 0 < a < 1: The subject of this dissertation is the initial-boundary value problem for the subdiffusion equation and its approximation by finite differences. At the beginning, the one-dimensional equation is observed. The existence and the uniqueness of weak solution is proved. The stability and the convergence rate estimates for implicite and the weighted scheme are obtained. The main focus is on two-dimensional subdiffusion problem with Laplace operator as well as problem with general second-order partial differential operator. It is assumed that the coefficients of the differential operator satisfy standard ellipticity conditions that guarantees existence of solution in appropriate spaces of Sobolev type. In that case, apart from above mensoned, we constructed the additive and the factorized difference schemes. We investigated their stability and convergence rate depending on the smoothness of the input data and of generalized solution.
false
Ispitivanje svojstava kompleksnih mreža sa diskretnom dinamikom
Analysis of properties of complex networks with discrete dynamics
Kompleksne mreže su se u toku poslednje dve decenije pokazale kao izuzetno koristan koncept u proučavanju karakteristika kompleksnih sistema. Potreb- no je da se prvo izdvoje pojedinačni elementi sistema i interakcije između njih predstave pomoću odgovarajuće mreže, a zatim se proučavanje organizacije sistema svodi na analizu strukture i dinamičkih procesa na mreži primenom odgovarajuće metodologije. Kako je raznovrsnost realnih sistema za koje su dostupni podaci koji omo- gućavaju uvid u strukturu mreže interakcija sve veća, potrebno je neprekid- no razvijati nove tehnike i teorijske modele koji mogu da objasne ponašanje specifičnih sistema. U ovoj disertaciji izučavane su kompleksne mreže sa diskretnom dinamikom, na osnovu podataka vezanih za socijalne sisteme koji su organizovani kroz kolektivne aktivnosti. Karakteristike mreža koje opi- suju ovakve sisteme nedovoljno su proučene. Jedan od glavnih razloga za to je dostupnost empirijskih podataka. U navedenim sistemima interakcije iz- među pojedinaca su direktne, zbog čega je znatno teže doći do podataka nego u slučaju socijalnih sistema gde pojedinci komuniciraju posredstvom nekog uređaja. Posebna pažnja posvećena je ispitivanju dinamike učestvovanja pojedinaca u kolektivnim aktivnostima. Na osnovu rezultata statističke analize empi- rijskih podataka, pokazano je da učestvovanje pojedinaca u kolektivnim ak- tivnostima nije nasumično i razmatrani su matematički modeli koji mogu da objasne aktivnost pojedinaca na događajima koja je izrazito heterogena. Utvrđeno je da se primenom generalizovanog binarnog Pojinog modela mogu uspešno reprodukovati posmatrani empirijski rezultati. Koristeći ansambl bipartitnih mreža sa zadatim parametrima koji ima maksimalnu entropiju, identifikovane su značajne veze u otežinjenoj mreži koje odgovaraju relevantnim socijalnim interakcijama. Kako bi se dobio uvid u evoluciju strukture mreže u vremenu, proučavane su promene lokalnih struk- turnih parametara nakon svakog učešća na događaju. Pokazalo se da članovi sistema tokom prvih pojavljivanja na kolektivnim događajima obrazuju nove veze sa susedima, a zatim, sa povećanjem broja učešća na događajima, interak- cije sa susedima i učvršćivanje postojećih zajednica postaju dominantni u poređenju sa formiranjem novih veza u mreži. U cilju ispitivanja uticaja koji pojedinačni događaji imaju na strukturu mreže, ovde je predložen pristup gde se događaji izdvajaju na nekoliko raz- ličitih načina i posmatraju se promene do kojih dolazi u strukturi mreže. Dobijeni rezultati pokazuju da su na događajima sa malim brojem učesnika karakteristične interakcije između članova sa jakim vezama, za razliku od velikih događaja, gde su uglavnom zastupljene interakcije između pojedinaca sa vezama koje se lako raskidaju.
In the last two decades, complex networks have been proven as very useful concept for examination of properties of complex systems. The first step within this framework is to extract individual elements of the system and to represent interactions between these elements in the form of complex network. After this step, the study of complex system organization is reduced to the analysis of structure and dynamical processes on network with the use of suitable methodology. The increase in a variety of real systems with available data, which enable insight into the structure of network of interactions, requires constant development of new techniques and theoretical models that could explain behavior of specific systems. In this thesis, we studied complex networks with discrete dynamics using data on event-based social systems. There has been very little previous research on properties of networks representing these systems. One of the main reasons is availability of data. In the given systems individuals interact face-to-face, wherefore it is more difficult to get the data, than in the case of social systems where individuals use some communication device to communicate with each other. Special attention was paid to examination of activity of individuals in group events. According to the results of statistical analysis of empirical data it has been shown that individuals do not attend events randomly. We analysed mathematical models that can explain member’s participation patterns on events which turned out to be strongly heterogeneous. It has been shown that generalized binary P´olya model can reproduce given empirical results successfully. Using bipartite networks ensemble with maximum entropy, we identified significant connections in weighted network that represent relevant social interactions. In order to get the insight into evolution of the network structure, we analyzed change of local structural parameters after each event attendance. It has been shown that members of the system establish new connections with neighbors during member’s early involvement in the group activities, while later, as number of attended events increase, the interactions with neighbors and strenghtening of existing communities become preferred in comparison to forming new connections in network. In order to analyse the influence that particular event has on network structure, we proposed an approach based on event removal according to different criteria and examination of resulting structural changes in network. The results showed that interactions between individuals with strong connections are dominant on events with small number of members, while during the large events typically individuals with weak connections that could be easily broken interact.
true
Akustičke karakteristike govora kao prediktor uspešnosti vokalne rehabilitacije osoba sa laringektomijom
Acoustic characteristics of speech as success predictor of vocal rehabilitation of persons with laryngectomy
Sprovedeno istraživanje je usmereno na otkrivanje prediktora uspešnosti rehabilitacije govora kod laringektomiranih bolesnika. Cilj istraživanja je da se ispita povezanost između akustičkih i perceptualnih parametara, kao i razumljivosti govora kod osoba sa laringektomijom, koji koriste različite oblike govorne komunikacije: ezofagealni govor, govor uz pomoć traheoezofagealne vokalne proteze i govor uz pomoć elektronske laringealne proteze. Krajnji cilj istraživanja bio je da se utvrdi koji je od ta tri oblika govorne komunikacije najefikasniji, odnosno najrazumljiviji u svakodnevnoj komunikaciji. Istraživanje je izvršeno na uzorku od 125 laringektomiranih bolesnika, oba pola, starosne dobi od 48 do 85 godina. Instrumenti koji su bili korišćeni u ovom istraživanju su: a) Kompjuterizovana laboratorija za akustičku analizu, “Kay Elemetrics“ korporacije, model 4300, kojom smo ispitivali akustičke parametre glasa i govora–parametre varijabilnosti frekvencije (Fo, To, PFR, Jita, Jitt, RAP, PPQ), parametre varijabilnosti intenziteta (ShdB, Shim, APQ, vAm), parametre procene šuma i tremora u glasu (NHR, VTI, SPI), parametar MPT– maksimalno vreme fonacije vokala, parametar WPM–broj izgovorenih reči u minuti, parametre spektralne analize govora–formantna struktura samoglasnika i suglasnika; b) GRBAS skala za perceptualnu procenu kvaliteta glasa; v) Test za procenu razumljivosti govora. U statističkoj obradi dobijenih podataka korišćene su: mere centralne tendencije (aritmetička sredina), mere varijabilnosti (standardna devijacija), jednofaktorska analiza varijanse, t-test za nezavisne uzorke, Spirmanov koeficijent korelacije i višestruka linearna regresija. Najbolje rezultate na testovima akustičke i perceptualne procene glasa, kao i razumljivosti govora ostvarili su pacijenti koji koriste traheoezofagealne vokalne proteze. Kod pacijenata sa ezofagealnim govorom, parametri procene hrapavosti i slabosti u glasu, GRBAS skale, pokazali su se kao najznačajniji prediktori razumljivosti govora (β=0,000; R²=0,715; F=26,10, p=0,000). Kod pacijenata koji koriste vokalne proteze, parametar procene napetosti u glasu, GRBAS skale, predstavlja najznačajniji prediktor razumljivosti govora (β=0,000; R²=0, 458; F=18,33, p=0,000).
Conducted research is directed to identify predictors of success in speech rehabilitation of laryngectomized patients. The aim of research is to examin connection between acoustic and perceptive parameters as well as speech intelligibility of laryngectomized persons that use different types of vocal communication: esophageal speech, speech with tracheoesophageal prosthesis and speech with electonic laringeal prosthesis. The final aim of this research was to determine which of the three mentioned vocal communication types is the most effective and the most intelligible in everyday communication. The research was done on the sample of 125 laryngectomized patients of both sexes, from 48 to 85 years old. Instruments that were used in this research are: a) Computer laboratory for acoustic analysis, of «Kay Elemetrics» Corporation model 4300 to examin acoustic parameters of voice and speech –parameters of frequency variation (Fo, To, PFR, Jita, Jitt, RAP, PPQ), parameters of intensity variation (ShdB, Shim, APQ, vAm), assessment parameters of voice breathiness and strain), MPT parameter - maximum phonation time of vocal, WPM – number of words pronounced per minute, speech spectral analysis parameters – formant structure of vowels and consonants; b) GRBAS scale for a perceptive evaluation of voice quality; c) Speech intelligibility test. In statistical data analysis we used: measures of central tendency (arithmetic mean), measures of variability (standard deviation), one factor analysis of variance, t-test for independent semples, Spearman's correlation coefficient and Multiple linear regression. Patients that use tracheoesophageal vocal prosthesis had the best achievement on acoustic and perceptual voice assesment tests as well as on intelligibility test. Parametar of voice roughness and asthenia of GRBAS scale were the most significant predictors of speech intelligibility of patients with esophageal speech (β=0,000; R²=0,715; F=26,10, p=0,000). Parameter of strain in voice of GRBAS scale was the most significant predictor of speech intelligibility of patients with vocal prosthesis (β=0,000; R²=0, 458; F=18,33, p=0,000). On the subsemple of patients with electrolarynx parameter of voice hoarsness of GRBAS scale was the most significant predictor of speech intelligibility (β=0,004; R²=0,399; F=7,46, p=0,000).
false
Experimental characterization of magneto-optical properties of Faraday crystal applied in magnetic field sensor
Eksperimentalna karakterizacija magneto-optičkih osobina Faradejevog kristala namenjenog za konstrukciju senzora magnetskog polja
Disertacija razmatra pripremu za proizvodnju i magneto optička svojstva Bizmut germanijum oksidnih kristala (B12GeO20) koje se ozračuju femtosekundnim laserskim impulsima zracima sa rastućom snagom. Analize izvršene nad ozračenim i neozračenim uzorcima kristala pokazale su značajne promene u transmitansi, transmisionom spektru, boji, Ramanovom spektru, rezultatima difrakcije X zraka, Verdeovoj konstanti, magneto-optičkim svojstvima i koeficijentu apsorpcije. Posle ozračivanja došlo je do porasta transmitivnosti i izostanka anizotropije u transmisionom spektru u odnosu na neozračene uzorke. Izmena boje uzoraka primetna je čak i golim okom. XRD merenja pokazala su strukturne promene indukovane laserskim zracima tako da je npr. strana uzorka neposredno ozračena laserom postala skoro potpuno amorfna dok se suprotna strana uzorka indeksira na B12GeO20 jedinjenje. Ozračivanje je izazvalo porast vrhova u Ramanovom spektru sa izuzetkom vrhova za tip E koji se nisu pojavili. Femtosekundni laserski impulsi mogu se iskoristiti da unaprede optička svojstva kristala.
The thesis explores preparation and magneto properties of bismuth germanium oxide single crystal (B12GeO20) which is irradiated by pulsed femtosecond laser of increasing power. Analyses performed on irradiated and un irradiated samples showed significant changes in transmittance, transmission spectra, sample color, Raman spectra (RS), X-ray diffraction (XRD) pattern, Verdet constant (VC), magneto-optical property and absorption coefficient. After irradiation, the transmission spectra values increased whereas anisotropy detected in the transmission spectra of unirradiated samples disappeared. The change of color caused by irradiation was noticeable to the naked eye. The XRD measurements confirmed structural changes induced by laser irradiation, i.e., the laser- beam-incident side of the sample became almost amorphous, whereas the side opposite to the incident can be indexed to the Bi12 Ge O20 compound. Irradiation caused increase of Raman spectra peaks with the exception of crystal peaks of type E, which disappeared. The femtosecond pulsed laser irradiation can be used to improve bismuth germanium oxide single crystal optical properties. Femtosecond pulsed laser irradiation can improve optical properties of Bi12 GeO20 single crystals.
true
Rješavanje nekih problema u nastavi primjenom metoda kombinatorne optimizacije
Solving some problems in teaching by using combinatorial optimization methods
U ovom radu se istražuju neki aktuelni problemi kombinatorne optimizacije...
In this work some actual combinatorial optimization problem are investigated...
false
Softverski sistem za daljinsko upravljanje i nadzor robota baziran na android operativnom sistemu i bežičnoj komunikaciji
Software system for remote control and monitoring of robots based on Android operating system and wireless communication.
Robotski sistemi su složeni sistemi čiji korisnici moraju biti dobro obučeni. Upravljanje se najčešće sprovodi blizu samog robota ili upotrebom žične konekcije. Današnje brze promene i napredak u razvoju tehnologije zahtevaju implementaciju novih tehnoloških rešenja u oblasti programiranja i upravljanja industrijskih robota i drugih robota specijalne namene. U savremenoj robotskoj industriji postoji stalna potreba za unapređenjem pristupačnosti i interakcije u radu sa robotskim sistemima, pa je sve veći značaj razvoja korisničkih interfejsa koji su intuitivni i laki za korišćenje. U skladu sa tim, prilikom projektovanja grafičkog korisničkog interfejsa za programiranje i upravljanje robota nije dovoljno napraviti fokus samo na obezbeđivanju detaljnih i kompletnih funkcionalnosti upravljanja, već se ergonomija upotrebe takođe treba uzeti u obzir. Pored toga, masivni i nedostupni roboti sve više zahtevaju upravljanje nezavisno od lokacije operatera. Resursi vezani za robotske laboratorije često nisu dostupni korisnicima koje tek treba obučiti za rad sa ovakvim sistemima, a rad sa robotskim sistemima često zahteva detaljno poznavanje softverskih paradigmi za programiranje robota. Zbog navedenih razloga studenti koji tek uče o osnovnim principima robotike uglavnom nisu u mogućnosti da se upoznaju sa programiranjem i upravljanjem kompleksnih robotskih sistema na praktičan način. U oblasti edukacije u robotici, bežična virtuelna okruženja, koja daju realnu sliku o radu robota, bi bila od velikog značaja. Posebno bi se izdvojio njihov značaj u slučaju primene takvog rešenja za edukaciju u oblasti upravljanja slabo dostupnih robota kao što su, na primer, uređaji za adaptaciju i trening pilota. Takođe, pored edukativne svrhe, upotreba lako dostupnih, široko rasprostranjenih i ekonomičnih uređaja bi bila jako isplativa. Istraživanje o generalnoj upotrebi i znanju studenata o tehnologiji pametnih telefona, koje je sprovedeno u okviru ove doktorske disertacije, potvrđuje da današnji brz razvoj tehnologije zahteva unapređenje metoda za učenje i edukaciju...
Robot control systems are complex systems whose users must be well trained to use them. Control process is mainly carried out near the robot or by using wired connections. Because of rapid advances in technology, it is of great importance to implement new technology solutions in training environments, in robotics and aerospace industry. In contemporary robotics industry there is a constant need to improve accessibility of robotic systems and human-robot interaction, so the significance of developing new user interfaces that are intuitive and easy to use is growing. Therefore, for designing robot control user interfaces it is not sufficient to simply focus on providing detailed and complete control capability, but ergonomy of use should also be taken into account. Besides, massive and unreachable robots require control that is independent of the user’s location. Resources in the robotics field are limited for a common user to access and demand deep knowledge of robot-specific programming paradigms. For these reasons, it is not often possible for inexperience students to become familiar with programming and control of complex robotic systems in a practical manner. Virtual wireless environments that provide realistic experience in robot control would be of great significance in the robotics education. It would be especially important in the case of application of such a solution for education in control of poorly accessible robots, such as devices for pilots’ training and flight simulations. Also, in addition to educational purposes, the use of easily accessible, widespread and cost-effective device would be very profitable. The evaluation of the general engineering students’ use and knowledge of smart phone technology, which was conducted within this dissertation, confirms that today's rapid development of technology requires improved methods for learning and education. There are a number of scientific and research papers and reports that describe software solutions that the majority of robot manufacturers and independent software vendors offer to developers and system integrators for design and simulation of manufacturing projects. A certain number of science and research papers report the implementation and the use of smart phones for robot monitoring and control. The remote monitoring of robot and machine performance is implemented by using camera, in textual form, or by using 3D animation, and robot programming is performed by using accelerometer, gyroscope or by filling out the text fields with desired values. Such programming methods are timeconsuming for setting up and generating the desired robot performance, that is especially pronounced when frequently use the same or a similar program. Also, these methods are not sufficiently intuitive, especially for education in robotics and students that only begin to learn about the basic elements of robotics...
false
Nejednakosti izoperimetrijskog tipa u prostorima analitičkih funkcija
Isoperimetric type inequalities in spaces of analytic functions
Rad se sastoji od tri glave. Prva glava sadrži dobro poznate činjenice o Hardijevim klasama harmonijskih, analitičkih i logaritamsko subharmonijskih funkcija u disku i njihove primene. Zatim kratko govorimo o harmonijskim i minimalnim površima i klasičnoj izoperimetrijskoj nejednakosti, kao i o rezultatima koji su povezani sa ovom nejednako šću. Jedan od najelegantnijih načina da se ustanovi izoperimetrijska nejednakost je preko Karlemanove nejednakosti za analitičke funkcije u disku. U drugoj glavi prezentujemo rezultate našeg skorijeg rada [29] koji se odnose na harmonijska preslikavanja diska na proizvoljnu Jordanovu površ. U ovoj glavi su klasični rezultati Karateodoria i Smirnova za konformna preslikavanja razmotreni za prethodni tip preslikavanja. Na kraju glave prethodne rezultate primenjujemo u cilju dokaza izoperimetrijske nejednakosti za Jordanove harmonijske površi omedjene rektificijabilnom krivom. U trećoj, prema [35], izvodimo dokaz jedne nejednakosti izoperimetrijskog tipa, slične Karlemanovoj, za analiti čke funkcije više promenjivih. Prva verzija ove nejednakosti je za analitičke funkcije u proizvoljnoj Reinhardtovoj oblasti, a druga se odnosi na funkcije koje pripadaju Hardijevim prostorima na polidisku.
This work consists of three chapters. The first one contains some well known facts about Hardy classes of harmonic, analytic, and logarithmically subharmonic functions in the unit disk, as well as their applications. Then we briefly talk about the harmonic and minimal surfaces, the classical isoperimetric inequality, and the more recent results related to this inequality. One of the most elegant way to establish the isoperimetric inequality is via Carleman’s inequality for analytic functions in disks. In the second chapter we present the results from our recent work [29] for harmonic mappings of a disc onto a Jordan surface. In this chapter we establish the versions of classical theorems of Carath´eodory and Smirnov for mappings of the previous type. At the end of the head we apply these results to prove the isoperimetric inequality for Jordan harmonic surfaces bounded by rectifiable curves. In the third chapter, according to the author paper [35], we prove an inequality of the isoperimetric type, similar to Carleman’s, for functions of several variables. The first version of this inequality is for analytic functions in a Reinhardt domain. The second one concerns the functions that belong to Hardy spaces in polydiscs.
false
Auto-dualni simplicijalni kompleksi, njihova generalizacija i primene u kombinatorici i geometriji
Self-dual simplicial complexes, their generalization and applications in combinatorics and geometry
U ovoj tezi se istražuje kombinatorna struktura auto-dualnih simplicijalnih kompleksa i njihovih generalizacija, r−neizbežnih simplicijalnih kompleksa...
This thesis is devoted to the study of the combinatorial structure of self-dual simplicial complexes and their generalizations, r−unavoidable simplicial complexes...
false
Quantitative movement analysis in the rehabilitation of neurological disorders using vision and wearable sensors
Kvantitativna analiza pokreta u rehabilitaciji neuroloških poremećaja korišćenjem vizuelnih i nosivih senzora.
Neuroloska oboljenja, kao sto su Parkinsonova bolest i slog, dovode do ozbiljnih motornih poremecaja, smanjuju kvalitet zivota pacijenata i mogu da uzrokuju smrt. Rana dijagnoza i adekvatno lecenje su krucijalni faktori za drzanje bolesti pod kontrolom, kako bi se omogucio normalan svakodnevni zivot pacijenata. Lecenje neurolo skih bolesti obicno ukljucuje rehabilitacionu terapiju i terapiju lekovima, koje se prilagodavaju u skladu sa stanjem pacijenta tokom vremena. Tradicionalne tehnike evaluacije u dijagnozi i monitoringu neuroloskih bolesti oslanjaju se na klinicke evaluacione alate, tacnije specijalno dizajnirane klinicke testove i skale. Medutim, iako su korisne i najcesce koriscene, klinicke skale su sklone subjektivnim ocenama i nepreciznoj interpretaciji performanse pacijenta...
Neurological disorders, such as Parkinson's disease (PD) and stroke, lead to serious motor disabilities, decrease the patients' quality of life and can cause the mortality. Early diagnosis and adequate disease treatment are thus crucial factors towards keeping the disease under control in order to enable the normal every-day life of patients. The treatment of neurological disorders usually includes the rehabilitation therapy and drug treatment, that are adapted based on the evaluation of the patient state over time. Conventional evaluation techniques for diagnosis and monitoring in neurological disorders rely on the clinical assessment tools i.e. specially designed clinical tests and scales. However, although benecial and commonly used, those scales are descriptive (qualitative), primarily intended to be carried out by a trained neurologist, and are prone to subjective rating and imprecise interpretation of patient's performance...
false
Functional norm regularization for margin-based ranking on temporal data
Primena funkcionalnih normi za regularizaciju rangiranja nad temporalnim podacima
Kvantikovanje osobina (karakteristika) od interesa je vazan problem u mnogim domenima, npr. utvrdivanje tezine bolesti kod pacijenata, ocena rizika investicije ili relevantnost vracenih rezultata pretrage. Medutim, osobine od interesa su cesto latentne i tesko se mogu izmeriti direktno, sto otezava dobijanje klasikacionih oznaka (labela) ili ciljeva za regresiju, koji su potrebni za ucenje prediktivnih modela iz merljivih karakteristika. U takvim slucajevima obicno je mnogo lakse pribaviti relativno poredenje dva slucaja, tj. proceniti koji od dva je intenzivniji (iz ugla karakteristike od interesa). Jedna klasa algoritama koji mogu uciti iz ovakvih informacija je SVM za rangiranje i on ce biti osnova ovde predlozenog pristupa...
Quantifying the properties of interest is an important problem in many domains, e.g., assessing the condition of a patient, estimating the risk of an investment or relevance of the search result. However, the properties of interest are often latent and hard to assess directly, making it dicult to obtain classication or regression labels, which are needed to learn a predictive models from observable features. In such cases, it is typically much easier to obtain relative comparison of two instances, i.e. to assess which one is more intense (with respect to the property of interest). One framework able to learn from such kind of supervised information is ranking SVM, and it will make a basis of our approach...
false
High frequency noise approximation and adaptive reduction in the ECG signals
Aproksimacija i adaptivno smanjenje visokofrekventnog šuma u EKG signalima
EKG signali su korišćeni još početkom prošlog veka. Uglavnom su se koristili za dijagnozu određenih grupa srčanih, arterijskih i plućnih bolesti. Stoga, obrada EKG signala nije nova tema za istraživanje. Mnogo metoda koje se odnose na glavne izazove koji su prisutni u postupcima za obradu i analizu EKG signala, su predstavljeni u literaturi. Jedan od izazova koji je i dalje predmet istraživanja je prisustvo visokofrekventnog šuma u snimanju EKG signala. Zbog širenja upotrebe “internet of things” (IOT) rešenja i razvijene telekomunikacijske infrastrukture, nove primene za snimanje EKG singala su predstavljene i integrisane u telemedicinskim platformama. Tako da primena EKG snimanja više nije ograničena na standardna klinička i holter snimanja. Post-event i loop snimanja su postala popularnija, jeftinija i pristupnija pacijentima. U cilju zamene klasičnog snimanja sa adhezivnim elektrodama, pojavilo se i nekoliko različitih metodologija koje koriste suve elektrode i prenosive uređaje. Pitanje koje se često javlja u radu sa snimljenim signalima ovih uređaja, je pouzdanosti snimljenih signala bez obzira na metodu snimanja. Najvažniji fakor koji određuje pouzdanost EKG signala je količina visokofrekventnog šuma prisutnog u signalu. Međutim, prisustvo šuma (početak i kraj), njegov odnos u signalu, priroda i boja su varijabilni kroz vreme u celom signalu, zato što taj šum je nestacionaran u EKG signalima. Ovo se posebno odnosi na dugoročno snimanje, pogotovo u loop event recorder i holter. Imajući to u vidu, ne postoji mera koja može da se primeni na celi signal za estimaciju celokupnog kvaliteta signala. Tako da treba da se razmatra estimacija ili aproksimacija nivoa šuma na translation-invariant način. Razvijanje metode za aproksimaciju nivoa šuma kroz vreme može da se koristi za estimaciju celokupnog kvaliteta singala. U ovom kontekst, predstavljena je estimacija nivoa šuma kroz vreme na translation-invariant način. Zbog nestacionarnog prisustva šuma, njegove jačine i boje, upotreba termina aproksimacija je tačnija od termina estimacija. Tako da je navedena metoda nazvana aproksimacija nivoa šuma ili samo aproksimacija šuma. Stationary wavlet transformacija (SWT) je korišćena metoda za nalaženje translation- invariant aproksimacije visokofrekventnog šuma. Ovo je ostvareno u formi referentnog signala izdvojenog kao estimacija kvaliteta signala u dva režima rada. Prvi režim je posle izbacivanja potencijalnih QRS kompleksa, koji mogu da se detektuju sa multi-resolutional XIII analizom na detalje SWT transformacije. Drugi režim je isti kao prvi samo što nisu izbačeni QRS kompleksi iz kalkulacije referentnog signala. Različite srčane frekvencije i srčani ritmovi su korišćeni za normalizaciju referentnog signala u cilju povećanja robustnosti i pouzdanosti ekstrahovanog referentnog signala. Odgovarajući talasi, izazvani električnom aktivnošću srca u ovim srčanim ritmovima i srčanim frekvencijama, pokazuju brze promene sa neregularnom morfologijom. Uzimajući to u obzir važno je izbegavanja pogrešne klasifikacije između tih ritmova i zašumljenih intervala. To povećava pouzdanosti i primenljivosti aproksimacije šuma bez obzira na ritam koji postoji u ulaznom EKG signalau. “Smoothing“ referentnog singala daje prikladan vodeći singal koji može da se koristi za određivanje praga i kvantizaciju amplitude referentnog singala na nekoliko nivoa šuma. Ovo je i bio motiv za predlog nove metode za korišćenje ekstrahovanog signala u platformi za uklanjanje šuma kao što su adaptivno filtriranje, banka filtara, i adaptivno Wavelet Wiener filtriranje. Uzimajući aritmije u obzir, metoda za izdvajanje referentnog vodećeg signala, povećava pouzdanost rezultata filtriranja bez obzira na ritam koji postoji u signalu. Ovo poboljšava celokupnu tačnost signala posle filtriranja. Tačnost signala nakon obrade meri koliko je rezultat posle filtriranja sličan tačnom ulaznom signal. Nedostatak tačnosti algoritma za filtriranja je glavni razlog zašto lekari daju prednost isključivanju svih filtara pre interpretacije EKG signala. Dodatno, izbacivanje potencijalnih QRS kompleksa će proizvesti referentni signal koji reflektuje šum u S-Q intervalu između dva otkucaja za redom. Zbog toga će amplituda referentnog signala opasti za vreme QRS kompleksa. Ovo smanjuje slabljenje QRS kompleksa i minimizira distorziju izazvanu filtriranjem koje je vođeno sa izdvojenim referentim signalom. Ovo je važno zbog toga što su ti kompleksi asocirani sa najvišim frekvencijama za razliku od ostalih talasa koji postoje u EKG signalima a uglavnom postaju neprepoznatljiv posle filtriranja EKG signala. Na kraju su analizirani metode za apoksimaciju i adaptivno filteriranje šuma u EKG signalima u realnim primenama. Nova arhitektura EKG uređaja je predstavljena za snimanje EKG signala u različitim režimima rada. Takođe su predstavljene integracija navedenih algoritma i njihova implementacija u procesu analize snimljenih signala sa suvim i vlažnim adhezvnim elektrodama. XIV
The ECG signal has been used since the beginning of the last century. Mainly, ECG signals are recorded to help in the diagnosis of certain group of heart, artery, and pulmonary diseases. Therefore, ECG signal processing is not a new theme for research. Plenty of methods that address the main challenges in ECG signal preprocessing and analysis pipeline were presented in the literature. One of the challenges that are still a subject of research is the high frequency noise presence in ECG signal recording. Due to the overspread of the internet of things (IOT) solutions and developed telecommunication infrastructure, new application of the ECG signal recording were presented and integrated in telemedicine platforms. Additionally, ECG signal recording applications are not limited anymore to the standard clinical and holter recording. Post event and loop recording has become more popular, cheaper, and available to patients. Moreover, several recording methodologies, using dry and wearable fashion, have appeared to replace the classic adhesive electrodes recording. The question that often arises when dealing with signals recorded with these devices is the reliability of recorded signals regardless of the used recording method. The most important factor that determines the reliability of ECG signals is the amount of high frequency noise present in the signal. However, the noise presence (offset and onset), ratio to the signal, nature, and color are variable over the whole recording. This is because noise is not stationary in ECG signal, especially in the long-term recording which is the case of loop event recorders and holters. Hence, there is no measure that could be applied on the whole signal to estimate the overall signal quality. Thus, noise level estimation or approximation should be considered in translation invariant manner in the time domain. Developing a method to approximate the noise level over time can be utilized to estimate the reliability of ECG signal in sliding local windows of any length rather than estimating the signal quality in overall. In this context, a noise level estimation over time in translation-invariant manner is introduced. Due to the non-stationarity of noise presence, strength, and color, the word approximation is more accurate than estimation. So, the proposed method is called noise level approximation or just noise approximation. The stationary wavelet transform (SWT) is used to find the translation-invariant approximation of the high frequency noise. This is accomplished in the form of reference signal extracted as an estimation of the signal quality in two modes. In the first mode, the X reference signal (noise approximation signal) is extracted after the exclusion of possible QRS complexes candidates which could be found using multi-resolutional analysis on the details of SWT transform. The second mode is the same as the first but without the exclusion of QRS complexes when the reference signal is calculated. In order to increase the robustness and the reliability of the extracted reference signal, different heart rates and rhythms were used in the normalization of the reference signal. The corresponding waves, caused by heart electrical activity in these rhythms and heart rates, show fast changes with irregular morphology. Taking such arrhythmias into consideration is important for avoiding misclassification between them and noise intervals, because of high ratio of amplitude changes in the ECG signal waves in these arrhythmias. Thus, noise level approximation is reliable and applicable regardless of the rhythm included in the input signal. Smoothing of the reference signal provides suitable guiding signal that could be thresholded and quantized to several noise levels corresponding to the amplitude of the reference signal. This was the motivation behinds the proposal of the extracted signal in noise suppression framework such as adaptive filtering, filters banks, or adaptive Wiener wavelet denoising. Considering arrhythmias when the building the reference signal increases its reliability to give good filtering results regardless of input signal rhythm. This enhances the whole signal fidelity. Where, signal fidelity after processing measures how close the result of digital processing represents the "true" input signal. The lack of reliability and fidelity of filtering algorithms are the main reasons why physicians prefer to disable all ECG signal filtering methods before interpretation. Additionally, the exclusion of possible QRS complexes candidates will be translated in a reference signal that reflects the noise in the S-Q interval between each consecutive two beats and therefore, reference signal will drop during the QRS complex. This reduces the QRS complexes attenuation and minimizes the distortion of filtering methodology guided by the built reference signal. This is important due to the fact that these complexes are associated with higher frequencies than other segments or waves in the ECG signal and they are usually distorted when ECG signals are filtering. Finally, application of noise level approximation and adaptive reduction are discussed in real case where an ECG recorder was presented to record ECG signals in different recording modes. Algorithms integration is discussed as well as the proposed device design and implementation. Noise level approximation and adaptive reduction are integrated in the processing pipeline of the recorded signals using both dry and wet electrodes. XI
true
O P-temenima nekih stabala
About P-vertices of some trees
Ova doktorska disertacija izučava P-temena i P-skupove nekih nesingularnih acikličnih matrica i takođe nekih singularnih acikličnih matrica. Ranije je pokazano da svaka singularna aciklična matrica reda n ima najviše n ¡ 2 P- temena. Takođe je pokazano da ovo ne važi za nesingularne aciklične matrice, i to konstrukcijom takvih matrica čiji pridruženi graf T ima n ¡ 1 ili n P-temena (ove matrica i dostižu maksimalnu veliqinu P-skupa među acikličnim nesingularnim matricama čiji je graf T). U ovoj tezi je data klasifikacija stabala za koje postoji nesingularna matrica čije je svako teme P-teme. Posebno, pokazano je da ova stabla moraju imati paran broj temena. Oba rezultata daju odgovor na otvoreno pitanje koje su postavili I.-J. Kim i B.L. Shader. Na kraju, izvršena je i klasifikacija stabala sa ograničenim P-skupom. Takođe je pokazano da su dvostruke zvezde DSn sa n temena primer stabala takvih da svaka nesingularna matrica A čiji je graf DSn ima najvixe n¡2 P-temena. Ovaj primer obezbeđuje pozitivan odgovor na još jedno pitanje koje su nedavno otvorili Kim i Shader. Nedavno je izvršena klasifikacija stabala za koje svaka pridružena aciklična matrica ima različite sopstvene vrednosti kada su dijagonalni elementi različiti. U ovom radu data je analiza maksimalnog broja različitih dijagonalnih elemenata i njihov položaj koji je neophodan da sačuva traženu višestrukost...
This thesis concerns P-vertices and P-set of non-singular acyclic matrices A and also singular acyclic matrices. It was shown that each singular matrix of order n has at most n ¡ 2 P-vertices. Also, it is shown that this does not hold for non-singular acyclic matrices by constructing non-singular acyclic matrices whose graphs are T having n¡1 ( or n) P-vertices. These matrices also achieve maximum size of P-set over non-singular acyclic matrices whose graphs are T. In this thesis, there is classi¯cation of the trees for which there is non- singular matrix where each vertex is P-vertex. In particular, it is shown that such trees have an even number of vertices. Both results provide answer to questions proposed by I.-J. Kim and B. L. Shader. In the end, related classi¯cations on non-singular trees with the size of a P-set bounded are addressed. Also, it is shown that double star DSn with n vertices, is an example of a tree such that, for each non-singular matrix A whose graph is DSn the number of P-vertices of A is less than n¡2. This example provides a positive answer to a question proposed recently by Kim and Shader. A recent classi¯cation of those trees for which each of associated acyclic matrices has distinct eigenvalues whenever the diagonal entries are distinct was established. Here is analyze of maximum number of distinct diagonal entries, and corresponding location, in order to preserve that multiplicity characterization. Recently, the multiplicities of eigenvalues of ©-binary tree was analyzed. This paper carry this discussion forward extending their results to larger family of trees, namely, the wide double path, a tree consisting of two paths that are joined by another path...
false
Prekidački reluktantni motori sa bipolarnim strujama
Switched reluctance motors with bipolar currents
Glavni ciljevi ovog rada su razvoj novog nelinearnog analitičkog modela prekidačkog reluktantnog motora (Switched Reluctance Motor-SRM) koji uzima u obzir efekte međusobne interakcije faza i razvoj nove topologije energetskog pretvarača baziranog na polumostnim strukturama sa smanjenim brojem prekidača za napajanje četvorofaznih prekidačkih reluktantnih motora. Zanemarenje efekata međusobne interakcije faza prekidačkog reluktantnog motora dovodi do greške u predikciji njegovih performansi. Postojeći modeli koji uzimaju u obzir međusobnu interakciju uglavnom se baziraju na rezultatima koji su dobijeni metodom konačnih elemenata (Finite Element Method-FEM) ili eksperimentalno. Zbog toga oni nisu podesni za dinamičke simulacije koje su neophodne u procesu projektovanja i optimizacije SRM-a. Predloženi model uzima u obzir sve inherentne nelinearnosti i efekte međusobne interakcije faza SRM-a pri čemu se samo jedan ulazni parametar ovog modela određuje korišćenjem metode konačnih elemenata ili eksperimentalno. Detaljna komparativna analiza simulacionih rezultata novog analitičkog i baznog modela, koji predstavlja početnu tačku njegovog razvoja, pružila je mogućnost da se utvrdi priroda greške koja se javlja kada su efekti međusobne interakcije faza zanemareni i time naglasi važnost njihovog modelovanja. Pored toga, korišćenje novog analitičkog modela je omogućilo i da se utvrdi uticaj magnetskog polariteta susednih faza motora na njegove performanse. Na ovaj način su određene sekvence magnetskog polariteta faza za koje se postižu najbolje performanse najčešće korišćenih konfiguracija SRM-a. Verifikacija simulacionih rezultata i zaključaka koji su dobijeni novim analitičkim modelom izvršena je korišćenjem FEM modela. Korišćenjem ovog modela detaljno je analiziran uticaj dubokog magnetskog zasićenja na performanse motora kao i uticaj međusobne interakcije faza na gubitke u gvožđu. iv Jedna od glavnih prepreka u razvoju pogona sa prekidačkim reluktantnim motorom i njegovoj široj primeni jeste nepostojanje standardnih energetskih pretvarača koji su dostupni na tržištu. To se pre svega odnosi na napajanje četvorofaznih prekidačkih reluktantnih motora. U svrhu prevazilaženja ovog problema predstavljena je nova topologija energetskog pretvarača koja se u potpunosti bazira na standardnim polumostnim strukturama čiji je broj manji u odnosu na broj koji je potreban za realizaciju postojećih rešenja. Pri tome, nova topologija pretvarača ispunjava sve zahteve za napajanje SRM-a. Pored ove prednosti, novom topologijom pretvarača se postižu i bolje performanse motora skoro na celom opsegu eksploatacione karakteristike u odnosu na slučaj kada se on napaja iz referentne topologije pretvarača. Naime, pokazuje se da međusobna interakcija faza u korelaciji sa inherentnim osobinama nove topologije pretvarača dovodi do pojava koje unapređuju performanse motora. Napajanje motora iz nove topologije pretvarača se označava kao bipolarno napajanje i njegova prednost u odnosu na unipolarno napajanje koje se vezuje za referentnu topologiju pretvarača je predmet detaljne analize koja je sprovedena korišćenjem FEM modela. Analogno ovom postupku sprovedena je i komparativna analiza performansi unipolarno i bipolarno napajanog trofaznog prekidačkog reluktantnog motora. Bipolarno napajanje se u ovom slučaju vezuje za postojeću topologiju pretvarača koja se takođe bazira na minimalnom broju polumostnih struktura. Međutim, pokazuje se da ova topologija pretvarača narušava performanse motora u odnosu na unipolarno napajanje. Eksperimentalna verifikacija novog analitičkog modela pokazuje da su simulacioni rezultati veoma bliski realnim. Takođe, pojave koje su posledica bipolarnog napajanja idu u prilog verifikaciji simulacionih rezultata i zaključaka o uticaju međusobne interakcije faza i inherentnih osobina pretvarača baziranih na polumostnim strukturama i smanjenim brojem prekidača na rad motora.
The main аim of this work is the development of a new nonlinear analytical model of the switched reluctance motor (SRM) which takes mutual coupling effect between phases into account, as well as the development of a new power converter topology for the supply of a four phase SRM, which is based on half-bridge structures with a reduced number of switches. If the mutual coupling effect between the SRM phases is neglected, prediction of SRM performance can be incorrect. Existing SRM models, which take mutual coupling into account, are usually based on finite element method (FEM) or experimental results. This makes them inconvenient for dynamic simulations which are necessary in the process of SRM design and optimization. The new model proposed and developed in this work takes all inherent nonlinearities and effects of mutual interaction into account, whereby only one input parameter of this model has to be determined by FEM or experimentally. A previously developed basic analytical model was taken as a starting point for the development of this new improved model. A detailed comparative analysis of the simulation results of these two models provides a better insight in the nature of the error caused by neglecting mutual interaction, at the same time emphasizing the importance of its modeling. Apart from this, employment of the new analytical model helped recognize the influence of the magnetic polarity of successive phases on the SRM performance and thereby determine which sequences of magnetic polarities ensure the optimal performance for the most commonly used SRM configurations. Verification of the simulation results obtained by the new analytical model and the accompanying conclusions is performed using a FEM model. This model helped investigate the influence of deep magnetic saturation on the motor performance, as well as the influence of mutual interaction on the core losses. One of the greatest obstacles in development and wider application of SRM drives is the absence of suitable standardized power converter structures on the market. This mainly refers to the four-phase SRMs and their supply. In order to overcome this problem, a new power converter topology is developed in this thesis, which is based on vii half-bridge structures with fewer semiconductor components compared to the existing converters. In addition, this new converter topology fulfills all requirements related to the SRM supply. The new converter topology also produces better motor performance over almost the entire torque-speed range as compared to the classic converter topology. It is shown that mutual phase interaction, combined with inherent properties of the new converter topology, produces certain phenomena which improve the SRM performance. The new converter topology is referred to as bipolar supply, whereas unipolar supply is inherently associated with the reference converter topology commonly used for SRM supply. A detailed analysis regarding the advantages of bipolar over unipolar supply is conducted using a FEM model. Comparative analysis of unipolar and bipolar supply of three-phase SRM is also conducted in the same manner. In this case, bipolar supply is achieved by employing an existing converter topology which is also based on minimizing the number of half-bridge structures and thus the number semiconductor switches. However, it was shown that this converter topology impairs SRM performance when compared to the case of unipolar motor supply. Experimental verification of the new analytical model exhibits a very good match between simulation and experimental results. Additionally, certain characteristic phenomena that arise as a consequence of bipolar motor supply provide further verification of simulation results. These analyses helped confirm the conclusions regarding the influence of mutual interaction and inherent features of converters based on half-bridge structures on the motor performance.
true
Ispitivanje elemenata galaktocentričnih orbita zvezda tankog diska iz sunčeve okoline varijacijom oblika potencijala galaksije
Examination of elements of galactocentric orbits of the thin disk stars from the solar neighbourhood by varying potential of the galaxy.
Cilj ove disertacije je da se odrede vrednosti dinamičkih kon- stanti. Ovaj cilj se ostvaruje proučavanjem uzoraka sastavljenih od odabranih zvezda koje se nalaze blizu Sunca. Odabir se vrši najpre kinematički pri čemu se koriste ekscentričnost orbite i amplituda oscilacija normalno na ravan Mlečnog puta. Predloženo je rešenje za problem određivanja eksce- ntričnosti. U narednom koraku se obrazuje uzorak velikog obima tako što se definišu gornje granice za ekscentričnost orbite i amplitudu oscilacija normalno na ravan Mlečnog puta. Ovaj uzorak se dalje razbija na poduzorke sa dva niza poduzoraka: u prvom nizu se smanjuje gornja granica ekscentričnosti, a u drugom gornja granica amplitude. Analizom ovog materijala dobijaju se vrednosti dinamičkih konstanata.
The goal of this dissertation is to determine values of local dynamical constants. This goal is achieved through examination of multiple samples of selected stars near the Sun. The selection is done by using planar and vertical eccentricities as sampling criteria. The solution for calculating eccentricities is given. In the next step a large sample of stars is selected by defining upper limits for eccentricities and vertical amplitude. Then nested subsamples are formed in two ways: in the first one upper eccentricity limit is subjected to decreasing, in the other one this is the case with upper amplitude of oscillations perpendicular to the plane. The values of the local dynamical constants are deduced by analysing this material.
true
Uticaj hrapavosti zidova kanala pri strujanju mešavine gasa i čestica u unutrašnjim turbulentnim strujanjima
Influence of wall roughness in particle-laden turbulent channel flows
U disertaciji se proučava uticaj hrapavosti zidova pri strujanju mešavine če- stica i gasa. Pri ovakvim strujanjima, sudar čestica sa hrapavim zidom je uvek trodimenzionalan (3D) usled prilaznih uglova čestica prema zidu kao i geomet- rije hrapavih zidova, dok su prema ranijim istraživanjima, modeli korišćeni za ovaj sudar u numeričkim simulacijama uglavnom dvodimenzionalan (2D) ili uprošćeni 3D. Kako bi se omogućio tačniji proračun strujanja mešavine čestica i gasa, cilj disertacije je razvoj modela za 3D sudar čestice i hrapavog zida. U di- sertaciji su analizirani sudari čestica sa izotropnim i anizotropnim hrapavim zidovima, putem determinističkih simulacija i izvedeni su odgovarajući stoha- stički modeli za te sudare, primenljivi u Ojler-Lagranž simulacijama strujanja mešavine čestica i gasa. Nalazi se da sa porastom prilaznog ugla čestica u od- nosu na makroskopski gladak zid, dolazi do porasta standardne devijacije bočnih uglova odbijanja čestica i u slučaju izotropne i anizotropne površine. Mehani- zam višestrukih sudara čestica sa zidom ostaje nepromenjen u odnosu na 2D model za sudar čestice i zida. Izvedeni stohastički modeli se porede se raspoloživim eksperimentalnim rezultatima i determinističkim simulacijama. Izvršene su LES simulacije strujanja čistog fluida (bez prisustva disperzne faze) u horizon- talnom kanalu pravougaonog poprečnog preseka koristeći različite SGS modele: model Smagorinskog, WALE i dinamički jedno-jednačinski model. Simulacije su izvedene u programskom paketu OpenFOAM. Uticaj zidova tog kanala, različite hrapavosti, na disperznu fazu, pri strujanju mešavine gasa i čestica istražen je putem LES-DPS simulacija. Međusobni sudari čestica se izračunavaju prime- nom modela mekih sfera. Kada se ovi sudari zanemare, što bi bilo opravdano pri strujanjima sa veoma malom koncentracijom disperzne faze, nalazi se da hra- pavost zida može biti značajan mehanizam za produkciju poprečnih fluktuacije čestica. Uzimanjem u obzir međusobnih sudara čestica, dolazi do porasta popreč- nih fluktuacija brzine čestica i pokazuje se da je u pogledu ovih fluktuacija uticaj međusobnog sudara čestica veći u odnosu na uticaj hrapavosti zidova ka- nala.
Influence of wall roughness in turbulent particle-laden wall confined flows is studied in this thesis. In such a flow configuration, particle-wall collision is always three-dimensional (3D) owing to the particle incident angles and rough wall geometry, while according to the previous investigations, models for this collision in numerical simulations are mostly two-dimensional (2D) or simplified 3D. In order to enable more accurate calculation of wall confined particle-laden flows, the goal of the thesis is development of the model for 3D particle-rough wall collision. Particles collision with isotropic and anisotropic rough walls are analyzed using deterministic simulations and stochastic models for those collisions are developed in Euler-Lagrangian framework. It is found that with increase of particle incident angle a standard deviation of particle transverse angles increases, for cases of isotropic and anisotropic rough wall surface. Mechanism of multiple particlerough wall collision remains unchanged comparing to the 2D model for particle-rough wall collision. Derived stochastic models are compared to the available experimental measurements and results of deterministic simulations. LES of pure fluid flow (without disperse phase) in a horizontal channel are carried out, using different SGS models: Smagorinsky, WALE and dynamic one-equation model. Simulations are carried out in software package OpenFOAM. Also, LES-DPS is performed in turbulent particleladen flow in horizontal channel in order to estimate influence of walls roughness on disperse phase. Inter-particle collisions are calculated using soft-sphere model. When these collisions are neglected, which is justified at very diluted flows, it is shown that wall roughness can be significant mechanism for production of particle spanwise velocity fluctuations. When inter-particle collisions are accounted for, these fluctuations increase and the influence of inter-particle collisions on particle spanwise velocity fluctuations becomes larger than the influence of channel wall roughness.
false
Algoritam za određivanje fazne razlike sinusnih signala niskih učestanosti
The algorithm for phase difference determination of low frequency sinusoidal signals
Pouzdano merenje fazne razlike između električnih signala neophodno je u mnogim mernim sistemima kao što su: merenje aktivne i reaktivne električne snage i energije, faktora snage, harmonika u električnom naponu i struji, određivanje kompleksnog odnosa izlaznog i ulaznog signala mernih pretvarača (delitelji napona i strujni šantovi). O značaju merenja faznog ugla najbolje govori činjenica da su mogućnosti merenja i etaloniranja koje se kraće označavaju CMCs (engl. Calibration and Measurement Capabilities) faznog ugla, nacionalnih metroloških instituta pojedinih država, objavljene na internet stranici Međunarodnog biroa za tegove i mere (fr. Bureau International des Poids et Measures - BIPM) http://kcdb.bipm.org, pošto ih je prethodno odobrila regionalna metrološka organizacija EURAMET (engl. European Association of National Metrology Institutes). Danas se fazna razlika između dva kontinualna sinusna signala najčešće određuje koristeći uzorke signala koji se obrađuju nekim od algoritama za određivanje fazne razlike. U NMI (engl. National Metrological Institutes) se za uzorkovanje signala uobičajeno koriste DSV (engl. digital sampling voltmeters) koji obavljaju vremensku i amplitudnu diskretizaciju signala. Uz rezultate merenja obavezno se daje i procena merne nesigurnosti, u skladu sa međunarodnim dokumentima. Predmet istraživanja disertacije je algoritam za određivanje fazne razlike iz uzoraka dva sinusna signala. Na osnovu poznatih algoritama UQDE (engl. unbiased quadrature delay estimator) i QDE (engl. quadrature delay estimator) razvijen je i testiran novi algoritam MSAL (engl. modified simple algorithm) za određivanje fazne razlike iz uzoraka dva sinusna signala. U novom algoritmu, korišćenjem dva postojeća signala uvode se dva nova signala fazno pomerena u odnosu na postojeće signale. Testira se hipoteza da je izborom faznog kašnjenja moguće dobiti vrednost fazne razlike između dva sinusna signala koristeći po jedan uzorak svakog signala (algoritam nazvan SAL), a v takođe i da se primenom nešto složenije formule na veći broj uzoraka signala, može odrediti vrednost fazne razlike između dva sinusna signala sa znatno manjom mernom nesigurnošću (algoritam nazvan MSAL). Dat je matematički model, rezultati simulacija i rezultati primene novih algoritama na uzorke realnih signala. U literaturi postoji mnogo algoritama za određivanje fazne razlike između dva sinusna električna signala, a u disertaciji je dat pregled samo onih algoritama koji su od interesa za temu disertacije. Prvo je razmatran algoritam UQDE na kojem su zasnovani novi algoritmi predloženi u disertaciji. Dat je i kratak pregled algoritama 3PSF, 4PSF i 7PSF (engl. three, four and seven parameter sinus fitting) koji su primenjeni na uzorke realnih signala u cilju poređenja rezultata određivanja fazne razlike sa rezultatima dobijenim koristeći nove algoritme, kao i kratak pregled algoritama za određivanje fazne razlike zasnovanih na Furijeovoj transformaciji, maksimalnoj verodostojmosti i linearnoj predikciji. Upoređeni su rezultati dobijeni primenom novih algoritama za određivanje fazne razlike iz uzoraka realnih električnih signala sa rezultatima dobijenim primenom drugih postojećih algoritama na iste uzorke signala. To je omogućilo da se verifikuju i uporede karakteristike postojećih i novih algoritama za određivanje fazne razlike i dokažu pretpostavljene hipoteze istraživanja. U disertaciji su algoritmi primenjeni na uzorke realnih električnih signala niskih učestanosti jer je na raspolaganju bila takva oprema za uzorkovanje signala. Uz rezultate za faznu razliku date su i procene merne nesigurnosti tih rezultata u skladu sa Uputstvom za izražavanje nesigurnosti u merenju (engl. „Guide to the expression of uncertainty in measurement”- GUM). U disertaciji je dat i kratak pregled metoda merenja koje se danas koriste za etaloniranje fazne razlike električnih signala niskih učestanosti u nacionalnim metrološkim institutima.
The reliable measurement of phase difference between two signals is essential in many measurement systems such as: measurement of active and reactive electric power and energy, power factor, voltage and current harmonics, the determination of the complex ratios between the output and input signals of transducers (voltage dividers and current shunts). The importance of phase angle mesurements is the best illustrated by fact that calibration and measurement capabilities (CMCs) of the national metrological institutes of individual countries are published at website of International Bureau of Weights and Measures (Bureau International des Poids et Measures - BIPM), http://kcdb.bipm.org, as previously approved by the regional metrological organization EURAMET (European Association of National Metrology Institutes). Nowadays, the phase difference between two continuous sinusoidal signals is usually determined by means of discrete samples of signals processed by some of the algorithms for the phase difference estimation. DSV (digital sampling voltmeters) are usually used for the time and amplitude discretization of signals in NMI (National Metrological Institutes). The results of measurements must be accompanied with the estimates of uncertainties according to the international recommendations. The subject of the dissertation research is algorithm for phase difference determination using samples of two sinusoidal signals. On the basis of known algorithms UQDE (unbiased quadrature delay estimator) and QDE (quadrature delay estimator) the new algorithm MSAL (modified simple algorithm) for phase difference determination from the samples of two sinusoidal signals is developed and tested. Two new signals phase shifted to the existing signals are introduced in new algorithm. The hypothesis that it is possible to get the phase difference between two sinusoidal signals vii by choice of phase delay between two samples, one sample of each signal (algorithm called SAL) is tested, and also by applying the more complex formula to a larger number of samples phase difference between two sinusoidal signals can be determined with lower uncertainty (algorithm called MSAL). The mathematical model, results of simulations and results of applications of new algorithms on real samples, are given. There are many algorithms for estimation of the phase difference between two electrical sinusoidal signals, in literature. The short overview of those algorithms that are of interest for dissertation, is provided. The algorithm UQDE is first considered becouse new algorithms suggested in dissertation are based on it. Also, 3PSF, 4PSF and 7PSF (three, four and seven parameter sinus fitting) algorithms are considered, becouse those algorithms are applied on samples of real signals for phase difference determination in order to compare obtained results with results for the phase difference determined by new algorithms. The algorithms for the phase difference determination based on Fourier transform, maximum likelihood and linear prediction, are also short rewieved. The results obtained using a new algorithm MSAL for determination of the phase difference of samples of real electrical signals are compared with those obtained by other existing algorithms, applied on the same signals samples. That enabled to verify and compare the characteristics of the existing and proposed algorithm for determination of the phase difference and demonstrate assumed research hypotheses. The algorithms are applied to a low frequency real electrical signals becouse such equipment for signals sampling were available. Together with results of phase difference determination, the estimation of uncertainies of those results is given according to „Guide to the expression of uncertainty in measurement” – GUM. The dissertation provides also a brief overview of measurement methods that are nowadays used for calibration of the phase difference of electrical signals in the national metrological institutes (MNIs).
true
Fiksne tačke preslikavanja na prostorima sa nedeterminističkim rastojanjem
Fixed points of mappings on spaces with nondeterministic distances
Osnovu sadrˇzaja teze ˇcine uslovi postojanja nepokretnih taˇcaka preslikavanja definisanih na prostorima sa nedeterministiˇckim rastojanjem i sastoji se iz ˇcetiri dela. U prvoj glavi izloˇzen je pregled osnovnih pojmova i rezultata, na metriˇckim prostorima, koji su inicirali istraˇzivanja izloˇzena u ovoj tezi. U drugoj glavi izloˇzene su definicije i pored¯enja prostora na kojima ´ce presli- kavanja biti razmatrana, kao i osnovni stavovi koji karakteriˇsu ove prostore i nji- hove podskupove. Razmatraju se pitanja: metriˇcke i topoloˇske strukture prostora; neprekidnosti nedeterministiˇckih metrika; ograniˇcenosti (semi-ograniˇcenosti) i stroge ograniˇcenosti skupova u prostorima sa nedeterministiˇckim rastojanjem; navode se primeri prostora sa nedeterministiˇckim rastojanjem i osobine metrika na tim pros- torima. Posebno, paˇznja je posve´cena razjaˇsnjavanju odnosa pojmova kompletnosti i G-kompletnosti i njihovom znaˇcaju za rezultate u teoriji nepokretne taˇcke. U tre´coj glavi razmatraju se pitanja uopˇstenja pojma komutativnosti pojmovima R-slabe komutativnosti, kompatibilnosti i slabe kompatibilnosti funkcija. U ovoj glavi izloˇzena su tri rezultata, jedan autorski i dva koautorska. Prvi rezultat se odnosi na postojanje zajedniˇcke nepokretne taˇcke dva preslikavanja definisana na L -fazi metriˇckim prostorima koja ispunjavaju nelinearan kontraktivan uslov. Drugi rezultat se odnosi na postojanje zajedniˇcke nepokretne taˇcke dva neekspanzivna preslikavanja definisana na L -fazi metriˇckim prostorima sa konveksnom strukturom. Tre´ci, autorski rezultat odnosi se na postojanje nepokretne taˇcke preslikavanja defin- isanog na verovatnosnim Mengerovim prostorima koje ispunjava nelinearan kvazi- kontraktivan uslov odred¯en alterniraju´com Φ-funkcijom. U poslednjoj, ˇcetvrtoj glavi, prikazana su tri rezultata, jedan autorski i dva koautorska, o postojanju zajedniˇcke nepokretne taˇcke preslikavanja definisanih na verovatnosnim Mengerovim prostorima. Prvi od njih odnosi se na postojanje za- jedniˇcke nepokretne taˇcke familije preslikavanja koja su R-slabo komutativna sa nekim fiksiranim preslikavanjem, pri nelinearnom kontraktivnom uslovu. Drugi rezultat odnosi se na postojanje zajedniˇcke nepokretne taˇcke dva para preslikavanja, pri čemu je jedan od njih kompatibilan, a drugi od njih slabo kompatibilan, pri nelinearnom kontraktivnom uslovu. Ovaj rezultat, u izvesnom smislu, uopštava se u trećem, autorskom rezultatu ove glave, zamenom nelinearnog kontraktivnog uslova sa nelinearnim kontraktivnim uslovom određenim alternirajućom Φ-funkcijom.
Basic contents of the thesis are conditions of existence of fixed points of mappings defined on spaces with nondeterministic distances and it consists of four parts. In the first chapter an overview of basic definitions and results on metric spaces that have initiated the research in this thesis is given. The second chapter presents definitions and comparison of spaces on which the mappings will be considered, as well as basic propositions that characterize these spaces and their subsets. The following matters are considered: metrical and topo- logical structures of spaces; continuity of nondeterministic metrics; boundness (semi- boundness) and strong boundness of sets in spaces with nondeterministic distances; examples of spaces with nondeterministic distances and properties of metrics on these spaces are given. Special attention is paid to clarification of the relation be- tween concepts of completeness and G-completeness and their significance for results in fixed point theory. The third chapter considers generalizations of commutativity with terms of R- weakly commutativity, compatibility and weak compatibility of mappings. This chapter presents three results, one produced independently and two in cooperation. The first result relates to existence of common fixed point of two mappings defined on L -fuzzy metric spaces that satisfy nonlinear contractive condition. The second result relates to existence of common fixed point of two nonexpansive mappings defined on L -fuzzy metric spaces with convex structure. The third, independently produced result relates to existence of fixed point of mapping defined on Probabilistic Menger spaces that satisfies nonlinear quasi-contractive condition determined by alternating Φ-function. The last, fourth chapter considers three results, one produced independently and two in cooperation, on existence of common fixed points of mappings defined on Probabilistic Menger spaces. The first result relates to existence of common fixed point of family of mappings which are R-weakly commutative with a fixed mapping with nonlinear contractive condition. The second result relates to existence of common fixed point of two pairs of mappings, with one pair being compatible and the other weakly compatible, with nonlinear contractive condition. In a certain sense this result is generalized in the third, independently produced result by replacement of nonlinear contractive condition with nonlinear contractive condition determined by alternating Φ-function.
true
Upravljanje temperaturnom raspodelom u kotlovima termoenergetskih sistema na bazi extremum seeking strategije
Thermal power plant boiler temperature distribution control based on extremum seeking strategy.
Termoelektrane predstavljaju vec´inske proizvod¯acˇe elektricˇne energije u svetu, što za posledicu ima stalne pokušaje unapred¯enja u vidu smanjenja štetnih uticaja na životnu sredinu uz pove´canje efikasnosti, raspoloživosti i prihoda. Pomenuti efekti su u direktnoj vezi sa procesom sagorevanja, zbog ˇcega su upravljanje i optimizacija ovog procesa od suštinskog znaˇcaja. Ove potrebe su ˇcesto ograniˇcene mogu´cnostima nadgledanja procesa sagorevanja. Obiˇcno se za sagledavanje ovog procesa analizira neko od dostupnih merenja dobijenih iz hemijskih analizatora sadržaja ugljen-monoksida ili kiseonika u dimnim gasovima, što je korisno za ocenu kvaliteta sagorevanja fosilnih goriva. Med¯utim, ovo su globalni pokazatelji i ne govore ništa o lokalnoj raspodeli temperatura unutar ložišta...
Thermal power plants are the main producers of electrical energy in the world, and therefore there are constant attempts for improvement in terms of reducing harmful effects on the surroundings, while at the same time increasing eciency, availability and incomes. Aforementioned eects are directly related to the combustion process, which gives great importance to control and optimization of this process. These requirements are usually constrained by the possibilities of combustion process monitoring. This is mainly done by performing the acquisition of available measurements obtained from the chemical analysis of carbon dioxide and oxygen content in the flue gases, which is useful for the evaluation of fossil fuels combustion. However, these are only global indicators and they provide no information about the local temperature distribution inside the furnace...
false
Modelovanje radio-evolucije ostataka supernovih na osnovu hidrodinamičkih simulacija i nelinearnog difuznog ubrzavanja čestica
Modeling the radio-evolution of supernova remnants by using hydrodynamic simulations and non-linear diffusive shock acceleration
Eksplozije supernovih (SN) obogacuju Univerzum razlicitim teskim elementima. Ovi elementi su gradivni blokovi sveta oko nas i nas samih. Ostaci supernovih (OSN) su izvanredni astronomski objekti koji su takode od velikog naucnog znacaja, jer pruzaju uvid u mehanizme eksplozija supernovih i predstavljaju vazne izvore galakti ckih kosmickih zraka. Radio-posmatranja su medu najstarijim sredstvima za proucavanje ovih objekata. Radio-sjaj i spektar, posebno mladih OSN, predstavljaju rezultat aktivnog ubrzavanja elektrona na udarnim talasima. U ovoj doktorskoj disertaciji, analizira se radio-evolucija OSN primenom trodimenzionalnih hidrodinami ckih simulacija i nelinearnog difuznog ubrzavanja cestica. Hidrodinamicke simulacije, razvijane i prilagodavane u ovoj disertaciji, omogucavaju nam eksplicitno racunanje modikacije udarnih talasa od strane kosmickih zraka. Takode, ukljucujemo konzistentan numericki tretman pojacanja magnetnog polja usled rezonantnih i nerezonantnih nestabilnosti izazvanih kosmickim zracenjem. Modelovali smo radio-evoluciju najmladeg poznatog Galaktickog OSN G1.9+0.3 i zakljucili da je rastuca radio-emisija cesta pojava kod jako mladih OSN. Nas model nam je omogucio da izvedemo vazne zakljucke o sadasnjem i predvidanja o buducem ponasanju radio-emisije ovog OSN. Takode smo primenom simulacija razvili i opsti model radio-evolucije OSN, za sirok spektar relevantnih zickih parametara, kao sto su gustina okolne sredine, energija eksplozije supernove, ekasnost ubrzavanja i pojacanja magnetnog polja. Pouzdanost nasih radio-evolutivnih traka potvrdena je na posmatrackom uzorku, koji se sastoji od Galaktickih i vangalaktickih OSN...
Supernova (SN) explosions disperse the dierent heavy elements across the Universe. These elements are the building blocks which make up the world around and inside us. Supernova remnants (SNRs) are extraordinary astronomical objects that are also of high scientic interest, because they provide insights into aforementioned supernova explosion mechanisms, and because they are important sources of Galactic cosmic rays (CRs). Radio observations are among the oldest means to study these objects. The radio luminosity and spectra of SNRs, especially young ones, requires active acceleration of electrons by the SNR shocks. In this doctoral dissertation, radio evolution of SNRs is investigated by using three-dimensional hydrodynamic modelling and non-linear diusive shock acceleration of CRs in SNRs. Hydrodynamic simulations, developed and adopted in this dissertation, allow us to explicitly account for the shock modication by CRs. We also include consistent numerical treatment of magnetic eld amplication (MFA) due to CR resonant and non-resonant streaming instabilities. We modelled the peculiar nature of radio evolution of the youngest known Galactic SNR G1.9+0.3 and concluded that increasing radio emission is a common occurrence among very young SNRs. Our model enabled us to make important conclusions about the present and predictions about the future properties of radio emission from this SNR. We also developed more general model of the radio evolution of SNRs, by performing simulations for wide range of the relevant physical parameters, such as the ambient density, the supernova explosion energy, the acceleration eciency and the MFA eciency...
false
Uticaj sveobuhvatne kardijalne rehabilitacije na dijastolnu disfunkciju i funkcionalni status pacijenata lečenih perkutanom koronarnom intervencijom nakon akutnog koronarnog događaja
The impact of comprehensive cardiac rehabilitation on diastolic dysfunction and the functional status of patients treated with percutaneous coronary intervention after acute coronary event
Uvod: U savremenom svetu koronarna arterijska bolest srca (KABS) je vodeći uzrok obolevanja i umiranja, a akutni koronarni sindrom (AKS) je jedna od njenih najčešćih i najopasnijih kliničkih manifestacija. Dijastolna disfunckija leve komore često prati KABS i mogući je doprinosni faktor za loš klinički tok i ishod. Postavlja se pitanje u kom obimu je dijastolna disfunkcija leve komore udružena sa koronarnom arterijskom bolešću i da li savremeni programi ambulantne sveobuhvatne kardijalne rehabilitacije (ASKR) imaju uticaja na bolju prognozu ove grupe kardioloških bolesnika. Uprkos širokoj primeni revaskularizacionih procedura u svakodnevnoj kardiološkoj praksi i brojnih studija koje su ukazale na pozitivne efekte programa SKR na funkcionalni status pacijenata nakon AKS, još uvek postoji mnogo kontroverzi o efektima fizičkog treninga, na srčanu funkciju i poboljšanje funkcionalnog kapaciteta kod pacijenata sa KABS i pridruženom dijastolnom disfunkcijom. Cilj istraživanja: bio je da ispita uticaj superviziranih vežbi fizičkim opterećenjem (VFO) u okviru programa ambulantne sveobuhvatne kardijalne rehabilitacije (ASKR), na dijastolnu disfunkciju leve komore (DDLK) i funkcionalni status pacijenata (FS), nakon AKS, rešenog perkutanom koronarnom intervencijom (PCI), kao i da li, s druge strane, prisustvo i stepen dijastolne disfunkcije na početku istraživanja, utiče na funkcionalni status i pojavu neželjenih kardijalnih događaja, kod ove grupe pacijenata u okviru programa ASKR i van njega Materijal i metode: Istraživanjem je bilo obuhvaćeno ukupno 85 ispitanika, oba pola, starosti od 18-65 godina, koji su tokom indeksne hospitalizacije lečeni kao klinički dokazani AKS (APNS; NSTEMI; STEMI) i kod kojih je urađena neka od interventnih koronarnih procedura (pPCI; PCI; PTCA). Nakon 4 nedelje od otpusta sa hospitalizacije, zbog NSTEMI ili APNS, odnosno nakon 6 nedelja od otpusta sa hospitalizacije zbog STEMI, pacijenti sa EFLK ≥ 45%, bez značajnih valvularnih i drugih mana i sa nekim od poremećaja dijastolne funkcije, bili su kandidati za učešće u istraživanju. Svi ispitanici su podvrgavani „ulaznom“ehokardiografskom pregledu (EHO) u cilju procene sistolne funkcije i stepena dijastolne disfunkcije leve komore, kao i „ulaznom“ spiroergometrijskom testu (CPET) u cilju procene funkcionalnog statusa, na osnovu kojeg je vršena preskripcija vežbi fizičkim opterećenjem (VFO) u okviru programa ASKR. Program ASKR odvijao se u ukupnom trajanju od 12 nedelja, odnosno 36 pojedinačnih sesija VFO, 3 puta nedeljno u trajanju od po 30 minuta. Kontrolna grupa obuhvatila je grupu pacijenata koja nije živela u blizini IKVBV i koja nije bila u mogućnosti da dolazi redovno na VFO u sklopu ASKR. Njima je bilo pušteno na volju da na osnovu urađenog EHO pregleda i CPET, određuju stepen VFO po sopstvenom nahođenju, uz primenu optimalnog medikamentnog lečenja i ostalih mera sekundarne prevencije. Nakon 3 meseca obe grupe pacijenata bile su podrvrgnute novom –“izlaznom” ehokardiografskom i CPET pregledu u cilju komparacije sa rezultatima na početku istraživanja. Rezultati:Istraživanje je pokazalo da nakon 3 meseca superviziranog treninga VFO, u okviru programa ASKR, kod bolesnika nakon AKS, lečenih perkutanom koronarnom intervencijom, dolazi do poboljšanja stepena dijastolne disfunkcije leve komore, naročito kroz promene vrednosti ehokardiografskih parametara e’l i E/e’ l. U kontrolnoj grupi e’l se smanjio za (0,003 ± 0,025), a u osnovnoj se povećao za (0,011 ± 0,021). U kontrolnoj grupi e’l se nije značajno promenio (p = 0,515), a u osnovnoj grupi se značajno povećao (p < 0,0005). Na početku istraživanja u osnovnoj grupi e’l je bio (0,097 ± 0,027 m/sec), a na kraju (0,108 ± 0,031 m/sec). E/e’l se nije značajno promenio u kontrolnoj grupi (p = 0,226), a u osnovnoj grupi se značajno smanjio (p = 0,002). Na početku istraživanja u osnovnoj grupi E/e’l je bio (8,02 ± 2,98), a na kraju (6,97 ± 2,17). Takođe je utvrđeno da nakon 3 meseca superviziranog treninga u okviru programa ASKR, dolazi do poboljšanja funkcionalnog kapaciteta pacijenata sa KABS i dijastolnom disfunkcijom leve komore, kroz povećanje CPET parametara: peak VO2, VO2 predict i METs. U kontrolnoj grupi peak VO2 se smanjio za (1,79 ± 3,84), a u osnovnoj se povećao za (1,67 ± 4,29). U kontrolnoj grupi peak VO2 se značajno smanjio (p = 0,018), a u osnovnoj grupi se značajno povećao (p = 0,005).Na početku istraživanja u kontrolnoj grupi srednja vrednost peak VO2 iznosila je (23,01 ± 3,99 ml/kgTT/min), a u osnovnoj grupi je iznosila (23,15 ± 4,99 ml/kgTT/min). Na kraju istraživanja u osnovnoj grupi srednja vrednost peak VO2 iznosila (24,82 ± 5,77 ml/kgTT/min), dok je kod kontrolne grupe iznosila (21,21 ± 4,05 ml/kgTT/min). U kontrolnoj grupi ppVO2(%) se smanjio za (5,28 ± 14,39), a u ispitivanoj se povećao za (7,16 ± 18,51). U kontrolnoj grupi ppVO2(%) se nije značajno promenio (p = 0,058), dok se u osnovnoj grupi statistički značajno povećao (p = 0,005). Razlika srednjih vrednosti promena METs između osnovne i kontrolne grupe je bila statistički značajna (p < 0,0005). U kontrolnoj grupi METs se smanjio za (0,55 ± 1,12), a u osnovnoj se povećao za (0,58 ± 1,12). U kontrolnoj grupi METs se značajno smanjio (p = 0,013), a u osnovnoj grupi se značajno povećao (p < 0,0005). Zaključak: Program ambulantne sveobuhvatne kardijalne rehabilitacije, kod bolesnika nakon akutnog koronarnog sindroma, lečenih perkutanom koronarnom intervencijom, utiče na poboljšanje faktora rizika kardiovaskularnih bolesti, značajno utiče na poboljšanje stepena dijastolne disfunkcije leve komore i na poboljšanje funkcionalnog statusa pacijenata, u odnosu na početak istraživanja.
Background: In modern world, coronary artery disease (CAD) is one of the leading cause of morbidity and mortality, and acute coronary syndrome (ACS) is one of its most common and most dangerous clinical manifestations. Left ventricle diastolic dysfunction (LVDD) is often associated with CAD and is possibly a contributing factor for poor clinical course and outcome. The question arises as to the extent to which the left ventricle diastolic dysfunction is associated with CAD and whether modern ambulatory comprehensive cardiac rehabilitation programs (ACCR) have an impact on a better prognosis of this group of cardiac patients. Despite the wide use of revascularization procedures in everyday clinical practice and numerous trials that have indicated the positive effects of the ACCR on the functional status of patients after ACS, there is still much controversy about the effects of physical training, and its impact on cardiac function and the improvement of functional capacity in patients with CAD and associated diastolic dysfunction. Objective: to examine the influence of supervised physical training programe as a part of ambulatory comprehensive cardiac rehabilitation (ACCR), on diastolic dysfunction and the functional capacity of patients after ACS resolved by percutaneous coronary intervention (PCI), and whether, on the other hand, the presence and the degree of diastolic dysfunction at the beginning of the study, affects the functional capacity and occurrence of unwanted cardiac events (MACE), in this group of patients within the ACCR program and without it. Methods: The study include a total of 85 subjects, both sexes, ages 18-65, who were treated as clinically proven ACS during the index hospitalization and in which some of the interventional coronary procedures were performed (pPCI; PCI; PTCA). After 4 weeks of hospitalization due to NSTEMI or APNS, or after 6 weeks of hospitalization due to STEMI, patients with LVEF ≥ 45%, without significant valvular and other disorders and with some degree of lef ventricle diastolic dysfunction, were candidates for participation in the study. All patients were subjected to an initial, “entrance”, echocardiographic examination (ECHO) in order to evaluate the left ventricle systolic function and the degree of diastolic dysfunction, as well as the initial, “entrance”, spiroergometry test (CPET) in orther to asses the functional capacity. The ACCR program include a total duration of 12 weeks, or 36 individual exercise training sessions, 3 times a week, with the duration of 30 minutes. The control group included a group of patients who did not live near ICVDV and who was not able to come regularly to the exercise training sessions as a part of ACCR. They determine the degree of exercise trainig on their own will after they performed initial EHO and the CPET examination, using the optimal medical treatment and other secondary prevention measures. After 3 months, both groups of patients were subjected to a new, "exit", ECHO and CPET examinations in order to compare with the results at the start of the study. Results: The study showed that after 3 months of supervised exercise training program, as a part of the ACCR, in patients following ACS treated with percutaneous coronary intervention, there is an improvement in the degree of left ventricular diastolic dysfunction, especially through changes in the echocardiographic parameters e'l and E / e 'l . In the control group e'l decreased by (0,003 ± 0,025), and in the base group it increased by (0,011 ± 0,021). In the control group, e'l did not change significantly (p = 0,515), while in the base group significantly increased (p <0,0005). At the beginning of the study in the base group e'l was (0,097 ± 0,027 m/sec), and at the end (0,108 ± 0,031 m/sec). E / e'l did not change significantly in the control group (p = 0,226), and in the base group significantly decreased (p = 0,002). At the beginning of the study in the base group E / e'l was (8,02 ± 2,98), and at the end (6,97 ± 2,17). It has also been established that after 3 months of supervised training within the ACCR program, the functional capacity of patients with CAD and left ventricular diastolic dysfunction is improved, by increasing in peak VO2, VO2 predict and METs. In the control group, peak VO2 decreased by (1,79 ± 3,84), and in the base group it increased by (1,67 ± 4,29). In the control group, peak VO2 decreased significantly (p = 0,018) and significantly increased in the base group (p = 0,005). At the start of the study in the control group, the mean peak VO2 was (23,01 ± 3,99 ml/ kgTT / min), and in the base group it was (23,15 ± 4,99 ml / kgTT / min). At the end of the study in the base group, the mean peak VO2 was (24,82 ± 5,77 ml/ kgTT / min), while the control group was (21,21 ± 4,05 ml / kgTT / min). In the control group ppVO2 (%) it decreased by (5,28 ± 14,39), and in the examined it increased by (7,16 ± 18,51). The control group ppVO2 (%) did not significantly change (p = 0,058), while in the base group it increased significantly (p = 0,005). The difference in mean values of MET changes between base and control group was statistically significant (p <0,0005). In the control group METs decreased by (0,55 ± 1,12), and in the base group it increased by (0,58 ± 1,12). METs significantly decreased in the control group (p = 0,013), while in the base group significantly increased (p <0,0005). Conclusions: The program of ambulatory comprehensive cardiac rehabilitation in patients following acute coronary syndrome, treated with percutaneous coronary intervention, has an positive effect on the improvement of the risk factors for cardiovascular diseases, significantly influencing the improvement of the degree of left ventricle diastolic dysfunction and the improvement of the functional capacity of patients.
true
Efikasan mehanizam kriptografske sinhronizacije u algoritmima selektivnog šifrovanja multimedijalnih sistema nove generacije
An efficient mechanism of cryptographic synchronization within selective encryption algorithms of the new genration multimedia systems.
Brzi razvoj digitalne multimedije, dostupnost većih propusnih opsega u komunikacionim mrežama i porast procesorske snage prouzrokovali su svakodnevno korišćenje digitalnih multimedijalnih podataka na različitim uređajima i u različitim sferama života. Velika količina kako ličnih tako i poslovnih multimedijalnih podataka postaje javno dostupno i mogu biti lako ukradeni, kopirani ili modifikovani. HEVC (High Efficiency Video Coding) je najnoviji standard kompresije video podataka koji su zajedno razvile institucije ITU-T i ISO/IEC u toku 2013. godine. Ovaj standard kodovanja video podataka je razvijen kao odgovor na rastuće potrebe za većim rezolucijama video podataka, većim stepenom kompresije pokretnih slika i boljim iskorišćenjem računarskih arhitektura za paralelnu obradu podataka. Dizajn najnovijeg standarda kompresije video podataka obezbeđuje približno 30% -50% redukcije bitskog protoka (bitske brzine) u odnosu na ekvivalentni perceptualni kvalitet koji se postiže prethodnim standardom video kompresije - H.264/AVC High Profile. Ovakva karakteristika čini HEVC standard pogodnim za različite primene kao što su Internet striming, komunikacione tehnologije, konverzacija u realnom vremenu koja obuhvata i video ćaskanje, video konferencije i telepresence sisteme komunikacije. Osim navedenog, HEVC standard se može efikasno koristiti za skladištenje digitalnih video podataka i za emitovanje televizijskog signala visoke definicije (engl. HD TV - High deffinition) preko satelitskih, kablovskih ili zemaljskih sistema prenosa. Prethodno navedeno čini HEVC standard atraktivnim rešenjem za širok opseg mogućih primena video sadržaja kako i okviru različitih Internet servisa u komercijalnom sektoru tako i u vojnim komunikacionim sistemima. Navedeni standard predstavlja najefikasniji sistem kompresije video podataka koji postoji. Međutim, ovaj standard ne obezbeđuje bezbednosne mehanizme kojima se implementiraju kriptografski servis očuvanja tajnosti podataka. Postoji nekoliko javno dostupnih algoritama selektivnog šifrovanja za prethodni H.264/AVC standard kompresije i nekoliko novih za novi H.265/HEVC standard. Algoritmi selektivnog šifrovanja se koriste za zaštitu tajnosti video toka podataka. Mali deo video toka je iv kriptografski obrađen, sa minimalnim utroškom resursa (procesorska snaga, vreme obrade) i još uvek dovoljnim nivoom sigurnosti za različite oblasti primene. Selektivnim šifrovanjem video toka postižu se značajne uštede u vremenu obrade podataka. Ovakav način očuvanja procesorske snage je poželjan u komunikacionim sistemima sa ograničenim resursima (mrežne aplikacije koje rade u realnom vremenu, razmena slika i video sadržaja visokog kvaliteta i rezolucije, mobilni sistemi sa uređajima koji imaju ograničenu procesorsku snagu i ograničen vek baterije). Šifrovanje malog dela video toka na predajnoj strani sprečava ili ometa slučajni pristup u okviru selektivno šifrovanog HEVC video toka na prijemnoj strani. Slučajan pristup u okviru selektivno šifrovanog HEVC video toka na prijemnoj strani, podrazumeva da algoritam selektivnog šifrovanja i HEVC dekoder mogu da pokrenu proces dešifrovanja i dekodiranja u bilo kojoj tački slučajnog pristupa video toka podataka. To znači da su oni u mogućnosti da pristupe proizvoljnoj poziciji u okviru video fajla, proizvoljnoj poziciji u okviru video toka podataka, do obave operaciju spajanja video tokova ili operaciju promene kanala (promena izvora video toka podataka) u bilo kom vremenskom trenutku. Na prijemnoj strani, algoritam selektivnog šifrovanja treba da zna koji deo sintaksnih elemenata HEVC video toka podataka treba da dešifruje i koje je početno stanje od koga počinje operaciju dešifrovanja. Drugim rečima, prijemna strana (strana dekodera) mora biti kriptografski sinhronizovana sa predajnom stranom (enkoderom). Delovi sintaksnih elemenata koji trebaju da se dešifruju definisani su primenjenim algoritmom selektivnog šifrovanja dok je inicijalno stanje od koga počinje proces dešifrovanja definisano primenjenim simetričnim kriptografskim algoritmom. Suštinske karakteristike algoritama selektivnog šifrovanja su: sintaksni elementi video toka podataka koji su šifrovani i primenjeni simetrični kriptografski algoritam. Navedene karakteristike se razlikuju među različitim algoritmima selektivnog šifrovanja. Ovakva različitost među algoritmima selektivnog šifrovanja zahteva postojanje efikasnog mehanizma kriptografske sinhronizacije koji je nezavistan od primenjenog algoritma selektivnog šifrovanja. Glavni doprinos ove disertacije je definicija originalnog i efikasnog mehanizma kriptografske sinhronizacije u okviru selektivno šifrovanog HEVC video toka podataka koji je nezavistan od primenjenog algoritma selektivnog šifrovanja. Ovako efikasan v mehanizam postignut je definisanjem sintakse i semantike novog sintaksnog elementa u okviru HEVC video toka čijom primenom se implementira kriptografska sinhronizacija i omogućava slučajan pristup u okviru selektivno šifrovanog HEVC video toka. Efikasnost ponuđenog rešenja ogleda se u minimalnoj količini dodatih bita u HEVC video tok podataka. Veličina dodatnih sinhronizacionih podataka je direktno proporcionalna veličini bloka primenjenog simetričnog kriptografskog algoritma. Veličina sinhronizacionih podataka ne zavisi od parametara video toka podataka koji je selektivno šifrovan. Centralni deo ove disertacije je definicija dodatnog sintaksnog elementa u HEVC video toku, što predstavlja proširenje navedenog standarda video kompresije, sa ciljem da se dizajniraju efikasni H.265/HEVC enkoder i dekoder sa mogućnošću slučajnog pristupa.
Following advancement and rapid development of digital multimedia, larger bandwidths available within the communication network and increased processing power led to the everyday utilization of digital multimedia on different devices and in different areas of life. A large amount of both personal and business multimedia data has consequently become publicly available, and in turn can be more easily copied or modified. High Efficiency Video Coding (HEVC) is the newest video coding standard of ITU- T and ISO/IEC, proposed in 2013. This coding standard was developed in response to the growing need for increased video resolution support, higher compression of moving picture and greater use of parallel processing architectures. The design of the HEVC video coding standard provides approximately 30% - 50% bit-rate reduction for the equivalent perceptual quality relative to the performance of the previous standard H.264/AVC High Profile. This feature makes HEVC suitable for various applications such as Internet streaming, communication, real-time conversation comprising video chat, video conferencing and telepresence systems. Furthermore, HEVC can be efficiently used for digital storage media, and broadcasting of high definition (HD) television signal via satellite, cable and terrestrial transmission systems. This makes HEVC an attractive solution for a wide range of video applications, including various Internet services, commercial sectors, as well as military purposes. This standard represents the most efficient video compression system available nowadays. However, the standard does not provide security mechanisms that would ensure confidentiality and authenticity. There are several publicly available selective encryption algorithms for the previous H.264/AVC standard and for the new H.265/HEVC standard. Selective encryption algorithms are used to protect video stream – a small part of the video stream is encrypted, with minimal resource overhead and still a sufficient security level for most applications. Selective encryption of video stream achieved significant savings in time of data processing. This way of preserving CPU power is desirable in communication systems with limited resources (real time network vii application, exchange of images and video stream of high quality and resolution, mobile devices that have limited processing power and battery life). Encryption of a small part of the video stream at the transmiting side prevents or hinders random access within selectively encrypted HEVC video stream on the receiving side. Random access within selectively encrypted HEVC video stream, at the receiving side, means that selective encryption algorithm and HEVC decoder can start the deciphering and decoding process at any point of the video stream. This means that they are possible to jump to a particular position within the file, a particular position within the video stream, perform splicing operation or channel switching (change source of video stream) at any time. At the receiving side, selective encryption algorithm needs to know which parts of HEVC syntax elements necessary to decrypt and that the initial state from which starts when decrypt. In other words, the receiving side (decoder) must be cryptographically synchronized with a transmitting side (encoder). Parts that need to be decrypted are defined by selectively encryption algorithm and the initial state аrе defined by the selected symmetric cryptographic algorithm. The essential characteristics of a selective encryption algorithm are: video stream syntax elements that are encrypted and applied symmetric cryptographic algorithm. These characteristics are different among different selectively encryption algorithms. Such diversity between the selective encryption algorithms, requires the existence of the cryptographic synchronization mechanism that is independent from the applied selectively encryption algorithm. The main contribution of this disertation is defining an original and efficient cryptographic synchronization mechanism within the selectively encrypted HEVC video stream that is independent from the applied selective encryption algorithm. This mechanism is achieved by defining the syntax and semantics of the new syntax element in the HEVC bitstream that provides cryptographic synchronization and allows random access to the selectively encrypted HEVC video stream. The efficiency of the offered solutions is reflected in the resulting data overhead. The size of the synchronization parameter is directly proportional to the size of the block of the applied cryptographic algorithm. It also does not depend on the parameters of the encoded video. Defining an viii additional syntax element is the central part of the disertation and it is dedicated to the design of an efficient H.265/HEVC crypto encoder with random access capability.
true
Detekcija stanja rotacionih aktuatora zasnovana na analizi akustičkih signala
State detection of rotating actuators based on acoustic signal analysis
Danas se u industriji sve veći akcenat stavlja na razvijanje ekonomičnijih načina održavanja mašina i blagovremene detekcije otkaza. Ovo nije toliko iznenađujuće ako se uzme u obzir procena da je čak jedna trećina troškova održavanja prosečnog industrijskog postrojenja posledica suvišnog ili neefikasnog održavanja...
In recent years in modern industry great emphasis has been placed on the development of cost-eective approaches to machine maintenance and timely fault detection. This is not so surprising if one takes into account the estimate that one third of the maintenance costs of an average industrial plant are due to excessive or inecient upkeep. Modern predictive maintenance techniques were developed with goal to prolong life expectancy of machines and detect worn components in timely manner, thus significantly reducing maintenance costs. These methods are based on acquisition of appropriate signals (vibration measurements, thermographic images, etc.), processing and analysis of those signals and, finally, estimating the state of the machine and assessing the necessity of repairs. Rotating machines are especially common in industry and are frequently mentioned in modern predictive maintenance literature, with vibration signals being the most informative resource for their diagnostics. The usage of vibration signals for feature extraction and machine condition diagnostics is well founded in the literature and tested on various types of actuators, including industrial mills. However, in the last decade an increased emphasis has been placed on usage of acoustic signals with purpose of fault detection. It has been shown that they can be as informative as vibration signals and can enable the detection of certain changes even faster, due to their high sensitivity to certain changes in the environment. Also, acoustic sensors are cheaper, contactless and can record signals in an immediate vicinity of an actuator while it is in operation. There has been much success in using acoustic sensors for predictive maintenance; however, almost all those results are obtained in strictly controlled laboratory conditions. A major issue in using these signals in industrial surroundings is their high susceptibility to the surrounding noise which is always present in real-life conditions. Filtering the noise using traditional techniques is usually not possible without significant loss of useful information, so all the benefits of using sound signals for fault detection are often overshadowed by this one flaw. The main objective of the research conducted within this doctoral dissertation is the development of new methodology which, using acoustic signals in contaminated environment, searches for informative features in time and frequency domain. The special attention will be given to preprocessing of the signal for the purpose of detecting the existence of contaminating components. The research will be focused on problems in thermal power plants (state detection of impellers within mills), but it is our belief that the results can be generalized and expanded to broad family of rotating actuators like feedwater pumps, compressors, ventilators, etc...
false
Nove kombinatorne konstrukcije u vezi sa problemima iz hromatske teorije grafova, ekstremalne teorije skupova i teorije Bulovih matrica
New combinatorial constructions related to problems from the chromatic graph theory, extremal set theory and Boolean matrix theory.
Disertacija prikazuje originalne rezultate iz narednih podoblasti diskretne matematike: hromatske teorije grafova, ekstremalne teorije skupova i teorije Bulovih matrica. Iz hromatske teorije grafova razmatraju se bojenja grafova koja uključuju bojenje grana ili totalno bojenje, sa uslovom da susedni čvorovi grafa pose- duju različite parametre indukovane datim bojenjem (multi-skup, skup ili suma). Multi-skup sused-razlikujuće bojenje grana grafa je dodela boja granama grafa tako da se za svaku granu uv grafa razlikuju multi-skup boja grana incidentnih sa čvorom u i multi-skup boja grana incidentnih sa čvorom v. Do sada najbolji rezultat po pitanju ovakvog bojenja je da je za svaki graf bez izolovanih grana moguće dodeliti granama boje iz skupa od četiri elementa, tako da prethodno navedeno svojstvo bude zadovoljeno. Doprinos autora je dokaz da je ovo moguće postići sa svega tri boje, što je u opštem slučaju i optimalan broj boja. Konstruisan je graf, za koji je zatim dokazano da je potreban različit broj boja da bi se izvršilo multi-skup sused-razlikujuće i sused-razlikujuće bojenje po sumi. Koliko je nama poznato, ovo je prvi graf sa datim svojstvom. Izloženo je više rezultata po pitanju proširenog sused-razlikujućeg bojenja po sumama. Doprinos autora je dokaz da je za proizvoljan graf moguće izvesti totalno bojenje elementima skupa {1, 2, 3}, tako da za svaka dva susedna čvora, vrednost sume boja susednih čvorova i incidentnih grana bude različi- ta. Takođe, za određene klase grafova je pokazano da postoji ovakvo bojenje korišćenjem vrednosti iz skupa {1, 2}. Sused-razlikujuće bojenje grana grafa zahteva da susednim granama grafa budu dodeljene različite boje, pri čemu se razlikuju skupovi boja grana inci- dentnih sa čvorovima u i v, za svaku granu uv grafa. Prikazan je postupak dodele boja granama proizvoljnom grafu bez izolovanih grana, pri čemu se koristi manji broj boja nego u do sada poznatim rezultatima. Kod susednog čvor-razlikujućeg totalnog bojenja grafa G neophodno je da svaka dva susedna i incidentna elementa iz skupa V (G)∪E(G) imaju različitu boju, kao i da za svaku granu uv grafa G, skup boja dodeljenih granama incidentnim sa u zajedno sa bojom čvora u, bude različit od takvog skupa boja za čvor v. Doprinos autora je popravljena gornja granica minimalnog broja boja za proizvoljni graf G, u odnosu na maksimalan stepen grafa G. Iz ekstremalne teorije skupova razmatrana je Franklova hipoteza koja tvrdi da za svaku familiju skupova zatvorenu za uniju, postoji element koji je sadržan u barem polovini skupova te familije. Dokazano je da je Franklova hipoteza zadovoljena za svaku familiju sa- stavljenu od 12 elemenata, dok je od ranije poznato da ovo važi za familije sastavljene od 11 ili manje elemenata. Sam dokaz je zasnovan na efikasnom algoritmu iscrpljivanja mogućnosti, pri čemu se koriste rezultati o potfamili- jama skupova koje su dobijene u više koraka, a koje eventualni protiv-primer familije ne sme da sadrži. Familija skupova G je FC-familija ukoliko za svaku familiju F koja je sadrži postoji element iz � G koji se pojavljuje u barem polovini skupova familije F. NonFC-familija je svaka familija koja nije FC. Doprinos autora je potpuna podela svih familija sastavljenih od 6 ili manje elemenata, na FC i NonFC-familije. Iz teorije Bulovih matrica dati su rezultati po pitanju kardinalnosti pro- stora vrsta Bulovih matrica. Bulove matrice su matrice kod kojih su sve komponente iz skupa {0, 1}, dok je prostor vrsta matrice skup svih vektora koji se mogu dobiti disjunkcijom vrsta te matrice. Dobijen je skup svih vrednosti a iz intervala [2n−2 + 2n−3, 2n−2] za koje postoji matrica dimenzije n × n sa kardinalnošću prostora vrsta jednakom a. Neka je an najmanja pozitivna celobrojna vrednost takva da ne postoji matrica dimenzije n × n sa kardinalnošću prostora vrsta jednakom an. Do- bijena je veća donja granica za vrednost an, u odnosu na do sada poznate rezultate. Dokazi glavnih rezultata u disertaciji su konstruktivni, pri čemu neki od njih zahtevaju pomoć računara u ispitivanju svih mogućnosti. Za ostale dokaze ovo nije neophodno, iako je moguće implementirati algoritme za dobi- janje bojenja grafova, odnosno za dobijanje matrica, sa traženim svojstvima.
We present original results from the following fields of discrete mathematics: chromatic graph theory, extremal set theory and Boolean ma- trix theory. From the chromatic graph theory we investigate edge and total colorings satisfying the condition that neighboring vertices of a graph possess different values of multi-set, set or sum, induced by the giving coloring. Multi-set neighbor-distinguishing edge coloring of a graph is an assign- ment of colors to edges such that, for every edge uv of a graph, multi-set of the edges incident with the vertex u differs from the multi-set of the edges incident with the vertex v. The previous best result concerning the minimum number of colors required for such a coloring of an arbitrary graph states that four colors are sufficient. The author’s contribution is a proof that such a coloring is always possible with only three colors, which is in general case the optimal number of colors. We construct a graph for which we subsequently prove that a different number of colors is required to obtain a multi-set neighbor-distinguishing coloring and neighbor-distinguishing coloring by sum. As far as we know, this is the first example of such a graph. A few results concerning the neighbor expended sum distinguishing col- oring are given. The main contribution is a proof that for an arbitrary graph there exists a total coloring from the set {1, 2, 3}, such that every two adja- cent vertices have different sums of its adjacent vertices and incident edges. Also, for certain classes of graphs is proved that there exists such a coloring using only the colors from the set {1, 2}. Neighbor-distinguishing edge coloring of a graph G requires that every two adjacent edges receive different colors, while the sets of the edges incident with the vertices u and v differ for every edge uv of G. The author presents a procedure of edge coloring for an arbitrary graph without isolated edges, where we a smaller number of colors is used compared to all known results. For the adjacent vertex distinguishing total coloring of a graph G the condition is that every two adjacent and incident elements of V (G) ∪ E(G) receive different colors, while for every edge uv of G the set composed from the colors assigned to the edges incident with u together with the color of u, differs from such a set for v. The author improves the upper bound of the minimum number of colors needed for such a coloring, relative to the maximal degree of a graph. Frankl’s conjecture from the extremal set theory states that for every family closed under union there exists an element contained in at least half of the sets of the family. We give a proof that Frankl’s conjecture holds for every family contained from 12 elements, while it is known that this is true for families contained from 11 or less elements. Our proof is based on the efficient algorithm that exhausts all the possibilities, while using the results for subfamilies that eventual counter-example cannot contain, which we obtained in a number of consecutive steps. Family of sets G is an FC-family if for every family F containing G there exists an element from � G that appears in at least half of the sets of F. NonFC-family is every family that is not FC. The author’s contribution is the complete classification of all families consisting of 6 or less elements into FC and NonFC-families. From the Boolean matrices theory we present our results concerning the row space cardinality. Boolean matrices are the matrices whose all compo- nents are from the set {0, 1}, while the row space of a Boolean matrix is the set of vectors that can be obtained by disjunction from the rows of a matrix. We present the set consisted of all values a from the interval [2n−2 + 2n−3, 2n−2] such that there exists a matrix of dimension n×n having the row space cardinality equal to a. For the least positive integer an for which there exists no matrix of di- mension n × n having the row space cardinality equal to an, the author gives a lower bound that is an improvement over the previously known results. All proofs for the main results in the dissertation are constructive. Proofs of some of them require the use of computers where there is a calculation of a great number of possibilities. For other proofs this was not necessity, though algorithms following the steps of the proofs can be implemented to obtain a graph coloring or a matrix with the desired properties.
true
Preslikavanja kontraktivnog tipa i njihove primene u nelinearnoj analizi
Mappings of contractive type and its applications in nonlinear analysis
U ovoj disertaciji su prikazane neke nove osobine izvesnih preslikavanja kontraktivnog tipa, i date su primene tih rezultata u nelinearnoj analizi. Disertacija sadrži šest poglavlja. Prvo poglavlje sadrži osnovne osobine polu - metričkih prostora. U njoj je dato i jedno proširenje Borel - Lebegove teoreme za tu klasu prostora. U drugom poglavlju je dokazano jedno uopštenje teoreme o nepokretnoj tački Niemitckog. U trećem poglavlju je pojam mere nekompaktnosti proširen na klasu polu – metričkih prostora. Takođe je dokazana teorema o fiksnim tačkama za kondenzujuća preslikavanja definisana na toj klasi prostora. Sledeća tri poglavlja sadrže nove stavove o zajedničkim fiksnim tačkama Majr - Kelerovog tipa, nove rezultate o akretivnim preslikavanjima i nove teoreme o zajedničkim fiksnim tačkama za preslikavanja definisana na verovatnosnim metričkim prostorima.
The aim of this dissertation is to present new properties of some classes of contractive mappings, and applications of this results to nonlinear analysis. It contain six chapters. The first chapter gives basic properties of semi-metric spaces. New result is extension of Borel - Lebesque theorem to semi metric spaces. The second chapter contain generalization of Niemytzki’s fixed point theorem. In third chapter the notion of measure of non-compactness are extended to class of semimetric spaces. New fixed point theorem for condensing mappings defined on semi-metric spaces is presented. Further three chapters contains new common fixed points theorems of Mair - Keeler type, new results for accretive mappings and new common fixed point results for mappings defined on probabilistic metric spaces.
false
Rekonfigurabilni i podesivi efikasni pojačavači snage za predajnike telekomunikacionih uređaja
Reconfigurable and tunable efficient power amplifiers for transmitters in telecommunication devices
Disertacija se bavi unapređenjem telekomunikacionih sistema i u njoj su predložena rešenja koja se mogu iskoristiti u sadašnjim i budućim telekomunikacionim sistemima. Nakon kratkog uvoda u osnovne klase rada pojačavača na mikrotalasnim frekvencijama, predstavljena su tri pojačavača snage. Sva tri pojačavača su namenjena modernim telekomunikacionim sistemima. Bliže rečeno, sva tri pojačavača su koncipirana i napravljena da rade u više frekvencijskih opsega. U dizajniranju pojačavača, akcenat je bio stavljen na efikasnost pojačavača. Stoga su prva dva pojačavača dizajnirana i realizovana da rade u klasi AB. Kod prvog, rekonfigurabilnog, pojačavača jedna od dve moguće radne frekvencije bira se pomoću uključenja/isključenja PIN diode u ulaznom kolu za prilagođenje. Drugi, podesivi (tjunabilni), pojačavač je dizajniran i realizovan tako da može kontinualno da menja frekvenciju rada, između dve udaljene frekvencije, adekvatnom polarizacijom varikap diode smeštene u ulazno kolo za prilagođenje. Treći pojačavač je dizajniran i realizovan da radi u klasi J. U vezi s njim je najpre predstavljena teorija koja se do sada nije mogla naći u literaturi. Ta teorija predstavlja dopunu postojeće teorije vezane za pojačavače u klasi J. U okviru novorazvijene teorije predložen je način dizajniranja pojačavača, gde se na početku određuju parametri pojačavača za unapred zadatu efikasnost. Prateći predloženu teoriju, dizajniran je i realizovan pojačavač u klasi J. Rezultati dobijeni simulacijom i mereni rezultati su se odlično poklopili, što je verifikovalo valjanost predložene teorije. Na kraju su dati rezultati simulacija nelineranosti projektovanog pojačavača u slučaju da mu se na ulaz dovedu neki od signala koji se koriste u modernim telekomunikacionim sistemima...
This dissertation addresses the improvement of telecommunication systems. The proposed solutions can be implemented in both present and future telecommunication systems. After a brief introduction regarding the classes of amplifiers operating at microwave frequencies, three power amplifiers are presented. All three amplifiers are designed for modern telecommunication systems. More precisely, they are able to operate in several frequency bands. In the process of the amplifier design, the focus was set on the amplifier efficiency. Hence, the first two amplifiers are designed and fabricated to operate in class AB. In the first, reconfigurable amplifier, one of two possible operating frequencies can be chosen by switching on or off a PIN diode in the input matching network. The second, tunable, amplifier is designed and fabricated in such a way that the operating frequency can be continuously changed, between the lower and upper frequency boundaries. This is achieved by controlling the bias of the varicap diode, positioned in the input matching network. The third amplifier is designed and fabricated to operate in class J. Firstly, a novel theory describing the operation and design procedure of a class-J power amplifier is presented as an extension of the existing theory given in the literature. In the process, novel simple formulas are derived which govern the amplifier efficiency. An amplifier in class J is designed and fabricated following the proposed theory. Results obtained by numerical simulations and measurements on fabricated prototypes are in excellent agreement, thus verifying the proposed theory. Finally, simulation results are presented that characterize the amplifier performance when it is excited by some of the wideband signals used in modern telecommunication systems. The obtained results, which give insight in the linearity of the designed amplifier, are comparable with other results published in the literature, thus they additionally validate the applicability of the newly proposed theory...
false
Asimptotska svojstva neparametarskih testova zasnovanih na U-statistikama i V-statistikama sa nedegenerisanim i slabo degenerisanim jezgrom
Asymptotic properties of non-parametric and V-statistics with non-degenerate and weakly degenerate kernel.
Testovi saglasnosti sa raspodelom i testovi simetrije zauzimaju značajno mesto u neparametarskoj statistici...
Goodness of t and symmetry tests occupy a signicant part of nonparametric statistic. Most of classical tests are based on the distance between the assumed distribution function and its consistent estimate, empirical distribution function. The symmetry tests are analogously constructed. A new approach that is especially attractive in recent years is making tests based on characterizations of dierent types. Those tests use U-empirical distribution functions (generalized empirical ones), U-empirical transforms (eg. Laplace transform, characteristic functions etc.) and U-empirical moments of distributions. The main advantage of these tests is that they are often free of some distribution parameters. Therefore they are suitable for testing composite hypothesis.
false
Model izbora transportnih sredstava vojnih jedinica angažovanih u multinacionalnim operacijama
Model of selection of transport vehicles used by military units engaged in multinational operations
Izbor transportnih sredstava za potrebe vojnih jedinica važan je faktor koji utiče na pokretljivost jedinica, kao i na kvalitet izvršenja različitih namenskih transportnih zadataka u okviru dodeljenih misija. Ovaj rad se bavi rešavanjem višekriterijumskog problema izbora transportnih sredstava za potrebe vojnih jedinica koje se angažuju multinacionalnim operacijama, primenom metode Analitičkog hijerarhijskog procesa i Analize obavijanja podataka (DEA). Za ocenu pet usvojenih alternativa (terenskih vozila) primenjen je hibridni AHP-DEA model sa osam kriterijuma i dvadesetčetiri potkriterijuma. Rezultati dobijeni pomoću AHP metode koriste se kao višestruki izlazi u okviru definisanog DEA modela za izbor transportnih sredstava, na bazi kojeg se definišu efikasnosti alternativa i određuje njihov rang. Poređenjem ranga alternativa dobijenih AHP metodom i ranga alternativa dobijenih DEA modelom definiše se referentni rang alternativa hibridnog AHP-DEA modela za izbor transportnih sredstava. Dobijeni rezultati definisanog AHP-DEA modela proveravaju se analizom osetljivosti formiranjem različitih scenarija, dok se dodatna provera dobijenih rezultata vrši postupkom validacije korišćenjem novih BWM, BWM-DEA, AHP-MABAC i BWM-MABAC modela. Određivanje konačnog ranga alternativa hibridnog AHP-DEA modela za izbor transportnih sredstava zasniva se na poređenu referentnog ranga alternativa AHP-DEA modela i rangova alternativa koji su dobijeni u okviru postupka validacije dobijenih rezultata. Za utvrđivanje relacije između dobijenih rangova alternativa koristi se Spearmanov koeficijent korelacije rangova.
Selection of transport vehicles used by army units is an important factor influencing unit mobility, as well as quality of performance of transportation tasks within the given missions. This paper deals with solving multicriteria problems of selecting transport vehicles used by army units engaged in multinational operations, by using the method of Analityc hieararchy process and Data envelopment analysis (DEA). For assesing five adopted alternatives (terrain vehicles) a hybrid AHP- DEA model with eight criteria and twenty-four subcriteria has been used. The results obtained by AHP method are used as multiple outputs within the defined DEA model for transport vehicles selection, which is used as a basis for defining efficiency of alternatives and their ranking. By comparing the ranking of alternatives produced via AHP method and the ranking of alternatives produced via DEA model a preferential ranking of alternatives of a hybrid AHP-DEA model for transport vehicles selection is defined. The results obtained by defined AHP- DEA model are checked by sensitivity analysis through different scenarios, while additional checking of obtained results is conducted through the process of validation using new BWM, BWM-DEA, AHP-MABAC and BWM-MABAC models. Defining the final alternative ranking of the hybrid AHP-DEA model for transport vehicles selection is based on comparison between preferential ranking of alternatives of a AHP-DEA model and alternative ranking gathered through the process of validation of obtained results. Spearman's rank correlation coefficient is used for defining the correlation between obtained alternative ranking.
true
O prstenu trigonometrijskih polinoma sa primenom u teoriji analitičkih nejednakosti
On trigonometric polynomial ring with applications in the theory of analytic inequalities.
Ova doktorska disertacija se sastoji iz dva dela...
This doctoral dissertation comises two parts...
false
Topologija i kombinatorika kvazitorusnih mnogostrukosti i K stepena
Topology and combinatorics of quasitoric manifolods and the polyhedral product functors
Predmet proučavanja ove doktorske disertacije su kvazitorusne mnogostrukosti i K-stepeni. Ove topološki veoma zanimljive mnogostrukosti su uopštenja torus- nih varijeteta i predmet su proučavanja mnogih matematičkih disciplina: torusne geometrije, simplektičke geometrije, torusne topologije, algebarske geometrije, algebarske topologije, topološke kombinatorike, teorije konveksnih politopa, itd. U ovoj tezi je akcenat stavljen na mogućnost da ove objekte proučavamo kombinatornim metodama, uzimajući u obzir geometriju torusnog dejstva i kombinatoriku prostih politopa koji predstavljaju prostore orbita ovih mnogostrukosti. Objekti kojima se bavi ova disertacija su našli primenu kako u matematici, tako i u drugim primenjenim oblastima. Zato su oni predmet intenzivne naučne pažnje. Originalni naučni doprinos ove teze je u novim rezultatima o kvazitorusnim mno- gostrukostima u odnosu na klasična pitanja algebarske topologije: stepene preslikavanja među mnogostrukostima, ulaganjima i imerzijama mnogostrukosti u euklidske prostore. Metodi koji se koriste u ovoj tezi oslikavaju duboku vezu između onoga što smatramo klasičnom matematikom i torusnom topologijom. U tom smislu, metodi koje ovde koristimo imaju potencijal da u skoroj budućnosti prošire dosadašnja znanja o ovoj klasi mnogostrukosti, ali takođe i da se iskoriste u nekim drugim problemima, kao što su teorija bordizama. U prva tri poglavlja su na elementaran način izložene osnovne osobine simplici- jalnih kompleksa, politopa, torusnih varijeteta i kvazitorusnih mnogostrukosti, kao i najvažniji rezultati o njima. Akcentovan je kombinatorni pristup koji nam omogućava da ove apstraktne objekte shvatimo kao prirodne. Četvrto poglavlje disertacije proučava stepene preslikavanja među kvazitorusnim mnogostrukostima. Tehnika koja se koristi bazirana je na rezultatima Haibao Duan-a i Shicheng Wang-a koji su koristili presecnu formu za formulisanje konkretnog uslova za postojanje preslikavanja odredenog stepena. Dobijeni su veoma interesantni rezultati, a posebno je klasa kvazitorusnih 4-mnogostrukosti pogodna za proučavanje. Kohomološki prsten ovih mnogostrukosti je takav da je u mnogim slučajevima bilo moguće odrediti skup svih mogućih stepena preslikavanja. Rezultati su dobijeni u terminima elementarne teorije brojeva, oslanjajući se na klasične rezultate o brojevima koji se mogu zapisati kao suma odredenog broja potpunih kvadrata. Opisani su skupovi mogućih stepena preslikavanja među mnogostrukosti za veliki broj konkretnih mnogostrukosti kao što su CP^2, CP^2.CP^2, S^2 × S^2, itd. U petom poglavlju računaju se Stifel-Whitney-eve klase ovih mnogostrukosti, dualne klase, a potom se one koriste kao opstrukcije za totalno kosa ulaganja i imerzije.Glavni rezultat je da one prvenstveno zavise od dejstva torusa, jer su konstruisani primeri mnogostrukosti nad istim politopom od kojih jedna ima trivijalne, a druga netrivijalne karakteristične klase. Pažnja je posvećena mnogostrukostima nad kubom koju predstavljaju nove primere u kojima je donja granica bliska gornjoj granici za totalno kosa ulaganja, a za koje se tačno mogu odrediti euklidski prostori minimalne dimenzije u koje se ove mnogostrukosti mogu imerzovati i uložiti. U Dodatku je kratko izložena teorija o dejstvima grupa i osnovne osobine ekvivarijantnih preslikavanja.
The main objects studied in this doctoral thesis are quasitoric manifolds and spaces arising as the images of polyhedral product functors. Quasitoric manifolds are particularly interesting as topological generalization of non-singular toric varieties. They are a research topic of many mathematical disciplines including toric geometry, symplectic geometry, toric topology, algebraic geometry, algebraic topology, theory of convex polytopes, and topological combinatorics. These objects have already found numerous applications in mathematics and sciences and they continue to be intensively studied. In this thesis we put some emphasis on combinatorial methods, focusing on the interaction of the geometry of toric actions and combinatorics of simple polytopes. This connection of geometry and combinatorics is based on the fundamental observation that convex polytopes naturally arise as orbit spaces of toric actions on quasitoric manifolds. Our main original contributions in this thesis are related to classical topological questions about degrees of maps between manifolds as well as their embeddings and immersions into Euclidean spaces. We follow the general scheme characteristic for Algebraic Topology where a topological problem is reduced, often by non-trivial reductions, to a question of arithmetical, algebraic, or combinatorial nature. We believe that the novel applications of this scheme developed in the thesis, especially the new techniques and calculations, have a potential to be applied on other problems about quasitoric manifods. Here is a summary of the content of the thesis. For the reader’s convenience and for completeness, in the first three chapters we give an elementary exposition of the basic theory of simplicial complexes, convex polytopes, toric varieties and quasitoric manifolds. The emphasis is on the fundamental constructions and central results, however the combinatorial approach, utilized in the thesis, allows us present the theory in a direct and concrete way, with a minimum of topological prerequisites. The mapping degrees of maps between quasitoric manifolds are studied in Chapter 4 with a particular emphasis on quasitoric 4-manifolds. Utilizing the technique pioneered by Haibao Duan and Shicheng Wang, which is based on the intersection form and the cohomology ring calculations, we demonstrate that a complete information about mapping degrees can be obtained in many concrete situations. The theorems and the corresponding criteria for the existence of mapping degrees are formulated in the language of elementary number theory. It is amusing that the question whether a number appears as a mapping degree between concrete 4-manifolds is directly linked with classical results from number theory such as whether a number can be expressed as a sum of two or three squares, etc. This approach allows us to analyze many concrete 4-manifolds, including CP2, CP2♯CP2, S2×S2, etc. In Chapter 5 we calculate the Stiefel-Whitney classes of some concrete quasitoric manifolds and their duals. This information is used to determine cohomological obstructions to embeddings and immersions of these manifolds in Euclidean spaces. As an initial observation we showed that the calculations are highly dependent on the action of torus. Indeed, there are examples of quasitoric manifolds over the same polytope which exhibit a very different behavior and different complexity of the associated characteristic classes. Focusing on the quasitoric manifolds over the n-dimensional cube, we are able to produce quasitoric manifolds which are very complex in the sense that they almost attain the theoretical minimum dimension for their embedding or (totally skew) immersion in Euclidean spaces. The thesis ends with an appendix with an outline of the theory of group actions and equivariant topology.
false
Electronic properties of interfaces between domains in organic semiconductors
Elektronska svojstva organskih poluprovodnika na granicama domena
Cilj ove disertacije da je pruži vezu izmedu atomske i elektronske strukture različitih tipova granica izmedu domena u organskim poluprovodnicima. U polikristalnim organskim poluprovodnicima na bazi malih molekula granica se formira između homogenih kristalnih domena. Ustanovili smo da granica u polikristalnom naftalinu dovodi do pojave stanja zamki u energijskom procepu materijala, lokalizovanih na parovima molekula sa različitih strana granice izmedu kojih je rastojanje malo...
The aim of this thesis is to provide a link between atomic and electronic structure of different types of interfaces between domains in organic semiconductors. In polycrystalline small-molecule organic semiconductors interfaces are formed between single crystalline domains. We found that grain boundaries in polycrystalline naphthalene introduce trap states within the band gap of the material. Trap states are localized on closely spaced pairs of molecules from opposite sides of the boundary. Realistic conjugated polymers, such as poly(3-hexylthiophene) (P3HT), contain mixed crystalline and amorphous domains. We found that HOMO state of the interface between crystalline and amorphous domain in P3HT belongs to crystalline domains. States that belong to both domains and trap states were not found. Effects of thermal disorder are important in realistic conjugated polymers. Our results show that disorder in backbone chains of P3HT has strong effect on the electronic structure and leads to the localization of the wave functions of the highest states in the valence band, similar to the ones that occur in amorphous polymers. At the interfaces between two materials in organic electronic devices, effects of spontaneous polarization in one or both of them on electronic properties can be pronounced. We show that ordered P3HT exhibits spontaneous polarization along the backbone direction, which is caused by the lack of inversion symmetry due to head-to-tail side chains arrangement. We additionally show that spontaneous polarization in ordered P3HT keeps significant values even at room temperature when the effects of thermal disorder are important.
false
Doprinos zračenja galaksija i galaktičkih jata pozadinskom gama zračenju
Contribution of galaxies and galaxy clusters to the diffuse gamma-ray background
Izučavanje kosmičkog zračenja, veoma je bitno za razumevanje fizičkih procesa na visokim energijama, a energije do kojih čestice mogu da se ubrzaju u svemiru mnogo su više od energija koje čovek trenutno može da postigne u akceleratorima na Zemlji. Procesi u kojima nastaje kosmičko zračenje su još uvek nedovoljno objašnjeni, a modeli koji se predlažu čekaju na eksperimentalnu potvrdu. Prisustvo kosmičkog zračenja može da se meri kroz detekciju različitih produkata interakcija ovih visokoenergijskih čestica sa međuzvezdanom materijom kroz koju se kreću. Ova doktorska disertacija bavi se akrecionim udarnim talasima kao izvorom kosmičkog zračenja u jatima galaksija, ali i kosmičkim zracima ubrzanim u ostacima supernovih u samim galaksijama. Bez obzira koji od ovih mehanizama se razmatra, kosmičko zračenje u sudaru sa atomima i jonima međuzvezdane materije, između ostalog, proizvodi gama zračenje, neutrine, kao i lake elemente, od kojih se posebno razmatra litijum. U disertaciji se pre svega razvijaju modeli koji opisuju gama zračenje proizvodeno od strane kosmičkog zračenja ubrzanog na udarnim talasim nastalih u različitim procesima. Prvo se razmatra akrecije gasa na već virijalizovane strukture na najvećim skalama (na primer u jatima galaksija). U modelima prvi put je implementirana vremenska promena proizvodnje gama zračenja u toku istorije svemira, koja nastaje zbog evolucije akrecionih udarnih talasa u toku formiranja velikih struktura. Zbog toga modeli razvijeni u ovoj disertaciji realnije opisuju gama zračenje na velikim skalama, u odnosu na modele koji su ranije razvijani i u kojima se pretpostavljalo da gama zračenje potiče sa jednog crvenog pomaka. Modeli se koriste za predviđanje ukupnog gama zračenja nerazlučenih jata galaksija. Ovo modelirano gama zračenje se zatim poredi sa posmatranim izotropnim difuznim pozadinskim gama zračenjem, koje je izmerio teleskop Fermi-LAT, iz čega se zaključuje da ovaj tip kosmičkog zračenja ima nezanemariv doprinos pozadinskom gama zračenju (u zavisnosti od normalizacije modela, moguće je objasniti i skoro celu gama pozadinu) i da bi trebalo da se uzima u razmatranje pored drugih komponenti za koje se pretpostavlja da imaju veliki doprinos, kao što su na primer normalne nerazlučene galaksije ili blazari. Modeli za gama zračenje akrecionih udarnih talasa se u okviru disertacije takođe porede i sa merenjima visokoenergijskih neutrina detektovanih od strane detektora IceCube. Neutrini se koriste za normiranje modela, iz čega se zaključilo da ako su akrecioni udarni talasi pretežno jaki, neutrinska pozadina jače ograničava moguću gama emisivnost ovakvih objekata, u odnosu na pozadinsko gama zračenje koje smo pre toga koristili. Izučavanje neutrina kao produkata interakcija kosmičkog zračenja je od posebne važnosti, zbog toga što neutrini slabo interaguju sa drugim česticama, pa zbog toga u sebi kriju nepromenjenu informaciju o trenutku kada su nastali, tj. o kosmičkim zracima koji ih proizvode. Jedan deo disertacije se bavi i proizvodnjom kosmičkog zračenja u ostacima supernovih, konkretno u slučaju Malog Magelanovog oblaka, koji je inače detektovan u gama oblasti. U ovoj galaksiji prvi put je izmerena zastupljenost litijuma u međuzvezdanom gasu, izvan Mlečnog puta. S obzirom na to da i gama zračenje i litijum nastaju kroz interakcije kosmičkog zračenja sa međuzvezdanom materijom, njihovo zajedničko poreklo se može iskoristiti za procenu moguće proizvodnje litijuma i gama zračenja od strane bilo koje populacije kosmičkog zračenja. Pokazano je da galaktičko kosmičko zračenje, za koje se smatra da je dominantno u okviru Malog Magelanovog oblaka, može da objasni tek delić posmatrane zastupljenosti litijuma, ako se pretpostavi da je ukupna gama emisivnost ove galaksije koja se danas meri takođe potekla od interakcija galaktičkog kosmičkog zračenja sa gasom unutar galaksije. Ovaj zaključak je zanimljiv, jer navodi na mogućnost postojanja drugog izvora litijuma u okviru Malog Magelanovog oblaka. Takođe, koristeći zajedničko poreklo litijuma i gama zračenja procenjeno je i koliko nepravilne patuljaste galaksije mogu da doprinesu difuznoj gama pozadini. Razmatranja nekoliko različitih produkata interakcija kosmičkog zračenja sa međuzvezdanom materijom (gama zračenja, neutrina i litijuma), na manjim skalama (u okviru galaksija) kao i na veoma velikim skalama (jata galaksija), pokazala su da pored galaktičkog kosmičkog zračenja nastalog u ostacima supernovih i drugo, još uvek hipotetičko, kosmičko zračenje (nastalo npr. prilikom akrecije gasa na velikim skalama ili plimskih interakcija galaksija), takođe može imati nezanemariv doprinos merenjima.
Studying cosmic rays is very important for better understanding of high energy physical processes, since particles accelerated in space can reach energies far above what we can produce in accelerators on Earth, at the moment. Processes that produce cosmic rays are still unexplained to some extent, and models that have been proposed are awaiting cofirmation. Presence of cosmic rays can be measured by detecting dfferent products of interactions of these high energy particles with the interstellar medium through which they propagate. This thesis deals with the accretion shock as sources of cosmic rays in clusters of galaxies, as well as cosmic rays accelerated in supernova remnants inside galaxies. No matter which of these mechanisms is being considered, cosmic rays will collide with atoms and ions in the interstellar medium, and produce, among other things, gamma rays, neutrinos, as well as light elements, of which we will discuss lithium. In the thesis we primarily develop models that describe gamma rays produced by cosmic rays accelerated in shocks that can appear in dfferent processes. We first examine accretion of new gas onto already virialized structures (for example in galaxy clusters). For the first time, we include the change of gamma-ray production with time, through the history of the universe, that re ects the evolution of accretion shocks which appear during large scale structure formation. Therefore, the models developed in this thesis describe the gamma rays from large scale structures more realistically, compared to models which have previously been developed and which use single redshift approximation for the gamma-ray origin. Models are used to derive the gamma-ray ux of all unresolved galaxy clusters. These modeled gamma rays are then compared to the isotropic dffuse gamma-ray background, measured by telescope Fermi-LAT. This leads to the conclusion that these cosmic rays have non-negligible contribution to the isotropic diffuse gamma-ray background (depending on the normalization, they can even explain the whole isotropic diffuse gamma-ray background) and that this population of cosmic rays has to be taken into consideration in addition to other components that are thought to be major contributors, like for example, unresolved normal galaxies or blazars. In the thesis, models of gamma-ray production in accretion shocks are also compared to observations of high-energy neutrinos detected by IceCube detector. Neutrinos are used to normalize gamma-ray models, from which we conclude that if the accretion shocks are predominantly strong, neutrino background is more limiting to the possible gamma-ray emissivity of these objects, compared to the gamma-ray background we first used. Study of neutrinos as products of cosmic-ray interactions is very important, since neutrinos interact weakly with other particles, and therefore keep all of the information about the time they were produced and about cosmic rays that produced them. One part of the thesis deals with the production of cosmic rays in supernova remnants, in particular, the case of the Small Magellanic Cloud, which was detected in gamma rays. In this galaxy we also have the first measurements of the lithium abundances in the interstellar gas outside of the Milky Way. Since gamma rays and lithium are produced through interactions of cosmic rays with the interstellar medium, their same origin can be used to estimate the production of lithium and gamma rays by any cosmic-ray population. We show that galactic cosmic rays, which are considered to be dominant population of cosmic rays in the Small Magellanic Cloud, can only explain a very small part of the observed abundance of lithium, if we assume that the entire present gamma-ray emissivity that we observe also originates from the interaction of galactic cosmic rays with gas within the galaxy. This conclusion is interesting, because it leads to the possible existence of other sources of lithium in the Small Magellanic Cloud. Also, using the fact that gamma rays and lithium share the same origin, we estimate how much can irregular dwarf galaxies contribute to the diffuse gamma-ray background. Study of several different products of cosmic-ray interactions with the interstellar medium (gamma rays, neutrinos and lithium) on smaller scales (within the galaxy), as well as on the largest scales (galaxy clusters), showed that in addition to the galactic cosmic rays accelerated in supernova remnants, other still hypothetical cosmic rays (produced for example during accretion of gas on largest scales, or tidal interactions of galaxies) can have a non-negligible contribution to the measurements.
false
Difuziono-talasna jednačina razlomljenog reda sa koncentrisanim kapacitetom i njena aproksimacija metodom konačnih razlika
Fractional diffusion-wave equation with concentrated capacity and its finite difference approximation
Difuziono-talasna jednačina razlomljenog reda po veremenskoj promenljivoj dobija se iz klasične difuzione ili talasne jednačine zamenom prvog, odnosno drugog izvoda po vremenskoj promenljivoj izvodom razlomljenog reda...
The time fractional diusion-wave equation can be obtained from the classical diffusion or wave equation by replacing the rst or second order time derivative, respectively, by a fractional derivative of order 0 < 2. In particular, depending on the value of the parameter , we distinguish subdiusion (0 < < 1), normal diusion ( = 1), superdiusion (1 < < 2) and ballistic motion ( = 2). Fractional derivatives are non-local operators, which makes it dicult to construct ecient numerical method. The subject of this dissertation is the time fractional diusion-wave equation with coecient which contains a singular distribution, primarily Dirac distribution, and its approximation by nite dierences. Initial-boundary value problems of this type are usually called interface problems. Solutions of such problems have discontinuities or non-smoothness across the interface, i.e. on support of Dirac distribution, making it dicult to establish convergence of the nite dierence schemes using the classical Taylor's expansion. The existence of generalized solutions of this initial-boundary value problem has been proved. Some nite dierence schemes approximating the problem are proposed and their stability and estimates for the rate of convergence compatible with the smoothness of the solution are obtained. The theoretical results are conrmed by numerical examples.
false
Termalno zračenje ostataka supernovih u radio-području
Thermal radiation of supernova remnants in radio domen
Evolucija ostataka supernovih vezana je za kretanje jednog, bezsudarnog udarnog talasa, nastalog prilikom prostiranja odbacenog zvezdanog materijala nakon eksplozije supernove, velikom brzinom kroz med-uzvezdanu sredinu. Teorijska razmatranja udarnih talasa su neophodna radi boljeg razumevanja emisionih maglina i, generalno, procesa u med-uzvezdanoj sredini. U okviru ove disertacije, izmed-u ostalog, predstavljeni su rezultati magnetohidrodinamicke (MHD) teorije udarnih talasa sa posebnim osvrtom na moguca resenja u svetlu modela idealne radijativne MHD za opticki gust slucaj. Specijalno, diskutovana su moguca svojstva udarnih talasa kada je dozvoljen skok adijabatskog indeksa i/ili parametra koji odred-uje odnos gasnog prema ukupnom pritisku. Osnovna hipoteza, razmatrana u okviru ove disertacije, odnosi se na mogucnost znacajnog ucesca termalnog zakocnog zracenja u neprekidnom radio-spektru pojedinih Galaktickih ostataka supernovih. Pozitivan odgovor na ovo pitanje bi mogao predstavljati prirodno objasnjenje blago konveksnih neprekidnih radio-spektara pojedinih Galaktickih ostataka koji se prostiru kroz sredinu koja je nesto gusca od prosecne. U tom smislu je vazno utvrditi postojanje indikatora termalnog ansambla elektrona, na dovoljno niskim temperaturama i dovoljno velikim gustinama, kako bi se zakocno zracenje moglo detektovati u radio-podrucju (blizina, interakcija ili prostiranje ostatka kroz molekulski oblak, prisustvo ohlad-enih elektrona koji su ranije zracili u X-podrucju, u fazama nakon Sedov-Tejlorove, detekcija niskofrekventnog krivljenja radio-spektra usled termalne apsorpcije vezane za sam ostatak, detekcija u H, identikacija radio-rekombinacionih linija vezanih za sam ostatak, itd). Znacajno prisustvo termalne komponente u principu moze objasniti i radiospektralne indekse manje od 0.5 odred-ene kod nekolicine evolutivno starijih ostataka (uglavnom mesane morfologije), koji se sire kroz guscu sredinu. Zapravo, u tom slucaju bi manji spektralni indeksi, pod pretpostavkom prostog stepenog zakona, predstavljali prirodnu manifestaciju znacajnog udela termalnog zakocnog zracenja u neprekidnom radio-spektru ostataka supernovih...
The evolution of supernova remnants is linked to the propagation of a collisionless shock wave, formed during the initial expansion of high-velocity supernova ejecta through the interstellar environment. Theoretical studies of shock waves are very important for the analysis of supernova remnants, as well as processes in the interstellar medium in general. In this doctoral dissertation, some theoretical results based on the magnetohydrodynamical theory of shock waves are presented, with special emphasis on ideal radiative magnetohydrodynamics for the optically thick case. Particularly, solutions for the case when jump in adiabatic index and/or ratio of gas to total pressure is allowed, are discussed. The main hypothesis of this dissertation is that thermal bremsstrahlung radiation at radio continuum frequencies can provide a signicant contribution in the case of several Galactic supernova remnants. This hypothesis can give a natural explanation for nearly concave up radio continuum spectra of several Galactic supernova remnants that are expanding in the environment with higher than average density. In this context, it is important to identify the existence of the possible indicators of ensemble of thermal electrons at suciently low temperatures and suciently high densities so that the thermal bremsstrahlung radiation linked to a particular remnant could be observed at radio continuum frequencies (vicinity, interaction or expansion through the molecular cloud, presence of the cooled thermal X-ray electrons during the post Sedov-Taylor phases, detection of low-frequency turnovers associated with thermal absorption linked to the remnant, detection in H, identication of radio recombination lines linked to the remnant, etc)...
false
Analiza prebrojivih modela potpunih teorija linearno uređenih struktura
An analysis of countable models of complete theories of linearly ordered structures
U ovoj tezi izučavamo linearno uređene strukture i njihove potpune teorije. Glavni tehnički alat koji koristimo u našoj analizi su kondenzacije, tj. razlaganje uređenja u konveksne delove i izučavanje količničke strukture i strukture delova. Uvodimo uniformno definabilnu kondenzaciju cδ koja razlaže uređenje u najveće konveksne delove čije su teorije prvog reda jednostavne: oni su ili gusta ili diskretna uređenja. Izučavamo cδ količničke strukture koje su ekspanzije određenih prostih prebrojivih diskretnih uređenja i dajemo detaljan opis onih koje imaju Kantor-Bendiksonov rang 1. Takođe koristimo kondenzaciju cδ da dokažemo da je svako linearno uređenje prošireno sa konačno mnogo unarnih predikata i relacija ekvivalencija sa konveksnim klasama interpretabilno u čistom linearnom uređenju. Uvodimo svojstva linearne i jake linearne binarnosti za linearno uređene strukture i njihove potpune teorije. U slučaju teorije, definicija opisuje osobinu grupe automorfizama njenog zasićenog modela. Dokazujemo da je svaka potpuna teorija linearnog uređenja sa unarnim predikatima i relacijama ekvivalencije sa konveksnim klasama jako linearno binarna. Glavni rezultat tvrdi da je jako linearno binarna struktura definiciono ekvivalentna linearnom uređenju sa dodatim unarnim predikatima i relacijama ekvivalencije sa konveksnim klasama. U dokazu dajemo opis definabilnih skupova proizvoljnog linearnog uređenja sa unarnim predikatima i relacijama ekvivalencije sa konveksnim klasama.
We study linearly ordered structures and their complete theories. The main technical tools used in the analysis are condensations, i.e. partitioning the ordering into convex parts and then studying the quotient structure and that of the parts. We introduce a uniformly definable condensation relation cδ that decomposes the ordering into largest convex pieces whose first order theory is simple: they are either dense or discrete orderings. We study cδ quotient structures that are expansions of certain simple countable discrete orderings and give a precise description of those having Cantor Bendixson rank 1. We also use the condensation cδ to prove that any linear ordering expanded by finitely many unary predicates and equivalence relations with convex classes is interpretable in a pure linear ordering. We introduce notions of linear and strong linear binarity for linearly ordered structures and their complete theories. In the case of a theory, the defining condition expresses a property of the automorphism group of its saturated model. We prove that any complete theory of a linear ordering with unary predicates and equivalence relations with convex classes is strongly linearly binary. The main result states that a strongly linearly binary structure is definitionally equivalent to a linear ordering with unary predicates and equivalence relation with convex classes added. In the proof we give a description of definable sets in any linear ordering with unary predicates and equivalence relations with convex classes.
false
Lokalno konačni varijeteti sa polu-distributivnom mrežom kongruencija
Locally finite varieties with semi{distributive congruence lattice.
Predmet ove disertacije je sintaksna karakterizacija kongruencijske polu{distributiv- nosti (u odnosu na inmum) lokalno konacnih varijeteta Maljcevljevim uslovima (posmatramo varijetete idempotentnih algebri). Dokazujemo da takva karakteri- zacija nije moguca sistemom identiteta koji koriste jedan ternarni i proizvoljan broj binarnih operacijskih simbola. Prvu karakterizaciju dobijamo jakim Maljcevljevim uslovom koji ukljucuje dva ternarna simbola: Lokalno konacan varijetet V zadovo- ljava uslov kongruencijske polu{distributivnosti (u odnosu na inmum) ako i samo ako postoje ternarni termi p i q (koji indukuju idempotentne term operacije) takvi da V zadovoljava: p(x; x; y) p(x; y; y) p(x; y; x) q(x; y; x) q(x; x; y) q(y; x; x). Ovaj uslov je optimalan u smislu da su broj terma, njihove visestrukosti i broj identiteta najmanji moguci. Druga karakterizacija koju dobijamo koristi jedan 4- arni simbol i data je jakim Maljcevljevim uslovom t(y; x; x; x) t(x; y; x; x) t(x; x; y; x) t(x; x; x; y) t(y; y; x; x) t(y; x; y; x) t(x; y; y; x) : Treca karakterizacija je data kompletnim Maljcevljevim uslovom: Postoje binarni term t(x; y) i wnu-termi !n(x1; : : : ; xn) varijeteta V tako za sve n > 3 vazi sledece: V j= !n(x; x; : : : ; x; y) t(x; y).
The subject of this dissertation is a syntactic characterization of congruence ^{ semidistributivity in locally nite varieties by Mal'cev conditions (we consider va- rieties of idempotent algebras). We prove that no such characterization is possible by a system of identities including one ternary and any number of binary opera- tion symbols. The rst characterization is obtained by a strong Mal'cev condition involving two ternary term symbols: A locally nite variety V satises congruence meet{semidistributivity if and only if there exist ternary terms p and q (inducing idempotent term operations) such that V satises p(x; x; y) p(x; y; y) p(x; y; x) q(x; y; x) q(x; x; y) q(y; x; x). This condition is optimal in the sense that the number of terms, their arities and the number of identities are the least possible. The second characterization that we nd uses a single 4-ary term symbol and is given by the following strong Mal'cev condition t(y; x; x; x) t(x; y; x; x) t(x; x; y; x) t(x; x; x; y) t(y; y; x; x) t(y; x; y; x) t(x; y; y; x) : The third characterization is given by a complete Mal'cev condition: There exist a binary term t(x; y) and wnu-terms !n(x1; : : : ; xn) of variety V such that for all n > 3 the following holds: V j= !n(x; x; : : : ; x; y) t(x; y).
false
Usmeravanje pretrage u automatskom dokazivanju teorema
Guiding search in automated theorem proving
U ovom radu se razmatra problem usmeravanja pretrage u automatskom dokazivanju teorema. Rad se sastoji od dva dela čija je dodirna tačka CDCL sistem pretrage, koji se intenzivno koristi kod modernih SAT rešavača. U prvom delu rada razmatran je problem jednostavnog usmeravanja pretrage | izborom rešavača, njegovih heuristika i njihovih parametara, a u zavisnosti od svojstava instance koju je potrebno rešiti. Osnova predloženih metoda za izbor algoritama je sintaksna sličnost formula koja se odražava na njihovu grafovsku strukturu. Ova sličnost je prvi put pouzdano ustanovljena i analizirana pomoću originalne mere sličnosti grafova (koja se pokazala korisnom i u drugim domenima). Praktični pristupi merenju sličnosti formula se zbog računske efikasnosti ipak zasnivaju na numeričkim atributima iskaznih formula. Predložene su dve jednostavne metode izbora algoritma zasnovane na algoritmu k najbližih suseda. Prva tehnika, ArgoSmArT se zasniva na klasifikaciji instance u jednu od unapred zadatih familija za koje su poznati algoritmi koji ih efikasno rešavaju. Instanca se rešava algoritmom koji odgovara familiji u koju je instanca klasifikovana. Druga tehnika, ArgoSmArT k-NN se zasniva na nalaženju nekoliko sličnih instanci u trening skupu za koje je poznato vreme rešavanja pomoću svih algoritama kojima sistem raspolaže. Instanca se rešava algoritmom koji se najbolje ponaša na pronađenim instancama. Tehnika ArgoSmArT je pogodnija za izbor konfiguracije SAT rešavača, a ArgoSmArT k-NN za izbor samog SAT rešavača. Tehnika ArgoSmArT k-NN se pokazala značajno efikasnijom od najvažnijeg i pritom vrlo složenog sistema za izbor SAT rešavača i sistema SATzilla. Pored problema izbora KNF SAT rešavača i njihovih heuristika, razmatran je i problem izbora ne-KNF SAT rešavača u kojem fokus nije bio na tehnikama izbora rešavača, pošto se predložene tehnike direktno primenjuju i na taj problem, već na atributima kojima se ne-KNF instance mogu opisati, a koji do sad nisu predloženi. Rezultati u ovom domenu su pozitivni, ali za sada ograničeni. Osnovni razlog za to je nedostatak veceg broja ne-KNF resavaca raznovrsnog ponasanja, sto ne iznenaduje s obzirom da je ova vrsta resavaca tek u svom povoju. Pored konstrukcije ekasnog sistema za izbor SAT resavaca, prikazana je i metodologija poredenja SAT resavaca zasnovana na statistickom testiranju hipoteza. Potreba za ovakvom metodologijom proizilazi iz velike varijacije vremena resavanja jedne formule od strane jednog SAT resavaca, sto moze dovesti do razlicitih redosleda SAT resavaca prilikom poredenja njihovih performansi ili rangiranja, sto je i eksperimentalno demonstrirano. Predlozena metodologija pruza ocenu statisti cke znacajnosti testiranja i ocenu velicine efekta, poput verovatnoce da jedan SAT resavac bude brzi od drugog...
In this thesis the problem of guiding search in automated theorem proving is considered. The thesis consists of two parts that have the CDCL search system, the system intensively used by modern SAT solvers, as their common topic. In the rst part of the thesis a simple approach to guiding search is considered | guiding by the selection of the solver, its heuristics, and their parameters, based on the properties of an instance to be solved. The basis of the proposed methods for algorithm selection is syntactical similarity of formulae which is re ected in their graph structure. This graph similarity is established and analyzed by using an original graph similarity measure (which turned out to be useful in other contexts, too). Yet, practical approaches to measuring similarity of formulae are based on their numerical features due to the computational complexity issues. Two simple methods for algorithm selection, based on k nearest neighbors, were proposed. The rst technique, ArgoSmArT is based on classication of instance in one of the predened families for which the ecient algorithms are known. The instance is solved by algorithm corresponding to the family to which the instance was classied. The second technique, ArgoSmArT k-NN is based on nding several similar instances in the training set for which the solving times by all considered algorithms are known. The instance is solved by the algorithm that behaves the best on those instances. ArgoSmArT technique is better suited for conguration selection of a SAT solver, and ArgoSmArT k-NN for SAT solver selection. ArgoSmArT k-NN technique showed to be more ecient than the most important and very complex system for SAT solver selection | SATzilla system. Apart from CNF SAT solver selection, the problem of non-CNF SAT solver selection is considered. The focus was not on solver selection techniques, since the proposed techniques are directly applicable, but on the attributes that can be used to describe non-CNF SAT instances, which have not been proposed earlier. The results in this domain are positive, but still limited. The main reason for that is the lack of greater number of non-CNF SAT solver of dierent behaviour, which is not surprising, having in mind that this kind of solvers is in its early stage of development. Apart from construction of ecient SAT solver selection system, the methodology of SAT solver comparison, based on statistical hypothesis testing is proposed. The need for such a methodology comes from great run time variations of single instance solving by a solver, which can result in dierent SAT solver orderings when one tries to compare their performance or rank them, as experimentally demonstrated. The proposed methodology gives the estimate of statistical signicance of the performed test and the estimate of the eect size, for instance the probability of a solver being faster than another...
false
Razvoj algoritma za ispitivanje pripadnosti asteroida familijama
Development of an algorithm for examination of the association of asteroids to families
Glavni cilj ove disertacije je unapredenje pristupa identikaciji clanova familija asteroida metodom hijerarhijskog grupisanja. U proceduru se dodaje novi korak koji omogucava smanjenje tzv. lancanog efekta. Uvedeni algoritam se sastoji od cetiri glavna koraka. U prvom se primenom metode hijerarhijskog grupisanja na inicijalnom katalogu sopstvenih elemenata odreduje preliminarna lista clanova. U drugom koraku, koristeci poznate zicke karakteristike potencijalnih clanova, a na osnovu denisanih kriterijuma, identikuju se uljezi. U trecem koraku, identikovani uljezi se izbacuju iz inicijalnog kataloga sopstvenih orbitalnih elemenata cime se dobija novi, modikovani katalog. Konacno, u cetvrtom koraku ponovo se primenjuje metoda hijerarhijskog grupisanja, ali ovaj put na modikovanom katalogu sopstvenih elemenata. Pokazano je da se na ovaj nacin znacajno smanjuje broj potencijalnih uljeza u familijama asteroida. Osim toga data je kriticka analiza dobijenih rezultata, kao i detaljan prikaz nekih specicnih slucajeva. Takode, razvijen je i onlajn portal (Asteroids Families Portal; AFP) koji omogucava dostupnost ove metode siroj naucnoj zajednici. Drugi cilj disertacije je odredivanje granice za primenu metode integracije unazad, kojom se procenjuje starost mladih familija asteroida. Taj cilj je postignut upotrebom numerickih simulacija evolucije test familije. Odredivanjem trenutka grupisanja sekularnih uglova i $, ustanovljena je linearna zavisnost izmedu dubine minimuma grupisanja i starosti familije. Prema dobijenim rezultatima, metoda integracije unazad moze se uspesno primeniti na familije koje nisu starije od 18 miliona godina.
The main goal of this dissertation is improvement of an approach for identi- cation of the members of asteroid families, based on the hierarchical clustering method. An additional step is introduced in order to reduce a so-called chaining eect. The introduced algorithm consists of four main steps. In the rst step, the hierarchical clustering method is applied to the initial catalogue of proper elements in order to obtain a preliminary list of family members. In the second step, using available physical properties, and based upon the dened criteria, the interlopers are identied. In the third step, objects identied as interlopers in the second step, are excluded from the initial catalogue of proper orbital elements, producing a modied catalogue. Finally, in the fourth step, the HCM analysis is performed again, but this time using the modied catalogue of proper elements. It is shown that in this way a number of potential interlopers among family members is signicantly reduced. Moreover, an on-line based portal (Asteroids Families Portal; AFP) to apply this procedure is developed, and is freely available to all interested researchers. The second goal of the dissertation is to determine the limitations of the backward integration method, used for estimation of ages of young asteroid families. This aim is achieved through numerical simulations of the evolution of a ctitious family. By determining instants of secular angles i $ clustering, a linear relationship is found between the depth of a clustering and the age of a family. According to the obtained results, the backward integration method could be successfully applied to families not older than 18 Myrs.
false
Data mining on protein sequences: n-gram analysis of ordered and disordered protein regions
n-gramska analiza uređenih i neuređenih regiona proteina
Proteini koji imaju neuređene regione učestvuju u velikom broju ćelijskih procesa kao što su prenos signala, transkripcija i remodelovanje funkcija hromatina. Sa druge strane, pojava takvih proteina je uočena kod osoba koje boluju od neuroloških i kardiovaskularnih bolesti, kao i različitih oblika maligniteta. Eksperimentalno određivanje neuređenih regiona protiena je vrlo skup i spor proces. Zbog toga su razvijeni i stalno se usavršavaju različiti računarski programi za predviđanje pozicija neuređenih regiona u proteinu. U radu je prikazana nova metoda za određivanje niski amino kiselina koje karakterišu neuređene i uređene regione proteina. Materijal nad kojim je vršeno istraživanje obuhvata 4076 virusa sa preko 190000 proteina. Metoda je zasnovana na ispitivanju osobina n-grama (koji obuhvataju ponavljajuće i palindromske niske) i njihove pripadnosti uređenim i neuređenim regionima proteina. Pozicije neuređenih /uređenih regiona u proteinima su određene korišćenjem tri programa za predviđanje. Osobine ponavljajućih niski koje su korišćene u istraživanju uključuju molske frakcije, frakcijske razlike i z-vrednost. Takođe, na ponavljajuće niske kao i na palindromske niske primenjene su određivanje pravila pridruživanja i klasifikacija, kao tehnike istraživanja podataka. Rezultati dobijeni svim tehnikama pokazuju visok nivo saglasnosti, za niske dužine manje od 6, dok nivo saglasnosti rezultata raste sve do maksimalnog sa porastom dužine niski. Visoka pouzdanost rezultata dobijenih tehnikama istraživanja podataka, pokazuje da postoje n-grami, kako ponavljajuće sekvence tako i palindromi, koji jednoznačno karakterišu neuređene/uređene regione proteina. Dobijeni rezultati su provereni upoređivanjem sa rezultatima zasnovanim n- gramima iz DisProt baze koja sadrži pozicije eksperimentalno verifikovanih neuređenih regiona proteina, i mogu da budu korišćeni kako za brzu lokalizaciju neuređenih/uređenih regiona u proteinima tako i za dalje poboljšanje postojećih programa za njihovo predviđanje. iv
Proteins with intrinsically disordered regions are involved in large number of key cell processes including signaling, transcription, and chromatin remodeling functions. On the other side, such proteins have been observed in people suffering from neurological and cardiovascular diseases, as well as various malignancies. Process of experimentally determining disordered regions in proteins is a very expensive and long- term process. As a consequence, a various computer programs for predicting position of disordered regions in proteins have been developed and constantly improved. In this thesis a new method for determining Amino acid sequences that characterize ordered/disordered regions is presented. Material used in research includes 4076 viruses with more than 190000 proteins. Proposed method is based on defining correspondence between n-grams (including both repeats and palindromic sequences) characteristics and their belonging to ordered/disordered protein regions. Positions of ordered/disordered regions are predicted using three different predictors. The features of the repetitive strings used in the research include mole fractions, fractional differences, and z-values. Also, data mining techniques association rules and classification were applied on both repeats and palindromes. The results obtained by all techniques show a high level of agreement for a short length of less than 6, while the level of agreement grows up to the maximum with increasing the length of the sequences. The high reliability of the results obtained by the data mining techniques shows that there are n-grams, both repeating sequences and palindromes, which uniquely characterize the disordered/ordered regions of the proteins. The obtained results were verified by comparing with the results based on n-grams from the DisProt database which contains the positions of experimentally verified disordered regions of the protein. Results can be used both for the fast localization of disordered/ordered regions in proteins as well as for further improving existing programs for their prediction. ii
true
Evolucija patuljastih galaksija u jatima galaksija male mase
Evolution of dwarf galaxies in the lower mass clusters of galaxies
U okviru ove disertacije će biti predstavljeni rezultati analize formiranja, rasprostranjenosti i evolucije patuljastih galaksija iz kosmoloških simulacija. Biće predstavljena prva kosmološka simulacija koja je u potpunosti izvedena i analizi- rana u Srbiji. Iz dobijenih rezultata je pokazano kako se formirala struktura na velikim skalama, pre svega kako se formiraju jata galaksija. Kako su jata galak- sija dominantne strukture u svemiru, koje se relativno sporo formiraju u odnosu na postojanje svemira, korišćenjem Ilustris simulacije, u kojoj je uključena kom- pletna fizika, biće pokazano da značajan deo jata galaksija nije u stanju dinamičke ravnoteže, kako neki autori tvrde, odnosno da su u procesu formiranja i danas. Dobijeni rezultati će biti iskorišćeni da se jata galaksija podele na tri tipa: jata u interakciji, jata koja interaguju sa manjim grupama (slabo interagujuća jata) i jata galaksija koja nisu u interakciji. Kako patuljaste galaksije po svom broju čine najveći broj galaksija u svemiru, a pogotovu u jatima, ovo postavlja pitanje kako trenutni procesi u kojima se jato nalazi utiču na strukturu i rasprostranjenost ra- zličitog tipa patuljastih galaksija. Nakon kratkog pregleda danas poznatih tipova patuljastih galaksija, pokazaćemo da postoji osetljivost rasprostranjenosti patulja- stih nepravilnih galaksija u odnosu na to da li je matično jato galaksija u interakciji ili ne. Kako se pretpostavlja da su patuljaste nepravilne galaksije prethodnici za formiranje najbrojnijih tipova patuljastih galaksija, ovo bi moglo da ukaže na do- minantan kanal dopremanja tih galaksija u jata galaksija gde se očekuje da dožive morfološku transformaciju. Na kraju će biti predstavljeno otkriće dva kanala for- miranja za kompaktne eliptične galaksije u jatima galaksija. Formiranje ovog tipa patuljastih galaksija je bilo predmet diskusije zbog ograničenja u posmatranjima i njihove kompaktne prirode iz koje nije moglo da se zaključi koji proces dovodi do njihovog formiranja. Ovde je pokazano da nastaju ili plimskim ogoljavanjem velikih spiralnih galaksija nakon njihovog upada u jato ili formiranjem direktno u jatu iz oblaka gasa koji je formiran najverovatnije akrecijom hladnog gasa iz okoline.
Within this dissertation results from the analysis of formation, abun- dance and evolution of the dwarf galaxies from cosmological simulations will be presented. We will present results from first cosmological simulation which was performed and analyzed in its entirety in Serbia. From the obtained results it was shown how the large scale structure formed, primarily the formation of clusters of galaxies. Since the clusters of galaxies are dominant structures in the universe, which are relatively slowly forming in relation to the existence of the universe, using the Illustris cosmological simulation, with included hydrodynamics, it will be shown that a significant part of the clusters of galaxies are not in the state of dynamic equilibrium, as some authors claim, therefore their formation process continues. The obtained results will be used to divide the clusters of galaxies into three types: interacting clusters, clusters interacting with smaller groups (weakly interacting clusters), and clusters of galaxies that are not interacting. As the dwarf galaxies constitute largest number of galaxies in the universe, notably largest num- ber in the clusters of galaxies, this raises the question how cluster’s interactions affect the structure and abundance of a diverse type of dwarf galaxies within them. After a brief review of today’s known types of dwarf galaxies, we will show that there is susceptibility of the abundance of dwarf irregular galaxies in relation to whether the parent cluster of galaxies is in interaction or not. As it is assumed that the dwarf irregular galaxies are predecessors for the formation of the most numerous types of dwarf galaxies, this could point to the dominant channel of de- livering these galaxies to clusters of galaxies where they are expected to undergo morphological transformation. In the end, the discovery of two formation channels for compact elliptical galaxies in the clusters of galaxies will be presented. The formation of this type of dwarf galaxies has been the subject of discussion due to limitations in observations and their compact nature from which it could not be concluded which process leads to their formation. It has been shown here that they occur either by tidal stripping of large spiral galaxies after their spiraling into the cluster of galaxies or forming directly inside clusters from the clouds of gas which is most probably formed by the accretion of cold gas from the environment.
true
Evolution, adaptation and speciation in Anthroherpon Reitter, a genus of subterranean Coleoptera
Evolucija,adaptacija i specijacija visoko evoluiranih pećinskih koleoptera roda Anthroherpon Reitter
Doktorska disertacija predstavlja studiju evolucije, adaptacije i specijacije u podzemnim staništima troglobiontnih tvrdokrilaca roda Anthroherpon. Pomenuti rod pripada tribusu Leptodirini (Leiodidae, Cholevinae), grupi koja je prošla intenzivnu diverzifikaciju u uslovima podzemnih staništa. Sve vrste pomenutog roda poseduju tipične troglomorfne osobine, kao što su: anoftalmija, apterizam, ekstremno izduženi telesni nastavci, glava i pronotum, i fizogastrija. Radi razumevanja evolucione istorije grupe, troglomorfne adaptacije su studirane u filogenetskom kontekstu. U analizi evolutivne radijacije roda Anthroherpon korišćena je datirana molekularna filogenija kao okvir za razumevanje diverzifikacije roda, evolucije troglomorfnih karaktera i rekonstrukciju predačkog areala. U svetlu novih nalaza predložena je nova taksonomska organizacija grupe.
The PhD research project focus on the study of evolution, adaptation, and speciation in the subterranean environment using troglobitic Coleoptera of the genus Anthroherpon as a model organism. Genus Anthroherpon belongs to the tribe Leptodirini (Coleoptera: Leiodidae: Cholevinae), a group that has undergone extensive diversification in the subterranean environment. All species of this genus have developed typical troglomorphic modifications: complete anophthalmy, apterism, extreme elongation of appendages, head, and pronotum, and physogastric elytra. To understand the evolutionary history of this group, the troglomorphic adaptations need to be studied in a phylogenetic framework. The thesis provide a comprehensive evolutionary analysis of the Anthroherpon radiation, using a dated molecular phylogeny as a framework for understanding Anthroherpon diversification, reconstructing the ancestral range, and exploring troglomorphic diversity. In light of these findings, a new taxonomical organisation of the group has been proposed.
false
Predlog podrške za spekulativno izvršavanje niti u smr procesorima
A proposal of support for thread level speculation in CMPs
U ovom radu se pošlo od pretpostavke da se resursi višejezgarnih (CMP) procesora, uz dodatnu hardversku i/ili softversku podršku za spekulativno izvršavanje niti (TLS), mogu iskoristiti i za izvršavanje danas preovlađujućih sekvencijalnih aplikacija. Međutim, pojedina postojeća rješenja CMP procesora sa TLS podrškom zahtijevaju korišćenje znatnih dodatnih hardverskih i softverskih resursa za obavljanje specifičnih operacija tokom spekulativnog izvršavanja sekvencijalnih aplikacija. Stoga je u ovom radu urađena opsežna i veoma detaljna komparativna analiza postojećih CMP sistema sa TLS podrškom pri čemu su posebno razmotreni elementi hardverske i softverske podrške za spekulaciju na registarskom i memorijskom nivou. Sa stanovišta podrške za registarsku komunikacije postojeći sistemi su klasifikovani po načinu organizacije skupa registara i načinu povezivanja, a nakon toga su podrobno analizirani registarski komunikacioni mehanizmi, tehnike oporavka nakon pogrešne spekulacije, performanse i skalabilnost. Klasifikacija sistema sa stanovišta spekulativne memorijske komunikacije je urađena prema organizaciji memorijske hijerarhije, a zatim je razmotrena složenost hardverske i softverske podrške za spekulaciju. Uočeni su neki potencijalni uzroci za degradaciju performansi kao što su neujednačen saobraćaj pri završetku niti i neprilagođen algoritam zamjene u keš memoriji. Na osnovu sprovedene analize postojećih spekulativnih CMP sistema, a u cilju ostvarivanja što boljeg odnosa cijena/performanse i otklanjanja uočenih neefikasnosti, u ovoj disertaciji je definisan predlog rješenja CMP procesora sa jednostavnijom TLS podrškom, u kojem se komunikacija između spekulativnih niti obavlja i na registarskom i na memorijskom nivou. Predloženi CMP sistem za spekulativno izvršavanje niti pripada grupi CMP sistema sa osnovnom generičkom arhitekturom nadgrađenom podrškom i za registarsku i za memorijsku komunikaciju. On se sastoji od četiri procesorska jezgra sa privatnim L1 keš memorijama za podatke i instrukcije povezanim zajedničkom magistralom sa zajedničkom L2 keš memorijom. Spekulativna komunikacija na registarskom i memorijskom nivou se obavlja preko posebnih zajedničkih magistrala na principima „snoopy“ protokola, dok se spekulativna paralelizacija sprovodi na nivou petlji tako da spekulativne niti odgovaraju pojedinim iteracijama petlje. Identifikacija niti u predloženom sistemu se vrši uz pomoć binarnog anotatora, posebne softverske komponente koja radi nad sekvencijalnim izvršnim kodom tako da se nije potrebno ponovno prevođenje izvornog koda. Pored toga, binarni anotator za potrebe protokola koji kontroliše registarsku komunikaciju vrši klasifikaciju instrukcija upisa u registre čije se vrijednosti tokom spekulativnog izvršavanja mogu mijenjati u petlji. Spekulativno proizvedene memorijske vrijednosti se drže u privatnoj...
In this research it was assumed that resources of chip multiprocessors (CMP), with additional hardware and/or software support for Thread Level Speculation (TLS) can be exploited for the execution of sequential applications which prevail today. However, some existing CMPs with TLS support require substantial additional hardware and software resources to perform specific operations during speculative execution of sequential applications. Therefore, an extensive and very detailed comparative analysis of existing speculative CMPs is performed focusing on hardware and software support for the speculation on both register and memory levels. According to the register level communication support, existing systems are classified by the organization of register file and interconnection topology. Then, a comparative analysis of registration communication mechanisms, misspeculation recovery techniques, and performance and scalability, are presented. According to the memory level communication support, existing systems are classified by the organization of the memory hierarchy, and then, the complexity of hardware and software support for speculation is discussed. Some potential causes for performance degradation such as burst traffic on thread commit and inappropriate replacement algorithm for cache memories are considered. Based on the analysis of existing speculative CMPs, in order to achieve a better cost/performance and to eliminate the observed inefficiencies, a CMP with a simpler TLS support, which enables the communication between speculative threads on both register and the memory level, is proposed in this dissertation. The proposed speculative CMP belongs to the systems with basically generic architecture enhanced with support for register and memory speculative communication. It consists of four processing cores with private data and instruction L1 caches which are connected by a shared bus to a shared L2 cache. The speculative communication at the register and memory level is done through corresponding shared buses following the principles of snoopy protocols. Speculative parallelization is implemented at the loop level and speculative threads correspond to the loop iterations. Speculative threads are identified using the binary annotator, the special software component that operates on a sequential executable code without need for source code recompilation. In addition, this tool annotates write instructions to loop-alive registers for the purpose of protocol that controls the register level communication. Speculatively produced memory values are kept in private L1 caches, while shared L2 cache stores the sequential state. Memory values can be transferred to the L2 cache only after a thread becomes non-speculative. The speculative thread can read data from non-speculative thread or some preceding or subsequent speculative thread. The proposed solution implements a support for dynamic identification of data dependence violations and recovery from misspeculation. Two protocols are proposed for register communication: the basic version – the SIC protocol, and its extended version - the ESIC protocol which is, unlike the SIC, based on speculative forwarding of potentially safe register values. Qualitative comparison of...
false
Metod za određivanje efikasnosti sistema aktivne katodne zaštite plovnih objekata od korozije
The method for determing the efficiency of the impressed current system of cathodic protection of the vessels from the corrosion
Korozija podvodnog dela trupa broda je jedan od izuzetno važnih problema u brodogradnji. Pored osnovnog načina zaštite dobijenog organskim premazima, postoji tendencija primene aktivne katodne zaštite pomoću spoljašnjeg izvora jednosmerne struje. Time se obezbeđuje znatno veći stepen zaštite čelika od korozije. U radu su navedeni faktori koji utiču na koroziju kao i njihovo međusobno dejstvo. Predstavljena je električne shema dvojnog električnog sloja i polarizacionih krivih, pri koroziji metala u morskoj i rečnoj vodi. Izvršeno je matematičko modelovanje anode, čelika koji se štiti, kao i sredine u kojoj se obavlja korozija. Prikazani su rezultati merenja oštećenja na brodu usled korozije a koji je proveo preko 20 godina u eksploataciji. Dati su osnovni podaci o sistemu aktivne katodne zaštite. Data su analitička rešenja nekih dvodimenzionalnih i trodimenzionalnih rešenja za idealizovane slučajeve. Prikazan je numerički metod za rešavanje eliptičkih parcijalnih diferencijalnih jednačina koji se koristi za računanje potencijalnog polja oko trupa broda, kao i numerički dobijeni rezultati raspodele potencijala na trupu broda sa jednom, ili obe elektrode. Navedeni su rezultati merenja zaštitnog potencijala na brodu dizalici koji se nalazi u bazenu brodogradilišta "Beograd". Prikazani su rezultati merenja prirodnog potencijala korozije, brzine porasta kao i pada zaštitnog potencijala. Na kraju je data mreža mernih tačaka sa izmerenim potencijalom trupa broda u ustaljenom stanju. Izvršeno je poređenje numeričkih i analitičkih rešenja sa eksperimentalno dobijenim podacima. Istaknut je praktičan značaj primene rezultata istraživanja datih u ovoj disertaciji. Posebno treba istaći činjenicu da je eksperiment izveden na realnom objektu i da su pri tome rešeni gotovo svi praktični problemi prilikom ugradnje i primene ovog sistema zaštite od korozije. Treba istaći anode i okoloanodne ekrane, referentne elektrode, ispravljače za ovu namenu, kablovske prolaze i drugo. Na osnovu ovog rada stvorena je podloga da se ovaj sistem može ugrađivati na realnim objektima i praktično koristiti.
The corrosion of the underwater part of the ship hull is one of the very important problems in shipbuilding. Beside the elementary way of protection, made by the organic paint, there is a tendency of applying the impressed current system (active cathodic protection) with help of exterior source of the direct current. There is a way to achieve the higher degree of protecting the steel from the corrosion. The factors that have an influence to the corrosion and their mutual action are presented in this paper. It is also presented an electric scheme of double electric layer an polar curve, by corrosion of the metal in sea and river water. It is made the mathematic model of the anode, the protecting steel and of the water. This paper includes also, the results of measuring the damages on the ship, which is more than 20 years in exploitation, caused by the corrosion. There are also given the elementary data about the impressed current system of cathodic protection. The analytical solutions of some two dimensional or three dimensional solutions for ideal cases are also presented. It is shown the numerical method for elliptical partial differential equations, which serves for calculating the potential field around the ship hull. The numerically received results of the potential distribution on the ship hill with one or both electrodes is also given. There are quoted also the results of measuring the protecting potential on the ship, located in the Shipyard "Beograd". The results of measuring the natural potential of the corrosion, the changing of the protecting potential are presented. It is shown the screen of the measuring potential of the ship hull in steady condition. The numerical and analytical solutions with experimentally received data are compared. It is emphasized the practical importance of the use of researching results, that are presented in this dissertation. The experiment is done on the real object and that way, almost all practical problems about the application of this system are solved. There are anodes, dielectric shield, reference electrode, rectifier for this use, cable coffer dam, etc. This paper shows that, based on the facts that are presented, this system could be built-in on the real objects and it could be practically used.
false
Robusna obrada signala u elektroenergetskim sistemima
Robust signal processing in electrical power systems.
Obrada signala u elektroenergetskim sistemima, a posebno robusni pristup u ovom segmentu, predstavlja značajnu i aktuelnu temu jer u okviru današnjih elektroenergetskih sistema implementacija savremenih tehničkih rešenja zahteva i uslovljava sve manje prisustvo ljudskog faktora kako bi se postigao maksimalan kvalitet u proizvodnji, prenosu i distribuciji električne energije i kako bi se minimizirali rizici od kvarova i havarija ne elementima i postrojenjima sistema. Da bi se obezbedila najveća raspoloživost u svim segmentima elektroenergetskog sistema, koja u značajnoj meri zavisi ne samo od primenjenih upravljačko računarskih sistema nego suštinski i od upotrebljene opreme u polju (primarna oprema u elektroenergetskim objektima, senzori, aktuatori) potrebno je raspolagati i sa unapređenim sistemima za ranu detekciju i prevenciju neželjenih režima i otkaza ključne primarne opreme da bi se predupredila teža havarijska stanja i kako bi se blagovremeno reagovalo na realne poremećaje u sistemu i da bi se na taj način sprečili veći zastoji u radu svih podsistema elektroenergetskog sistema. U osnovi pobrojanih zahteva i primenjenih metodologija i rešenja je kvalitetna obrada velikog broja veličina i signala, kako električnih tako i neelektričnih, koji egzistiraju u elektranama, prenosnoj elektroenergetskoj mreži i u elektrodistributivnom sektoru a za najrazličitije segmente primene: merenja, upravljanja, dijagnostike stanja primarne opreme, zaštite, ali i za veliki broj sistemskih aplikacija koje rade u on- line (u realnom vremenu) ili off-line modu. Disertacija je pokazala da robusna metodologija obrade signala u savremenim elektroenergetskim sistemima predstavlja optimalno, adekvatno i tehnički prihvatljivo rešenje čijom se primenom efikasno otklanjaju ograničenja i nedostaci klasične metodologije obrade signala u vremenskom domenu, u kontekstu imunosti na prisustvo loših merenja manifestovanim kroz impulsni šum i outlier-e, čiji se broj u današnje vreme rapidno povećava, koja su redovna pojava u praksi i eksploataciji visokonaponskih elektroenergetskih objekata. Klasične metodologije obrade signala u elektroenergetskim sistemima u vremenskom domenu, dominantno do sada primenjivane u inženjerskoj praksi, zasnovane su na normalnoj raspodeli šuma merenja što ih čini neefikasnim i praktično neprimenljivim u slučajevima prisustva impulsnih poremećaja i outlier-a. U disertaciji razvijeni adaptivni rekurzivni M- robusni algoritam pomenuta ograničenja klasičnih metodologija rešava polazeći od činjenice da šum merenja praktično podleže kontaminiranoj normalnoj raspodeli, odnosno normalnoj raspodeli sa otežalim krajevima koja u svom centralnom delu odgovara Gausovoj raspodeli a na krajevima podleže Laplasovoj raspodeli. Shodno tome predloženi adaptivni rekurzivni M-robusni algoritam je izveden iz principa maksimalne verodostojnosti kojim je efikasna estimacija, odnosno identifikacija, nepoznatih parametara signala realizovana maksimiziranjem uslovne funkcije gustine verovatnoće merenja. Robusnost predloženog rešenja je bazirana na adekvatno izabranoj funkciji uticaja koja se procenjuje na pogodno odabranim sekvencama merenja i kojom se klasifikuju i suzbijaju komponente loših merenja. Suboptimalnost predloženog rešenja je bazirana na usvajanju određene klase raspodele iz koje stohastički proces šuma merenja može poticati. Razvijena metodologija omogućava efikasnu obradu realnih signala i praktičnu implementaciju u okviru savremenih mikroprocesorskih uređaja široko rasprostranjenih u brojnim primenama u elektroenergetskim sistemima: monitoringu i dijagnostici stanja ključnih elemenata ovog sistema, ranoj detekciji kvara i opasnih radnih režima kao i otkaza na visokonaponskoj opremi, merenjima, nadzoru nad kvalitetom električne energije, brojnim upravljačkim funkcijama u nadređenim dispečerskim centrima različitih hijerarhijskih nivoa upravljanja. Implementacija predloženog adaptivnog rekurzivnog M-robusnog algoritma se pokazala u potpunosti kompatibilnom i svrsishodnom u okviru sigurnosnog koncepta upravljanja savremenim elektroenergetskim sistemima jer se razvijeni algoritam efikasno nosi sa pojavama velikih realizacija loših merenja, odnosno sa pojavama outlier-a i impulsnog šuma visokog intenziteta, i omogućava zadovoljavajući kvalitet u obradi signala što je imperativ u okviru proklamovanog koncepta upravljanja. Najvažniji ostvareni doprinosi u disertaciji su: validacija dokaza proisteklog na bazi analitičkog kritičkog pristupa stanovištu funkcionalne realizacije sigurnosnog koncepta upravljanja elektroenergetskim sistemima da tradicionalno korišćeni klasični estimatori parametara signala u ovom sistemu ne omogućavaju adekvatnu obradu signala za potrebe ovog koncepta u prisustvu grešaka merenja velikog intenziteta, koje se označavaju kao outlier-i, čija raspodela značajno odstupa od podrazumevane normalne raspodele; razvoj novog koncepta, koji omogućava efikasan tretman outlier-a; novi pristup u obradi signala u elektroenergetskim sistemima, u vidu robusne metodologije koncipirane na estimatoru tipa maksimalne verodostojnosti, kada imamo prisustvo apostrofiranih grešaka merenja velikog intenziteta čija raspodela značajno odstupa od podrazumevane normalne raspodele što je situacija koja se ima u najvećem broju realnih slučajeva, kojom se u potpunosti prevazilazi nedostatak klasičnih estimatora; razvoj i implementacija novog optimalnog adaptivnog rekurzivnog M-robusnog algoritma za obradu signala u elektroenergetskom sistemu u vremenskom domenu, koji efikasno rešava problem prisustva loših merenja kod kojih distribucija greškaka merenja značajnije odstupa od normalne raspodele; validacija efikasnosti i prednosti predloženog algoritma, u odnosu na klasične, do sada standardno korišćene algoritme za obradu signala u elektroenergetskom sistemu u vremenskom domenu; razvoj generalizovanog M-robusnog estimatora za identifikaciju parametara signala elektroenergetskog sistema kao multivarijabilnog sistema za potrebe sistemskih aplikacija koje su od praktičnog interesa za rad sistema u realnom vremenu; razvoj i validacija rutine za zaštitu sinhronih generatora velikih snaga, baziranoj na proračunu inverzne komponente struje, od svih vrsta asimetričnih radnih režima; razvoj i evaluacija robusnog pristupa za statičku estimaciju parametara sinhronih mašina kao nove dijagnostičke funkcije realnog vremena; razvoj i evaluacija specijalne rutine koja je bazirana na analizi zvuka ARMR algoritmom za dijagnostiku stanja energetskih transformatora; razvoj i evaluacija novog pristupa u koncipiranju funkcije usmerene prekostrujne zaštite i frekventne zaštite; razvoj i evaluacija novog pristupa u realizaciji funkcije distantne zaštite i lokacije mesta kvara na visokonaponskim vodovima; razvoj i evaluacija novog pristupa u detekciji kvarova preko visokoomske impedanse u elektroenergetskim sistemima, prevashodno u srednjenaponskim mrežama; razvoj i evaluacija novog pristupa u realizaciji funkcije diferencijalne zaštite energetskih transformatora velikih snaga.
Signal processing in electrical power system, especially the robust signal processing in this segment, presents significant and state-of-the-art topic since implementation of modern technical solutions in power systems requires less and less human involvement in order to achieve maximum quality in the power generation, transmission and distribution and to minimize the risk of faults and damage of power system elements and power substations. In order to achieve maximum availability in all segments of power system, which depends not only on the applied control systems but on the applied equipment (primary equipment in power systems, sensors, actuators) as well, it is necessary to have improved systems for the early detection and prevention of undesired regimes and main primary equipment failure in order to prevent difficult breakdowns and to react on time on real system disturbance, and all in order to prevent higher interrupts in operation of all power system subsystems. The essence of mentioned requirements and applied methodologies and solutions is high quality huge scale signal processing of values and signals, both electric and non-electric, which exist in power plants, transmission and distribution systems. High quality signal processing has various segments of application: measuring, control, primary equipment diagnostic, protection and for various system applications which works in on-line as well as in off-line mode. The dissertation has shown that robust methodology of signal processing in state-of-the-art power systems presents optimal, adequate and technically acceptable solution. Limitations and disadvantages of classical methodology of signal processing in time domain, in the context of immunity to the presence of bad measurements that are manifested through the impulse noise and outliers, are efficiently eliminated with this robust methodology. It is important to notice that the presence of impulse noise and outliers is constantly growing, and its occurrence is very common in today’s practice and exploitation of high voltage electrical power systems. Classical signal processing methodologies in electrical power systems in the time domain, which are mostly applied in the engineering practice, are based on normal distribution of measuring noise. This makes them non-efficient and practically non-applicable when impulse noise and outliers occur. Adaptive recursive M-robust algorithm presented in the dissertation resolves early mentioned limitations by acknowledging a fact that measuring noise represents contaminated normal distribution – normal distribution with weighted ends, which can be modeled by Gaussian distribution in the middle and Laplace distribution at its ends. Based on this fact, described adaptive recursive M-robust algorithm is derived from principle of maximal likelihood and provides efficient estimation and identification of unknown signal parameters by maximizing conditioned measurement probability density function. The robustness of the proposed solution is based on adequately chosen influence function which is evaluated with suitable selected sequences of measurements and which classify and restrain components of bad measurements. Sub-optimization of proposed solution is based on the assumption of certain classes of distribution from which stochastic process of measuring noise could be derived. Developed methodology enables efficient real signal processing and practical implementation within modern microprocessor devices widely installed in various applications in electric power systems: monitoring and diagnostic of vital power system elements, early fault and dangerous operation modes detection as well as high voltage equipment malfunction, measurements, power quality monitoring, various control functions in main control centers for different hierarchical levels of control. Implementation of the described adaptive recursive M- robust algorithm has been confirmed as compatible and practical within state-of-the-art electrical power system control security concept since developed algorithm is efficient in presence of high-intensity bad measurements – outliers and high-intensity impulse noise and, what is the most important, it provides sufficient signal processing quality. The most important contributions presented in this thesis are the following: Validation of argument, based on the analytical critical approach to functional realization of secure concept of electrical power system control, that commonly used classical estimators of signal parameters in this system don’t provide adequate signal processing when high-intensity measuring errors – outliers occurs; Development of the novel concept which treats outliers efficiently; Novel approach in signal processing in electrical power systems based on usage of robust methodology, conceptualized on maximal likelihood estimator, which resolves disadvatages of classical estimators in presence of high-intensity measuring errors (distribution of these errors significantly differs from assumed Normal distribution in the most of the real cases); Development and implementation of the novel optimal adaptive recursive M-robust algorithm for signal processing in electrical power systems in the time domain, which efficiently resolves the problem of the presence of bad measurements which have a distribution significantly different than Normal distribution; Validation of efficiency and advantages of the proposed algorithm compared to the classic ones which are commonly used for signal processing in electrical power system in the time domain up to now; Development of general M-robust estimator for signal parameters identification in electrical power system as one multivariable system for the purposes of system applications which are practically needed for the system operation in the real time; Development and validation of the routine for the high- power synchronous generator protection, based on negative sequence current calculation, for all kind of asymmetrical operating regimes; Development and validation of the robust approach for statical synchronous machine parameter estimation as novel diagnostic function in the real time; Development and validation of the special routine based on sound processing with ARMR algorithm for power transformers condition diagnostic; Development and evaluation of novel approach in conceptualizing directional overcurrent and frequency protection; Development and evaluation of novel approach in realization of distance protection and detection of fault location on high voltage lines; Development and evaluation of novel approach for high impedance fault detection, mostly in middle voltage networks; Development and evaluation of novel approach in realization of high-power power transformer differential protection;
true
Razvoj kompjuterske vizije za upravljanje funkcionalnom električnom stimulacijom
Development of computer vision based control of functional electrical stimulation.
Sistemi za funkcionalnu električnu stimulaciju (FES) se u terapiji gornjih ekstremiteta, kod osoba koje su usled oštećenja centralnog nervnog sistema uzrokovanog povredom kičmene moždine ili moždanim udarom delimično izgubili moć voljne kontrole pokreta, izdvajaju kao potencijalno najadekvatnija rehabilitaciona tehnika. Putem sistema za FES se električni impulsi odgovarajućih oblika (amplitude, trajanja i frekvencije) dovode do motornog sistema (mišića) koji su zaduženi za pokrete koje subjekat nije u stanju samostalno da uradi, sa ciljem da se datoj osobi omogući izvršavanje nekog funkcionalnog zadataka. Da bi to bilo moguće, sam FES sistem u sebi mora da uključi interpretaciju komandi pacijenta ili terapeuta, regulaciju stimulacionih parametara, generisanje strujnih impulsa i interfejs za prenos energije do senzorno-motornog sistema subjekta koristeći implantibilne ili površinske elektrode. Kada je reč o gornjim ekstremitetima funkcionalni zadaci koji se postavljaju pred subjekta podrazumevavaju najčešće manipulaciju predmetima, pri čemu korisnik bira tip hvata koji odgovara predmetu i kontroliše trenutke zatvaranja i otvaranja ruke kako bi obavio željenu funkciju. Definisanje i realizacija upravljanja kakvo bi subjektu pružio dovoljno fleksibilnosti uz odgovajući nivo jednostavnosti se i dalje smatra jedinim od otvorenih pitanja. U postojećim FES sistemima korisnik manuelno bira funkciju koristeći prekidač ili sličan periferal računara. Ovaj metod je nepogodan, jer je neophodno da se korinik koncentriše na funkciju koju želi da obavi, a ne na komandu koju treba da pošalje sistemu da bi obavio datu funkciju. U literaturi su prikazana idejna rešenja koja zamenjuju ovaj konvencionalni način komunkiacije i primenjuju neki oblik automatizacije. Rešenja su ponudjena kroz komande glasom, udisanjem i izdisanjem, primenom elektrofizioloških signala (elektromiografija - EMG), senzora pokreta i sile, međutim svi ovi sistemi su i dalje podrazumevali interakciju od strane subjekta koja ga zapravo kroz taj dodatni korak odvraća od zadatka koji je pred njega postavljen, a osim toga ometa pravilnu senzomotornu spregu pri obavljanju odabrane funkcije. Primena mozak-računar (brain-computer ili brain- machine) interfejsa je tema kojoj je u današnje vreme posvećena velika pažnja, a podrazumeva direktno prepoznavanje namere iz elektroencefalograskih (EEG) signala. Ovo, logično rešenje, medjutim, ne pruža odgovarajući nivo pouzdanosti kako bi sistemi postali deo kliničke prakse. Istraživanja u motornoj kontroli su pokazala da je prirodna kontrola bazirana na vizuelnoj percpeciji, i da na osnovu iskustva osoba bira način hvata i korišćenja objekta od interesa. Ova činjenia je našla primenu u istraživanjima u domenu proteze šake i ruke, i razvijeni su sistemi koji su zasnovani na pojednostavljenom modelu percepcije. Predmet istraživanja predstavljenog u ovoj tezi je unapređenje ove metodologije za automatsku kontrolu sistema viii za FES gornjih ekstremiteta zasnovanog na kompjuterskoj viziji. U skladu sa tim, definisane su postavke sistema za kompjutersku viziju koja koristi mogućnosti Kinect kamere za odabiranje tipa hvata, ali i povratnu spregu u toku stimulacije. Cilj istraživanja je razvoj sistema (od ideje do inovacije) koji bi se mogao primeniti u kliničkom okruženju za restoraciju hvata i to sa minimalnim angažmanom korisnika rehabilitacionog pomagala. U radu su prikazani novi algoritmi za procenu pozicije i orijentacije ruke, u realnom vremenu, na osnovu slike sa Kinect kamere i rezultati testiranja sistema na zadatku kontrole pozicije podlaktive u odnosu na nadlakticu (stimjulacija fleksora i ekstenzora lakta). Uveden je koncept artificijelne percepcije i razvijeni su algoritmi kompjuterske vizije na bazi kojih se prepoznaju predmeti koje subjekat može da uhvati, a zatim i procenjuje namera subjekta, kako bi se omogućio automatski izbor odgovarajućeg tipa hvata tokom FES terapije. Na kraju je na bazi razvijenih algoritama putem bežičnog komunikacionog interfejsa zatvorena sprega u sistemu za hvatanje asistirano FES-om. Omogućeno je da se parametri stimulacije automatski biraju kroz algoritam artificijelne percepcije, dok se na osnovu estimacije pozicije i orijentacije ruke pravovremeno pokreće adekvatna stimulaciona sekvenca za otvaranje ruke, odnosno formiranje hvata. Kroz testove na tri hemiplegična pacijenta potvrdjeno je da se korišćenjem predložene metodologije smanjuje vreme potrebno za formiranje hvata tokom vežbanja i samim tim skraćuje i olakšava proces rehabilitacije. Direktan rezultat istraživanja predstavljenih u ovoj tezi je sistem za kompjutersku viziju koji omogućava automatski izbor stimulacione sekvence i njenu pravovremenu aktivaciju prilikom vežbanja hvata asistiranog FES-om. Algoritmi koji su razvijeni tokom ovog istraživanja mogu se pokazati korisnim u raznim drugim aspektima rehabilitacionih tehnologija (npr. u proceni oporavka i kroz aplikacije proširene veštačke realnosti). Glavni naučni doprinos ostvaren kroz istraživanja je generisanje modela upravljanja oponašanjem bioloških sistema (sekvencijalna integracija artificijelne vizuelne percepcije i kompjuterske integracije internog i eksternog referetnog sistema). Osnovni rezultat primenjenih istraživanja je prenos rezultata bazičnih istraživanja u sistem koji kotisti elektrostimulaciju za neruorehabilitaciju.
Functional electrical stimulation (FES) based systems for the treatment of upper extremity in people who have, due to the central nervous system leasion, caused by spinal cord injury or stroke, to some extent lost the voluntary control offer great potential as a rehabilitation technique to improve the lost functionality. In a FES system electrical pulses of appropriate electrical parameters (shape, amplitude, duration and frequency) are transmitted to the motor system (muscles) that are responsible for the movements that the subject is not able to perform in a satisfactory manner, with an aim to enable the person perform some functional task with the help of this external muscle activation. In order to function, FES system needs to include a mechanism for interpretation of the patient or therapist commands, control of stimulation parameters, generation of adequate electrical pulses and an interface for transfer of electrical energy to the sensory-motor system of the subject through the implantable or surface electrodes. Considering the upper extremity treatment, the functional tasks the subject is performing are mostly related to reaching, grasping and manipulating with objects of daily living. In order to perform desired function with the help of electrostimulation, user must, therefore, select the adequate grasp type in respect to the object of interest and timely trigger the stimulation to perform hand opening and closing. The implementation of the control mechanism that would provide the user with the adequate flexibility and the desired level of simplicity is still considered an open question. In conventional FES systems user needs to select the grasp type and trigger the stimulation manualy via keybord, joystick or similar peripheral device. This method is inappropriate because the user should concentrate solely on performing the desired function and not on communicating the desired command to the stimulation system. A variety of methods for ease and automation of the control mechanism can be found in litterature. Control mechanisms based on voice commands, electromyographic (EMG) control, motion sensors and force transducers were developed and tested, however, all these systems still required interaction from the subject that considers additional steps which may distract the subject from the task at hand and could result with the development of the incorrect sensorimotor loop while practicing the desired function. The methodology of the brain-computer interface relys on identifying the intentions of the subject directly from electroencephalography (EEG) signals and, thus, represents the most appropriate solution to the problem. However, this solution still does not provide the adequate level of reliability in order to be used in daily clinical practice. Previous research in motor control has shown that natural control is based on visual perception, and that the choice of adequate grasp type and appropriate object interaction is based on experience. This fact was first exploited in the field of prostheses hands and arms, x where several systems that are based on simplified model of perception were developed. The main topic of the research presented in this thesis is the further improvement of this methodology to enable automatic control of the upper extremity FES system. With this in mind, we have defined the system setup for computer vision that guarantees easy integration of Kinect camera in the standard clinical environment and requires minimal interaction from the user. We developed novel algorithms for the real-time assessment of the position and orientation of the hand based on images from the Kinect camera and we have tested their work for closed loop control of hand position through electrical stimulation of the elbow flexor and extensor muscles (biceps and triceps). We have introduced the concept of artificial perception and developed computer vision algorithms for detection and identification and objects that the user can interact with, along with the estimation of the intention of the subject, with the aim to enable automatic selection of the appropriate grasp type during the FES therapy. Using the developed algorithms and the wireless communication interface we managed to close the loop in the system for FES assisted grasping. Automatic selection of stimulation parameters by the artificial perception algorithm and timely triggering of the chosen stimulation pattern based on the estimation of the position and orientation of the hand was implemented. Performed tests on three hemiplegic patients confirmed that the use of the proposed methodology reduced the time required for performing the grasping exercises and could thereby shorten and simplify the rehabilitation process. Computer vision system that enables automatic selection of the appropriate stimulation sequence and can timely activate stimulation during the FES assisted grasping exercise can be considered a direct result of the research presented in this thesis. Most of the developed computer vision algorithms that are result of this research and are in detail described in this thesis could prove useful in various other aspects of the rehabilitation technologies (e.g. assessment of recovery and the augmented reality applications). The main scientific contribution of the research described in this thesis is the design of the control models that are based on biological system and consider sequential integration of artificial perception. Application of this methodology for external control of electrostimulation in the neurorehabilitation system should be considered a significant outcome of applied research.
true
Sistem za interaktivnu proveru sličnosti konceptualnih i logičkih modela relacionih baza podataka
System for Interactive Similarity Comparison of Conceptual and Logical Relational Database Models
Razvoj konceptualnog i logičkog modela predstavlja osnovne korake u dizajnu relacione baze podataka. Postupak dizajna, iako metodološki dobro definisan, veoma često u praksi polazi od dvosmislene i nepotpune specifikacije zahteva. Rezultat toga je da ne postoji jedno najbolje rešenje, već često nekoliko različitih modela zadovoljava iste zahteve. Iz tog razloga, obrazovanje studenata, budućih inženjera, u oblasti baza podataka u velikoj meri zavisi od praktičnog rada na kome bi stekli potrebno znanje i veštine. Postojanje sistema koji bi bio podrška praktičnom radu, u vidu laboratorijskih vežbi, bilo bi od velike koristi studentima i predavačima. Osnovni problem u razvoju jednog takvog sistema jeste u veoma zahtevnim postupcima provere ispravnosti konceptualnih i logičkih modela. Ova disertacija predstavlja rezultate istraživanja i razvoja novog sistema, nazvanog ADVICE, koji služi kao podrška na kursevima iz baza podataka i pomaže studentima da savladaju razlike između teorije i prakse. U radu su takođe opisani novi postupci provere sličnosti konceptualnih i logičkih modela, razvijeni za potrebe ADVICE sistema. Upotreba sistema je predstavljena kroz skup laboratorijskih vežbi razvijenih tako da pomognu studentima da savladaju konceptualno i logičko modelovanje, SQL, formalne upitne jezike i normalizaciju. Koristeći ADVICE, studenti mogu da pristupaju bazama podataka, rešavaju konkretne probleme, a sistem im pruža podršku u vidu poruka o njihovim rešenjima. Iz perspektive predavača, sistem omogućava laku organizaciju laboratorijskih vežbi i kontinuirano praćenje napretka studenata. U radu su takođe opisana praktična iskustva prilikom upotreba sistema ADVICE na kursu iz baza podataka u toku trogodišnjeg perioda.
Development of conceptual and logical models represents basic design activities for a relational database. Design process in practice, although methodologically thoroughly defined, very often starts from an ambiguous and incomplete specification of requirements. For such specification, there is no one best solution, but often several different models meet the same requirements. That is the reason why, the education of students, future engineers in the field of databases, depends largely on a practical work on which they acquire the necessary knowledge and skills. The existence of a system that would be a support for practical work, in the form of laboratory exercises, would be of great benefit to both students and lecturers. The core problem in developing the system is that it depends on very demanding procedures needed for similarity comparison of both conceptual and logical models. The dissertation presents the results of a research that led to the development of a new system, named ADVICE, which can be used as a support on database courses to help students to bridge the gap between database management theory and practice. The dissertation also defines the new procedures, specifically developed for the ADVICE system, that are used for similarity comparison of both conceptual and logical database models. The usage of the system is presented through a set of laboratory exercises developed to teach students conceptual and logical modeling, SQL, formal query languages, and normalization. While working on the exercises, students use the system to access real databases, and the system provides them with feedback about their solutions. From the perspective of an instructor, the system allows easy exercise management and continual progress monitoring. The dissertation also describes a practical experience with the use of ADVICE on a database course over a three-year period.
true
Model za procenu efikasnosti izbora tehnologija digitalne televizije
Simulation model for digital broadcasting technologies efficiency estimation
Digitalizacija zemaljskog emitovanja televizijskih programa prva je velika promena u TV sistemima još od uvoĊenja kolor televizije. Prelazak sa analognog na digitalno emitovanje predstavlja jedan od najvećih izazova u razvoju video i audio tehnologije i desiće se u celom svetu u veoma bliskoj budućnosti. Pored ozbiljnih tehniĉkih zahvata, ovaj proces zahteva znaĉajne finansijske investicije vezane za enorman broj standardnih TV prijemnika i široko rasprostranjene televizijske mreţe. Zbog toga, gde god je moguće u distributivnoj mreţi, potrebno je obezbediti kompatibilnost sa postojećim sistemima. Nasuprot tome, digitalizacija video signala visoke rezolucije zahteva visoko sofisticirane algoritme sa velikom kompresijom i veoma efikasnom tehnologijom prenosa. Trenutno u svetu koriste se tri meĊunarodna standarda za digitalno zemaljsko emitovanje TV programa - Advanced Television Systems Committee (ATSC) u upotrebi u Severnoj Americi i Juţnoj Koreji, Digital Video Broadcasting Terrestrial (DVB-T) pored Evrope upotrebljava se u Australiji, Indiji i nekim zemljama Azije i Afrike, Integrated Services Digital Broadcasting Terrestrial (ISDB-T), koristi se u Japanu i Juţnoj Americi. Fokus ove disertacije je na analizi performansi i razvoju metoda za procenu efikasnosti tehnologija za fiksni prijem koje se koriste u Evropi. Za isporuku TV servisa krajnjim korisnicima postoje tri glavna sistema prenosa, satelitski, kablovski i preko zemaljskih predajnika. Digitalni zemaljski sistem je sposoban za isporuku TV programa ĉak i mobilnim korisnicima. Servisi prenosa podataka i televizije visoke rezolucije (HD) raspoloţivi su preko digitalnih satelitskih i kablovskih mreţa, ali sve do sada korisnici zemaljskih mreţa nisu primali ove servise jer je za njihov prenos potreban veći kapacitet kanala u odnosu na onaj korišćen za televiziju standardne rezolucije (SD) - Digital Video Broadcasting Terrestrial (DVB-T) i Moving Picture Expert Group MPEG-2 standard kompresije. Novije tehnologije, kao što su standard kompresije MPEG-4 verzija 10 i druga generacija standarda za zemaljsko emitovanje (DVB-T2), obezbeĊuju povećan kapacitet i robusnost u zemaljskim mreţama. MeĊutim, pored tehniĉkih prednosti evidentno je da najnovije tehnologije sa sobom donose i više cene opreme i sistema. Sprovodljivost poslovnog plana uvoĊenja digitalnog emitovanja je pod direktnim uticajem ovih tehno-ekonomskih pitanja.
The first major change within the TV system since the introduction of color is the digitalization of television terrestrial broadcast network. The digital switchover is one of the main challenges in video and audio technology developments and it is going to happen in the very near future throughout the world. Besides the severe technical requirements, this process is driven by significant financial investments, regarding to an enormous number of standard TV sets and widespread broadcast networks. Consequently, it is necessary, wherever is possible in distribution network, to ensure the compatibility with existing systems. On the contrary, digitization of high-resolution video and audio signals needs highly sophisticated algorithms with high compression and very efficient transmission technology. There are three international standards currently used world-wide for digital terrestrial broadcasting - Advanced Television Systems Committee (ATSC) used in North America and South Korea, Digital Video Broadcasting Terrestrial (DVB-T) used apart from Europe in Australia, India and some other Asian and African countries, Integrated Services Digital Broadcasting Terrestrial (ISDB-T), used in Japan and South America. The focus of this dissertation is the performance analysis and efficiency estimation method applied to the technologies for fixed reception used in our region – Europe. The three main transmission systems for delivering TV service to end-users are satellite, cable and terrestrial networks. The digital terrestrial television (DTT) system is capable of distribution of television content even to the mobile devices. Data-casting services and high-definition (HD) television content is available via digital satellite and digital cable networks but, until now, terrestrial users have not received data and HD content because its transmission requires a higher channel capacity than that used for standard definition (SD) TV - Digital Video Broadcasting Terrestrial (DVB-T) and Moving Picture Expert Group MPEG-2 coding and compression standard. The latest technologies, like compression standard MPEG-4 part 10 and the second generation of digital terrestrial broadcasting standard (DVB-T2), provide increased capacity and ruggedness in the terrestrial transmission networks. But besides technical benefits it is evident that the latest technologies bring also a higher system cost. The viability of digital broadcasting business case is directly influenced by this techno-economic issue.
false
Primena multifraktalne analize mikroskopskih slika u klasifikaciji intraosealnih metastatskih karcinoma
Application of multifractal analysis on microscopic images in the classification of metastatic cancer intraoseal
mikroskopskih medicinskih slika s ciljem njihove klasifikacije, u smislu utvrđivanja primarnog karcinoma u slučajevima intraosealnih metastatskih karcinoma. Dosadašnja saznanja ukazuju da postoji razlika u morfologiji zdravih i kancerogenih ćelija. Prethodni rezultati, izloženi u magistarskom radu kandidatkinje „Klasifikacija digitalnih medicinskih slika primenom multifraktalne analize“ [ANDJ06], i primenjeni u klasifikaciji slika karcinoma kolona, otvorili su mogućnosti za dalja istraživanja i moguće primene multifraktalne analize na mikrofotografije uzoraka drugih tkiva. S obzirom da su, u slučaju različitih vrsta primarnih karcinoma, ćelije često neregularnog ("haotičnog") oblika, utvrđivanje eventualne korelacije između multifraktalnih parametara koji opisuju oblik ćelija i primarnih karcinoma, moglo bi imati veliku primenu. Očekuje se da multifraktalna analiza omogući definisanje objektivnih pokazatelja kojima se može izvršiti klasifikacija tkiva. Multifraktalna analiza u klasifikaciji primarnog karcinoma, kao pomoćni alat, mogla bi ubrzati proces odlučivanja i doprinela smanjenju subjektivnog faktora i verovatnoće greške pri dijagnozi.
The aim of this study was to determine statistically significant differences between microscopic medical images in order of their classification in terms of determining primary cancer in cases of metastatic carcinoma intraoseal. Past findings indicate that there is a difference in the morphology of healthy and cancerous cells. Preliminary results, presented in the magisterium thesis of the candidate, „Classification of digital medical images using multifractal analysis“ [ANDJ06], and applied to classify images of colon cancer, opened the possibility for further investigation micrographs of samples of other tissues. Since, in the case of different types of primary cancer cells are often of irregular ("chaotic") forms, determining the possible correlation between multifractal parameters describing the shape of primary cells and the cancer, could have broad application. It is expected that multifractal analysis enables the definition of objective indicators that can performed tissue classification. In the classification of primary tumors, the multifractal analysis, as an assistant tool, could accelerate the decision-making process and contribute to reduction of subjective factors and error probability in medical diagnosis.
true
Uprošćavanje modela elektroenergetskih sistema u analizama pouzdanosti
Approximation of power systems in the adequacy evaluation.
Ovaj rad prikazuje DC mrežni ekvivalent za uprošćavanje interkonektivnih sistema u analizama pouzdanosti posmatranog elektroenergetskog sistema. Procena pouzdanosti složenih interkonektivnih elektroenergetskih sistema često je težak zadatak zbog velikog broja stanja ispada koje treba analizirati u proračunima. Potrebno računarsko vreme može biti značajno smanjeno ako se neki delovi elektroenergetskog sistema uproste ekvivalentima sa ograničenim brojem stanja koja predstavljaju samo ona najuticajnija stanja sa aspekta pouzdanosti sistema. Rad predlaže jedan alternativni DC model koji ima jedinstven izvorni čvor koji je dobijen paralelnim povezivanjem svih izvornih čvorova aproksimiranog dela sistema. Takav pristup zadržava osnovnu strukturu modelovanog sistema. Ekvivalentna mreža je formirana paralelnim povezivanjem svih izvornih čvorova kako bi se dobio jedan izvorni čvor. Mreža je oformljena na takav način da je transformisana u mrežni ekvivalent koji sadrži samo ekvivalentni izvorni čvor i interkonetivne čvorove prema sistemu koji se ispituje. Takođe je razmatrana mogućnost dodatne redukcije prostora stanja koja se zasniva na ugovorima o međusobnoj havarijskoj podršci između razmatranog sistema i aproksimiranog dela sistema i na ograničenju prenosnog kapaciteta interkonektivnih vodova. Prekoračenja u uprošćenom ekvivalentiranom sistemu vezana za simultane ispade proizvodnih jedinica i prenosnih vodova su modelovana preko jednog ekvivalentnog stanja izvora i preko ekvivalentiranja admitansi ekvivalentne mreže prema verovatnoćama ovih prekoračenja. Na ovaj način je obuhvaćena redukcija viškova energije obezbeđena preko interkonektivnog sistema zbog ispada samo vodova ili simultanih ispada vodova i proizvodnih jedinica. Ekvivalentna mreža formirana na ovaj način ima jednostavnu strukturu i redukovan broj stanja u poređenju sa originalnom mrežom, čime se smanjuju računski problemi u proceni pouzdanosti sistema od interesa. Primena predloženog modela je demonstrirana na dva poznata sistema, i to na Garverovom i IEEE RTS sistemu, kao i na realnim elektroenergetskim sistemima Srbije, Crne Gore i Albanije. Kao što može da se vidi iz prikazanih rezultata, predstavljanje interkonektivnih sistema ekvivalentom daje praktično iste rezultate kao kod potpunog modelovanja sistma i za parametre pouzdanosti po potrošačkim čvorovima sistema i za sistemske parametre celokupnog sistema. Jedini izuzeci u ovom pogledu su potrošački čvorovi u neposrednoj blizini graničnih čvorova. Ukoliko se koristi ekvivalentni model interkonektivnog sistema, računsko vreme je samo nekoliko procenata od vremena kada su sistemi predstavljeni u potpunosti. Ovo je rezultat redukcije broja komponenti sistema i broja stanja sistema postignutih aproksimacijama interkonetivnog sistema. DC mrežna aproksimacija interkonektivnih sistema prikazana u ovom radu ima jednostavnu strukturu i smanjen broj stanja u poređenju sa orginalnom mrežom. Ovo značajno smanjuje proračunske probleme u proceni pouzdanosti sistema od interesa. Kvalitet predloženog pristupa je uspešno dokazan na test primerima i na realnim sistemima.
This paper presents a DC equivalent network for the approximation of interconnected systems in the adequacy evaluation of the system under consideration. The evaluation of adequacy of composite interconnected power systems is in many cases a cumbersome task because of the enormity of failure states that should be considered by enumeration. The computation time can be significantly reduced if some portions of the power system are approximated by equivalents with a limited number of states that reflect only the most effective adequacy performances of these portions. This paper suggests an alternative DC model having a single source obtained by parallel connecting of all sources of the approximated system portion. Such an approach retains the basic structure of the modeled system. The equivalent network is built by parallel connecting all source buses to obtain a single source bus. The network constructed in such a way is than transformed in an equivalent network containing only equivalent source bus and the interconnection buses to the system under investigation. The possibility of the additional reduction of the state space, based upon the agreements between the approximated portion and the remaining system concerning the mutual emergency support, and due to the limited loading capacity of interconnection lines, is also considered. The contingencies in the approximated system associated with simultaneous outages of generating units and transmission lines are modeled by a single equivalent source state and by modifying the admittances of the equivalent network in accordance with the probabilities of these contingencies. In such a way, the reductions of surplus power provided by interconnected system due to the outages of lines alone or in combination with generating units are encompassed. The equivalent network formed in such a way has a simple structure and a reduced number of states when compared to the original network, which considerably reduces the computational burden in the evaluation of the adequacy of the system under consideration. The application of the proposed model is demonstrated on two well-known systems Garver, IEEE RTS and real electrical systems of Serbia, Montenegro and Albania. As can be seen from presented results, the equivalent representaion of interconnected systems has provided practically the same results as the complete model of this system both for the adequacy indices of load nodes and for the total indices for the entire system. Load node being in the closest vicinity of tie node, appears as a single exception in this regard. The calculation time if the equivalent model of interconnected systems is used was only few proccent of this time with comlplete representaion of the systems. This was due to the reduction of the number of system components and system states achieved by the approximation of the interconnected system. The DC network approximating interconnected systems presented in this paper has a simple structure and a reduced number of states when compared to the original system. This considerably decreases the computational burden in the evaluation of the adequacy of the system of interest. The quality of the proposed approach has been succesfully verified for test system examples and real electical systems.
false
Model implementacije servisno orijentisanih mreža pametnih pretvarača
Implementation model for service oriented smart transducers networks
Pametni pretvarač je uređaj koji pored pretvaračkih elemenata i pratećih kola za obradu signala, poseduje i lokalnu inteligenciju i mogućnost komunikacije. Veliki broj načina na koji se takav uređaj može realizovati, dovodi do nekompatibilnosti i niskog nivoa interoperabilnosti između uređaja različitih proizvođača, zbog čega je uveden standardni model pametnog pretvarača definisan familijom standarda IEEE 1451. Dodatno unapređenje interoperabilnosti se postiže nadogradnjom standarda, kroz uvođenje posebnih Web servisa za komunikaciju sa pametnim pretvaračima. Analiza postojećih rešenja je pokazala da i pored uvedene standardizacije ne postoji opšti model mrežne konfiguracije i komunikacije. Zbog toga je u ovoj tezi razvijen model implementacije koji omogućava uniformnu konfiguraciju svih mrežnih čvorova od strane centralnog serverskog čvora i koji uvodi dva obrasca komunikacije: (i) komunikaciju posredstvom centralnog servera i (ii) obrazac zasnovan na upotrebi servisnih agenata. Servisni agenti su uvedeni kao aktivne komponente, zadužene za transfer podataka između pasivnih komponenata sistema kojima se pristupa putem servisnih interfejsa. Servisni agenti se mogu nalaziti na bilo kom mrežnom čvoru, čime se eliminiše centralizovani model komunikacije i omogućava proizvoljna mrežna topologija. Za potrebe integracije entiteta koji nisu konkretni pametni pretvarači, poput algoritama za obradu podataka i ulazno/izlaznih uređaja, predstavljen je koncept virtuelnog pretvaračkog modula. Dati mrežni entiteti u formi virtuelnog pretvaračkog modula se na nivou interfejsa vide isto kao i konkretni pametni pretvarači, što omogućava uniforman pristup od strane centralnog menadžerskog čvora i servisnih agenata. Verifikacija modela je data putem studija slučaja mreža za osmatranje parametara okoline, procenu uslova za pojavu leda, predikciju signala upotrebom neuralnih mreža i kontrole temperature sušare. Zaključeno je da predloženi model ima praktičnu primenu, pri čemu je pokriven širok spektar mreža pametnih pretvarača, podržava upotrebu i drugih funkcionalnosti koje nisu date IEEE 1451 specifikacijom i različite komunikacione protokole, što ga čini pogodnim za dalji razvoj servisno orijentisanih mreža pametnih pretvarača.
Smart transducer is a device possessing the local intelligence and communication capability, apart from the basic transducer elements and accompanying signal processing circuits. A large number of different design possibilities leads to the incompatibility and poor interoperability between devices of different manufacturers, which are solved by the smart transducer model established by the IEEE 1451 family of standards. Additional improvement of interoperability is achieved by an upgrade of the standard, through introduction of the customized Web services for communication with smart transducers. An analysis of existing solutions showed that the general model of network level configuration and communication, actually, does not exist, although the standardized smart transducer model does. Because of that, this work establishes an implementation model which enables uniform configuration of all network nodes by the central server node, and introduces the following two communication patterns: (i) communication over the central server intermediary and (ii) communication based on the service agents deployment. Service agents are introduced as the active components that transfer data between the passive system components accessed through service interfaces and can be deployed to an arbitrary network node, which eliminates the centralized communication model and enables arbitrary network topology. In order to integrate the entities which are not physical smart transducers in the network, like processing algorithms and input/output devices, concept of the virtual transducer interface module is introduced. The given network entities in form of the virtual transducer interface module are seen at the interface level the same as the physical smart transducers, and are accessed uniformlly by the central manager node and service agents. Model verification is given through the case studies of networks for environmental monitoring, ice formation condition estimation, neural network signal sequence prediction, and dryer chamber temperature control. It is concluded that the proposed model is applicable to a wide range of smart transducers networks, it enables functionalities other than IEEE 1451 and different communication protocols, which makes it suitable for further development of the service oriented smart transducers networks.
true
Razdvajanje faza na nanoskali u superprovodnicima na bazi gvožđa korišćenjem Ramanove spektroskopije
Nanoscale phase separation in iron-based superconductors investigated by Raman spectroscopy.
U okviru ovog rada izlozeni su rezultati istrazivanja vibracionih svojstava materijala iz grupe superprovodnika na bazi gvozda i niskodimenzionalnih magnetnih materijala, metodom Ramanove spektroskopije. Ova spektroskopska tehnika omogucava i analizu uticaja strukturnih, elektronskih i magnetnih svojstava i njihovih promena na dinamiku resetke kristalnih materijala. Eksperimentalni rezultati podrzani su numerickim proracunima dinamike kristalne resetke...
In this work the results on vibrational properties of unconventional iron-based superconductors and low dimensional magnetic materials, by means of Raman spectroscopy, are given. This spectroscopic technique allows also an analysis of the impact of structural, electronic and magnetic properties and their changes on lattice dynamics of crystalline materials. Experimental results are supported by the numerical calculations of lattice dynamics...
false
Electroencephalographic control signals of brain-computer interface in neurorehabilitation
Електроенцефалографски сигнали за управљање рачунарским интерфејсом у неурорехабилитацији
Mozak-računar interfejs (MoRI) sistemi mogu iskoristiti karakteristične promene moždane aktivnosti korisnika kao kontrolne signale uređaja (računara). Različiti mentalni zadaci ili spoljašnji stimulusi (vizuelni, auditivni ili somatosenzorni) indukuju promene koje su kodirane u spontanoj neuralnoj aktivnosti. Generisane promene se mogu identifikovati merenjem moždanih signala koji predstavljaju direktnu ili indirektnu meru električne aktivnosti mozga...
Brain Computer Interface (BCI) systems can use characteristic brain neural alterations as control signals of the device/computer. Various mental tasks or external stimulation (visual, auditory or somatosensory) induce changes which are embedded in the spontaneous neural activity. Generated changes can be extracted and identified from the brain-signal recordings that represent the (direct or indirect) measure of electrical neural activity...
false
Mikrotalasni filtri nepropusnici opsega sa rezonantnim pregradama u E- i N-ravni talasovoda
Microwave bandstop filters using resonant inserts placed in Е-plane and H-plane of waveguide
Predmet proučavanja ove disertacije predstavljaju minijaturizovani filtri nepropusnici opsega učestanosti koji koriste rezonantne pregrade u E i N-ravni pravougaonog talasovoda. U disertaciji se polazi od osnovne teorijske pretpostavke, da se umetanjem štampanog diskontinuiteta u pravougaoni talasovod može realizovati filtar nepropusnik opsega učestanosti. U skladu sa tim, diskontinuiteti u vidu štampanog kola se pozicioniraju u E- ili u N-ravan talasovoda. Za bazične elemente štampanih pregrada su izabrani četvrttalasni i polutalasni rezonatori. Predložena tema doktorske disertacije ima značajno utemeljenje u pogledu rešavanja problema minijaturizacije mikrotalasnih komponenti koje predstavljaju sastavni deo komunikacionih sistema. Usled neprekidnog razvoja komunikacionih servisa, dolazi do porasta eksploatacije spektra gde je potrebno efikasno izdvojiti korisni signal od šuma. Prednosti filtara, implementiranih u tehnici talasovoda u odnosu na planarne, se ogledaju u faktoru dobrote i snazi, a nedostaci u veličini komponente. Obzirom da nalaze značajnu primenu u sistemima gde su potrebne velike snage i mali gubici, projektovanje ove klase filtara predstavlja zahtevan proces jer mora da istovremeno zadovolji više kriterijuma. Cilj istraživanja je usmeren na projektovanje kompaktnih talasovodnih filtara višeg reda, sa više nepropusnih opsega učestanosti. Rezonantne pregrade se postavljaju u E- ili u N-ravan pravougaonog talasovoda. Uporedo, istraživanje se bavi razvojem generalizovanih ekvivalentnih električnih šema za svaki od ova dva tipa filtara. Razvoj ekvivalentnih šema je svojevrstan doprinos projektovanju filtara, obzirom da se pomoću simulacija kola može pratiti uticaj izbora vrednosti parametara kola na odziv filtra, čime se v omogućava optimizacija filtra a da pritom simulacije nisu vremenski i računarski zahtevne kao trodimenzionalne elektromagnetske simulacije. Prilikom razvoja ekvivalentne šeme, poseban naglasak je stavljen na modelovanje spregnutih rezonatora i simboličko određivanje parametara sprege. Naučni doprinos doktorske disertacije su novi metodi projektovanja kompaktnih mikrotalasnih filtara višeg reda sa više nepropusnih opsega učestanosti sa štampanim rezonantnim pregradama postavljenim u E-ravni ili N-ravni pravougaonog talasovoda. Okosnica predloženog metoda projektovanja filtara sa više nepropusnih opsega je činjenica da ne postoji neželjena sprega između rezonatora koji učestvuju u realizaciji specificiranih opsega. Osnovni koncept postizanja kompaktnosti je meandriranje štampanih vodova rezonatora u cilju smanjenja površine koju zauzimaju na dielektričnom nosaču. U pogledu postizanja kompaktnosti filtara, višeg reda sa više nepropusnih opsega, tehnike minijaturizacije se primenjuju u zavisnosti od ravni u kojoj se pregrada pozicionira. U cilju smanjenja dužine filtra, sa pregradom u N-ravni pravougaonog talasovoda, predlaže se minijaturizacija invertora. Predstavljene kompaktne implementacije karakterišy se nezavisnom kontrolom svakog od projektovanih opsega. Uporedna analiza karakteristika predloženih novih realizacija sa karakteristikama rešenja objavljenih u literaturi predstavlja značajan deo ove disertacije, neposredno ilustrujući njen doprinos. Dodatni doprinos se ogleda u unapređenju karakteristika filtara propusnika opsega učestanosti. Tradicionalna implementacija koja koristi spregnute polutalasne rezonatore je značajno poboljšana upotrebom dodatnih četvrttalasnih rezonatora. U zavisnosti od toga da li je rezonantna učestanost četvrttalasnog rezonatora podešena na vrednost iz propusnog opsega filtra ili van njega, moguće ostvariti različite funkcionalnosti – bolje potiskivanje u nepropusnom opsegu filtra i podelu širokog propusnog opsega na uže podopsege.
The scope of the study of the dissertation are miniaturized bandstop filters based on resonant inserts placed within E-plane or H-plane of rectangular waveguides. Starting point is the hypothesis that bandstop filter can be designed by placing discontinuities within rectangular waveguide. Accordingly, printed-circuit inserts are positioned in E- plane or H-plane of waveguide. Quarter-wave resonators and half-wave resonator are chosen as fundamental elements of considered inserts. The topic of the dissertation has solid foundation in miniaturization of microwave components problem solving. Since the rapidly developing communications services, frequency spectrum is extensively exploited imposing efficient filtering as a requirement, so useful signals are distinguished from noise and unwanted frequency components. The advantages of the waveguide filters compared to planar filters are in quality factor and power handling capability, and major disadvantage is the size of the waveguide component. Being core devices in systems where low losses and high power levers are required, waveguide filter design is demanding process since several criteria should be satisfied. The aim of the research is focused towards compact waveguide high order filter design with multiple stopbands. Resonating inserts are positioned in E-plane or H-plane of rectangular waveguide. Simultaneously, the research is oriented towards the development of generalized equivalent electrical circuit of the multiband bandstop filter. Development of the equivalent electrical circuit represents contribution intended for fast filter design since circuit simulations provide insight how circuit parameters affect filter response thus enhancing filter optimization, while simulations are not time and memory consuming as three-dimensional electromagnetic simulations. The main focus is towards the modelling of coupled resonators and symbolic analysis of coupling parameters. The scientific contribution of the dissertation are novel design methods for compact microwave higher order E-plane or H-plane waveguide filters with multiple viii independently controlled stopbands. The core of the proposed method for design compact higher order filter are multiple stopbands that are independently tuned, since there is no unwanted coupling between resonators intended for different stopbands. The concept behind miniaturization relies on the technique of segmentation of straight-line resonator strip based on folding, so that resonator occupies the smallest possible area. Several miniaturization techniques are considered regarding compactness of higher order filter with multiple stopbands. They extend the basic concept and are applied in accordance with the type of the filter to be implemented, whether it is E-plane or H-plane filter. With the aim of reducing the length of the H-plane filter, strategy of inverter miniaturization based on replacement of the conventional quarter-wave inverter by the shorter one is presented. Presented compact solutions feature flexible control of specified stopbands. Comparative analysis of proposed solutions and corresponding filters found in literature is conducted and exposed. Additional contribution is made regarding E-plane bandpass waveguide filter design. Traditional implementation which use coupled half-wave resonators is outperformed by using quarter-wave resonators that exhibit bandstop characteristics. Depending whether the resonant frequency of the quarter-wave resonator is located within passband or stopband of the bandpass filter, it is possible to achieve different functionalities – advanced stopband attenuation and division of wide band to narrow sub- bands.
true
Primena tehnike formiranja slika proširenog dinamičkog opsega u monitoringu loše osvetljene scene
High dynamic range imaging implementation in scene monitoring under bad illumination
Unapređenje kvaliteta slike širenjem dinamičkog opsega u poslednje vreme se intenzivno koristi. Ovo za posledicu ima prisustvo znatno više detalja na slici, što je jako bitno u većini primena. Širenje dinamičkog opsega ima svoje granice i one su definisane fizičkim limitima senzora koji se koristi, tj. ograničenjima njegovog A/D konvertora. Kada je dinamički opseg scene značajno širi od dinamičkog opsega senzora, mnogi detalji neće biti adekvatno prikazani na slici. Međutim, ukoliko senzor inherentno podržava široki dinamički opseg, jasno se može uočiti da je snimljena slika kvalitetnija od one koja se dobija sa standardnog senzora. Širenje dinamičkog opsega slike za sobom povlači i korišćenje novog formata koji je namenjen za čuvanje više informacija o svakom pikselu na slici. Sliku ili video sadržaj u ovom formatu moguće je integrisati u postojeće sisteme, ali rešenje može biti veoma skupo. Postoji način da se izbegne uvođenje novog formata, a da se pri tom poveća broj detalja na slici, tj. da se snimi scene koja ima široki dinamički opseg. Tada se koriste standardni senzori sa kojih se sekvencijalno uzima više slika iste scene sa različitim ekspozicijama i koriste tehnike fuzije snimljenih slika. Opisanim principom dobija se slika u standardnom formatu koja sa sobom nosi znatno više informacija. Bitno je napomenuti da se ovakva slika može koristiti u svim postojećim sistemima. Glavni cilj ove doktroske disertacije predstavlja razvoj metoda namenjenog za implementaciju u postojeće industrijske kamere sa namerom unapređenja njihove funkcionalnosti. Potreba za ovakvim unapređenjima postoji u situacijama kada se kamere koriste za snimanje nejednako osvetljenih scena, pa upotrebom standardnog načina akvizicije slike nije moguće videti sve detalje. Ideja vodilja prilikom projektovanja metoda bila je da se on izvršava potpuno automatski, bez ikakvog uticaja čoveka. Takodje, bilo je bitno da nema nikakvih izmena u postojećem softveru kamere, a da se hardver modifikuje tako što se samo umetne nova ploča sa implementiranim metodom izmedju procesora i senzora, jednostavnim premošćavanjem postojećeg kabla koji ih povezuje. U ovoj disertaciji biće detaljno opisan postupak realizacije metoda, zajedno sa svim algoritmima od kojih je sastavljen. Za potrebe vii evaluacije rezultata, pored sopstvene baze slika, korišćena je i baza slika grupe za multimedijalnu obradu signala sa EPFL u Lozani.
Improving image quality by expanding the dynamic range is extensively used recently. This results in the presence of significantly more details in the picture, which is very important for most applications. Expanding the dynamic range has its limits, and they are defined by the physical limits of sensor used, i.e. the limits of its A / D converter. When the dynamic range of the scene is significantly wider than the dynamic range of the sensor, many details will not be shown properly in the picture. However, if the sensor inherently supports wide dynamic range, it can be clearly noticed that the recorded image quality is higher than the one obtained with the standard sensors. Expanding the dynamic range of the image means using of a new format that is designed to store more information about each pixel in the image. Image or video content in this format can be integrated into existing systems, but the solution can be very expensive. There is a way to avoid a new format introduction, and thereby to increase the amount of details in the image, i.e. to record the scene that has a wide dynamic range. In this case the standard sensors are used, by sequentially taking multiple images of the same scene at different exposures and applying the fusion to the images recorded. Described principle will produce an image in a standard format which carries significantly more information than capturing single image. It is important to note that this image can be used in all existing systems. The main objective of this doctoral dissertation is the development of the method which will be used for implementation in existing industrial cameras with the intention of improving theirs functionality. The need for such improvements exists in situations where the camera is used to record unevenly lit scene, thus using a standard image acquisition tecniques is not possible to see all details. The main idea in the designed method was that it runs completely automatically, without any human influence. Also, it is important to stress that there are no modifications in the existing camera software and hardware. Is is only nedeed to insert a new board with the method implemented between processor and sensor, simply bypassing the existing cable that connects them. In this thesis will be described in detail the process of realization of the method, together with all the algorithms of which it is composed. For the purposes of the evaluation results the image database from multimedia signal processing group from the EPFL in Lausanne is used, in addition to our own database files. ix
true
Risken-Nummedal-Graham-Haken instabilities and self-pulsing in quantum cascade lasers
Рискен-Нумедал-Грахам-Хакен нестабилности и само-пулсирање у квантним каскадним ласерима.
Prikazano je teorijsko istraživanje multimodnih nestabilnosti sa niskim pragom u kvantnim kaskadnim laserima (KKL). Prethodno su u više eksperimentalnih studija o KKL-u uočene nestabilnosti Risken- Numedal-Graham-Haken (RNGH) tipa malo iznad laserskog praga...
A theoretical study on low-threshold multimode instabilities in quantum cascade lasers (QCLs) with Fabry-Pérot cavity is presented. Previously, low threshold Risken-Nummedal-Graham-Haken (RNGH) instabilities were reported in several experimental investigations of QCLs...
false
Unapređenje inteligentih industrijskih mernih instrumenata sa MEMS senzorima
Improvement of intelligent industrial measurement instruments with MEMS sensors
Primena senzora je u naglom porastu u svim oblastima života i rada. Sve je veći broj fizičkih, ali i drugih (hemijskih, bioloških) veličina koje je potrebno meriti, uz rastuće zahteve u pogledu funkcionalnosti i mernih performansi uređaja i uz istovremeno smanjenje dimenzija, cene i potrošnje energije. Pojavom mikroelektromehaničkih tehnologija i sistema (MEMS), počelo je novo razdoblje u razvoju senzora, koje i danas traje. Senzori koje je tim tehnologijama moguće realizovati po karakteristikama su uporedivi ili bolji od konvencionalnih, a malih su dimenzija i pogodni za jeftinu masovnu proizvodnju. Konvergencijom senzorskih, računarskih i komunikacionih tehnologija nastali su inteligentni instrumenti i distribuirani senzorski sistemi, čime se dostiže novi nivo funkcionalnosti i performansi. Naučna istraživanja u oblasti senzora, mernih metoda i instrumenata su zbog toga veoma aktuelna i imaju sve izraženiji multidisciplinarni karakter. Cilj istraživanja obuhvaćenih ovom disertacijom je unapređenje funkcionalnosti i performansi inteligentnih industrijskih mernih instrumenata zasnovanih na postojećim i novim MEMS senzorima, posebno na onima koji se razvijaju i proizvode u Centru za mikroelektronske tehnologije Instituta za Hemiju, tehnologiju i metalurgiju. Pošto jedino sveobuhvatni pristup, od teorijskih istraživanja, preko razvoja senzora i obrade signala do gotovog mernog instrumenta, može da pruži kvalitetan rezultat upotrebljiv u praksi, ova disertacija je koncipirana tako da sadrži sve te elemente. U disertaciji je dat pregled MEMS tehnologija i senzora, gde je težište na MEMS senzorima mehaničkih veličina, a naročito na piezootpornim senzorima pritiska. Kao doprinos značajan za tu oblast, prikazani su rezultati eksperimentalnog istraživanja električnih karakteristika MEMS struktura izrađenih na fleksibilnim supstratima, od posebnog interesa za primene MEMS senzora u medicinskim implantima, protezama, odeći, kao i u robotici. Od novih MEMS senzora koji se istražuju u Centru i drugim naučnim ustanovama u svetu, kao posebno značajni se izdvajaju senzori hemijskih i bioloških veličina, čiji se princip rada zasniva na adsorpciji čestica. U ovoj VI disertaciji je istraživan uticaj procesa kompetitivne adsorpcije čestica i prenosa mase na fluktuacije mase MEMS senzorskih struktura. Izveden je analitički izraz koji opisuje taj uticaj i predstavlja koristan alat za optimizaciju hemijskih i bioloških MEMS senzora zasnovanih na selektivnoj adsorpciji. U disertaciji je dat pregled inteligentnih mernih instrumenata, gde se posebno razmatraju oni koji su predviđeni za industrijske primene. Objašnjeni su koncepti i podele i prikazano je aktuelno stanje nauke i tehnike u toj oblasti. Formiran je nov parametarski matematički model silicijumskog piezo- otpornog MEMS senzora pritiska, na osnovu kojeg je razvijena nova merna metoda za istovremeno merenje pritiska i temperature pomoću istog senzora, uz postizanje visokih mernih performansi. Takva metoda nije opisana u dostupnoj literaturi. U cilju primene metode u praksi, projektovan je novi inteligentni industrijski instrument modularnog tipa. On omogućuje korišćenje i drugih vrsta MEMS senzora, kao što je npr. višenamenski termalni MEMS senzor, takođe razvijen u Centru. Kao rezultat primene opisane metode i projekta, razvijena su dva nova inteligentna industrijska transmitera: unapređeni transmiter nivoa tečnosti i transmiter vakuuma, koji su takođe prikazani u disertaciji. Osim originalnog naučnog doprinosa u oblasti senzora i mernih instrumenata, koji je potvrđen objavljivanjem recenziranih publikacija, izradom ove disertacije su ostvareni rezultati čija je upotrebljivost dokazana u praksi. Autor je razvio više tipova inteligentnih industrijskih instrumenata koji se proizvode u Centru za mikroelektronske tehnologije. Oni su po svojim karakteristikama uporedivi sa skupljim inostranim proizvodima ili ih čak prevazilaze. U redovnoj su upotrebi u većem broju energetskih i drugih postrojenja u Srbiji, gde se pokazalo da ispunjavaju sve zahteve korisnika u smislu funkcionalnosti, pouzdanosti i mernih performansi.
The use of sensors is rapidly growing in all fields of industry, as well as in everyday life. There is an increasing number of physical and other (chemical, biological) quantities that have to be measured with increasing demands regarding the instrument functionality and measurement performance, and with simultaneous decrease in their dimensions, cost and power consumption. The emergence of micro-electro- mechanical (MEMS) technologies has started a new era in sensor development. Compared to conventional sensors, the sensors based on these technologies are of comparable or superior characteristics, but of miniature dimensions and suitable for inexpensive mass-production. The convergence of sensor, computer and communication technologies has led to the development of intelligent instrumentation and distributed sensor systems, enabling a new level of functionality and performance. The research topics in the fields of sensors, measurement methods and instruments are therefore of great scientific interest, and have an increasingly interdisciplinary character. The objective of the research described in this dissertation is the improvement of both the functionality and the performance of intelligent industrial measurement instruments based on the existing and novel MEMS sensors, especially those developed and produced by the Center of microelectronic Technologies (Institute of Chemistry, Technology and Metallurgy, University of Belgrade, Serbia). Since only a comprehensive approach, from theoretic research, sensor development and signal processing to the finished instrument, can yield high-quality and practically viable results, the dissertation is composed so that it includes all these elements. In this dissertation, an overview is given of MEMS technologies and sensors, with an emphasis on mechanical MEMS sensors, and especially on piezoresistive pressure sensors. As a contribution relevant to this field, the results are presented of the experimental research of electrical characteristics of MEMS structures fabricated on flexible substrates. They are of particular interest for MEMS sensors applications in medical implants, prostheses, clothing, and robotics. Of the new MEMS sensors in development at the Center, as well as in other research centers worldwide, chemical and VIII biological sensors based on particle adsorption are regarded as the most significant. In this dissertation, the research is performed of the influence of both competitive adsorption and mass transfer processes on mass fluctuations in MEMS sensor structures. An analytical expression is derived that describes this influence and thus represents a useful tool for optimization of chemical and biological MEMS sensors based on selective adsorption. The dissertation includes an overview of intelligent instruments, with an emphasize on those intended for industrial applications. The concepts and divisions are discussed, as well as the current state-of-the-art in the field. A new parametric mathematical model of a silicon piezoresistive MEMS pressure sensor is devised, and then utilized as a base for a newly developed measurement method that enables high-performance simultaneous pressure and temperature measurement using a single sensor. Such a method has not been previously described in the available literature. In order for the method to be practically applied, a new modular intelligent industrial instrument is designed. It can be used with some other MEMS sensor types, including the multipurpose thermal MEMS sensor, also developed at the Center. The use of both the method and the project resulted in the development of two new intelligent industrial transmitters: the improved liquid level transmitter and the vacuum transmitter, both described in the dissertation. Apart from the original scientific contribution in the fields of sensors and measurement instrumentation, also presented in peer-reviewed publications, the work on this dissertation yielded results proven in practical applications. The author developed several types of intelligent industrial instruments that have been produced at the Center of Microelectronic Technologies. They are comparable with, or even superior to more expensive products on the market. The experience from their regular use in various industrial plants in Serbia shows that they fulfill all the user requirements regarding functionality, reliability, and measurement performance.
true
Unapređenje metoda za procenu napora u softverskim projektima
Enhancing methods for effort estimation in software projects
Procena napora potrebnog da se implementira softverski sistem je jedan od najtežih ali i najbitnijih zadataka koji se postavlja pred softverske timove. Procena napora je često ključni faktor koji će odrediti da li će projekat početi, pod kojim ograničenjima će se raditi i da li će se uopšte uspešno završiti i koristiti. Realnost je da se procene često vrše na osnovu nedovoljno definisanih zahteva pomoću različitih heuristika, pretpostavki i iskustava, pri čemu se za ograničeno vreme traže što je moguće bolje procene. I pored značajnog napretka u različitim oblastima softverskog inženjerstva tokom poslednjih decenija, procene napora se uglavnom i dalje svode na subjektivne procene timova ili eksperata u nekoj oblasti sa greškama procene koje često nisu zadovoljavajuće. Globalne statistike govore da se samo trećina softverskih projekata uspešno završi, trećina se završi uz probijanje budžeta i rokova, dok trećina projekata propadne. Jedan od glavnih razloga za ovakve loše rezultate su nedovoljno dobre procene napora. Cilj ovog rada je analiza postojećih metoda procene i predlaganje novih metoda kojima se procene napora mogu olakšati i poboljšati. Osnovna ideja koja će biti predstavljena u radu je mogućnost uvođenja novih predikcionih modela i tehnika kojima će se postojeće metode poboljšati. U radu će biti predstavljeni predlozi rešenja za unapređenje postojećih metoda procene kao i konkretni rezultati koji se mogu dobiti primenom tih rešenja na realnim softverskim projektima. Unapređenja se svode na identifikaciju parametara kojima se mogu opisati projekti koji najviše utiču na napor potreban da se implementira projekat, kao i zavisnosti među parametrima i naporom.
Estimating the effort required to implement the software system is one of the most difficult, but also the most important tasks in software engineering. Effort estimate is often a key factor that will indicate whether the project will begin, under which restrictions will be implemented and would it be successfully completed and used. The reality is that the estimates are often made based on the incomplete requirements using various heuristics, assumptions and experience, where best possible estimates are required in a limited timeframe. Besides significant advances in various domains of software engineering in recent decades, effort estimation methods are mostly based on the subjective assessment by the project teams or experts in a certain field, with the effort estimate errors often higher than the expected. Global statistics show that only one third of software projects is successfully concluded, one third is not finished within the budget and deadlines, while the rest of projects fail. One of the main causes for these results are inaccurate effort estimates. The aim of this thesis is to analyze the existing effort estimation methods and propose new methods that can improve effort estimation. The basic idea that will be presented in the thesis is the usage of new models and techniques that will improve the existing effort estimation methods. In the thesis, proposals for the improvement of existing assessment methods and concrete results that can be obtained using these solutions to real software projects will be presented. Improvements are confined to the identification of project parameters describing projects that influence the effort the most, as well as determining dependencies between the parameters and the effort.
true
Uticaj negativne modulacije benzodiazepinskog mesta vezivanja GABA-A receptora na ispoljavanje različitih bihejvioralnih efekata benzodiazepina kod pacova
The influence of negative modulation of benzodiazepine binding site of GABA-A receptor on the different bihevioural effects expression of benzodiazepines in rats
GABA (γ-aminobuterna kiselina) predstavlja glavni inhibitorni neurotransmiter centralnog nervnog sistema i većinu svojih inhibitornih efekata ostvaruje putem jonotropnih GABA-A receptora. Receptori koji sadrže α1, α2, α3, ili α5 subjedinicu, u kombinaciji sa γ2 i nekom od β subjedinica, predstavljaju dominantnu frakciju GABA- A receptora S obzirom na svoju kompleksnu strukturu, GABA-A receptori su označeni u ovom radu kao GABAA-BDZ-Cl receptorski kompleks. GABA-ergička sinapsa je mesto delovanja nekoliko različitih klasa lekova koji moduliraju inhibitornu neurotransmisiju i koriste se u farmakoterapiji anksioznih poremećaja, poremećaja sna, epilepsije, obustave alkohola, kao i za uvođenje i održavanje anestezije. Nedavno je utvrđeno da lekovi sa antidepresivnim efektima i stabilizatori raspoloženja u značajnoj meri utiču i na GABA-ergičku transmisiju, a da promene u inhibitornoj neurotransmisiji mogu imati ključnu ulogu u patofiziologiji depresije. Ova hipoteza je potkrepljena i klinčkim nalazima u kojima je utvrđen redukovan nivo GABA-e u korteksu pacijenata koji boluju od depresije, a koji se normalizuje nakon terapije antidepresivima. Benzodiazepini (BDZ), kao najčešće propisivani psihofarmaci, deluju kao pozitivni modulatori GABA-ergičke neurotransmisije i agonisti specifičnih BDZ receptora. Međutim, istraživanja lokalizacije i funkcionisanja BDZ receptora su pokazala postojanje nove, drugačije klase supstanci koje deluju na nivou ovih receptora ali kao negativni modulatori, odnosno inverzni agonisti. Naime, pokazano je postojanje supstanci, najčešće iz klase β-karbolina, koji stimulišu BDZ receptor, ali u suprotnom smislu od BDZ. Dok BDZ pokazuju anksiolitičke i amnezijske efekte, od inverznih agonista se očekuje da pokažu drugačije bihejvioralne efekte, anksiogene i potencijalno promnestičke efekte, odnosno uticaj na kogniciju i motivaciju. S jedne strane, genetska istraživanja, uz korišćenje modifikovanih linija životinja, ukazuju na korelaciju između poremećaja raspoloženja i promene nivoa GABA-ergičke neurotransimisije i gustine pojedinih podtipova GABA receptora, dok su, s druge strane, komplementarna bihejvioralna ispitivanja usmerena ka ispitivnju novosintetisanih liganada, kao i povezivanju pojedinih efekata BDZ i drugih liganada sa konkretnim podtipom GABA- A receptora. Imajući u vidu, često kontradiktorne, podatke u literaturi o delovanju inverznih agonista na benzodiazepinsko mesto vezivanja GABAA–BDZ-Cl receptorskog kompleksa, cilj istaživanja bio je da se, primenom dva komplementarna, i u oblasti bihejvioralne farmakologije široko korišćena testa (aktivno izbegavanje averzivne draži i forsirano plivanje) ispita uticaj jednokratne i ponavljane primene rastućih doza metil 6,7-dimetoksi-4-etil-β-karbolin-3-karboksilata (DMCM), inverznog agoniste benzodiazepinskog mesta vezivanja GABA-A receptora, na ispoljavanje različitih bihejvioralnih efekata benzodiazepina kod pacova i da se, u okviru analize mehanizma dejstva, utvrdi da li se efekti inverznih agonista ostvaruju preko specifičnih mesta vezivanja u sklopu GABAA-BDZ-Cl receptorskog kompleksa. U testu aktivnog izbegavanja (AA), statistička analiza nije pokazala značajne razlike između grupa koje su tretirane rastvaračem i fiziološkim rastvorom, dok flumazenil nije značajno menjao bihejvioralne parametre, odnosno broj uspešnih odgovara i parametar motorne aktivnosti u odnosu na kontrolnu grupu. Statistička analiza je pokazala da jednokratna primena DMCM-a drugog dana u testu aktivnog izbegavanja utiče na broj odgovora AA i da doza od 0,1 mg/kg DMCM-a pokazuje facilitirajući uticaj na učenje, dok je flumazenil u potpunosti antagonizovao ovaj efekt. Promnestički efekt DMCM-a u dozi 0,1 mg/kg pokazan je i pri ponavljanoj primeni supstance. U testu forsiranog plivanja (FST), statistička analiza nije pokazala značajne razlike između grupa koje su tretirane rastvaračem i fiziološkim rastvorom u testu forsiranog plivanja. Takođe, flumazenil nije značajno menjao bihejvioralne parametre, odnosno vreme imobilnosti životinja u odnosu na kontrolnu grupu. Motorni parametri, u smislu plivanja i gnjurenja, nisu se značajno razlikovali ni kod jedne grupe. Jednokratna primena DMCM-a, drugog dana u testu forsiranog plivanja, utiče na vreme imobilnosti životinja i doza od 0,1 mg/kg DMCM-a smanjuje vreme imobilnosti, dok flumazenil u potpunosti antagonizuje ovaj efekt. Statistička analiza je pokazala i da ponavljana primena doze DMCM-a od 0,1 mg/kg, tokom 14 dana, statistički značajno smanjuje vreme imobilnosti životinja u odnosu na kontrolnu grupu tretiranu rastvaračem. U testu spontane lokomotorne aktivosti, statistička analiza nije pokazala značajne razlike prilikom testiranja spontane lokomotorne aktivnosti između grupa koje su tretirane rastvaračem, fiziološkim rastvorom i flumazenila, dok je najveća primenjena doza DMCM-a (1,0 mg/kg) pokazala hipolokomotorni uticaj, koji je antagonizovan flumazenilom. Naša istraživanja su pokazala da DMCM, inverzni agonista benzodiazepinskog mesta vezivanja ima značajan uticaj na učenje reakcije aktivnog izbegavanja. DMCM u najmanjoj primenjenoj dozi povećava broj AA odgovora, odnosno poboljšava učenje. Tokom ponavljane primene, pokazano je da postoji statistički značajna korelacija između ove doze i efekta DMCM-a na učenje, što je u skladu sa nalazima drugih autora koji su pokazali da male doze negativnih modulatora GABA-ergičke neurotransmisije mogu biti korisne za podizanje vigiliteta i bolje učenje kod eksperimentalnih životinja. S druge strane, rastuće doze DMCM-a ne dovode do povećanja spontane lokomotorne aktivnosti, što ukazuje da učenje reakcije AA nije bila posledica povećane aktivnosti životinja. Naša istraživanja upućuju i na značajne antidepresivne efekte inverznog agoniste DMCM-a u najmanjoj primenjenoj dozi, ali se takav trend gubi pri većim dozama. Koliko nam je poznato, naša studija je prvi izveštaj o uticaju inverznih agonista BDZ receptora na ponašanje pacova u testu forsiranog plivanja, kao i o uticaju ovih supstanci na motivaciju. Rezultati koje smo dobili su, na prvi pogled, iznenađujući i u suprotnosti sa postavljenim hipotezama da smanjenje GABA-ergičke neurotransmisije i aktivnosti GABA-A receptora doprinosi razvoju depresije. S druge strane, akutni antidepresivni efekti, pokazani u našem istraživanju, pokazuju značajnu sličnost sa nedavno objavljenim rezultatima efekata blokade glutamatskih NMDA receptora. S obzirom na hronični karakter depresivnih ispoljavanja, posebno je značajan rezultat da je antidepresivni efekt prisutan i nakon ponavljane primene DMCM-a tokom 2 nedelje. Pokazan uticaj inverznih agonista benzodiazepinskog mesto vezivanja GABA-A receptora na antidepresivnu aktivnost i učenje, u eksperimentalnim uslovima, u skladu je sa kliničkim nalazima, da su depresivni poremećaji često u komorbiditetu sa anksioznim poremećajima i kognitivnim oštećenjima. Stoga, sveukupni povoljni učinak negativne modulacije GABA-ergičke neurotransmisije može odražavati povoljne uticaje na kogniciju i motivaciju, istovremeno.
GABA (γ-aminobutyric acid) is the major inhibitory neurotransmitter of central nervous system and most of its inhibitory effects achieve through ionotorpic GABA-A receptors. Receptors containing α1, α2, α3, or α5 subunit, in combination with γ2 and one of the β subunits, represent the dominant fraction of GABA-A receptors Due to its complex structure, GABA-A receptors have been identified in this paper as a GABAA- BDZ-Cl receptor complex. GABA-ergic synapse is the site of action of several different classes of drugs that modulate inhibitory neurotransmission and are used in the pharmacotherapy of anxiety disorders, sleep disorders, epilepsy, alcohol cessation, and for the introduction and maintenance of anesthesia. It is increasingly recognized that drugs with antidepressant/mood-stabilizing effects profoundly alter GABAergic transmission, and that changes in inhbititory neurotransmission may play a key role in the pathophysiology of depression. This hypothesis is supported by clinical findings demonstrating reductions in cortical GABA levels in patients suffering with depression, which are restored to normal levels with antidepressant treatment. Benzodiazepines (BDZ), as the most commonly prescribed psychotropic drugs, act as positive modulators of GABA-ergic neurotransmission and specific BDZ receptor agonists. However, studies of localization and functioning of BDZ receptor have shown the existence of a new, different classes of substances that act at the level of the receptor but as negative modulators, or inverse agonists. Namely, it has been shown the existence of substances, mostly from the class β-carbolins, that stimulate BDZ receptor, but in the opposite sense of BDZ. While BDZ show anxiolytic and amnesic effects, inverse agonists are expected to show different behavioural effects, and potentially anxiogenic and promnesic effects, i.e. effects on cognition and motivation. On the one hand, genetic research, using modified animal lines, indicating a correlation between mood disorders and changes in levels of GABA-ergic transmission and density of certain subtypes of GABA receptors, whereas, on the other hand, a complementary behavioural tests aimed to test the newly synthesized ligands, and connect the individual effects of BDZ and other ligands with a specific subtype of GABA-A receptors. Bearing in mind, often contradictory information in the literature about the effect of inverse agonist of benzodiazepine site of the GABA-benzodiazepine receptor complex, this research has an aim that, by using two complementary, both in behavioural pharmacology widely used test (active avoidance of aversive stimuli and forced swimming test) examine the effect of single and repeated application of increasing doses of methyl 6,7-dimethoxy-4-ethyl-β-carbolineum-3-carboxylate (DMCM), an inverse agonist of benzodiazepine binding sites of GABA-A receptors, the expression of different behavioural effects of benzodiazepines in rats and that, in the analysis of the mechanism of action, determine whether the effects of inverse agonists were achieved through specific binding sites within the GABAA-BDZ-Cl receptor complex. In a study of active avoidance (AA), statistical analysis showed no significant difference between the groups treated with solvent and saline, while flumazenil was not significantly altered behavioural parameters, i.e. the number of successful matches and parameters of motor activity compared to the control group. Statistical analysis showed that a single application of DMCM, a second day in active avoidance test, affects the response of AA, and DMCM at the dose of 0.1 mg shows facilitating impact on learning, while flumazenil completely abolish this effect. Promnesic effect of DMCM at a dose 0.1 mg/kg was also shown with repeated application of the substance. In the forced swim test (FST), statistical analysis showed no significant difference between the groups treated with solvent and saline in the forced swimming. Also, flumazenil did not significantly alter behavioural parameters, i.e. immobility time of animals compared to the control group. Motor parameters, in terms of swimming and diving, did not differ significantly in none of the group. A single dose of DMCM, the second day of the forced swimming, affects the immobility time of animals, and DMCM at a dose of 0.1 mg/kg reduced immobility time, while flumazenil completely antagonised this effect. Statistical analysis showed that repeated application of DMCM 0.1 mg/kg for 14 days, significantly reduced the time of immobility of animals compared to the control group treated with solvent. In a test of spontaneous locomotor activity, statistical analysis showed no significant differences when testing spontaneous locomotor activity between the groups treated with solvent, saline and flumazenil, while the highest applied dose of DMCM (1,0 mg/kg) showed hypolocomotor effect, which was antagonized by flumazenil. Our research has shown that DMCM, a benzodiazepine binding site inverse agonist has a significant impact on active avoidance learning. DMCM, at a lowest applied dose, increased the number of AA responses and improved learning. During repeated application, it is shown the statistically significant correlation between the dose and the effect of DMCM on learning, which is consistent with findings of other authors who have shown that small doses of negative modulators of GABA-ergic neurotransmission may be useful to raise vigilance and better learning in animal experiments. On the other hand, higher doses of DMCM did not lead to an increase in spontaneous locomotor activity, suggesting that learning reactions AA was not a consequence of increased activity of animals. Our findings also pointed to significant antidepressant effects of inverse agonists DMCM at the lowest applied dose, but this trend was lost at higher doses. To the best of our knowledge, our study one of the first reports the influence of a BDZ receptor inverse agonist DMCM on rat behaviour in forced swim test, as well as the impact of these substances on the motivation. At first glance, our results seem surprising and contradict the certain hypothesis that reduction of GABA-ergic neurotransmission and decreased GABA-A receptor activity contribute to the development of depression. On the other hand, the acute antidepressant effects observed in our study, showed striking similarity to the recently published results of NMDA glutamate receptors blocking effects. Given the chronic nature of depressive manifestations, particularly significant result that the antidepressant effect is present even after repeated use of DMCM during two weeks. The demonstarted impact of BDZ receptor inverse agonists on antidepressant activity and learning in experimental conditions is consistent with clinical findings that depressive disorders often comorbid with anxiety disorders and cognitive impairments. The overall effect of negative modulation of GABA-ergic neurotransmission may thus reflect better peroformance and effects on both cognition and motivation enhancements.
true
Insulinska rezistencija i parametri inflamatornog odgovora tokom abdominalnih hirurških intervencija
Insulin resistance and inflammatory response following abdominal surgery
Pojava prolazne i vremenski ograničene insulinske rezistencije tokom hirurškog zahvata je povezana sa vrstom operacije kao i sa veličinom povrede tkiva. Cilj naše studije je bio da procenimo insulinsku rezistenciju merenu matematičkim modelom HOMA-IR i faktore inflamacije IL-6, TNF-α i CRP-a pre i posle laparoskopske i otvorene holecistektomije i otvorene rekonstrukcije ingvinalnih kila tokom sedmodnevenog praćenja, kao i ispitivanje prediktivne vrednosti preoperativnog odnosa neutrofila i limfocita u proceni težine bolesti kod bolesnika sa holecistitisom i njegove eventualne povezanosti sa stepenom insulinske senzitivnosti.. Metode: Istraživanje koje je sprovedeno po tipu ne-randomizovane prospektivne studije obuhvatilo je 192 bolesnika sa kamenom u žučnoj kesi (operisanih primenom otvorene ili laparoskopske holecistektomje) ili sa ingvinalnom kilom (operisanih otvoreno metodom) koji su operisani na Klinici za Urgentnu hirurgiju Urgentnog centra Kliničkog centra Srbije u toku sprovodjenja studije. Glukoza, insulin, CRP, IL-6, TNF- α, Le, Ly i Ne su određivani pre operacije (0 dan) i 1., 3. i 7. dan posle operacije. Rezultati: Veće vrednosti indeksa HOMA-IR su nađene kod pacijenata nakon otvorene operacije žučne kese u poređenju sa laparoskopskom operacijom u trećem postoperativnom danu. (p=0,036). HOMA-IR je bio značajno veći nakon otvorene rekonstrukcije kila u poređenju sa grupom nakon laparoskopske operacije žučne kese prvog postoperativnog dana (p=0,045). Vrednost CRP je bila značajno veća u prvom postoperativnom danu kod bolesnika nakon otvorene operacije žučne kese u odnosu na laparoskopsku operaciju (p=0,046). IL-6 se značajno povećao 1. i 3.dana u odnosu na 0.dan (p<0,0005) u grupi nakon laparoskopske operacije žučne kese i nakon otvorene rekonstrukcije kile. Povećanje IL-6 je bilo statistički značajno veće 1.postoperativnog dana nakon otvorene operacije kile u odnosu na laparoskopsku operaciju žučne kese (p=0,044). Pozitivna korelacija između CRP i HOMA-IR (r=0,46; p=0,025) i između IL-6 i HOMA-IR (r=0,44; p=0,030) postojala je u grupi sa otvorenom rekonstrukcijom kila prvog dana. Grupa ispitanika sa visokim NLR imala je znatno češće težak holecistitis (p<0,0001) i značajno veći nivo CRP i broj leukocita (p<0,0001). Vrednost NLR od 4,18 može predvideti težak holecistitis. Pokazana je značajna povezanost preoperativnog NLR i HOMA-IR prvog postoperativnog dana (r=0,254, p=0,030). Zaključak: Utvrđeno je značajno povećanje perioperativne insulinske rezistencije procenjeno primenom indeksa HOMA-IR kod bolesnika nakon otvorenih operacija žučne kese i kile u poređenju sa laparoskopskim operacijama žučne kese. Pozitivna korelacija između CRP-a i IL-6 sa HOMA-IR u grupi bolesnika koji su operisani zbog kile otvorenom metodom prvog postoperativnog dana ukazuje na mogući uticaj ovih medijatora na oštećenje insulinske senzitivnosti. Preoperativni odnos neutrofila i limfocita kod bolesnika koji se spremaju za operaciju žučne kese zbog holecistitisa mogao bi da bude koristan pokazatelj težine upale. Ustanovljena korelacija između preoperativnog NLR sa HOMA-IR u prvom postoperativnom danu ukazuje na mogućnost korišćenja ovog indeksa u predviđanju poremećaja insulinske senzitivnosti sa mogućim postoperativnim komplikacijama. Naše istraživanje je pokazalo manje oštećenje insulinske senzitivnosti i manji porast parametara inflamacije kod laparoskopske holecistektomije u odnosu na otvorenu operaciju žučne kese i otvorenu operaciju preponske kile.
The development of transient and temporal insulin resistance after surgical intervention is associated with the type and magnitude of operation and tissue injury. The aim of our study was to evaluate insulin resistance assesed by homeostatic model assessment of insulin resistance (HOMA-IR) and inflammation factors IL-6, TNF-α and CRP before and after laparoscopic and open cholecystectomy and open hernia repair during seven days follow-up, as well as the predictive value of preoperative neutrophil-to-lymphocyte ratio (NLR) to asses the severity of cholecystitis and its possible association with the degree of insulin resistance. Methods: a non-randomized prospective study included 192 patients with stones in gallblader (treated with laparoscopic or open cholecystectomy) and inguinal hernia repair treated with open method. Study was conducted at Clinic for emergency surgery, Emergency center, Clinical Center of Serbia during the course of the investigation. Results: Significantly higher values of HOMA-IR were found in patients with open cholecystectomy in comparison with patients with laparoscopic cholecystectomy on third postoperative day (p=0.036). HOMA-IR was significantly higher after open hernia repair in comparison with laparoscopic cholecystectomy at first postoperative day (p=0.045). CRP was significantly higher in patients after open cholecystectomy in comparison with patients with laparoscopic cholecystectomy, at day one (p=0.046). Statistically higher values of IL-6 were found at day 1 and 3 in comparison with day 0 in group with laparoscopic cholecystectomy and in group with open hernia repair (p<0.0005). IL-6 was significantly higher at day 1 after open hernia repair in comparison with laparoscopic cholecystectomy (p=0.044). Positive correlation between CRP and HOMA-IR (r=0.46; p=0.025) and between IL-6 and HOMA-IR (r=0.44; p=0.030) was found at day one in group with open hernia repair. More severe cholecystitis was found among patients in high NLR group (p<0.0001) and higher level of CRP and white blood cells counts (WBC) (p<0.0001). The NLR of 4.18 could predict severe cholecystits. Significant correlation between preoperative NLR and HOMA-IR values at first postoperative day was established (r=0.254, p=0.030). Conclusions: There was significant increase in perioperative insulin resistance expressed by HOMA-IR in patients with open cholecystectomy and open hernia repair in comparison with laparoscopic cholecystectomy. Positive correlation between CRP and IL-6 with HOMA-IR were found in patients with open hernia repair at day 0, indicating possible impact of these mediators on insulin sensitivity impairment. Preoperative NLR in patients prepared for cholecystectomy due to cholecystitis could be usseful predictor of severity of inflammation. The established correlation between preoperative NLR with HOMA-IR in the first postoperative day indicates the possible use of this index for prediction of insulin sensitivity impairment with consecutive postoperative complications. Our study demonstrated less damage in insulin sensitivity and a smaller increase in parameters of inflammation in the course of laparoscopic cholecystectomy in comparison to open gallbladder operation and open inguinal hernia repair.
true
Metodologija projektovanja simulatora arhitekture i organizacije računara
Methodology for designing simulators of computer architecture and organization
U ovom radu se razmatra metodološki pristup dizajnu simulatora iz oblasti arhitekture i organizacije računara koji treba da omogući razvoj simulatora digitalnih sistema proizvoljnog nivoa složenosti sposobnih za rad u konkurentnom i distribuiranom okruženju. Da bi se omogućio formiranje metodologije na početku rada je prikazan pregled nastave u oblasti arhitekture i organizacije računara na osnovnim studijama, kao i pregled oblasti projektovanja simulatora gde je poseban akcenat bio stavljen na oblasti konkurentnog i distribuiranog programiranja koje studenti treba da poznaju kao bi mogli da razviju simulatore koji omogućavaju rad u takvom okruženju. Na osnovu sprovedene evaluacije simulatora koji se koriste u nastavi iz oblasti arhitekture i organizacije računara a koji imaju raspoloživ izvorni kod predloženo je rešenje koje se zasniva na korišćenju slojevite arhitekture kod koje je svaki sloj odgovoran za drugi vid obrade i komunikacije. Predloženo rešenje se sastoji iz korišćenja pet slojeva: logičkog, izvršnog, prezentacionog, simulacionog, i sloja fizike. Detalji vezani za procedure i objašnjenja tehnika koje se koriste za realizaciju ovih slojeva su prikazani u radu. Za svaki sloj predloženog rešenja je dat analitički model procene vremena izvršavanja simulacije u zavisnosti od ulaznih parametara prilikom rada u konkurentnom i distribuiranom okruženju. Centralni deo rada opisuje simulator diskretnih događaja opšte namene razvijen prema opisanoj metodologiji kao simulator arhitekture i organizacije računara koji je sposoban za rad u konkurentnom i distribuiranom okruženju. Opis simulatora i njegovih delova je dat sa stanovišta detalja implementacije gde su predstavljeni paketi realizovani na osnovu predložene metodologije, kao i sa stanovišta korišćenja gde su opisane karakteristične situacije u kojima se simulator može koristiti. Na osnovu implementacije simulatora i pratećih biblioteka razvijene su laboratorijske vežbe i projekti iz predmeta konkurentno i distribuirano programiranje, koje su predstavljene u nastavku rada kao i evaluacija postignutih rezultata u nastavi. Pored ove evaluacije na kraju rada je predstavljena i evaluacija simulatora sa stanovišta eksperimentalnih rezultata i sa stanovišta analitičkog modela kao bi se utvrdilo u kojim slučajevima i u kom obimu se mogu koristiti simulatori razvijeni shodno opisanoj metodologiji. Razvijena je metodologija koja se zasniva na korišćenju slojevite arhitekture kod koje je svaki sloj odgovoran za različite oblike komunikacije i obrade. Odluka da se koristi slojevita arhitektura prilikom projektovanja simulatora arhitekture i organizacije računara sposobnih za izvršavanje u konkurentnom i distribuiranom okruženju omogućila je da se na jednostavan način razdvoje postupci obrade od postupaka komunikacije, sinhronizacije i interakcije između slojeva bez ulaženja u načine na koje se podaci koriste; Na osnovu predložene metodologije je bez poteškoća razvijen SLEEP, simulator arhitekture i organizacije računara sa slojevitom arhitekturom koji se može koristiti i kao simulator diskretnih događaja opšte namene. Pokazalo se da predloženi pristup razvoju SLEEP simulatora dovodi do stvaranja mogućnosti za izvršavanje simulacije koristeći rad sa jednom niti, rad sa više niti i rad u distribuiranom okruženju; Na osnovu evaluacije rezultata dobijenih tokom korišćenja simulatora kao jedan od zaključaka se izdvaja činjenica da se postupak projektovanja simulatora arhitekture i organizacije može koristiti u nastavi iz konkurentnog i distribuiranog programiranja kao polazna osnova za postupak paralelizacije aplikacija sa jednom tokom kontrole. Pored toga simulator se može koristiti i prilikom testiranja tako paralelizovanih aplikacija jer na konzistentan način stvara veliko opterećenje koje u velikoj meri pokriva ponašanje programa sa većim brojem tokova kontrole; Na osnovu evaluacije rezultata dobijenih na osnovu analitičkog modela i merenja performansi može se zaključiti da se vreme izvršavanja simulacije u oblasti nastave arhitekture i organizacije računara može značajno smanjiti u konkurentnom i distribuiranom okruženju u slučaju niskog stepena interakcije sa korisnikom. U slučaju prosečnog nivoa interakcije ograničavajući faktor prilikom izvršavanja simulacija predstavlja vreme potrebno za prikaz rezultata simulacije koje je uporedivo sa vremenom potrebnim za obradu same simulacije.
This paper presents methodological approach to the computer architecture and organization simulator design. The methodological approach should help students to bridge the gap between theory and practice in the domain of computer architecture and organization simulator design, and to design simulators capable to work in a concurrent and distributed environment. In order to achieve this goal at the beginning of the paper a survey of undergraduate level courses was given. The courses include those in the field of computer architecture and organization and in the field of concurrent and distributed programming. One possible solution to the problem is given after an analysis of simulators that have available source code. Proposed solution is based on a multi- layer design where each layer is responsible for different type of processing and communication. The solution includes five different layers: logic, simulation, execution, presentation, and physics. For each proposed layer an analytical model that estimates execution time of simulation in a concurrent and distributed environment is given. The central part of the paper presents a discrete event simulator names SLEEP developed by using the proposed methodology. The design is presented through explanations of simulator structure and simulator usage. Based on the developed simulator and supporting libraries a set of laboratory exercises and project assignment was presented. At the end of the paper evaluation of the proposed analytical model and laboratory excursuses and project assignments was given. Methodology, for designing simulators of computer architecture and organization, based on the layered architecture where each layer is responsible for different forms of communication and processing was developed. The decision to use a layered architecture for designing simulators of computer architecture and organization that are capable to work in a concurrent and distributed environment enabled strict separation between simulation execution and processes communication, process synchronization and interaction between the layers; Based on the proposed methodology computer architecture and organization simulator named SLEEP wad developed. The simulator adopts layered approach and can be uses as a general purpose discrete event simulator. The proposed approach enables the SLEEP simulator to execute simulations within single-thread, multi-thread and distributed environment; Based on the evaluation obtained during the SLEEP simulator usage can be concluded that the process of designing simulators for computer architecture and organization can be used for teaching parallelization within a concurrent and distributed programming. Based on the evaluation of results, obtained from the analytical model and from the performance measurement, can be concluded that the execution time of simulations, which are used for teaching computer architecture and organization, can be significantly reduced in a concurrent and distributed environment in cases of low interaction with the user.
true
Light propagation in deterministic aperiodic waveguide arrays
Пропагација светлости у детерминистичким апериодичним низовима таласовода
U prethodne dve decenije, prostiranje svetlosti u regularnim (periodič- nim) i neuređenim spregnutim talasovodima je intenzivno proučavano. Ova dva, na neki način suprotna tipa uređenja, odgovorna su za dva različita fenomena - diskretnu difrakciju u regularnim nizovima, gde se energija svet- losti širi linearno duž pravca propagacije, i transverzalnu lokalizaciju u neuređenim nizovima, gde je svetlosna energija raspoređena u konačnom broju susednih talasovoda. Još uvek je otvoreno pitanje ponašanja svetlosti u oblasti između sasvim periodičnih i sasvim neuređenih struktura - determinističkim strukturama bez periodičnosti. U ovoj tezi je eksperimentalno i numerički proučavan uticaj determin- ističkih aperiodičnih jednodimenzionih (1D) i dvodimenzionih (2D) nizova spregnutih talasovoda (fotonskih rešetki) na transverzalnu propagaciju svet- losti...
Over the past two decades, the light transport properties in regular (periodic) and disordered coupled waveguides have been extensively scrutinized. These two, in a sense, opposite types of ordering lead to two dierent phenomena - discrete diraction in regular arrays, where light energy spreads linearly along propagation direction, and transverse localization in disordered, where light energy is distributed between nite number of adjacent waveguides. Still an open question remains whether light behavior in intermediate domain is between fully periodic and fully disordered structures - deterministic structures with no periodicity. In this thesis the impact of deterministic aperiodic one-dimensional (1D) and two-dimensional (2D) coupled waveguide arrays (photonic lattices) on transverse light transport has been investigated both experimentally and numerically. Several types of lattices were considered: 1D and 2D position modulated Fibonacci lattices, 1D refractive index contrast modulated Fibonacci lattice, 1D regular lattice and regular lattice with defect. The aim is to explore the possibility of using these structures as a tool for light diraction suppression. In the experiment, waveguide arrays were generated in photorefractive crystals, lithium-niobate and strontium-barium-niobate, by direct laser writing and incoherent Bessel beam induction technique, respectively. Gaussian and Airy beam were used as excitation beams. A series of numerical simulations were conducted using Beam Propagation Method. It is found that the position modulated Fibonacci lattice can suppress light diraction in comparison to the regular one. This is more pronounced in 1D than in 2D lattice. In 1D case, for relatively high refractive index contrasts, the propagating beam becomes almost completely localized in transverse plane. In 1D refractive index contrast modulated Fibonacci lattice the similar result is obtained - diraction is suppressed considerably comparing to regular lattice. Furthermore, the stronger modulation leads to strongly localized states, so the light propagation in Fibonacci lattice resembles the propagation in disordered one. Besides, it is shown that the slightly disordered Fibonacci lattice with relatively small refractive index contrast modulation enhances diraction. In addition, it is shown that a regular lattice can reduce a self-bending of Airy beam trajectory. The research presented in the thesis provides additional standpoint in understanding of light transport in complex optical structures and imposes new ways for control of light diraction and engineering of the optical response of devices.
false
Mehanizmi bioloških efekata jedinjena dobijenih alkalnom hidrolizom jabučnog pektina
Mechanisms of biological effects of compounds obtained by alkaline hydrolysis of apple pectin
Jabuka (Malus domestica Borkh) je od davnina poznata po svojim pozitivnim nutritivnim i lekovitim svojstvima pa je njen hemijski sastav intenzivno proučavan. Za pektin iz jabuka, kao i iz drugih izvora npr. citrusnog voća, je poznato da pozitivno utiče na gastrointestinalni trakt, dominantno putem uticaja na gastrointestinalnu mikrobiotu. Pretpostavka je da se taj uticaj uglavnom ostvaruje preko njegove polimerne strukture i slabe rastvorljivosti. Mi smo derivatizovali pektin primenom alkalne hidrolize u prisustvu vodonik peroksida. Dobijeni proizvod okarakterisali smo FTIR spektroskopijom i pokazali da u odnosu na polazni materijal ima simetrične i asimetrične vibracije istezanja karboksilatnih grupa. Ovo indukuje mogućnost da dobijeni derivat ima specifične reakcije sa hidroksilnim radikalima koji nastaju u Fentonovoj reakciji in vitro, a uslovi za postojanje Fentonove reakcije in vivo postoje u gastrointestinalnom traktu u prisustvu Fe i vitamina C. Ispitivanja EPR spektroskopijom pokazala su da derivat jabučnog pektina u sistemu Fentonove reakcije podstiče stvaranje ugljen-dioksid radikalskog anjona (•CO2- radikala) umesto hidroksidnog radikala. Za ovaj radikal su prethodna ispitivanja pokazala da ima duži poluživot od hidroksidnog radikala i da specifično reaguje sa određenim proteinima. To ukazuje na mogućnost da derivat pektina koji smo mi dobili u gastrointestinalnom traktu ima dodatno pozitivne efekte putem direktnog uticaja na mikrobiotu. Stoga smo ispitali antimikrobnu aktivnost pektina i njegovih derivata i pokazali, da derivati pektina dobijeni alkalnom hidrolizom jabučnog pektina (APDO) sprečavaju rast Saccharomyces cerevisiae i Candide albicans, dok nemaju uticaj na rast Escherichia coli i Staphylococcus aureus. Analizirali smo takođe i efekte derivata pektina (jabuke i citrusa) i oligosaharida izvedenih iz poligalakturonske kiseline (APDO, CPDO i PGDO) na rast jednog Gram- negativnog reprezentativnog soja Escherichia coli ATCC 25922 i jednog Gram- pozitivnog reprezentativnog soja Staphylococcus aureus u prisustvu supstanci koje imitiraju uslove u gastrointestinalnom traktu koji favorizuju Fenton reakciju. Pokazano je da derivat jabučnog pektina inhibira rast E. coli i S. aureus u tim uslovima, najverovatnije putem ugljendioksid radikal inhibicije ključnih enzima antioksidacionog sistema (kod E. coli SOD i GSH-Px, a kod S. aureus GSH-Px). Na kraju ispitali smo i uticaj derivata pektina na imuni sistem. APDO nakon 24 h nije u statistički značajnoj meri uticao na vijabilitet CLNC, dok je nakon 48 h izazvao značajno smanjenje vijabiliteta ovih ćelija, pri čemu je efekat dozno zavisan. Na osnovu ovih rezultata može se zaključiti da je zapaženo smanjenje vijabiliteta CLNC pod uticajem APDO nakon 48 h verovatno posledica antiproliferativnog delovanja ovih molekula. Naši rezultati pokazuju da dobijeni derivat jabučnog pektina ima izraziti potencijal za primenu u regulaciji gastrointestinalne mikrobiote jer sprečava rast patogenih mikroorganizama bez značajnog uticaja na imuni sistem domaćina.
Apple (Malus domestica Borkh) has long been known for its positive nutritive and medical properties, so its chemical composition has been intensively studied. For apple pectin, as well as from other sources e.g. citrus fruit, is known that have positive effect on gastrointestinal tract, predominantly through the influence to gastrointestinal microbiota. The assumption is that this effect is mainly achieved through its polymeric structure and poor solubility. We have derivatized pectin using alkaline hydrolysis in the presence of hydrogen peroxide. The obtained product was characterized by FTIR spectroscopy and showed symmetric and asymmetric stretching vibrations of the carboxylate groups in comparing with the starting material. This induces the possibility that the resulting derivative has specific reactions with the hydroxyl radicals formed in the Fenton reaction in vitro, and the conditions for the existence of a Fenton reaction in vivo exist in the gastrointestinal tract in the presence of Fe and vitamin C. EPR spectroscopy studies have shown that the apple pectin derivative in the Fenton reaction system induces the formation of carbon dioxide radicals anion (•CO2- radicals) instead of hydroxyl radicals. Preliminary studies for this radical have shown that it has a longer half-life than hydroxyl radicals and specifically reacts with certain proteins. This indicates the possibility that the pectin derivative that we get has additional positive effects in the gastrointestinal tract through the direct impact on the microbiota. Therefore, we examined antimicrobial effect of pectin and its derivatives and showed that the pectin derivatives obtained by alkaline hydrolysis of apple pectin (APDO) inhibited the growth of Saccharomyces cerevisiae and Candide albicans and had no influence on the growth of Escherichia coli and Staphylococcus aureus. We also analyzed the effects of pectin derivatives (apple and citrus) and oligosaccharides derived from polygalacturonic acid (APDO, CPDO and PGDO) on the growth of one Gram-negative representative strain of Escherichia coli ATCC 25922 and one Gram-positive representative strain of Staphylococcus aureus in the presence of substances that imitate conditions in a gastrointestinal tract that favors the Fenton reaction. It has been shown that the apple pectin derivative inhibits the growth of E. coli and S. aureus in these conditions, most likely through inhibiting key enzymes of the antioxidant system with carbon dioxide radical (E. coli SOD and GSH-Px, S. aureus GSH-Px). At the end we also examined the effect of pectin derivatives on the immune system. After 24 h, APDO did not significantly affect the CLNC viability, and after 48h caused a significant decrease in the viability of these cells, with the effect being dose dependent. Based on these results, it can be concluded that the observed decrease in CLNC viability, after 48h, under the influence of APDO is probably a consequence of the antiproliferative action of these molecules. Our results show that the obtained apple pectin derivative has a pronounced potential for application in regulation of gastrointestinal microbiota because it prevents the growth of pathogenic microorganisms without significant influence on the host's immune system.
true
Detekcija namerno izazvanih promena u sadržaju slike
Detection of intentionally made changes in image content
Multifraktalna analiza se pokazala kao dobar alat za analizu postojećih slika, kao i segmentaciju određenih regiona, izdvajanje ivica, uglova slike i slično. Kako kopirani i nalepljeni delovi imaju sličnu strukturu, može se primeniti multifraktalna analiza, koja u osnovi analizira samosličnost. Multifraktalni spektar daje globalni opis slike (ili, opštije, fenomena koji se ispituje). Vrednost Hölder-ovog eksponenta zavisi od položaja u strukturi i opisuje lokalnu regularnost signala. Naime, različiti objekti na slici imaju različite spektre, različite pozicije maksimuma, minimuma, prve nule itd, što se pokazalo kao interesantan skup različitih parametara pomoću kojih se mogu detektovati namerne promene na slikama. U ovoj disertaciji je opisana mogućnost korišćenja multifraktalnog spektra, kao i njegovih karakterističnih parametara, za segmentaciju i klasterovanje digitalnih slika. Predložena je nova metoda za detekciju CMFD promena na slikama, zasnovana na blokovskoj analizi, kao i na parametrima multifraktalnog spektra i nekim statističkim parametrima. Pre same detekcije, slika od interesa se deli na nepreklapajuće blokove kvadratnog oblika, čija dimenzija može varirati. Za svaki blok se računa karakteristični vektor. Za potrebe klasifikacije posmatranih blokova u izmenjene i neizmenjene, predložena je nova metaheuristika i semi-metrika, za analizu sličnosti između razmatranih blokova. Eksperimentalni rezultati pokazuju da se predložena metoda može uspešno koristiti za detekciju izmenjenih regiona slike, sa velikim procentom ispravne detekcije falsifikovanih regiona i malim procentom lažne detekcije. Pri tome, kompleksnost računanja samog algoritma je manja u poređenju sa postojećim metodama.
Digital images and video signals represent the most frequently transmitted contents. Namely, with the development of modern digital cameras and smartphones, the use of multimedia content increases every day. They are used in everyday life, for getting information and also as authenticated proofs or corroboratory evidence in different areas like: forensic studies, law enforcement, journalism and others. With the development of digital imaging software, it became very easy to change the content of an existing image: add or remove a specific content, or even make a new image from two or more images. These changes occur without significant degradation in image quality, and most often do not leave clearly visible traces. If the image doesn’t have a digital signature or a watermark embedded in it, we cannot be sure of the originality and authenticity of the image. It is therefore very important to develop methods that, without the existence of additional embedded information, will be able to determine if an image is original or forgery. The most popular method in image forensics is copy-move forgery detection (CMFD), where content changes are deliberately made by copying a part of an image and pasting it to another location within the image. The aim of such changes is to hide the specific content of the image, or to duplicate the content of the image. Such deliberately made changes, e.g. copied parts, can be: translated, rotated, scaled, multiple copied, and so on, and the modified image can be compressed subsequently. In the publicly available literature, there are many algorithms that have been developed for the purpose of detecting these types of changes. However, there are some lacks of existing methods: high sensitivity to texture features, high computational complexity and only tested with JPEG compression, high sensitive if copied and pasted parts are different size, unable to detect copy move with scaling, have a problem with detection if the pasted parts belong to different blocks, high sensitive on the low contrast in the image, give a small percentage of the accuracy detection of the forgery regions and a large percentage of false detected regions. Multifractal analysis has been proved to be a good tool for analyzing existing images, as well as segmentation of certain image regions, edges, angles and other. As copied and pasted parts have a similar structure, multifractal analysis can be applied, which basically analyzes self-similarity. The multifractal spectrum gives a global description of the image (or, more generally, the phenomenon being examined). The value of the Hölder exponent depends on the position in the structure and describes the local regularity of the signal. Namely, different objects in the image have different spectra, different positions of maximum, minimum, first zero, etc., which proved to be an interesting set of different parameters that can detect intentional changes in the images. This dissertation describes the possibility of using the multifractal spectrum, as well as its characteristic parameters, for the analysis and segmentation of digital images. A new block-based method for detecting CMFD changes, based on the multifractal spectrum parameters, as well on some statistical parameters is proposed. Before the analysis, images are divided into non-overlapping blocks of fixed dimensions. For each block, the characteristic features are calculated. In order to classify observed blocks, new metaheuristic and semi-metric are proposed, for similarity analysis between blocks. Experimental results show that the proposed method can be successfully used to detect forgery regions of the image, with a high percentage of correct detection of falsified regions and a small percentage of false detection. In addition, the algorithm’s computational complexity is smaller compared to existing methods.
true
Model radio-mreže za pristup u slučaju velikih skupova
Large events radio access network model
Analizi problema organizacije radio-pristupne mreže i analizi kapaciteta u slučaju velikih skupova do sada nije bila posvećivana dovoljna pažnja. Razlozi za to su višestruki, počev od neprepoznavanja značaja problema u praksi, istraživačkim i praktičnim fokusom na makro radio-pristupne mreže zbog velikog broja korisnika u njima, do nejasnih kriterijuma za kategorizaciju i opisivanje ovakvih skupova pa samim tim i odgoravarujućih radio-pristupnih mreža na ovakvim skupovima. Predmet ovog rada jeste upravo da predloži metodu i prikaže rezultate istraživanja ponašanja korisnika na velikim skupovima, kao i da ponudi odgovarajuće objektivne mere za kategorizaciju skupova, procenu potrebnih resursa uz zadržavanje određenog kvaliteta usluge i rizika da se takav kvalitet naruši...
Radio access network (RAN) organisation and capacity analysis in the case of special events has not been given proper attention. The reasons for this are manifold, ranging from failure to recognize the importance of the problem in practice, research and practice focus on the macro RAN due to the large number of users in them, the vague criteria for classification and description of such events and therefore appropriate RAN description for such events. The subject of this paper is to propose a method and present the results of user behaviour in large (special) events, as well as to provide appropriate objective measure for categorizing events, assessment of resources needed while keeping certain quality of service and the risk that such a quality will be distorted.
false
Adsorpciono-desorpcioni šum i vremenski odziv MEMS hemijskih i bioloških senzora
Adsorption-desorption noise and time response of MEMS chemical and biological sensors
Detekcija i identifikacija hemijskih supstanci i bioloških agenasa prisutnih u gasovitim i tečnim uzorcima uzetim iz okruženja ili organizma živih bića, od velikog su značaja za zaštitu životne sredine, poljoprivredu, medicinu i odbranu. Visokoosetljivi i selektivni hemijski i biološki senzori, malih dimenzija i prenosivi, koji omogućuju dobijanje informacija o prisustvu i koncentraciji specifičnih supstanci na bilo kom mestu u realnom vremenu, rešili bi mnoge probleme u ovim oblastima. Upravo za takvu vrstu primena pogodni su senzori izrađeni tehnologijama mikroelektromehaničkih sistema (MEMS), koje pored pomenutih karakteristika odlikuju mala potrošnja energije, rad sa malom količinom uzorka i niska cena zahvaljujući masovnoj proizvodnji. Oni su pogodni za realizaciju senzorskih mreža i prenosivih senzora sa bežičnom komunikacijom, kod kojih senzorski element, elektronski sklopovi za obradu signala i sklopovi bežičnog primopredajnika mogu biti integrisani u maloj zapremini. Princip rada velike grupe MEMS hemijskih i bioloških senzora zasnovan je na reverzibilnoj adsorpciji čestica supstance koju treba detektovati na senzorskom elemen- tu. Vremenski odziv senzora određen je trenutnim brojem adsorbovanih čestica. U sprezi sa drugim procesima bitnim za funkcionisanje senzora, stohastički adsorpciono- desorpcioni (AD) proces je izvor fundamentalnog adsorpciono-desorpcionog šuma koji može biti dominantan kod struktura mikrometarskih i nanometarskih dimenzija, a time i glavno ograničenje za minimalni detektabilni signal i druge performanse senzora. Cilj ove disertacije je razvoj i unapređenje teorijske osnove neophodne za projekto- vanje MEMS hemijskih i bioloških senzora nove generacije. Originalni naučni doprinos ostvaren je razvojem matematičko-fizičkih modela vremenskog odziva i AD šuma ovih naprava. Polazeći od formulisanja opštih teorijskih modela za jednokomponentnu i više- komponentnu adsorpciju, definisani su modeli za slučajeve od praktičnog značaja, koji se razlikuju po procesima bitnim za generisanje odziva senzora i njegovih stohastičkih fluktuacija. Obuhvaćeni su i slučajevi sprege AD procesa sa većim brojem drugih pro- cesa. Izvedeni su analitički izrazi za spektralnu gustinu srednje snage fluktuacija odziva senzora u ovim slučajevima i definisani su opsezi važenja aproksimativnih rešenja u funkciji parametara sistema. Razvoj svakog od ovih modela predstavlja važan korak ka sveobuhvatnoj teoriji vremenskog odziva i šuma MEMS hemijskih i bioloških senzora. VII U disertaciji su prikazani rezultati istraživanja kvalitativnog i kvantitativnog uticaja različitih procesa, parametara senzora i eksperimentalnih uslova na vremenski odziv i AD šum MEMS hemijskih i bioloških senzora. Ta istraživanja su zasnovana na primeni razvijenih teorijskih modela i analitičkih izraza. Primenom računarskih simulacija i teorijski definisanih kriterijuma potvrđena je opravdanost primene aproksimativnih modela u širokom rasponu vrednosti parametara sistema. Zaključcima proisteklim iz analize koja je izvršena u višedimenzionom prostoru parametara unapređeno je znanje o AD šumu i kinetici odziva adsorpcionih senzora. Osim što je prikazanom analizom dat nov doprinos naučnoj oblasti disertacije, ona je takođe ukazala na praktični značaj teorijskih modela vremenskog odziva i AD šuma MEMS adsorpcionih senzora. Poznavanje zavisnosti vremenskog odziva i šuma senzora od parametara sistema omogućuje optimizaciju dizajna senzora i eksperimentalnih metoda, kojom se postiže unapređenje performansi MEMS hemijskih i bioloških senzora i stvaraju uslovi da se iskoristi njihov veliki potencijal u praktičnim primenama. Razvijeni teorijski modeli omogućuju da se analiza izvrši tokom projektovanja senzorskog sistema i planiranja eksperimenata, što doprinosi ekonomičnosti i bržem razvoju mikro/nanosistema optimalnih performansi. Rezultati istraživanja takođe omogućuju pravilnu interpretaciju mernih rezultata i poboljšanje postojećih mernih metoda, ali i razvoj novih metoda, zasnovanih na analizi fluktuacija signala u frekvencijskom domenu, koje omogućuju dobijanje većeg broja podataka o adsorbovanim supstancama i njihovoj interakciji sa senzorskim elementom od konvencionalnih metoda. One uključuju metode za identifikaciju i detekciju supstance u složenom uzorku, zatim metode za istovremenu detekciju većeg broja supstanci u jednom uzorku, kao i metode sa inherentnom selektivnošću, koje omogućuju detekciju većeg broja supstanci jednim nefunkcionalizovanim senzorom. Svi pomenuti rezultati imaju širi naučni značaj, jer su, osim kod MEMS senzora, primenljivi i kod drugih hemijskih i bioloških senzora zasnovanih na adsorpciji, kao i kod drugih MEMS naprava (npr. radiofrekvencijskih rezonatora i oscilatora).
Detection and identification of chemical substances and biological agents in gaseous and liquid samples is very important for environmental protection, agriculture, healthcare and defense. Highly sensitive and selective chemical and biological sensors which are small and portable, and provide information about the presence and concentration of specific substances at any place in real time, would solve many problems in these fields. Especially suitable for this kind of applications are the sensors fabricated using the technologies of microelectromechanical systems (MEMS). Apart from the mentioned features, they are energy efficient, capable of operation with small sample quantities, and affordable due to the mass-production. They are suitable for sensor networks and wireless portable sensors, as the sensing element, signal processing circuits and wireless transceiver can be integrated in a small volume. The principle of operation of a large group of MEMS chemical and biological sensors is based on reversible adsorption of particles on a sensing element. The sensor response depends on the number of adsorbed particles. In conjunction with other processes relevant for sensors' operation, stochastic adsorption-desorption (AD) process is the source of fundamental adsorption-desorption noise which can dominante in structures of micrometer or nanometer dimensions, and thus can be the main limitation for the minimal detectable signal and other sensor performance parameters. The objective of this dissertation is to develop and improve a theoretical basis necessary for the design of new generation MEMS chemical and biological sensors. The original scientific contribution has been made through the development of mathematical- physical models of AD noise and time response of such devices. Starting from the formulation of the general theoretical models valid for single- and multi-component adsorption, simplified models are devised for cases of practical relevance, differing in processes which generate both the sensor response and its stochastic fluctuations. This includes the cases of AD process coupled with several other processes. Analytical expressions are derived for the AD noise power spectral density in these cases. The ranges of system parameters are defined, within which the approximate solutions are valid. The development of each of the models is an important step towards a comprehensive theory of both the time response and the noise in MEMS chemical and biological sensors. IX In the dissertation the results are presented of the research of both the qualitative and quantitative effects of various processes, sensor parameters and experimental conditions on time response and AD noise of MEMS bio/chemical sensors. The research is based on the developed theoretical models and analytical expressions. By using computer simulations and theoretically defined criteria, it has been proven that the application of the developed approximate models is justified in a wide range of system parameters. Conclusions from the analysis conducted in the multidimensional parameter space enhance the knowledge about sensor AD noise and the sensor time response. Apart from giving the contribution to the scientific field of the dissertation, the presented analysis highlights the practical value of the theoretical models of time response and AD noise in MEMS adsorption sensors. The knowledge about the dependence of the sensor time response and noise on system parameters enables optimization of both the sensor design and experimental methods, resulting in improved MEMS chemical and biological sensor performance and ensuring that the great potential of such sensors can be utilized in practice. The developed theoretical models enable performance optimization during the process of sensor design and experiment planning, leading to faster and more economical development of micro/nanosystems. The performed research is also useful for the correct interpretation of measurement results, the improvement of existing measurement methods and the development of new methods based on the frequency domain analysis of signal fluctuations. These methods yield more data about the adsorbed substances and their interaction with the sensing element, compared with conventional methods. They include identification and detection of a substance in a complex sample, simultaneous detection of multiple substances, and inherently selective methods for detection of multiple substances by using non-functionalized sensors. All the mentioned results are of a wider scientific significance as they are, apart from MEMS sensors, also applicable for other chemical and biological sensors based on adsorption, as well as for other MEMS devices (e.g. RF resonators and oscillators).
true
Modelovanje i dizajn optičkih rezonatora u integrisanoj silicijumskoj tehnologiji
Modeling and design of optical resonators in integrated silicon tehnology
Optički sistemi prenosa preuzimaju primat u telekomunikacijama, zahvaljujući sve većoj brzini prenosa signala, većem propusnom opsegu i mogućnostima koje imaju tehnike multipleksiranja. Na prostiranje svetlosti kroz gusto integrisane fotonske komponente ključnu ulogu imaju dimenzije i oblici talasovoda, kao i vrsta materijala od kog su napravljeni. Od načina prostiranja i specifičnosti fizičkih procesa će zavisiti i funcionalnost svake integrisane komponente, te je neophodno proučiti sve aspekte procesa i parametre koji na njih utiču. U okviru ove disertacije opisano je prostiranje elektromagnetskih talasa u fotonskim komponentama sa osvrtom na rebraste talasovode, pravolinijske i zakrivljene, kao i direkcione sprežnjake, što čini strukturu optičkog rezonatora. Objašnjeni su modeli koji opisuju fizičke procese u rezonatoru, sa posebnim naglaskom na teoriju spregnutih modova (CMT, Coupled Mode Theory). Dat je i kratak osvrt na silicijumske optičke strukture i proces izrade integrisanih silicijumskih rezonatora. Detaljno su opisana dva osnovna fizička procesa, sprezanje talasa iz susednog talasovoda i promena polarizacionog stanja u zakrivljenju, na osnovu kojih su izvedene transfer matrice, te prenosna funkcija optičkih rezonatora. Korišćenjem standardnog CMT modela su proračunati frekvencijski odzivi rezonatora različitih dimenzija i na taj način je proučen uticaj koji na odziv rezonatora imaju njegovi geometrijski parametri. Od geometrijskih parametara razmotreni su dužina područja kaplovanja i poluprečnik zakrivljenja talasovoda, koji definišu oblik rezonatora, te širina talasovoda, inklinacioni ugao i debljina oksidnog v sloja, koji određuju izgled poprečnog preseka talasovoda koji sačinjava taj rezonator. Proračuni su izvedeni pomoću COMSOL mode solver-a koji je korišćen za izračunavanje raspodele polja po poprečnom preseku talasovoda i sa njim integrisanog MATLAB paketa, pomoću koga je računat frekvencijski odziv rezonatora. Na osnovu proračuna su detaljno opisani uticaji većeg broja relevantnih parametara na funkciju rezonatora. Ovaj postupak otvara mogućnosti predikcije i projektovanja odziva rezonatora u realnim uslovima. Zbog odstupanja koja primenjen CMT model pokazuje u odnosu na eksperimentalne rezultate za istu geometrijsku strukturu rezonatora, postojeći model je modifikovan tako da simulira realan odziv uvođenjem pojma višestrukih prolaza elektromagnetskih talasa u rezonatorskoj strukturi. Modifikovani model doslednije opisuje realne fizičke procese i kvalitativno smanjuje odstupanje od eksperimentalnih rezultata. Inovirani model je dodatno korigovan podešavanjem geometrijskih parametara i proveren poređenjem sa eksperimentalno dobijenim odzivima optičkog rezonatora različitih dimenzija. Na ovaj način je izvršena korekcija polaznog CMT modela, adekvatnije simuliran realan fizički proces prostiranja talasa i izvršena provera početnih pretpostavki.
Optical systems are becoming more and more important in the field of telecommunications, due to their higher transportation speed, greater bandwidth and multiplexing techniques possibilities. Dimensions and shapes of waveguides play a major role in the physical aspect of light propagation through integrated photonic devices. Functionality of each integrated component depends on light propagation and specificity of physical processes. That is why it is necessary to examine all the aspects of propagation process and parameters that affect it. In this thesis, the propagation of electromagnetic waves in photonic components is described, especially in straight and curved rib waveguides and directional couplers, which represents the main structure of optical resonators. Models which describe physical processes in an optical resonator are fully explained, with special attention is devoted to Coupled Mode Theory (CMT). A short insight to silicon optical structures is given, as well as a description of integrated silicon resonator’s manufacturing process. Two fundamental physical processes are minutely described: wave coupling in adjacent waveguides and polarization rotation in curved parts of waveguide. Based on those processes, transfer matrices and transfer functions of optical resonators are derived. Frequency responses of resonators with various dimensions are calculated using standard CMT model. Geometric parameters influence on resonator’s response is studied, as well. Considered geometric parameters are: coupling length and radius of waveguide curvature (ring shape parameters), as well as waveguide width, sidewall angle, and oxide cover thickness (cross-section parameters). Calculations are performed by vii COMSOL field solver, used for cross-section field distribution calculation, and associated MATLAB software used to calculate resonator’s frequency response. As a result, influences of numerous relevant parameters on resonator’s function are fully described. This procedure gives possibility to predict and project resonator’s response in various working conditions. Because of divergences that applied CMT model shows compared to experimental results for the resonator with same geometric structure, existing model is modified by introducing repeated passing principle of electromagnetic waves through resonating structure. In that manner, modified model describes real physical process more consistently and qualitatively reduces divergences from experimental results. Innovated model is additionally corrected by adjusting geometrical parameters of resonator. The modified model suggested in this thesis simulates the physical processes during the propagation of electromagnetic waves in a more adequate way, which is showed by comparing its results with experimentally obtained responses of racetrack resonators of various dimensions. Consequently, the correction of initial CMT model is accomplished and verification of initial assumptions is implemented.
true
Analiza metoda za zaštitu vetroelektrana od direktnog atmosferskog pražnjenja
Analysis of the methods for wind turbine protection against direct atmospheric discharges
Tema doktorske disertacije je veoma aktuelna zbog intenzivnog širenja primene vetroelektrana širom sveta sa jedne strane, a sa druge strane zbog njihove osetljivosti na atmosferske udare i teških posledica koje pri tome nastaju. Vetroturbine su visoke nestandardne konstrukcije sa rotirajućim elementima, koje se obično instaliraju na lokacijama na kojima je keraunički nivo veoma visok i loši uslovi uzemljenja, i kao distribuirani izvori električne energije imaju povećan rizik od udara groma. Najistaknutiji delovi vetroturbina, kao što su dugačke elise i gondola, tipično su napravljeni od kompozitnih materijala, koji ne mogu uspešno da izdrže direktne udare groma sa strujama intenziteta do 200 kA, niti proticanje struja pražnjenja po njihovoj površini, pri čemu značajan deo ove struje, prilikom sprovođenja u zemlju, ima putanju preko ili u blizini praktično svih vitalnih komponenti vetroturbine. Kontinuiran razvoj zaštitnih sistema vetroelektrana od atmosferskih pražnjenja ima izuzetan značaj kako bi se smanjile štetne posledice zbog mehaničkih oštećenja elisa, kvarova na električnoj i elektronskoj opremi i ispada vetroelektrana iz pogona. Originalna teorijska razmatranja, primena softversko-numeričkih postupaka, i veoma izazovno i finansijski zahtevno eksperimentalno ispitivanje, omogućili su da se u doktorskoj tezi na kvalitetan način izvrši analiza metoda za zaštitu vetroelektrana od direktnog atmosferskog pražnjenja i dobiju značajni doprinosi. Istraživanja koja su sprovedena u toku izrade doktorske disertacije su dosta složena, zbog prirode pojava koje se ispituju i nestandardnih načina ispitivanja, a istovremeno i multidisciplinarna, jer obuhvataju dobro poznavanje sledećih oblasti: tehnika visokog napona, vetroenergetika, fizika atmosferskih pražnjenja i gromobranska zaštita. U prvom delu disertacije objašnjen je sve veći značaj koje vetroelektrane imaju u savremenom decentralizovanom EES-u, dat je pregled aktuelnog stanja i standarda u vezi sa atmosferskim pražnjenjima u vetroturbine i ukazano je na specifičnosti metoda njihove gromobranske zaštite. Prikazani su glavni delovi savremene vetroturbine, koja je veoma kompleksan uređaj, pri čemu je posebna pažnja posvećena analizi karakteristika i načinima zaštite komponenti kroz koje protiče glavni deo struje pražnjenja. Razmotreni su različiti principi za zaštitu kompozitnih materijala od atmosferskog pražnjenja i objašnjeno je koje vrste kompozita i u kojoj meri mogu da se koriste za izradu elisa vetroturbina. Dat je pregled relevantnih parametara i karakteristika direktnih atmosferskih pražnjenja koja se tipično dešavaju na vetroturbinama koje su u pogonu. U tezi je razmotren opšti princip gromobranske zaštite vetroelektrana, zasnovan na određivanju potrebnog nivoa zaštite gromobranske instalacije. Analizirana su četiri nivoa zaštite, pri čemu su za svaki od njih definisani kriterijumi koje treba da zadovolji svaki od elemenata gromobranskog zaštitnog sistema. Ispitivanje ovih kriterijuma treba da obezbedi da verovatnoća efikasnog delovanja zaštite bude u okviru unapred definisanih granica, koje su karakteristične za svaki od nivoa zaštite. Predloženi su postupci za izračunavanje računske efikasnosti gromobranske instalacije vetroelektrane i udarnog rastojanja na osnovu kojih se jedinstveno određuje nivo zaštite štićene vetroelektrane...
The subject of this PhD Thesis is a very current subject due to intensive progress in applying wind turbines around the world on one hand, and on the other hand due to their sensitivity to atmospheric discharges and serious consequences that can arise. Wind turbines are tall and nonstandard constructions with rotating elements, usually installed in locations at which the keraunic level is very high and grounding conditions are bad, and as distributed sources of electrical energy they are more prone to lightning stikes. The most prominent parts of the wind turbines, e.g. long blades and nacelle are typically made of composite materials, which cannot successfully withstand direct lightning strikes with currents up to 200 kA, neither current disharge flow on their surface, during which a significant part of this current at grounding has a path over or in vicinity of practically all vital components of the wind turbine. Continuous development of protective wind turbine systems for atmospheric discharges has a remarkable significance in order to reduce negative consequences of mechanical damages of blades, failure of electrical and electronic equipment and outage rate. Original theoretical discussions, applying software numerical procedures and very chalenging and financially demanding experimental testing enabled a high quality analysis to be preformed in the PhD Thesis regarding direct atmospheric discharges and to significally contribute to the subject. The research conducted during work on this PhD Thesis are rather complex due to nature of the events that are the subject of the research and non-standard ways of testing and at the same time multidisciplinary because they involve a good knowledge of the following fields: highvoltage technique, wind energy, atmospheric discharge physics and lightning protection. In the first part of this thesis the increasing significance that wind turbines have in contemporary decentralized EES, an overview of actual state and standards with regards to atmospheric discharges into wind turbines, and the specifics of the methods of their lighting protections have been pointed out. The main parts of the modern wind turbine are shown, because the wind turbine is a very complex device, and special attention is paid to the analysis of characteristics and means of protection of components through which the main part of the discharge current flows. Different principles for protection of composite materials from atmospheric discharges are discussed and explanations on which types of composites can be used to construct wind turbine blades and to what extent. An overview is given of the relevant parameters and characteristics of direct atmospheric discharges, which typically happen on wind turbine in service. The general principle of lightning protection of wind turbines is discussed in the thesis, based on determination of the necessary level of protection of electrical installation for lightning protection. Four levels of protection are analyzed, and for each of them criteria are defined which need to be meet by each element of the lightning protection system. The testing of these criteria needs to enable the probability of efficient work of protection is within predefined limits which are characteristic for each level of protection. The proposed procedures for calculating evaluation efficiency of lightning protection installation of wind turbine and striking distance based on which the level of protection of the protected wind turbine is uniquely determined...
false
Sistem za funkcionalnu električnu stimulaciju baziran na elektrodama sa većim brojem provodnih polja
System for functional electrical stimulation bsed on multi-pad electrodes
Funkcionalna električna stimulacija (FES) je tehnika artificijelne aktivacije inervisanih skeletnih mišića pacijenata nakon povrede ili bolesti centralnog nervnog sistema (CNS) putem ekscitacije motoneurona električnim impulsima. Ova metoda omogućuje pacijentima koji su izgubili moć voljne kontrole pojedinih mišića ili kompletnih udova da korišćenjem elektronskog uređaja ( stimulatora) povrate neke od funkcija neophodnih za povećanje kvaliteta života. Sam proces električne stimulacije obuhvata interpretaciju komandi pacijenta ili terapeuta, definisanje parametara stimulacije, generisanje strujnih impulsa precizno određenih parametara (oblik impulsa, amplituda, trajanje i frekvencija), konvertovanje struje elektrona u struju jona koja stvara uslove za ekscitaciju neurona. Primena električne stimulacije je započela i razvije se intenzivno od šezdesetih godina dvadesetog veka, ali su unapređenja u svim aspektima neophodna u cilju širokog prihvatanja i upotrebe sistema za FES. Najveći problemi pri projektovanju i primeni FES sistema su definisanje i realizacija upravljačkog mehanizma, nefiziološko aktiviranje mišića koje rezultuje zamorom i neselektivnošću stimulacije, odnosno potreba za nezavisnim upravljanjem pojedinačnim mišićima uprkos prostornoj blizini njihovih motornih neurona i aktiviranju refleksnih odgovora ekscitacijom senzornih neurona. Rešavanjem ovih otvorenih pitanja bavi se više naučnih grupa širom sveta i koristi različite pristupe. Tako je razvijena tehnologija implantiranih sistema kojom je moguće postići selektivnu stimulaciju motoneurona, ali koja sa sobom nosi rizike vezane za hiruršku operaciju . Visoka cena, invazivnost, potreba za održavanjem sistema koja uključuje mogućnost ponovljene operacije kao i tehnička nesavršenost implantiranih uređaja su, uz činjenicu da dobitak za pacijenta nije dovoljan, primarni razlozi malog broja ovakvih sistema u svetu. Sa druge strane, površinska stimulacija ne konkuriše implantiranom sistemu u pogledu selektivnosti stimulacije, estetike i jednostavnosti pripreme sistema pri svakoj upotrebi. Kao unapređenje tehnike površinske stimulacije, u poslednjoj deceniji razvijena je metodologija korišćenja površinske elektrode sačinjene od većeg broja nezavisnih polja (matrične elektrode) čime su otklonjeni neke od nedostataka konvencionalne površinske stimulacije. Takođe, korišćenje matričnih elektroda otvara potencijal uvođenja novih obrazaca stimulacije namenjenih približavanju veštački generisanih i prirodnih pokreta...
Functional electrical stimulation (FES) is a technique for activation of innervated skeletal muscles in patients after injury or disease of the central nervous system (CNS) by controlled bursts of electrical impulses. The FES allows patients who have lost their volitional control of movement to restore some of the functions necessary for the improved quality of life. Аn effective FES comprises the following: interpretation of command signal originating from a patient or a therapist, selection of the appropriate set of parameters of stimulation for a given patient and her/his status, generation of electrical pulses with precisely defined parameters (pulse shape, amplitude, duration, and frequency) that are converted into ionic current in the tissues where the neurons will be excited. FES has been introduced in sixties of the XX century for the assistance of the dropfoot and closing of the hand. New designs and improved knowhow in the domain of neurosciences contributed to the improved performance of the FES; yet, it still has not reached its maturity and it is not used widely in the clinical and home environment. Major problems are the non-physiological activation of muscles that results in fast occurrence of fatigue, lack of selectivity (the independent management of individual muscles, despite their spatial proximity innervations zone) and control which would allow the user to concentrate on what she/he is doing and not on how to use the FES. One approach presented in literature and brought to the market suggests implanted systems that provide selective stimulation of motoneurons but introduce risks associated with surgical procedure. More precisely, the high cost, invasiveness, the need to maintain such a system which includes the possibility of repeated surgery and technical imperfection of implanted devices are the reasons that only a small number of such systems have been accepted by the users. On the other hand, conventional surface stimulation cannot achieve sufficient level of selectivity, aesthetics and ease of the system set-up for application, which are the reasons for slow translation of the technology to practice. During the last decade methodology based on electrode comprising multiple, independent pads (multi-pad electrode) succeeded to resolve some of drawbacks associated with conventional surface electrical stimulation. In parallel, the utilization of multi-pad electrodes permits application of new stimulation paradigms designed for generating "natural like" artificial movements...
false
Analiza obeležja u govornom signalu za potrebe prepoznavanja multimodalnog govora
Analysis of the speech signals features for multimodal speech recognition.
U sistemima autpmatskpg preppznavaoa gpvpra sve znašajniju ulpgu igra i multimpdalni gpvpr. Razlišiti mpdpvi gpvpra npse sa spbpm i razlišita akustiška pbeležja. Pd ppsebnpg interesa su mpdaliteti nprmalnpg (prirpdnpg) gpvpra i gpvpra generisanpg щapatpm kap i oihpve kpmbinacije (kada je sistem treniran nprmalnim gpvprpm, a preppznaje se щapat kap i pstale ppcije). Stpga ppstpje razlišiti scenariji i mpgu se ppdeliti na usaglaщene („nprmalan/nprmalan“ i „щapat/щapat“) i neusaglaщene („nprmalan/щapat“ i „щapat/nprmalan“). Za eksperimentalni dep pvpg rada kpriщćenp je jedanaest vektprskih pbeležja (LPCC, LFCC, TELFCC, MFCC, TEMFCC, GFCC, TEGFCC, PLPCC, TEPLPCC, RASTACC i TERASTACC) kpja na razlišite našine ppisuju akustiške karakteristike multimpdalnpg signala. Urađena je statistiška analiza gde je nivp ppuzdanpsti ppstavljen na 95% i na bazi tpga su rašunati intervali ppuzdanpsti. Ppsebnp je istaknuta pptreba za upptrebpm nprmalizacije (Cepstral Mean Subtraction i RelAtive SpecTrA) i ppkazanp je da se kpriщćeoem iste uspeщnpst preppznavaoa mpže znatnp pppraviti. Tp je ppsebnp izraženp kada su u pitaou neusaglaщeni scenariji (npr. kpd scenarija „щapat/nprmalan“ ppbpljщaoe mpže ići i dp 47%). Primenpm nelinearnpg Teager Energy pperatpra kpji je dpbar za ppis turbulentnpg kretaoa vazduha, a kpje se javlja prilikpm щapata, dpdatnp se mpže ppvećati uspeщnpst preppznavaoa (npr. i dp 21% kpd scenarija „щapat/nprmalan“ za vektprskp pbeležje TERASTACC). Ppkazanp je da i u usaglaщenim scenarijima primena Teager Energy pperatpra daje ppbpljщaoe uspeщnpsti preppznavaoa. iv Sa aspekta razlišitih vektprskih pbeležja isppstavlja se da TERASTACC skprp u svim slušajevima daje najbplje rezultate. Zatim pp uspehu slede TEPLPCC, TEMFCC, MFCC i takp dalje. Štp se tiše dužine vektpra, rezultati sa vektprpm pd 24 kpeficijenta pd kpjih su 12 kepstralni i 12 delta kepstralni (sa primenpm nprmalizacije) su se ppkazali najbpljim. Psim pvpg tipa vektpra kpriщćena su jpщ tri (jedan pd 12 kpeficijenata bez nprmalizacije i dva sa nprmalizacijpm - pd 12 i 36 kpeficijenata). Metpda za preppznavaoe, bazirana na dinamiškpm prpgramiraou i DTW algpritmu ppkazala se kap efikasana i ppuzdana za brzp nalažeoe pptimalne staze pp kpjpj se rašuna distanca između vektpra kepstralnih kpeficijenata. Dpbijeni rezultati ukazuju na nepphpdnpst kpriщćeoa nprmalizacije i Teager Energy pperatpra pri preppznavaou multimpdalnpg gpvpra sa akcentpm na preppznavaoe щapata.
In automatic speech recognition systems, multimodal speech is playing an increasingly important role. Different modes of speech involve different acoustic features. Particular importance is given to normal (natural) speech, whisper and their combinations (when the system is trained by normal speech and then used to recognize whisper, and vice versa). The match scenarios: „normal/normal“ and „whisper/whisper“, and mismatch scenarios: „normal/whisper“ and „whisper/normal“ are considered here. The experiment was performed using eleven vector features (LPCC, LFCC, TELFCC, MFCC, TEMFCC, GFCC, TEGFCC, PLPCC, TEPLPCC, RASTACC and TERASTACC) which provide different descriptions of acoustic properties of multimodal speech. A statistical analysis was conducted at a 95% confidence level, based on which confidence intervals were calculated. Particular emphasis was placed on the use of normalization (Cepstral Mean Subtraction and RelAtive SpecTrA) which considerably increased the word recognition rate. This was especially evident in mismatch scenarios (e.g. in „whisper/normal“ scenarios, the increase was up to 47%). The use of the nonlinear Teager Energy operator in describing turbulent airflow occurring during whisper production additionally increased the word recognition rate (i.e. up to 21% in a „whisper/normal“ scenario for the TERASTACC vector feature). The word recognition rate in match scenarios was also improved by the Teager Energy operator. As regards vector features, the best results were achieved by TERASTACC in almost all cases, followed by TEPLPCC, TEMFCC, MFCC etc. As for vector length, very good performance was obtained by the vector consisting of 24 coefficients i.e. 12 cepstral and 12 delta cepstral coefficients (with normalization included). Three additional types of vectors were used (one non- vi normalized vector containing 12 coefficients, and two normalized vectors consisting of 12 and 36 coefficients, respectively). Recognition was performed by the DTW method based on dynamic programming. The method proved effective and reliable in finding the optimal path to be used in computing the distance between feature vectors. The final results showed that normalization and the Teager Energy operator were beneficial for multimodal speech recognition with a specific focus on whisper.
true
Efekti jonizujućeg zračenja u fazno promenljivim memorijama
Radiation effects in phase change memories.
Fazno promenljive memorije ubrajaju se u neke od najaktuelnijih postojanih memorija zbog relativno jednostavnog načina proizvodnje, velike izdržljivosti, velike brzine upisa/brisanja i male potrošnje. One koriste jedinstvenu osobinu fazno promenljivih materijala da poseduju veliku razliku otpornosti između kristalnog i amorfnog stanja i uz to imaju mogućnost veoma brze kristalizacije. Ova disertacija se bavi ispitivanjem efekata jonizujućeg zračanja u fazno promenljivim memorijama, što je od značaja usled sve veće primene memorija u različitim radijacionim sredinama. Numeričkim proračunima ispitivani su efekti koje snopovi protona, alfa čestica, jona gvožđa i argona, različitih energija, proizvode u fazno promenljivih memorijama sa fazno promenljivim materijalom različitih debljina. Kao fazno promenljivi materijali korišćeni su halkogena jedinjenja: Ge2Sb2Te5 i AgSbSe2. Prolaskom protona, alfa čestica, Fe i Ar jona kroz ćeliju FPM-a, oni gube svoju energiju na jonizaciju, izmeštanje atoma ili fononsko pobuđivanje rešetke. Svi ovi tipovi energetskih gubitaka su posebno analizirani. Proračunate su vrednosti jonizacionih doza i doza izmeštanja, kao i fononski gubici za ispitivane naelektrisane čestice i fazno promenljivu memoriju. Dobijeno je da su jonizacioni gubici u AgSbSe2 materijalu veći u odnosu na Ge2Sb2Te5 materijal, za jone posmatranih energija. Analizom izmeštanja atoma iz njihovog položaja u rešetki, ustanovljeno je da je AgSbSe2 materijal otporniji na izmeštanje atoma usled prolaska ispitivanog zračenja. Najveća doza izmeštanja je dobijena u slučaju prolaska jona gvožđa kroz fazno promenljivu memorijsku ćeliju. Izmeštanje atoma u fazno promenljivim materijalima može dovesti do amorfizacije materijala, jer izmešteni atomi mogu narušiti uređenu rešetku kristalnog stanja. Uticaj brzih teških jona velikog linearnog prenosa energije na fazno promenljive materijale je od posebnog značaja, i ustanovljeno je da bi usled sve veće miniturizacije fazno promenljivih memorija moglo doći do odstupanja od redovnog rada fazno promenljive memorije usled amorfizacije materijala oko trajektorije ovih upadnih jona, što može dovesti do promene stanja ćelije. Detaljnom analizom, koja je obuhvatala ispitivanje efekata na različitim temperaturama, ustanovljeno je da se u AgSbSe2 materijalu proizvodi manji sloj amorfizovanog materijala nego u slučaju Ge2Sb2Te5 materijala, dakle on je za nijansu otporniji na uticaj ove vrste zračenja. Dobijeni rezultati su od značaja u inženjerskoj praksi zbog sve veće primene postojanih memorija u nuklearnoj tehnici i svemirskoj tehnologiji. Dalje, prikazani rezultati mogu biti od značaja proizvođačima komercijalnih fazno promeneljivih memorijskih komponenata, kako bi se izradile fazno promenljive memorijske ćelije optimalne strukture za korišćenje u poljima jonizujućeg zračenja.
Phase change memory is an emerging non-volatile memory technology due to relatively easy fabrication of these memories, high endurance, fast read/write and low energy consumption. It uses the unique property of phase change materials to have a pronounced contrast in resistivity between their crystalline and amorphous state with very fast crystallization kinetics. In this dissertation, radiation effects in phase change memories are being analysed, corresponding to greater interest in this subject, due to widely spreed usage of these memories in different radiation fields. Numerical calculations are being carried out in order to estimate the effects of proton, alpha, Fe and Ar ions of different energies, produced in Ge2Sb2Te5 and AgSbSe2 phase change materials with various thicknesses. When protons and ions traverse the material, they lose their energy on ionization, atom displacements and the production of phonons and all of these losses have been fully analysed. Ionizing and nonionizing energy losses, phonon losses, as well as absorbed dose and displacement damage dose in the phase change memories for investigated charged particles and examined cells have been calculated. It is shown that the ionizing losses are greater in AgSbSe2 material compared to Ge2Sb2Te5 material. By analyzing the displacements of atoms from their lattice sites, it has been found that AgSbSe2 is more robust to this kind of damage than Ge2Sb2Te5 material. The highest damage has been observed in the case of heavy ions (Fe) interactions. Radiation could induce the amorphization (crystalline to amorphous phase transition) in the phase change memory cell by creating a large number of atomic displacements which disrupt the highly ordered lattice structure of the crystalline phase. The effects of ions with high linear energy transfer (i.e. swift heavy ions) on phase change memories have been of special interest, due to theirs unique properties. It has been established that due to constant reduction of dimensions of phase change memory cells, amorphization in the vacancy of ions tracks, called latent tracks, could disrupt the functionality of memory cells. Comprehensive analysis, including calculations involving different temperatures, showed that radius of latent track is smaller in AgSbSe2 material, compared to Ge2Sb2Te5. All obtained results are of great importance for engineering practice due various applications of non-volatile memories in nuclear technologies and space environment. Furthermore, these results can also be of interest for commercial purposes, allowing the manufacturers of phase change memories to develop the phase change memories of optimal design, considering radiation hardness of the memory.
true
Efikasna paralelna kompresija polja visina
Efficient parallel compression of height fields
Disertacija predlaže novu metodu za brzu kompresiju i dekompresiju pravilnih polja visina, sa ili bez gubitaka. Pravilna polja visina su najčešće usvojen način predstavljanja površi uzorkovanih na pravilnim rastojanjima duž dve ose kartezijanskog koordinatnog sistema u horizontalnoj ravni. Zahvaljujući napretku u tehnici detekcije na daljinu (eng. remote sensing) u protekloj deceniji, koja je dostigla horizontalne i vertikalne rezolucije uzorkovanja reda metra i decimetra, respektivno, potreban je značajan prostor za smeštanje opisa površi realnih terena. Pri navedenim rezolucijama, površ neke planete veličine Zemlje zahteva nekoliko desetina terabajta. Kompresija polja visina je potrebna ne samo za kompaktno smeštanje, već i za efikasan prenos i manipulaciju tako velikim količinama podataka. Razvijena metoda pogodna je za SIMD paralelnu implementaciju, pa samim tim i za arhitekture modernih grafičkih procesora, koji značajno nadmašuju centralni procesor u brzini računanja, a već su dostupni u kućnim računarima. Kompresija sa gubicima postiže se aproksimacijom polja visina skupom kvadratnih Bezjeovih površi. Bezjeove površi su izabrane zato što su relativno jednostavne za računanje i poseduju nekoliko poželjnih osobina, koje su od suštinskog značaja za primenu predložene metode u inženjerstvu. Kompresija bez gubitaka postiže se dodavanjem dva sloja reziduala na aproksimaciju. Dodatno, predlaže se opcioni paralelni algoritam, specijalizovan za kompresiju reziduala. Mogućnost dodatne primene ovog algoritma dozvoljava balansiranje između stepena kompresije i brzine dekompresije metode. Predloženu metodu odlikuju jedinstvene osobine, koje nisu prisutne ni u jednoj drugoj metodi kompresije polja visina. Metoda omogućava nezavisnu dekompresiju pojedinačnih tačaka, kao i progresivnu dekompresiju. Čak i u slučaju dekompresije sa gubicima, dekomprimovana površ je inherentno bez pukotina. U poređenju sa savremenim kompetitivnim metodama, koje se izvršavaju na GPU (grafičkom procesoru), predložena metoda, u kombinaciji sa II široko rasprostranjenom opštenamenskom metodom kompresije bez gubitaka (poput DEFLATE, koja se koristi u popularnom alatu za kompresiju podataka ZIP), ili kombinovana sa specijalizovanom metodom za kompresiju reziduala bez gubitaka, postiže poredive stepene kompresije. Efikasnost metode potvrđena je kroz CUDA implementaciju algoritama kompresije i dekompresije. Implementacija osnovne metode (bez dodatne kompresije reziduala), po cenu razumno manjeg stepena kompresije, blago nadmašuje brzinu savremene kompetitivne metode za velika opterećenja i značajno za mala, ostvarujući sve gore navedene jedinstvene karakteristike. U cilju provere praktične upotrebljivosti i efikasnosti metode u realnoj primeni, algoritam za dekompresiju integrisan je u sistem za vizuelizaciju terena u realnom vremenu. Rezultati pokazuju da, zahvaljujući brzoj dekompresiji polja visina, upotreba komprimovanih podataka ostvaruje nešto kraće vreme crtanja slike u odnosu na upotrebu nekomprimovanih podataka. Međutim, čak i malo ubrzanje crtanja slike, u kombinaciji sa gotovo potpunim rasterećenjem centralnog procesora i smanjenim saobraćajem između sistemske memorije i memorije grafičkog procesora zahvaljujući manjoj veličini komprimovanih podataka, opravdava primenu predložene metode u sistemima za vizuelizaciju terena u realnom vremenu.
This thesis presents a novel method for fast lossy or lossless compression and decompression of regular height fields. Regular height fields are a commonly adopted solution for representing surfaces sampled at uniform rates along the two Cartesian axes in the horizontal plane. Due to the improvements in the remote sensing technology during the last decade, which has reached horizontal and vertical sampling resolutions of the order of a meter and a decimeter, respectively, significant storage space is needed for the surface of real terrains. At these resolutions, the surface of a planet of the size similar to Earth requires a few dozens of terabytes of storage space. Compression of height fields is needed not only for efficient storage, but also for efficient transfer and manipulation of such large amounts of data. The developed method is suitable for SIMD parallel implementation and thus inherently suitable for modern GPU architectures, which significantly outperform modern CPUs in computation speed, and are already present in home computers. Lossy compression is achieved by approximating the height field with a set of quadratic Bezier surfaces. Bezier surfaces have been chosen because they are relatively simple to compute and have several desirable features, essential for the application of the method in engineering. Lossless compression is achieved by superimposing two layers of residuals over the lossy approximation. In addition, an optional parallel algorithm, specialized for compression of residuals is proposed. The possibility of additional application of this algorithm allows for a tradeoff between the compression ratio and the decompression speed of the method. The proposed method is characterized with unique properties, which are not present in any of the height field compression methods. The method allows independent decompression of individual data points, as well as progressive decompression. Even in the case of lossy decompression, the decompressed surface is inherently seamless. In comparison with the GPU-oriented competitive state-of-the-art method, the proposed method, combined with a widely available general purpose lossless compression method (such as DEFLATE, used in the popular ZIP data compression utility), or combined with a specialized method for lossless compression of the residuals, achieves IV comparable compression ratios. The method's efficiency was confirmed through a CUDA implementation of compression and decompression algorithms. The implementation of the basic method (without additional compression of residuals), at the expense of reasonably worse compression ratio, slightly outperforms the competitive state-of-the-art method for very high workloads and considerably for lower workloads, achieving all of the above mentioned unique features. To verify the method's practical usability and efficiency in a real application, the decompression algorithm was integrated into a real-time terrain visualization system. The results show that, thanks to the fast height field decompression, using compressed data gives a slightly higher frame rate than using uncompressed data. However, even the small increase in frame rate, along with the facts that the CPU is almost completely offloaded and that lower traffic between the system and the GPU memory is achived due to smaller size of the compressed data, justifies the usage of the proposed method in real-time terrain visualization systems.
true
Ocjena ustavnosti akata i radnji sudske vlasti
Predmet doktorske disertacije odnosi se na problematiku ocjene ustavnosti akata i radnji sudske grane vlasti. Cilj doktorske disertacije baziran je na pronalaženju adekvatnog modus vivendi između sudstva, koje predstavlja primarni i najefikasniji institucionalni sistem zaštite prava pojedinaca od uzurpacije državnih struktura i ustavnog suda - ultimativnog tumača i zaštitnika ustava, koji putem kontrole ustavnosti obezbjeđuje neposrednu zaštitu ustava od povreda bilo koje vrste, pričinjene od strane bilo kog državnog organa, anuliranjem i obesnaženjem svih akata i postupanja suprotnih tekstu i duhu primarnog prava. Osnovno pitanje cjelokupnog rada jeste da li je sudskoj vlasti potrebna ustavnosudska kontrola i korektiv, i ukoliko jeste, u kom obimu odluke kontrolora ustavnosti mogu da preispituju i ograničavaju pravnu dejstvenost odluka redovne judikature, imajući u vidu jedan od osnovnih postulata pravnog sistema – konačnost i vanistancionu nepromenjivost sudske presude? U tom pravcu, autor je postavio hipoteze na kojima je zasnovano istraživanje pri izradi disertacije: - ustavni sud je garant i zaštitnik ustavnosti i funkcionalnosti pravnog poretka od bilo koje vrste povrede ili uzurpacije, pa samim tim i i one koja nastaje postupanjem i aktima organa sudske grane vlasti. Shodno tome, ustavni sud ima ingerencije nad sudskom granom vlasti, te može preispitivati i ograničavati pravnu dejstvenost odluka redovne sudske vlasti; - postupak po ustavnoj žalbi je vid provjere kvaliteta rada sudske vlasti; - zahvaljujući ovoj nadležnosti ustavnog suda mijenja se klasična fizionomija ovog organa i on se približava sudskoj vlasti. Metodološki posmatrano, istraživanje je zasnovano na teorijsko- analitičkom pristupu i kritičkoj analizi. Ove tri grupe pitanja prvo su kritički analizirane sa aspekta različitih poimanja i stavova koji su o njima iznijeti u pravnoj doktrini, a koji su nerijetko imali i značajnog praktičnog odjeka, a zatim su praćene stavovima autora, koji potvrđuju naučnu osnovanost hipoteza na kojima je ova disertacije bazirana.
The subject of the doctoral dissertation concerns the issue of reviewing the constitutionality of laws and actions of the judicial branch of government. The aim of the dissertation is based on finding the proper modus vivendi between the judiciary, which is the primary and most effective institutional system of protection of individual rights against usurpation of state structures and the constitutional court - the ultimate interpreter and defender of the constitution, which through control of the constitutionality of the constitution provides immediate protection from harm of any kind, caused by any governmental authority, annihilating and invalidation of laws and acts contrary to the text and spirit of the primary law. The main question of the entire work is whether the legal authority needs the constitutional control and a corrective, and if so, to what extent the decisions of controllores of constitutionality can review and limit the legal effectiveness of the decisions made by regular judicature, bearing in mind one of the basic principles of the legal system - the finality and non-jurisdictional immutability of a court ruling? In this regard, the author has set hypothesis on which the study is based on whilst preparing the thesis : - The constitutional court is the guarantor and protector of the constitutionality and the legal system functionality from any kind of injury or usurpation, and therefore those resulting from actions and acts of authority of the judicial branch of government. Accordingly, the constitutional court has jurisdiction over the judicial branch of government, and may review and limit legal effect of decisions of the regular judicial authorities; - Procedure by constitutional complaint is a form of quality assurance of the judiciary; - Thanks to the competence of the constitutional court, the classical physiognomy of this organ is changed and it is approaching the judiciary. In methodological point of view, the study is based on theoretical and analytical approach and critical analysis.
false
Istraživanje termomehaničkih procesa prilikom razgradnje tečnih goriva velikih gustina u fluidizovanom sloju
Research of thermomechanical proces during decomposition of high density liquid fuel in the fluidized bed
U zemljama, poput naše, sa ograničenim izvorima primarne energije, sve je izraženija potreba za korišćenjem sopstvenih raspoloživih goriva, ne bi li se delimično supstitusali uvezeni energenti. U tom slučaju, nekonvencijalna tečna goriva velikih gustina (TTG) predstavljaju jedno od mogućih izbora goriva, koji podrazumeva i primenu tehnologija sagorevanja u fluidizovanom sloju (FS). Činjenica da se u FS kotlovima mogu sagorevati goriva i sa 85% inertnog materijala, sa efikasnim odsumporavanjem i nižom emisijom NOx koja zadovoljava ekološke standarde, ovoj tehnologiji sagorevanja u odnosu na druge daje izrazitu prednost. U okviru teze urañena su obimna teorijska i eksperimentalna istraživanja složenih i meñusobno spregnutih strujno-termičkih procesa koji se odvijaju pri termičkoj dezintegraciji TTG-a u fluidizovanom sloju (FS), u cilju razvoja tehnologije za energetski efikasno sagorevanje ovakvih materija na ekološki prihvatljiv način. Predstavljeni su rezulti eksperimenata sagorevanja različitih tečnih nekonvencijalnih goriva u FS, na poluindustrijskoj laboratorijskoj aparaturi sa akcentom na ostvarivanje niske emisije štetnih materija u dimnim gasovima, na potpunost sagorevanja i lociranje zone intenzivnog sagorevanja. Detaljno poznavanje i razumevanje procesa doziranja goriva, njegovog mešanja sa inertnim materijalom i vazduhom u okviru radnog prostora ložišta je od izuzetnog značaja za definisanje optimalnih radnih parametara FS, a numeričke simulacije predstavljaju efikasno sredstvo za ostvarenje tog cilja. U tezi su predstavljena dva 2D matematičko-numerička (CFD) modela za detaljnu simulaciju prodora bočnog mlaza u fluidizovani sloj: sopstveni fluid-porozni medijum i granularni Euler-Euler model simulacije FS, koja koriste dvo-fluidni pristup modeliranja gde se gas i gusta faza FS posmatraju kao dva fluida sa različitim karakteristikama. Fluid-porozni medijum model prodora mlaza u FS, je zasnovan na postupku proračuna kod koga se gusta faza FS posmatra kao fiksni porozni medijum. Interakcija izmeñu gasa i čestica, kao i pojava mehuraste faze, je modelirana na osnovu bilansa sila trenja izmeñu gasa i čestica u gustoj fazi FS, pri čemu su uzete u obzir interakcije ii izmeñu čestica i efekti sudara čestica. Ovakav pristup modeliranju FS pogodan je za brze inženjerske proračune stacionarnog stanja fluidizacionog medijuma u postrojenjima sa FS. Drugi numerički postupak simulacije procesa u FS, granularni Euler-Euler model, koristi profesionalan CFD kod (FLUENT 6.3.26) za modeliranje interakcije izmeñu fluida i čestica fluidizovanog sloja. Koeficijent otpora izmeñu fluida i granularne faze je definisan Syamlal-O'Brienovim modelom. Rezultati proračuna primenom predloženih CFD modela su pokazali vrlo dobro meñusobno slaganje, kao i slaganje sa eksperimentima iz literature i to sa rezultatima i korelacijama Yatesa i Honga, kojima su modeli verifikovani. Numeričkim simulacijama mlaznog doziranja goriva razmatrana su dva uticajna parametra na sveukupni proces sagorevanja u sloju: dužina prodora mlaza i intenzitet mešanja hemijskih komponenti u sloju. Analizom dobijenih rezultata je zaključeno da na dužinu prodora mlaza u FS značajno utiče brzina fluida na izlazu iz mlaznice, kao i ugao nagiba mlaza, a da na intenzitet mešanja hemijskih komponenti u FS, pored brzine i nagiba mlaza, značajno utiče dubina uranjanja mlaznice i to tako da je mešanje intenzivnije kada se bočni mlaz sa gorivom postavi u nižim zonama sloja. Efekat boljeg mešanja goriva i oksidatora pri dubljem postavljanju mlaznice u sloj potvrñen u eksperimentima na poluindustrijskoj eksperimentalnoj instalaciji sa FS za sagorevanje TTG-a. Pored potvrde rezultata numeričkih modela, eksperimenti na pilot postrojenju sa FS sa sagorevanjem nekoliko vrsta tečnih goriva različitih po sastavu i termo-fizičkim karakteristikama, su potvrdili značaj adekvatne pripreme fluidizacionog ložišta i načina doziranja tečnih goriva u fluidizovan sloj. Uz to, u eksperimenatima je registrovan efekat povlačenja zone intenzivnog sagorevanja ka oblastima dublje ispod površine sloja prilikom sagorevanja tečnog goriva sa značajnim sadržajem vode. Pored pretpostavke nekih autora, koji ovaj efekat objašnjavaju time što se mešanjem goriva sa vodom odlaže trenutak devolatilizacije, pa gorivo i gasovi duže ostaju u sloju, u ovoj tezi se razmatrao i drugi aspekt ove pojave. Naime, voda u gorivu intenzivira mešanje goriva i oksidatora, usled nagle ekspanzije pare na ulazu u ložište, čime se povećava dužina prodora mlaza u FS u kome se gorivo dozira bočno u sloj. Ova pretpostavka je potvrñena numeričkim simulacijama prodora mlaza trofaznih sistema: gas (volatili, oksidator i vodena para), čestice sloja i tečna iii voda/tečno gorivo, primenom Fluentovog granularnog modela FS. Pri tome je slučaj doziranja lako isparljivog goriva analiziran je simuliranjem prodora mlaza koji sadrži isparljive komponente goriva i tečnu vodu, a slučaj teško isparljivog goriva razmatran je simulacijom mlaza sa mešavinom ulja i vode. Na osnovu izvedenih eksperimentalnih i numeričkih ispitivanja predloženo je tehnološko idejno rešenje sagorevanja TTG-a u fluidizovanom sloju.
In countries like ours, with limited primary energy sources, there is a need, which will be even more pronounced in the future, to use own fuels available, in order to partially substitute imported fuel. In that case, unconventional liquid large density fuels (heavy liquid fuels - HLF) are one of the possible choices of fuel, which implies the application of fluidized bed (FB) combustion technology. The fact that fluidized bed boilers can burn fuels with 85% of inert materials, with effective retention of SO2 by adding limestone into the furnace, and lower NOx emissions to meet environmental standards, gives significant advantages to this technology compared to other combustion technologies. As part of the thesis extensive theoretical and experimental studies of complex and mutually coupled flow-thermal processes occurring during thermal disintegration TTG in a fluidized bed (FB) were performed, in order to develop a technology for energy-efficient combustion of these materials in an environmentally sound manner. The results of combustion experiments of different unconventional liquid fuels in the fluidized pilot furnace apparatus were presented, with emphasis on achieving low emissions of pollutants in flue gases, completeness of combustion and locating zones of intense combustion. Detailed knowledge and understanding of the process of fuel feeding and its mixing with inert bed material and air in the working space of furnace is of great importance for defining optimal operating parameters of the FB, and numerical simulations are an effective tool for achieving this goal. In the thesis two 2D mathematical-numerical (CFD) models for detailed simulation of lateral jet penetration into the fluidized bed are presented: a previously developed fluid-porous media model, and granular Euler-Euler model, which both simulate the FB and use the two-fluid modeling approach, where gas and dense phase of the FB are considered as two fluids with different characteristics. v The developed fluid-porous medium model of jet penetration into the fluidized bed is based on the calculation procedure, in which the dense phase is considered as a fixed porous medium. Gas-particle interactions and the occurrence of bubbling phase are modelled on the basis of a balance of gas-particles friction forces in the dense phase of the FB, taking into account inter-particle interactions and the particle collisions. This kind of approach to FB modeling is suitable for computationally efficient engineering calculations of a fluidizing medium in a stationary condition in FB facilities. Another numerical simulation procedure of the process in the FB, granular Euler- Euler model, uses a professional CFD code (FLUENT 6.3.26) for modeling the interaction between the fluid and fluidized bed particles. The drag coefficient between the fluid and the granular phase is defined by the Syamlal-O'Brien model. The results of calculations using both proposed CFD models showed very good mutual agreement, as well as agreement with experiments from the literature, especially with the results and correlations of Yates and Hong, with which the models were verified. By numerical simulations of jet fuel feeding, two influential parameters on the overall combustion process in the bed were considered: jet penetration length and the intensity of mixing of chemical components in the bed. By analysis of obtained results it was concluded that the jet penetration length into the FB is significantly affected by fluid velocity at the nozzle outlet, as well as the jet inclination angle. The intensity of mixing of chemical components in the FB, in addition to being influenced by the velocity and jet inclination, is significantly affected by the nozzle immersion depth, in such a way that more intense mixing occurs when lateral fuel jet is placed in the lower zones of the bed. The effect of better mixing of fuel and oxidizer when the nozzle is set deeper in the bed have been confirmed by experiments in a semi-industrial FB installation for combustion of HLF. In addition to the confirmation of the results of the numerical models, experiments in a FB pilot facility with combustion of several types of liquid fuels with different compositions and thermo-physical properties, confirmed the importance of adequate preparation of fluidization furnaces and methods of feeding of liquid fuel into the fluidized bed. vi In addition, in the experiments the effect of withdrawal of the intense combustion zone towards the bed bottom was noticed, during the combustion of liquid fuels with significant water content. This effect was more stressed during combustion of fuels with higher contents of water. In addition to the assumptions of some authors, who explain this effect by the fact that mixing of fuel with water delays devolatilization, causing the fuel and the gases to stay longer in the bed, in this thesis another aspect of this phenomenon has been considered. Namely, water in the fuel intensifies the mixing of the fuel and the oxidizer, due to sudden expansion of steam at the entrance into the furnace, thus increasing the jet penetration length of the laterally fed fuel into the FB. This assumption is confirmed by numerical simulations of jet penetration of three-phase systems: gas (volatiles, oxidizer and steam), bed particles and liquid water / liquid fuel, using Fluent's granular model of the FB. Here, the case of feeding a light volatile liquid fuel was analyzed by simulating the penetration of a jet containing volatile components of the fuel and liquid water, and the case of heavy volatile liquid fuel is considered by simulation of a jet with a mixture of engine oil and water. On the basis of the experimental and numerical investigations, technological conceptual solution for the combustion of HLF in the fluidized bed is proposed.
true
Nedestruktivno ispitivanje termofizičkih osobina materijala IC termografijom
Nondestructive evaluation of the thermophysics properties materials by IR thermography
Cilj ovog rada je poboljšanje, razvoj i primena impulsne termografije kao eksperimentalnog metoda za otkrivanje defekata u materijalu. Da bi se ostvario taj zadatak, rad je usmeren ka tri sledeća pravca: 1. Eksperimentalna istraživanja koja obuhvataju razvoj i poboljšanja uslova za merenje i akviziciju podataka. 2. Teorijska istraživanja koja obuhvataju analitičko i numeričko rešavanje sistema parcijalnih diferencijalnih jednačina provođenja toplote kondukcijom uz odgovarajuće početne i granične uslove. 3. Teorijska istraživanja koja podrazumevaju primenu metode i procedure u impulsnoj infracrvenoj termografiji za date teorijske i eksperimentalne uslove. U toku prvog dela istraživanja, izrađeni su posebni uzorci pipremljeni za eksperiment, gde su na jednoj strani ploče podjednako razmaknuti žljebovi praveći periodičnu defektnu strukturu sa određenom prostornom frekfencijom i određenom dubinom. Druga strana ploče bila je ozračena kratkim svetlosnim impulsom i temperaturna raspodela na istoj strani bila je analizirana impulsnom infracrvenom termografijom. U drugom delu testirani su različiti modeli koji opisuju dvodimenzionalnu kondukciju toplote kroz planparalelnu ploču sa defektom u obliku žljeba sa ravnim dnom sa opštim i posebnim početnim i graničnim uslovima. U trećem delu izvršeno je poređenje eksperimentalnih rezultata i rezultata simulacije u cilju utvrđivanja uslova i granica primenljivosti metoda za periodičnu strukturu defekata. Izvršena je teorijska analiza i eksperimentalna provera metoda u slučaju defekata u aluminijumskoj ploči.
The aim of this paper is improving, development and application of pulsed thermography as experimental method for detection of defect in material. To accomplish this, the paper has 3 aspects: 1. Experimental research including оf development and improving conditions for measuring and data acquisition. 2. Theoretical research containing of analitycal and numerical solving the system of partial differential equations for heat convection by conduction, with corr esponding initial and boundary conditions. 3. Theoretical research considering method and procedure application in pulsed infrared thermography for given theoretical and experimental conditions. During the first part of research, specific samples were prepared for experiment where at one side of the plate equally spaced slots are forming the periodical structure of defects with defined spatial frequency and depth. The other side of plate was irradiated with short light pulse and temperature distribution on that side was analized by pulsed infrared thermography. The second part of the research, considers different models describing two- dimensional heat conduction through planparallel plate with defect in form of flat-bottom slot, with general and specific initial and boundary conditions. In the third part of research, comparison of experimental results and results of simulation was made to determine the conditions and limits of method applicability for periodically structured defects. The theoretical analysis and experimental testing of the method was done for the defects in aluminum plate.
true
Karakterizacija teksture impulsivnog odziva prostorija primenom multifraktalne analize
Characterization of rooms impulse response texture using multifractal analysis
U ovoj disertaciji eksperimentalno je proverena mogućnost za primenu multifraktalne analize za kvantifikovanje pojava u akustici prostorija koje do sada nisu imala odgovarajuću (zadovoljavajuću) meru. Pored u naslovu pomenute teksture, analizom pomoću fraktalne teorije obuhvaćeni su i fenomeni koji su tesno povezani sa teksturom impulsnog odziva, a to su: detekcija ranih refleksija, difuznost zvučnog polja i granica između oblasti ranih refleksija i reverberacije. Analiza impulsnih odziva sa multifraktalnog stanovišta obavljena je korišćenjem MFDFA (Multifractal Detrended Fluctuation Analysis) metoda i metod histograma. Osnova za primenu multifraktalne teorije u analizi impulsnog odziva leži u samoj prirodi impulsnog odziva i procesu putem koga se on formira. Dobijeni rezultati pokazuju da postoje osnovi za tumačenje impulsnog odziva prostorije sa multifraktalnog stanovišta. Karakterizacija teksture primenom multifraktalne analize kretala se u dva pravca, kroz analizu lokalne i globalne regularnosti. Eksperimentalni rezultati potvrđuju da se difuznost uspostavljenog zvučnog polja u prostoriji i tekstura impulsnog odziva prostorije mogu izraziti kroz multifraktalni spektar impulsnog osnova. Kvantifikovanje ovih karakteristika i klasifikaciju impulsnih odziva, a samim tim i prostorija moguće je ostvariti poređenjem njihovih multifraktalnih spektara, za šta su u radu date konkretne smernice. Analiza lokalne regularnosti impulsnog odziva kroz detekciju refleksija, osim načina za karakterizaciju, tj. kvantifikovanje teksture daje i predlog metoda za detekcije ranih refleksija u impulsnom odzivu.
In this dissertation experimentally is tested the possibility of applying multifractal analysis to quantify the phenomena in acoustics that don’t have adequate (satisfactory) measures. In addition to the title mentioned texture, analysis using fractal theory also covered and phenomena that are closely linked with the texture of the impulse response, such as: detection of early reflections, sound field diffuseness and the boundary between the early reflections and reverberation part of impulse response. The analysis of room impulse response from multifractal standpoint was performed by the MFDFA (Detrended Multifractal Fluctuation Analysis) method and the histogram method. The basis for the application of multifractal theory to analyse room impulse response lies in the very nature of the impulse response and the process through which it is formed. The results show that there are grounds for the interpretation of the room impulse response from multifractal standpoint. Texture characterization using multifractal analysis ranged in two directions, through the analysis of local and global regularity. Experimental results confirm that the diffusivity of sound field established in room and impulse response texture can be expressed through impulse response multifractal spectrum. Quantifying these characteristics and classification of impulse responses, and consequently rooms is possible by comparing their multifractal spectra, for which guidelines are defined. The analysis of impulse response local regularity through detecting reflections, in addition to the method for texture characterization, also gives the suggestion for early reflection detection in the impulse response.
true
Robotski sistem za vežbanje i procenu funkcionalnosti gornjih ekstremiteta tokom rehabilitacije osoba posle moždanog udara
Robotic system for training and assessment of upper extremity function during rehabilitation of stroke survivors.
Moždani udar je oboljenje centralnog nervnog sistema koje je, prema podacima Svetske zdravstvene organizacije, najveći uzrok invaliditeta i među prva tri uzroka smrtnosti. Kod osoba koje prežive moždani udar mogu se javiti delimična ili potpuna motorna oduzetost jedne ili obe strane tela. Ponovno uspostavljanje motorne kontrole gornjih ekstremiteta kod ovih osoba je veoma važno za njihovu samostalnost i poboljšanje kvaliteta života. Rezultati mnogih kliničkih studija ukazuju da asistencija pri intenzivnom treningu funkcionalnih pokreta (fokusiranih na konkretan zadatak) u prisustvu stimulišuće povratne informacije značajno podstiče oporavak motorne kontrole. Tumačenje ovih rezultata je da se reorganizacija kortikalnih struktura, koja je ključna za dugoročni oporavak, intenzivira povećanjem broja ponavljanja funkcionalnih pokreta. Međutim, ovakav tip vežbanja nije uvek moguć, pogotovo u ranim fazama oporavka. Pacijenti često nisu u mogućnosti da izvedu funkcionalan pokret, već im je potreban neki vid asistencije. Rešenje ovog problema jeste asistencija od strane terapeuta što predstavlja naporan i monoton zadatak i utiče na smanjenje motivacije pacijenata. Alternativno rešenje koje privlači veliku pažnju istraživača je primena personalnih robota u rehabilitaciji. Prednost ovakvih uređaja je što mogu da asistiraju pri izvođenju pokreta na način koji je ponovljiv i može da se prilagodi potrebama pacijenta. Još jedan značajan segment je što ovi sistemi pružaju mogućnost uključenja virtuelne realnosti i video igara u proces rehabilitacije. Pri ovome robotski sistemi omogućavaju objektivnu kvantifikaciju trenutnih sposobnosti pacijenta merenjem kinematičkih i dinamičkih karakteristika pokreta. Intenzivnija istraživanja primene ovakvih sistema u rehabilitaciji počela su krajem osamdesetih i početkom devedesetih godina, a poslednjih nekoliko godina ovakvi roboti su našli primenu u rehabilitacionim klinikama. Iako su u poslednjih trideset godina dizajn i kontrola ovih uređaja značajno unapređeni, efikasno programiranje robota u kliničkom okruženju i pružanje asistencije na najbolji način za oporavak pacijenta su i dalje otvorena pitanja bez adekvatnog rešenja. iv Cilj ovog istraživanja je razvoj metode za vežbanje i procenu kvaliteta pokreta ruke uz pomoć robotskog uređaja i komercijalno dostupnih video igara, koja bi se mogla efikasno koristiti u kliničkoj praksi, ali i kućnim uslovima. Razvili smo metodu za reprezentaciju pokreta ruke koja na osnovu snimljenih ekspertskih pokreta kreira stohastički model zasnovan na probabilističkim primitivima. Ova metoda, nazvana Probability Tube, razvijena je na osnovu merenja pokreta zdravih ispitanika i pacijenata u različitim fazama oporavka od moždanog udara, i validirana je u kliničkom okruženju. Na osnovu ove reprezentacije pokreta formirane su metode za upravljanje rehabilitacionim robotskim sistemom, procenu motorne kontrole pacijenata i implementaciju video igara u proces rehabilitacije. Razvijena je inovativna metoda za upravljanje rehabilitacionim robotskim sistemom. Ova metoda omogućava jednostavno "programiranje" pokreta koji treba da se vežba putem demonstracije, kao i prilagođavanje pokreta trenutnim sposobnostima pacijenta. Zbog primene Probability Tube, stohastičke reprezentacije pokreta, metoda omogućava varijabilnost pri izvođenju pokreta, odnosno adaptciju kontrole tokom izvođenja pokreta. Pored toga ova metoda omogućava progresivno otežavanje zadatka koji se vežba, čime se povećava aktivnost pacijenta i efikasnost terapije. Važan segment u rehabilitaciji jeste estimacija trenutnog stanja, kao i oporavka motorne kontrole pacijenata. U okviru ovog istraživanja razvijene su tri kompatabilne metode koje na osnovu objektivnih merenja kvantifikuju sposobnost pacijenta da izvede pokret koji mu je zadao terapeut, kao i opšte stanje motorne kontrole manipulacije rukom. Važan faktor u terapiji zasnovanoj na vežbanju je motivacija pacijenta. U okviru ove teze prikazano je bazno istraživanje u kome je utvrđen uticaj dodavanja stimulišuće povratne sprege u obliku video igara na motivaciju pacijenata za vežbanjem. Ovo istraživanje pokazalo je da je vežbanje u prisustvu video igara osim na motivaciju, imalo uticaj i na efikasnost terapije. Iako je terapija primenjivana tokom relativno kratkog perioda studije, napredak grupe koja je koristila video igre, meren objektivnim merama, bio je značajno veći od napretka kontrolne grupe pacijenata koji su primali identičnu terapiju, samo bez elementa video igara. Na osnovu v zaključaka ovog istraživanja formiran je sistem koji omogućava implementaciju komercijalnih video igara za konzolu Nintendo Wii u proces rehabilitacije. Na osnovu rezultata dobijenih u okviru ovog istraživanja formirana je inovativna metoda koja pacijentima omogućava da vežbaju pokrete zadate od strane terapeuta dok igraju komercijalne video igre uz robotsku asistenciju, pri čemu se sve vreme objektivno procenjuje njihova motorna kontrole.
Stroke is a neurological deficit of cerebrovascular origin, which is the leading cause of disability and one of the leading causes of mortality in the world, according to reports of the World Health Organization. Stroke survivors may suffer partial or complete paralysis of one or both sides of the body. Reestablishing motor control of upper extremities is essential for achieving independence and improvement quality of life of such patients. Results of many clinical trials indicate that intensive assisted training of functional (task oriented) movements, with stimulating feedback significantly improves restoring of motor control. Interpretation of these results is that the reorganization of cortical structures, which is instrumental for long term recovery, benefits from dosage increase of functional movement exercise. Unfortunately, such exercise is not always possible, especially in the early stages of rehabilitation. Patients are often unable to perform functional movements and require some form of assistance. One of the approaches is that therapists assist their movements, which is laborious and monotonous task, often leading to decrease of patients’ motivation for exercise. Alternative solution which is becoming increasingly popular in the scientific circles is the use of personalized robots in rehabilitation. Advantage of such systems is ability to assist movement- execution in a repeatable way, which can be adapted to patients’ needs. Another important aspect is the possibility of including virtual reality and video games in the rehabilitation process. Robotic systems also enable objective quantification of patient’s current abilities by measuring movement kinematics and dynamics. Intensive research in the domain of rehabilitation robotics was initiated in late 80’s and early 90’s, while in recent years these robots have become present in rehabilitation clinics as commercial systems. Even though in the last 30 years both design and control of these devices have significantly improved, efficient programming in clinical environment, as well as providing assistance in the manner most beneficial for patient’s recovery, remain open questions, without adequate solutions. vii Objective of this research is development of a method for arm movement exercise and assessment with the use of robotic device and commercially available video games, which may be efficiently used in the clinical environment and at home. We have developed a method for arm movement representation, which uses the recordings of movements performed by an expert in order to generate a stochastic model based on the probabilistic primitives. This method, termed Probability Tube, was developed based on measurements performed on healthy individuals and stroke survivors in different stages of recovery, and was tested in the clinical environment. On the grounds of this movement representation, we have developed methods for control of rehabilitation robotic system, assessment of patient’s motor control and implementation of video games in the rehabilitation process. An innovative method for control of rehabilitation robotic system was developed. This method enables simple programming of the movement to be exercised through demonstration. It also enables movement adjustment to patient’s current abilities. As it is based on Probability tube, which is a stochastic movement representation, the method enables variability during the movement execution, i.e. interactive movement control. Furthermore, this method enables progressive enhancement of the task difficulty, which stimulates patient’s effort and improves the therapy efficiency. An important segment in rehabilitation is the assessment of patient’s motor control ability. Three compatible methods for objective quantification of patient’s ability to perform a given movement, as well as overall arm manipulation ability are developed within this research. An important factor in exercise-based therapy is patient’s motivation. Part of this thesis presents an original research which analyses the effects of adding a stimulative feedback in form of a video game on patient’s motivation during exercise. Results of this research show that presence of video games improves motivation, as well as the therapy efficiency. Despite the relatively short duration of the study, the progress of the group which used video games was significantly higher that the control group, which was performing the exact same exercises, just without video games. Based on the conclusions of this study, we have devised a system which enables the implementation of commercial (Nintendo Wii) video games into the rehabilitation process. viii Based on the results obtained in this research, we have created an innovative rehabilitation method which enables patients to exercise movements presented by therapists while playing stimulating video games with robotic assistance. It also enables continuous assessment of patient’s motor control.
true
Numerička simulacija i analiza trodimenzionalnog dvofaznog strujanja gasa i čvrste faze u separatoru ugljenog praha
Numerical simulation and analysis of 3-dimensional two-phase flow of gas and solid phase in a coal powder separator
Cilj ovog istraživanja je razvoj diferencijalnog modela trodimenzionalnog dvofaznog strujanja gasa i ugljenog praha i separacija krupnijih čestica uglja, od zadate, u separatorima postrojenja za pripremu ugljenog praha. Finoća ugljenog praha je definisana granulometrijskom karakteristikom (ostatak na situ x R ), koja podliježe RRS (Rozin-Ramler-Sperling) raspodjeli. Rezultati ovog istraživanja su doprinos boljem razumijevanju procesa koji se dešavaju u separatorima ugljenog praha i poboljšavanje optimizacije sagorijevanja ugljenog praha u ložištima generatora pare, odnosno postizanje boljeg iskorišćenja goriva (stabilnije sagorijevanje i veći stepen korisnog djelovanja). Istraživanje je obavljeno ugljenom meljavom ventilatorskih mlinova na inercijalnom separatoru ugljenog praha. Kao gorivo je korišćen lignit. Za rješavanje matematičkog modela korišćena je numerička metoda kontrolnih zapremina. Verifikacija modela i primijenjene numeričke metode, izvršena je poređenjem sa rezultatima ispitivanja na realnom postrojenju. Naglasak je na trodimenzionalnosti, složenosti geometrije i problematici dvofaznog turbulentnog toka. Model se zasniva na Ojler-Lagranžeovom pristupu dvofaznom strujanju, k  modelu turbulentne gasne faze, stohastičkom modelu disperzije čvrstih čestica i PSI-CELL postupku kojim se uzima u obzir uzajamno dejstvo faza. Prikazani su rezultati numeričke simulacije strujanja dvofazne smješe gasa i čvrstih čestica ugljenog praha (aerosmješe) kroz inercijalni mlinski separator. Rezultati su prikazani u obliku: vektorskog i skalarnog polja brzine, trajektorija gasne faze i trajektorija čestica čvrste faze (po klasama, odnosno II veličini čestica) i distribucija čestica u izlaznom presjeku zavisno od položaja...
The basic goal of presented investigation is development partial differential three dimensional model of two-phase gas flow and pulverized coal as well as separation of rough coal particles into the separation facilities for pulverized coal preparation. The finest of pulverized coal is defined by granulation curve (residual on screen R); that is aligned to RRS (Rozin-Ramler-Sperling) distribution. The investigation results are contribution to the better understanding of processes occurring into the separation facilities for pulverized coal preparation, optimization of pulverized coal combustion into the furnaces of steam generator, e.g. better usefulness of coal (better combustion stability and greater combustion efficiency). The investigation has been preformed by pulverized coal obtained into ventilator mils on inertial separator. The coal was lignite. For numerical solution of developed model, the finite volume numerical method has been used. The verification of developed model and used numerical method has been performed by experimental data obtained on actual separation facility. It has been emphasized on the three dimensionality, actual complexity of separator geometry and two-phase turbulence flow. The model is based on Eulerian-Lagrangian approach for two-phase flow, k  turbulence model for gas phase, stochastic model of solid particles dispersion, and PSI-CELL method for interphone coupling of phase’s interaction. The results of numerical simulation of two-phase gas flow and pulverized coal particles into inertial separator are presented. The results are presented in the forms of: vector and scalar velocity fields, gas trajectories, coal particle trajectories (by size/diameter class) and coal particle size distribution in outlet plane of separator, depending on position of regulation plates into the separator. Additionally, relation of pulverized coal particle class in outlet plane and circulation number has been presented by diagrams, quantifying influence of opaque angle of regulation plates. All results are presented for the nominal load and 10 % reduced mass flow rate of gas phase. Quantitative results are presented by tables...
false
Metodologija za uvažavanje nesavršenog kontakta između elektroda i tla pri proračunu otpornosti rasprostiranja konturnih uzemljivača
Methodology for taking into account imperfect contact between grounding electrodes and soil when calculating the loop grounding resistance
Otpornost rasprostiranja uzemljivačkog sistema je osnovni parametar za proračune svih ostalih relevantnih veličina pomoću kojih se proverava njegov kvalitet. Međutim, pošto postoje tla koja formiraju loš kontakt sa elektrodama uzemljivača, u praksi se često pojavljuje značajna nesaglasnost između izmerenih vrednosti tih veličina i njihovih vrednosti izračunatih primenom konvencionalnih formula, koje važe uz pretpostavku o savršenom kontaktu između elektroda i okolnog tla. U takvim slučajevima, u praksi se ponekad koriste dodatni materijali koji se postavljaju između elektroda i tla sa ciljem da se eliminiše kontaktna otpornost, kao i da ukupna otpornost rasprostiranja uzemljivača bude manja od željene (standardima zahtevane) vrednosti. Cilj ove disertacije je razvoj opšte metodologije za određivanje optimalnog uzemljivača, sa uvažavanjem, kako nesavršenog kontakta između elektroda i okolnog tla, tako i opcije upotrebe dodatnog materijala za smanjenje kontaktne otpornosti. Razvijen je i prikazan opšti algoritam za projektovanje optimalnog uzemljivača u opisanim uslovima, zasnovan na tehno-ekonomskoj analizi u okviru koje se porede instalacioni troškovi uzemljivača čija je inicijalno planirana konstrukcija modifikovana na razne načine (upotrebom samo dodatnih elektroda, samo dodatnog materijala ili i jednog i drugog), pod uslovom da su postignute približno iste otpornosti rasprostiranja, manje od zahtevanih. U dostupnim domaćim i stranim standardima i naučno-stručnoj literaturi nisu postojale niti metode niti uprošćeni analitički izrazi pomoću kojih bi navedeni uticaji mogli da se uvaže pri proračunu otpornosti rasprostiranja uzemljivača. Zato su u ovoj disertaciji razvijeni neophodni algoritmi i uprošćeni izrazi za sprovođenje takve tehno-ekonomske analize pri projektovanju optimalnog uzemljivača sastavljenog od jedne ili više galvanski iii spojenih kvadratnih kontura (konstrukcije koje odgovaraju tipičnom uzemljivaču dalekovodnog stuba). Algoritmi i metode za razvijanje uprošćenih izraza, koji su prikazani u ovoj disertaciji, opšti su i bazirani su na 3D FEM (finite element method – metoda konačnih elemenata) modelovanju i numeričkoj analizi dobijenih rezultata. Nesavršen kontakt je modelovan „postavljanjem“ vazdušnih džepova između elektroda i okolnog tla. Razvijena je i detaljno predstavljena metoda za uvažavanje uticaja nesavršenog kontakta između elektroda i okolnog tla pri proračunu otpornosti rasprostiranja razmatranog konturnog uzemljivača. Izrađen je opšti algoritam za određivanje neophodnih karakteristika dodatnog materijala za eliminaciju kontaktne otpornosti, kao i opšti algoritam za procenu optimalne količine izabranog dodatnog materijala namenjenog potrebnom smanjenju otpornosti rasprostiranja uzemljivača. Za razmatrani slučaj izvedeni su svi uprošćeni izrazi neophodni za sprovođenje navedenih algoritama. Na primeru uzemljivača sastavljenog od jedne ili više galvanski spojenih kvadratnih kontura predstavljena je metoda za razvijanje uprošćenih izraza za izračunavanje teorijske (bazne) otpornosti uzemljivača, koja se može primeniti na razne tipove uzemljivača. Izvedene su nova formula za izračunavanje teorijske otpornosti rasprostiranja usamljene konture u homogenom tlu i formule za izračunavanje teorijske otpornosti rasprostiranja uzemljivača sastavljenog od više galvanski spojenih kvadratnih kontura položenih u homogeno tlo (one su bazirane na koeficijentima kojima se opisuje međusobni uticaj pojedinačnih elemenata uzemljivača). Zatim je razvijen i prikazan opšti algoritam za izračunavanje otpornosti rasprostiranja kompleksnih uzemljivačkih sistema, pomoću koga se uvažava postojanje kontaktne otpornosti. Predloženi algoritam je baziran na dodatnom test merenju, skupu uprošćenih izraza za uvažavanje uticaja nesavršenog kontakta između elektroda i okolnog tla pri proračunu otpornosti rasprostiranja složenog uzemljivača i skupu uprošćenih izraza za izračunavanje njegove teorijske otpornosti rasprostiranja. Provera tačnosti izvedenih formula i pouzdanosti predloženog algoritma izvršena je poređenjem dobijenih rezultata sa rezultatima dobijenim primenom konvencionalnih metoda proračuna datih u standardima i priručnicima, iv računarskim simulacijama sprovedenim na 3D FEM modelima i merenjima u okviru eksperimenta sa umanjenim žičanim modelima i elektrolitičkom kadom. Korišćenjem metoda, algoritama i uprošćenih izraza razvijenih u ovoj disertaciji broj ponavljanja procedure prepravke uzemljivača položenih u tlo koje formira nesavršen kontakt sa elektrodama sigurno će se smanjiti (u odnosu na broj ponavljanja karakterističnih za primenu konvencionalnih metoda proračuna), što će uticati na redukciju povećanja investicionih troškova predviđenih inicijalnim projektom.
Grounding resistance is a basic parameter for calculation of all other relevant parameters influencing the quality of a grounding system. However, since the grounding system can be installed in soils forming a bad contact with the grounding electrodes, significant discrepancies between the measured values of those parameters and the values calculated using conventional formulas often appear in practice, because conventional formulas are derived assuming perfect contact between the electrodes and surrounding soil. In such cases, backfill materials placed between the electrodes and soil are sometimes used in practice, with the aim to eliminate contact resistance, as well as to reduce the overall grounding resistance below the desired value (required by standards). The aim of the research work presented in this dissertation is the development of a general methodology for determining the optimal grounding system, taking into consideration the influence of imperfect contact between the electrodes and surrounding soil, as well as the influence of the use of backfill material, on the grounding system resistance. A general algorithm for the optimal design of a grounding system installed in the described conditions is developed and presented, based on a techno-economic analysis, where the installation costs of grounding system structures obtained after various modifications (using only the additional electrodes, only backfill material or both) are compared under an assumption that all grounding systems have approximately the same grounding system resistance (lower than the one required by standards). The available national and international standards, as well as scientific and professional literature, do not offer any method or simplified analytical expressions for calculating the grounding resistance, which would take into account the described grounding system modifications. Hence, all algorithms and simplified formulas necessary for conducting such a techno-economic analysis related to the design of the optimal grounding system composed of one or more galvanically connected square vi loops (structures that correspond to typical transmission tower grounding system) are developed in this dissertation. All of the algorithms and methods for the derivation of simplified formulas, which are presented in this dissertation, are general and based on 3D FEM (Finite Element Method) modelling and numerical analysis of the obtained results. Imperfect contact is modelled by air gaps placed between the grounding loop electrodes and surrounding soil. A method for taking into account the influence of imperfect contact between the electrodes and surrounding soil when calculating considered grounding loop resistance is developed and presented in detail. A general algorithm for determination of necessary characteristics which backfill materials must possess in order to eliminate the contact resistance, as well as a general algorithm for assessing the optimal quantity of the selected backfill material intended for the necessary reduction of the grounding resistance, are also developed and presented. For the considered case, simplified formulas necessary to implement the above algorithms are derived. A new method for deriving simplified formulas for calculation of theoretical (base) value of the grounding system resistance, which can be applied to various types of grounding systems, is presented and then applied to an example of a grounding system composed of one or more galvanically connected square loops. A new formula for calculating the theoretical value of the square loop grounding resistance, as well as a set of approximate formulas for calculation of theoretical value of the grounding resistance of a grounding system composed of several galvanically connected square loops (based on the use of coefficients taking into account the proximity effects), are also derived. Afterwards, a general algorithm for calculating the grounding resistance of complex grounding systems, including the contact resistance, is developed and presented. The proposed algorithm is based on an additional test measurement, a set of simplified formulas for taking into account the influence of imperfect contact between the electrodes and surrounding soil when calculating the grounding resistance, and a set of simplified formulas for calculating the theoretical grounding resistance of complex grounding systems. Validation of accuracy of the derived formulas and reliability of the proposed algorithm is performed (i) by comparing the results obtained using the proposed formulas with those obtained by conventional calculation methods given in standards vii and handbooks, (ii) by computer simulations conducted on 3D FEM models, and (iii) by measurements based on utilization of small scale wire models and the electrolytic tank. Using the methods, algorithms and simplified formulas developed in this dissertation, the number of necessary modifications after the installation of the initial grounding system in a soil forming imperfect contact with the electrodes will certainly be smaller than the number of necessary modifications if only conventional calculation methods are used), which will reduce the increase of the investment costs envisaged in the initial project.
true
Ostatak određenih tipova Gausovih kvadraturnih formula sa specijalnim klasama težinskih funkcija
The remainder term of certain types of Gaussian quadrature formulae with specific classes of weight functions.
Integracija ima xiroku primenu prilikom matematiqkog mode- lovanja mnogih pojava koje se javljaju u prirodnim, tehniqkim naukama, ekonomiji i drugim oblastima. Kada se vrednost integrala ne moe analitiqki izraqunati, potrebno je kon- struisati formulu koja aproksimira njegovu vrednost sa prih- vatljivom taqnoxu. Pored tradicionalnih formula koje se ko- riste, tendencije u razvoju ove oblasti odnose se na poveanje taqnosti formule i ocenu grexke nastale kada se integral za- meni konaqnom sumom...
Mathematical modeling of many phenomena which occur in the natural, technical sciences, economy requires signicant knowledge of the theory of numerical integration. In the situations where the integral cannot be determined analytically, it is necessary to construct the for- mula which approximates its value with acceptable error. Besides the traditional formulae, the tendencies in the development of this area refer to increment of algebraic degree of precision of the quadrature formula and its error estimation...
false
Modeling and control of the anthropomimetic robot with antagonistic joints in contact and non-contact tasks
Mоделирањe и управљањe антропомиметичког робота са антагонистичким погонима у контактним и бесконтактним задацима
U tezi je obrađena vrlo aktuelna tema u oblasti robotike - modeliranje i upravljanje antropomimetičkim robotima, posebno onim sa elastičnim antagonističkim pogonima. Pojam antropomimetički se odnosi na robota koji podseća na čoveka ne samo svojim izgledom, već i unutrašnjom strukturom i funkcionalnošću. Motivacija za izradu ove disertacije je proizišla iz aktivnosti autora na Elektrotehničkom fakultetu u Beogradu, a u okviru projekta Eccerobot - Embodied Cognition in a Compliantly Engineered Robot, finansiranog od strane sedmog okvirnog programa Evropske unije i projekta „Ambijentalno inteligentni servisni robota antropomorfnih karakteristika“, finansiranog od strane Ministartva za prosvetu, nauku i tehnološki razvoj Republike Srbije. S obzirom na nedostatke konvencionalnih krutih aktuatora u budućim robotskim aplikacijama koje zahtevaju visok stepen bezbednosti, razvoj i upravljanje novih aktuatora koji uključuju elastičnost je od velikog značaja. Posebno je značajan razvoj aktuatora sa promenljivom krutošću VSA (od engleskog naziva “variable stiffness actuator”) koji omogućavaju podešavanje krutostu tokom rada i samim tim daju kompromis između preciznosti rada robota i bezbednosti njegovog okruženja. Fokus našeg istraživanja su roboti sa antagonističkim pogonima kao posebna grupa biloški inspirisanih VSA. Disertacija je organizovana u šest poglavlja uz pregled literature. U prvom poglavlju su istaknute glavne ideje i motivacija za rad na ovoj temi, kao i pravci istraživanja u razvoju i upravljanju zglobova sa elastičnošću, sa naglaskom na antagonističke pogone. U ovom poglavlju se ističu bezbednosni izazovi budućih servisnih robota, navode se odgovarajuće tehnologije i trendovi: elastični aktuatori, pojmovi aktivne i pasivne popustljivosti i VSA. Kao grupa od posebnog interesa među VSA, prikazan je detaljan presek stanja na temu antagonističkih aktuatora u robotici. Razvoj simulacionog modela dinamike robota pogonjenog antagonistički uparenim elastičnim aktuatorima je detaljno prikazan u drugom poglavlju disertacije. U tom cilju, iskorišćen je Stepanjenkov metod za modeliranje vii robota u skladu sa njegovim superiornim karakteristikama kada se radi o simulaciji dinamike robota. Prikazani su modeli dva osnovna antagonistički pogonjena zgloba robota koji predstavljaju tipične strukture čovekovih zglobova pogonjenih antagonistički uparenim mišićima. Ovi modeli pojedinih zglobova su integrisani u model gornjeg dela antropomimetičkog robota (od struka na više) po uzoru na prototip Eccerobot-a. Uzimajući u obzir da je takav robot prevashodno namenjen za rad u neposrednom ljudskom okruženju, posebna pažnja je posvećena modeliranju kontakata. Ovo poglavlje daje efikasan alat za simuliranje dinamike antropomimetičkog robota koji je i primenjen u Matlab-u sa realizacijom pojedinih delova u C++. U tom cilju su prikazana i dva primera zasnovana na simulacijama. Prvi primer predstavlja antropomimetičkog robota na pokretnoj osnovi koji radi u nestrukturiranom okruženju koje ga kao takvo izlaže kratkim (impulsnim) i dugoročnim poremećajima. Ovde je razvijeni model iskorišćen za analizu dinamičkog balansa posmatranjem tačke nultog momenta. Drugi primer prestavlja robota koji namenski hvata objekat u prostoru. Na ovaj način su demonstrirane sve tri faze modeliranja kontaktnih zadataka između robota i objekta: približavanje, neelastični sudar i kretanje u kontaktu. Iako je razvijeni model već našao primenu i na realnom robotu Cassius [1] i studiji baziranoj na simulaciji na osnovu modela [2], postoji još mnogo mogućnosti za upotrebu modela: simuliranje i analiza biomehaničkih sistema ili delova sistema, ispitivanje mogućnosti korišćenja bioloških kocepata u bio-inspirisanoj robotici, simuliranje robota u kontaktnim zadacima, analiza i razvoj strukture antropomimetičkih robota, razvoj naprednih metoda upravljanja antropomimetičkim robotima na bazi modela, testiranje algoritama upravljanja antropomimetičkim robotima u kontaktnim i beskontaktnim zadacima, itd. Glavni rezultati koji obuhvataju razvoj modela antropomimetičkog robota i njegovu nadgradnju u pravcu analize dinamike kontaknih zadataka su obuhvaćeni u publikacijama zaključno sa radom u međunarodnom časopisu sa impakt faktorom [3]. U isto vreme, razvijeni simulacioni model je služio kao platforma za dalji rad i razvoj algoritama upravljanja prikazanih u disertaciji. U trećem poglavlju su prikazani doprinosi u oblasti upravljanja antagonistički pogonjenim, linearnim/nelinearnim, elastičnim zglobom robota viii sa žičanim prenosom. Prikazane upravljačke motode se oslanjaju na konvencionalnu teoriju upravljanja. Nakon preseka stanja u oblasti upravljanja antagonističkim aktuatorima i upotrebe povratne sprege za linearizaciju (eng. feedback linearization) u upravljanju robotima, unapređuje se tzv. „puler-folover“ pristup inicijalno uveden u radu [4]. Ovaj biološki inspirisan pristup se zasniva na energetski efikasnom istovremenom upravljanju pozicije zgloba robota i zatezne sile u jednom od antagonistički uparenih tendona. U ovoj tezi, puler-folover (eng. puller-follower) pristup je proširen za potrebe istovremenog upravljanja pozicijom zgloba i njegove krutosti. Zatim, inicijalno uvedeni pristup za kontrolu jednog zgloba robota koji se zasniva na fidbek linearizaciji je prilagođen za kompenzaciju gravitacionog opterećenja, efektivnog momenta inercije u svakom zglobu, kao i dinamičkog sprezanja u sistemima sa više povezanih tela. Pored robusne metode za podešavanje funkcije prenosa u otvorenoj sprezi (eng. loop shaping), metode nelinearne i multivarijbilne terorije upravljanja su primenjene na konačnu upravljačku šemu puler-folover metode. Ovaj deo disertacije se završava navođenjem ograničenja u primeni pomenutog metoda i otovorenih pitanja kao mogućih pravaca daljeg rada. Pored metoda upravljanja koje se zasnivaju na konvencinalnim inženjerskim principima, u disertaciji je prikazan razvoj nekoliko kognitivnih metoda upravljanja antagonističkim elastičnim aktuatorima u robotici. U skladu sa dizajnom antropomimetičkih robota, razvijene su i metode koje se u skladu sa ljudskim ponašanjem zasnivaju na iskustvu, učenju i heuristici. Razvijen je metod najbližeg suseda (eng. the nearest-neighbor) za upravljanje u povratnoj sprezi (eng. feedback) i direktnoj grani (eng. feedforward). Takođe, upravljanje u direktnoj grani je realizovano upotrebom neuralnih mreža tzv. radijalne osnove (eng. radial-basis neural networks), a upravljanje u povratnoj sprezi projektovano na bazi trenutne procene kinematičkih koeficijenata i fazi pravila (eng. fuzzy rules). Ove kognitivne metode predstavljaju značajan korak ka konačnoj šemi upravljanja antropomimetičkim robotom imajući u vidu da konvencionalne inženjerske upravljačke metode ne daju rešenja za kontrolu višeosnih robotskih zglobova ili tendona koji istovremenom deluju na više osa robota, gde se kognitivne metode upravljanja mogu primeniti bez dodatnih ograničenja. ix Demonstracija projektovanih metoda upravljanja je urađena kroz simulacione primere u programskom jeziku Matlab. Naučni doprinosi koji se odnose na ove metode su prikazani u četvrtom odeljku disertacije, dok se zbirni rezultati na temu upravljanja antropomimetičkim robotima dati u radu [5]. U petom odeljku disertacije prikazana su inicijalna istraživanja i stavovi na temu upravljanja antagonistički pogonjenim elastičnim zglobovima u kontaktnim zadacima. U tom cilju, predložena je upravljačka šema koja predstavlja modifikaciju opšte prihvaćene metode impedansnog upravljanja na bazi biološki inspirisanih obrazaca za podelu opterećenja između dva antagonistički uparena tendona. Glavna zapažanja, zaključci i doprinosi disertacije su sumirani u završnom odeljku. Možemo naslutiti da će konačna upravljačka šema antropomimetičkih robota koristiti kombinaciju različitih metoda upravljanja: upravljanje na bazi analitičkih modela dinamike, konvencionalnih metoda ili tehnika na bazi učenja i iskustva. U skladu sa tim očekujemo da će ova teza doprineti efikasnom modeliranju i upravljanju budućih humanoidnih i servisnih robota uopšte. Takođe, rezultati teze mogu biti iskorišćeni u oblastima: industrije video igara, biomehanike, motorne kontrole, itd. Doprinosi autora koji se direktno tiču tematike ove disertacije sumirani su u ukupno 25 publikacija (1 poglavlje u knjizi, 4 rada u međunarodnim časopisima sa impakt faktorom, 3 rada u domaćim časopisima, 8 radova na međunarodnim konferencijama i 9 radova na međunarodnim i domaćim konferencijama održanim u Srbiji).
The thesis considers a very popular and rapidly growing topic in robotics – modeling and control of anthropomimetic robots, in particular robots that feature compliant antagonistic actuation. The term ‘anthropomimetic’ refers to a robot of fully human-like appearance, but also human-like inner structure and functionality. The motivation for the thesis came from the involvement of the author and the School of Electrical Engineering, University of Belgrade in the project Eccerobot (Embodied Cognition in a Compliantly Engineered Robot), funded by European 7th Framework Programme, and the project Ambient Intelligent Service Robots of Anthropomorphic Characteristics funded by Serbian Ministry of Education, Science and Technological Development. Since conventional rigid actuators have become obsolete for inherently-safe future robotic applications, the design and control of novel compliant actuators have recently been under intensive investigation. In particular, variable stiffness (impedance) actuators (VSAs), which can trade-off between robot trajectory tracking and overall safety, are the focus of robotics research. We focus on antagonistically-driven joints, as a subgroup of bio-inspired VSAs. The thesis is organized in six sections and bibliography. The first section outlines the main ideas and motivation for the work. It presents research directions, points of view and initiatives in the design and control of compliant robots, with special emphasis on antagonistic drives. It also addresses safety issues of future service robots, and lists technologies, trends, and approaches towards achieving that goal: moving from stiff to compliant actuators, the benefits introduced by passive and active compliance, VSA, etc. As a special group of bio-inspired VSAs in robotics, antagonistic actuators are reviewed comprehensively. Such a detailed review of antagonistic actuator design aims to highlight the importance of the subject, on the one hand, while on the other hand it presents the technology that will be intensively exploited in future anthropomimetic robots. iii The second section presents a development of the simulation-based model of the robot driven by antagonistically-actuated compliant drives. To that end, Stepanjenko’s method to robot modeling is exploited since it has demonstrated superior performance in simulating robot dynamics. To begin with, basic models of antagonistically-actuated joints that resemble typical human-like antagonistic structures are introduced. These models are incorporated into the upper-body anthropomimetic robot model, which is a close approximation of the Eccerobot prototype. Since such a robot is intended to work in proximity to humans and directly interact with its surroundings, modeling of contacts is introduced. This section as a whole represents an efficient and accurate tool for simulating the anthropomimetic robot dynamic, which is then implemented in Matlab and C++. Two simulation case studies are presented. The first emulates an anthropomimetic robot on a wheeled-base, working in an unstructured environment and therefore exposed to impulse and long-term external disturbances. Here the model is used to observe the robot’s zero-moment-point and thus its balance. The second case study emulates intentional contact – grasping of an object. Models of all contact stages are demonstrated: approach, impact and in-contact-motion phases. Although some applications to the physical robot Cassius [1] and the simulation-base study [2] have already been implemented, there are numerous possibilities for prospective model utilization: simulation and analysis of biomechanical systems/subsystems, study of potential transfer from biological concepts to bio-inspired robotics, anthropomimetic robot system design and analysis, simulation of anthropomimetic robots in interaction tasks, development of model-based advanced control techniques for anthropomimetic robots, testing of anthropomimetic control approaches with additional possibilities of control in interaction tasks, etc. The main outcomes of this section, relating to anthropomimetic model development and extension to include analysis of contact dynamics, have already been published in [3]. At the same time, the simulation model served as a platform for the development of control algorithms elaborated in the following sections. The third section presents the contribution of this thesis to the control of antagonistically actuated, linear/non-linear, cable driven, compliant robot joints, using iv control methods that rely on the conventional control theory. After an overview of the background work on antagonistic actuator control and feedback linearization in robot control, an upgraded puller-follower approach, outlined in the Master’s thesis of Svetozarevic [4], is presented. This biologically-inspired and energy efficient approach simultaneously controls joint position and force in one of the antagonistic tendons. In the present thesis, the puller-follower approach extends to control joint position and joint stiffness. An initially demonstrated single-joint control algorithm, based on feedback linearization, is optimized to compensate for gravity load, changeable effective joint inertia and dynamic coupling in multi-joint systems, by introducing the robust control theory. In addition to loop-shaping robust control, methods based on the non-linear and multivariable control theories were exploited in the puller-follower control scheme. This section ends with limitations and issues that remain open and will be the topic of future research. In addition to control methods that rely on conventional engineering techniques, several cognitive approaches to the control of antagonistically coupled compliant drives in robotics were developed. In accordance with a bio-inspired background and fully human-like design of the anthropomimetic robot, the focus was on human-like control as well – control based on experience, learning and heuristics. To that end, nearest- neighbor algorithms for feedforward and feedback control of the anthropomimetic robot, an algorithm for neural network feedforward control using radial-basis networks, and an algorithm for feedback control based on on-line estimation of kinematic coefficients and fuzzy rules were developed. Although cognitive algorithms were not of primary importance in this thesis, they present a step forward in the control of a fully anthropomimetic robot since engineering-based control algorithms reach their limits if multi-articular muscles or multi-axes joints are considered. On the other hand, cognitive methods outlined in the thesis can be applied to such systems without restrictions. Finally, a comparative simulation study in Matlab was carried out to highlight the main features of the developed cognitive-based control schemes. The scientific contributions regarding cognitive-based control algorithms are presented in the fourth section of the thesis, while all contributions to control anthropomimetic robots are outlined in a paper [5]. v The fifth section provides the point of view and initial research on antagonistically- driven complaint joint control in contact tasks. The proposed control approach is a mixture of state-of-the-art conventional and widely adopted impedance control techniques and bio-inspired patterns of biological antagonistic structures. The main observations, findings, and contributions of the thesis are summarized once again in the final section. Ultimately, we could say that future service robots will not be controlled using exclusively analytical models, conventional control methods or experience-based learning. A combination of all these approaches needs to be elaborated to deal with numerous and diverse tasks, which will be required from robots. Therefore, this thesis is expected to contribute to modeling and control of future human-like service robots. Furthermore, there are areas that could benefit from exact analytical models of human dynamics or the developed control algorithms: gaming industry, biomechanics, motor control, etc. In summary, the author’s contributions presented in detail in this thesis have resulted in a total of 25 publications (one book chapter, four papers in international journals with an impact factor, three papers in domestic journals, eight international conference papers, and nine domestic conference papers).
true
Monte Karlo simulacija transporta pozitrona u realnim sistemima ispunjenim gasom
Monte Carlo simulation of positron transport in realistic gas filled systems
Napredak uređaja za akumulaciju i skladištenje pozitrona omogućava sve preciznija merenja preseka za interakciju pozitrona sa atomima i molekulima. Nedavno kompletirani setovi sudarnih podataka za molekule koji se najčešće javljaju u pozitronskim uređajima i primenama (najpre N2, CF4 i H2O ali i drugi) omogućili su detaljne simulacije transporta pozitrona u takvim sredinama, kao i detaljan opis uređaja koji se na njima baziraju. Osnovni cilj ove disertacije jeste da na osnovu izmerenih sudarnih podataka, pomoću računarskih Monte Karlo simulacija, detaljno analizira realne uređaje i sisteme ispunjene gasom koji svoju funkciju baziraju na elementarnim procesima u kojima učestvuju pozitroni. Disertacija se fokusira na tri konkretna sistema: pozitronski trap, kompresiju snopa pozitrona rotirajućim poljem i termalizaciju pozitrona u biološki relevantnoj sredini. Otkriće i razvoj pozitronskog (Surko) trapa početkom devedesetih godina, predstavlja prekretnicu u polju fizike niskoenergijskih pozitrona i dovelo je do velikog napretka u razumevanju interakcije niskoenergijskih pozitrona sa atomima i molekulima. Trapovi su omogućili akumulaciju velikog broja pozitrona (106-108), niske energije sa energijskom raspodelom koja odgovara sobnoj temperaturi ( <50 meV ) što omogućava formiranje intenzivnih snopova pozitrona sa precizno definisanom energijom. Kao takvi, pozitronski trapovi predstavljaju osnovi alat u brojnim eksperimentima poput merenja preseka za interakciju, proizvodnje antivodonika, proizvodnje i proučavanja pozitronijuma, pozitronske anihilacione spektroskopije i slično. Od trenutka kada je prvi put primenjen, princip rada i dizajn pozitronskog trapa se nije značajnije menjao i standardno je obezbeđivao efikasnost trapovanja od 10-20%. U zavisnosti od konkretne primene trapa, kroz empirijske pokušaje su optimizovani različiti parametri rada, kao što su: efikasnost trapovanja, vreme života akumuliranih pozitrona, maksimalna koncentracija pozitrona ili broj emitovanih sporih pozitrona u sekundi. U disertaciji koristimo preseke koji su izmereni zahvaljujući ovakvim trapovima, kako bismo kroz simulaciju analizirali njihov rad, identifikovali procese koji utiču na performanse trapa i našli način da se performanse optimizuju. Ujedno, simulacija omogućava lakše menjanje osobina trapa koje nije lako promeniti u eksperimentu, kao što su geometrija trapa, ili sastav gasa. Kako simulacija identifikuje proces formiranja pozitronijuma na molekulima N2 kao glavni ograničavajući faktor za ostvarivanje veće efikasnosti, predložen je model, i urađena je detaljna simulacija alternativnog načina trapovanja pozitrona koja kao glavni radni gas umesto N2, za trapovanje koristi vibracione eksitacije molekula CF4, u energijskom opsegu daleko ispod praga za formiranje pozitronijuma. U saradnji sa kolegama sa University College London, iz Grupe za atomsku, molekulsku, optičku i pozitronsku fiziku, pokušana je konverzija standardnog pozitronskog trapa u predloženu CF4 konfiguraciju i utvrđeni su, kroz eksperiment i simulaciju, konstrukcijski nedostaci koji onemogućavaju konverziju najvećeg broja postojećih trapova, kao i aspekti koji bi trebalo da se uzmu u obzir prilikom konstrukcije budućih trapova. Pozitroni akumulirani u trapu se, u najvećem broju modernih trapova, komprimuju primenom rotirajućeg električnog polja. Na taj način se sprečava difuzija pozitrona ka zidovima naprave i značajno produžava vreme života, postiže veća prostorna rezolucija izlaznog snopa i povećava koncentracija pozitrona u snopu. Dipolno ili kvadrupolno električno polje se ostvaruje segmentiranjem jedne od elektroda akumulacionog stepena trapa i primenom adekvatno fazno pomerenog sinusnog napona. Iako se rotirajuće polje odavno koristi za kompresiju jednokomponentnih plazmi, nedavno je pokazano da se na sličan način može komprimovati i oblak pozitrona u tzv. jednočestičnom režimu, kada gustina pozitrona nije dovoljna da se formira plazma, ali princip kompresije u ovom režimu nije bio do kraja objašnjen. U disertaciji je, kroz simulaciju, pokazano da je efekat kompresije u jednočestičnom režimu posledica kompleksne trajektorije pozitrona u uslovima rotirajućeg polja u trapu i oblika i veličine neelastičnih preseka, pre svih, preseka za vibracionu eksitaciju. Analiziran je uticaj jačine i frekvencije primenjenog električnog polja kao i doprinos pojedinih sudarnih procesa brzini kompresije. Prikazana je veza jačine primenjenog električnog polja, brzine kompresije i neizbežnog gubitka pozitrona kroz proces formiranja pozitronijuma usled grejanja roja. Jedna od najbitnijih oblasti primene pozitrona jeste medicinska primena, najpre u ovkiru PET dijagnostike. U ovoj disertaciji, kroz Monte Karlo simulaciju, analizirana je krajnja faza trajektorije pri termalizaciji visokoenergijskih pozitrona u vodenoj pari i jednostavnom modelu biološke sredine kao zamenama za tkivo, koristeći setove preseka bazirane na eksperimentalnim merenjima pozitronske interakcije sa molekulom H2O. Disertacija ukazuje na potrebu postojanja pozitronskih eksperimenata sa rojevima u vodenoj pari, kojima bi se ovakvi setovi normirali i validirali, i u tu svrhu predlaže nekoliko usrednjenih veličina poput vremena termalizacije, energije predate u sudarima sa gasom, ili domet pozitrona, koje bi mogle biti merene i iskorišćene za normalizaciju preseka. Takođe formulisan je jednostavan model u kojem je vodenoj pari dodata organska komponenta, kroz presek za disocijaciju molekula CH4 kao zamenu za proces oštećenja na DNK. Analizirane su trajektorije pozitrona u takvoj sredini, kao i učestanost i prostorna raspodela različitih sudarnih procesa u toku termalizacije. Kroz analizu prostornog profila energije predate sredini, i prostorne raspodele pojedinačnih događaja disocijacije CH4 utvrđeno je da ne postoji nužno korelacija između predate energije i pojedinačnih događaja oštećenja biološkog materijala što ukazuje na potrebu da se unaprede dozimetrijski modeli kako bi se uzela u obzir stohastička priroda procesa oštećenja molekula DNK.
Development and advancement of positron accumulation and storage devices has enabled increasingly accurate measurements of cross-sections for positron interaction with atoms and molecules. Recently assembled cross-section sets for molecules that most commonly occur in positron devices and applications (N2, CF4, H2O and others) have enabled detailed simulations of positron transport in such environments, and also for realistic geometries of devices that utilize such environments. Prime goal of this dissertation is to provide detailed analysis, through use of measured collision data and Monte Carlo simulations, of realistic devices and systems filled with gas, that base their operation on elementary processes involving positrons. Dissertation is focused on three systems in particular: positron buffer gas trap, positron cloud compression by rotating electric field and positron thermalization in biologically relevant environments. Discovery and development of positron (Surko) traps in the early 90's was a milestone in field of low-energy positron physics, and has led to significant advancement in understanding of low-energy positron interaction with atoms and molecules. Buffer gas positron traps have enabled accumulation of large numbers of low-energy positrons (106- 108), with room temperature energy distribution (<50 meV), and allowed formation of intense positron beams with finely controlled beam energy. As such, positron traps are the primary tool in numerous experiments such as cross-section measurements, antihydrogen production, production and study of positronium, positron annihilation spectroscopy etc. From their first application, the operating principles and basic design of positron traps has not changed significantly as it has regularly provided 10-20% trapping efficiency. Through empirical developments, depending on the particular trap application, people have been able to optimize different operation parameters, such as trapping efficiency, lifetime of accumulated positrons, maximum number density of accumulated positrons or number of slow positrons emitted per second. In this dissertation we use the cross-section data that was measured through use of such positron traps in a computer simulation, in order to analyze their operation, identify processes that govern their performance, and find a way to optimize it. At the same time, the simulation gives the benefit of easily modifying trap properties that are not easy to change in the experimental device, such as trap geometry or gas composition. As the simulation identifies the process of positronium formation on N2 molecule as the major limiting factor in achieving higher trapping efficiency, a model is proposed, and the detailed simulation is presented on an alternate scheme of trapping positrons, that exploits vibrational excitation of CF4 molecule, on the scales of energies well below the positronium formation threshold. In collaboration with the colleagues from the Atomic, Molecular, Optical and Positron Physics Group, University College London, we have attempted to convert a standard positron trap to the newly proposed CF4 configuration, and have determined, through both experiment and simulation, problems in current trap designs, that make conversion of most positron traps very difficult, as well as the aspects that should be considered in construction of future traps. In most modern traps, positron clouds that are accumulated are compressed using a rotating electric field. By doing so, diffusion of positrons to the walls of the device is negated which allows for a much longer positron lifetimes, higher spatial resolution of an outgoing positron beam as well as increased positron number density. Dipole or quadruple electric field is produced by segmenting one of the electrodes at the accumulation stage of the trap, and by application of adequately phase-shifted sine voltages. Even though the rotating field has been used for compression of single component plasmas for a long time, it has been shown recently that it is possible to, similarly, compress a cloud of positrons, in a so called single particle regime where positron number density is too low to form a plasma. However the origin of the compression effect has not been completely understood. This dissertation demonstrates through simulation that the compression effect in a single particle regime is a result of complex positron trajectory (in rotating field conditions inside the trap) coupled with the shape and magnitude of the cross-sections for inelastic collisions, first and foremost vibrational excitation collisions. Also presented is the analysis that the magnitude and frequency of the applied electric field, as well as the contribution of individual collision processes, have on the compression rate. Correlation of electric field magnitude, rate of compression and the inevitable loss of positrons through positronium formation (due to the heating of the swarm) is also analyzed. One of the most important fields of positron application is in the field medicine, primarily in the PET diagnostic technique. This dissertation studies, through Monte Carlo computer simulation, the final stages of positron trajectories during thermalization from high energies in water vapor and in a simple model of biological environment as a surrogate for tissue, by using cross-section sets based primarily on measured data for positron interaction with H2O molecule. The dissertation stresses the necessity of positron swarm experiments, which are currently lacking, that could be used to normalize and validate such cross-section sets and for that purpose proposes several averaged quantities such as thermalization times, energy transferred in collisions with gas or positron range, which can be measured and used to for normalization. Also, a simple model is described that adds an organic component to the water vapor, through the cross-section for CH4 dissociation as a proxy for damage to DNA molecules. Trajectories in such an environment are analyzed as well as the rates and spatial distribution of different collisions during thermalization. By analyzing the spatial profile of energy deposited to the medium and the spatial distribution of individual CH4 dissociation events, it can be concluded that the correlation between the deposited energy and the individual events of damage to biological material does not necessarily exist, which indicates the need to improve upon the current dosimetry models in order to fully take into account the stochastic nature of DNA damage.
true
Estimacija pokreta u video sekvencama prilagođena algoritmima za povećanje broja slika u sekundi
Motion estimation in video sequences adapted to algorithms for frame-rate upconversion.
Ova teza opisuje algoritam za estimaciju pokreta u video sekvencama prilagođenu algoritmima za povećanje broja frejmova u sekundi. Predložena metoda kombinuje rekurzivni blokovski pristup i prilagođenu faznu planarnu korelaciju. U poređenju sa alternativnim pristupima predloženi algoritam daje bolje subjektivne impresije i objektivna merenja (Peak Signal to Noise Ratio i Universal Image Quality Index). Računarska kompleksnost algoritma je smanjena i pogodna za izvršavanje u realnom vremenu na platformama sa limitiranim memorijskim i procesorskim resursima. Kao dokaz opisanog koncepta data je prva implementacija u realnom vremenu algoritma za povećanje broja slika (frejmova) u sekundi na mobilnoj platformi (Intel Merrifield). Ona omogućava procesiranje 1080p video sekvence sa 30 frejmova u sekundi i procesorom učestanosti 200 megaherca. Pored predloženog algoritma opisan je i značaj estimacije pokreta u video sekvencama, tipovi, kao i pregled postojećih rešenja, kroz istoriju.
This thesis describes a novel motion estimation algorithm adapted for frame- rate upconversion. The method combines recursive block-matching and customized phase plane correlation. In comparison with alternative approaches the proposed solution shows better subjective impressions and objective measurements, such as Peak Signal to Noise Ratio and Universal Image Quality Index. Computational complexity of the algorithm is constrained and suitable for real-time implementations on platforms with limited memory and processing resources. The proof of concept is the first real-time implementation of frame-rate up-conversion on a mobile platform (Intel Merrifield), which is capable of processing 1080p video at 30 frames per second while clocked at 200 MHz. Becides proposed solution, this thesis describes the importance and usages of motion estimation, its types and currently available solutions, throughout history.
true
Implantacija jona u kanale kristala silicijuma
Ion implantation in the silicon crystal channels
U ovoj doktorskoj disertaciji predmeti eksperimentalnog i teorijskog istraživanja su: (1) određivanje dubinskog profila koncentracije azota u silicijumu izazvanog implantaciom jona azota 14N2+, energije 4 MeV, u <100> i <110> kanale kristala silicijuma i u slučajno orijentisan kristal silicijuma, i (2) određivanje dubinskog profila amorfizacije kristala silicijuma izazvanog implantacijom jona aluminijuma 27Al2+, energije 6 MeV, u <110> kanale kristala silicijuma i slučajno orijentisan kristal silicijuma. U svim gore navedenim slučajevima fluens implantiranih jona je iznosio 1017 jona/cm2. Dubinsko profilisanje azota izvršeno je koristeći metoda nuklearnih reakcija (NRA) izučavanjem 14N(d,α0)12C i 14N(d,α1)12C nuklearnih reakcija i primenom SRIM 2010 i SIMNRA računarskih programa. Za slučajno orijentisani kristal silicijuma uočeno je neslaganje eksperimentalno dobijenih i simuliranih profila koncentracije azota, i, u skladu sa tim, pretpostavljeno je formiranje “mehura” azota u silicijumu i odgovarajuća promena gustine kristala silicijuma, kao moguće objašnjenje ovog neslaganja između eksperimentalnih rezultata i teorije. Pored toga, izmereni su spektri Raderfordovog povratnog rasejanja u kanalisanom slučaju (RBS/C) na implantiranim i neimplantiranim tačkama kristala silicijuma. Ovi spektri su zajedno sa dobijenim dubinskim profilima koncentracije azota korišćeni za određivanje dubinskog profila amorfizacije kristala silicijuma izazvanog implantacijom jona azota. Dubinski profil amorfizacije kristala silicijuma izazvan implantacijom jona aluminijuma određen je skenirajućim mikro-Ramanovim merenjem duž transverzalnog poprečnog preseka implantirane oblasti iv kristala. Koraci skeniranja su bili 0,2 i 0,3 μm, za slučaj implantacije u <110> i slučajno orijentisani kristal silicijuma, respektivno. Dobijeni rezultati su upoređeni sa odgovarajućim RBS/C spektrima. Takođe, snimljene su SEM fotografije tarnsverzalnog poprečnog preseka implantirane oblasti kristala silicijuma u slučaju <110> kristala silicijuma. Utvrđeno je odlično poklapanje maksimuma dubinskog profila amorfizacije kristala silicijuma sa maksimumom koncentracije implantiranog aluminijuma u silicijumu. Ovo jasno pokazuje da se skenirajuća mikro-Ramanova spektroskopija može koristiti kao novi metod u određivanju dubinskog profila amorfizacije kristala koja je izazvana visoko energijskom jonskom implantacijom.
In this PhD thesis, the subjects of the experimental and theoretical research are: (1) determination of the nitrogen depth profile of 4 MeV 14N2+ ions implanted in the <100> and <110> channels of silicon crystals and in a randomly oriented silicon crystal and (2) determination of the silicon crystal amorphization depth profile induced by 6 MeV 27Al2+ ions in the <110> channels of silicon crystal and in a randomly oriented silicon crystal. In all of the above mentioned cases the fluence of the implanted ions was 1017 ions/ cm2. Nitrogen depth profiling was obtained applying the Nuclear Reaction Analysis (NRA) method via study of 14N(d,α0)12C и 14N(d,α1)12C nuclear reactions and using the SRIM 2010 and SIMNRA computer programs. In the case of the randomly oriented silicon crystal pronounced difference between the experimentally obtained and simulated nitrogen profiles was observed, and, with respect to that, formation of the nitrogen “bubble” in silicon and the corresponding change of the silicon density were assumed, as a possible explanation of that difference between the experimental results and theory. Additionally, Raderford Backscattering Spectrometry in channeling mode (RBS/C) was performed on the implanted and virgin silicon crystal spots. Тhese spectra were used, together with the obtained concentration depth profiles, for determination of the amorphization depth profiles of the silicon crystals induced by the nitrogen ions implantation. Silicon crystal amorphization depth profile induced by the aluminum ions implantation was obtained via micro-Raman Spectroscopy (μRS) scanning measurements of the crystal transversal cross section along the implanted region. Scanning steps were 0.2 μm and 0.3 μm for <110> and randomly oriented silicon crystals, respectively. The obtained results were compared with the corresponding RBS/C spectra. Further, the Scanning Electron Microscopy (SEM) photograph of the vi transversal cross section of the implanted region for the case of <110> silicon crystal was taken. It was found an excellent agreement between maxima of the crystal amorphization depth profile and maxima of the implanted aluminum concentration depth profile. This clearly indicates that μRS scanning technique could be used as a novel method for obtaining of the crystal amorphization depth profile induced by the high-energy ion implantation.
true
Funkcije bazisa ultra visokog reda i singularne funkcije bazisa u analizi aksijalno simetričnih metalnih struktura
Ultra high order basis functions and singular basis functions in analysis of axially symmetric metallic structures
Rad elektronskih uređaja i sistema, kao što su mobilna telefonija, televizija, računari, satelitski komunikacioni sistemi, radarski sistemi i mnogi drugi, bazira se na elektromagnetskim fenomenima koji su predmet izučavanja teorije elektromagnetskih polja. Za razumevanje načina rada ovih sistema i za njihovo projektovanje potrebno je izvršiti elektromagnetsku (EM) analizu. U većini slučajeva EM analiza se sprovodi primenom neke od metoda numeričke elektromagnetike. Jedna od najčešće korišćenih numeričkih metoda za EM analizu je metoda momenata primenjena na rešavanje površinskih integralnih jednačina. Primenom ove metode struje raspodeljene po graničnim površima domena aproksimiraju se sumom poznatih funkcija bazisa pomnoženih nepoznatim koeficijentima, a površinske integralne jednačine transformišu se u sistem linearnih jednačina čijim se rešavanjem određuju ti nepoznati koeficijenti. Ova metoda se unapređuje već više desetina godina, a aktuelni tehnički i tehnološki trendovi postavljaju stalne zahteve za povećanjem tačnosti i efikasnosti EM modelovanja, za analizom sve većih i kompleksnijih struktura, koje uključuju sve sitnije detalje, kao i za proširenjem opsega učestanosti, idući od veoma niskih do veoma visokih učestanosti. Povećanje tačnosti, efikasnosti i domena primene ove metode mogu se ostvariti 1) implementacijom kompleksnijih elemenata za precizno modelovanje geometrije problema, 2) računanjem elemenata sistemske matrice sa velikom tačnošću, i 3) korišćenjem sofisticiranih funkcija bazisa za aproksimaciju struje. Cilj ove disertacije je razvoj nove metode, bazirane na metodi momenata, za analizu aksijalno simetričnih metalnih struktura korišćenjem egzaktnog jezgra integralne jednačine električnog polja i primenom sve tri gore pomenute tehnike. Za precizno geometrijsko modelovanje problema korišćeni su zarubljeni konusi, a struja je inicijalno aproksimirana modifikovanim funkcijama bazisa višeg reda. Impedansni integrali, koji predstavljaju ii elemente sistemske matrice su, uz pomoć simetrije, redukovani sa četvorostrukih na trostruke integrale, pri čemu prve dve integracije odgovaraju integralima potencijala. Svaka od tri integracije u okviru impedansnih integrala pokazuje singularno ili kvazi-singularno ponašanje u određenim situacijama. U tom slučaju se ovi integrali neefikasno računaju direktnom primenom Gaus-Ležandrove integracione formule, pa su za njihovo računanje predložene nove tehnike poništavanja singulariteta. Tehnika poništavanja singulariteta podrazumeva primenu pogodno odabrane smene kojom se integrand transformiše u sporo promenljivu funkciju pa se integral zatim može efikasno izračunati numerički, primenom Gaus-Ležandrove integracione formule. Nakon odabira optimalnih smena osmišljene su jednostavne formule za određivanje potrebnog broja tačaka integracije za ostvarivanje željene tačnosti. Računanje impedansnih integrala sa velikom tačnošću (do mašinske preciznosti), kao što je pokazano na nizu numeričkih primera, omogućava: a) veliku tačnost analize (relativna srednja kvadratna greška se spušta do 10–6), b) stabilnost rezultata za električki male strukture (do 10–6 λ, gde je λ talasna dužina EM talasa u vakuumu), v) efikasnu analizu električki velikih struktura (do 100λ), i g) dobijanje veoma tačnih i stabilnih rezultata u širokom opsegu vrednosti (do 100 dB). U narednom delu disertacije analizirana je mogućnost povećanja maksimalno upotrebljivog reda razvoja struje implementacijom maks-orto funkcija bazisa. Naime, sa povećanjem reda razvoja modifikovanih funkcija bazisa uslovljenost sistemske matrice se pogoršava, tako da za redove iznad n = 10 u običnoj, tj. n = 20 u duploj preciznosti, rezultati naglo počnu da divergiraju. Za razliku od toga, implementacija maks-orto funkcija bazisa može obezbediti stabilnu sistemsku matricu praktično bez obzira na korišćeni red razvoja, ali samo ukoliko se elementi sistemske matrice proračunaju sa dovoljno velikom tačnošću. Pokazano je da se sistemska matrica može izračunati sa relativnom srednjom kvadratnom greškom manjom od 10–11 za redove razvoja do 128, ako se matrični elementi računaju kao linearna kombinacija impedansnih integrala u kojima figurišu Ležandrovi polinomi i njihovi izvodi. Pri tome je potrebno da se Ležandrovi polinomi i njihovi izvodi u okviru impedansnih integrala iii računaju pomoću već poznatih rekurentnih formula. Time je omogućeno korišćenje ultra visokih redova razvoja (testiranih do n = 128), kao i smanjenje broja nepoznatih za oko 10 puta u odnosu na modifikovane funkcije bazisa. Mogućnosti metode su dodatno proširene primenom singularnih funkcija bazisa koje uzimaju u obzir ivične efekte. Naime, na spoju dva elementa čije generatrise zaklapaju ugao manji od 180o, kao i na otvorenoj ivici strukture, postoji teorijski beskonačno velika gustina površinskog naelektrisanja koja se teško može aproksimirati polinomskim funkcijama bazisa kao što su modifikovane i maks-orto. Stoga je predložena i primena singularnih funkcija bazisa, čiji je izvod beskonačan na ivici gde je posmatrani ivični efekat. Uvođenje singularnih članova u izraze za funkcije bazisa dodatno otežava računanje impedansnih integrala. Odabirom novih optimalnih smena za tehniku poništavanja singulariteta omogućeno je računanje impedansnih integrala usled singularnih funkcija bazisa do mašinske preciznosti. Prednosti implementacije singularnih funkcija bazisa ilustrovane su na primerima bikonične antene dužine kraka od 0,25λ do 10λ, pri čemu je pokazano da se za isti broj nepoznatih ostvaruje tačnost rezultata koja je za 3–5 reda veličine veća od slučaja kad se ove funkcije bazisa ne koriste. Nova metoda omogućuje veoma efikasnu EM analizu aksijalno simetričnih metalnih struktura sa velikom tačnošću (7–10 cifara tačnosti) u veoma širokom opsegu učestanosti, tako da se dimenzije strukture (radijalne i aksijalne) kreću u opsegu od 10–6λ do 100λ. Iskustvo i znanje stečeno pri implementaciji maks-orto funkcija bazisa ultra visokog reda, singularnih funkcija bazisa i tehnika poništavanja singulariteta za numeričko računanje impedansnih integrala može se iskoristiti i proširiti na analizu proizvoljnih metalno-dielektričnih struktura.
The operation of electronic devices and systems, such as mobile phone systems, television, computers, satellite communication systems, radar systems and many more, is based on electromagnetic phenomena, the subject of studies of electromagnetic field theory. In order to understand the functionality of these systems and to complete the design, it is necessary to perform electromagnetic (EM) analysis. In most of the cases EM analysis is carried out by applying one of the methods of computational electromagnetics. One of the most frequently used numerical methods for EM analysis is the method of moments applied to solving surface integral equations. By the use of this method, the currents distributed along boundary surfaces of various domains are approximated as the sum of known basis functions multiplied by unknown coefficients, and surface integral equations are transformed into a system of linear equations. The unknown coefficients are determined by solving this system. The moment method has been growing and improving for more than several decades. Current technological and technical trends are continuously expanding the demands for more accurate and efficient EM modeling to include complex and electrically large structures with very small details. Furthermore, required frequency range is widening, spanning from very low up to very high frequencies. The efficiency, precision, and the application area of the method can be increased 1) by implementing more complex elements for precise geometrical modeling of the problem, 2) by evaluating the system matrix elements more accurately, and 3) by the use of sophisticated basis functions for the current approximation. The goal of this dissertation is to develop a new method, based on moment methods, for the analysis of axially symmetric metallic structures utilizing exact kernel of electric field integral equation and all three above mentioned techniques. Right-truncated cones are used for precise geometrical modeling of the structures, and the currents are initially approximated by modified higher order basis functions. Impedance integrals, representing the system matrix elements, are reduced from fourfold to threefold integrals, in which the first two integrations correspond to potential integrals. All three v integrations within impedance integrals can potentially be singular or quasi-singular in some situations. In that case, the integrals cannot be efficiently evaluated by directly applying the Gauss-Legendre integration formula. Therefore, new singularity cancellation techniques are proposed in the dissertation for the efficient evaluation of the singular or quasi-singular integrals. By the use of a convenient change of variables, the singularity cancellation technique transforms the integrand into a slowly varying function, after which the integral can be efficiently evaluated numerically using the Gauss-Legendre integration formula. After selecting the optimum variable transforms, simple formulas are determined for estimating required number of integration points needed for prescribed accuracy. As shown on a number of numerical examples, accurate evaluation of the impedance integrals (up to machine precision) enables: a) high precision analysis (relative root-mean square error is brought down to 10–6), b) obtaining stable results for electrically small structures (with dimensions as small as 10–6λ), c) efficient analysis of electrically big structures (with dimensions up to 100λ), and d) obtaining accurate and stable results with a dynamic range of up to 100 dB. In the next part of the dissertation, the possibility of increasing maximally applicable expansion order for current distribution is examined by implementing maximally orthogonalized basis functions (max-ortho). Namely, with increasing the expansion order of modified basis functions the stability of system matrix worsens, so that for orders higher then n = 10 in single, and n = 20 in double precision, the results diverge rapidly. In contrast, the implementation of max-ortho bases provides stable system matrix regardless of the expansion order, but only if the system matrix elements are evaluated with sufficient precision. It is shown that the system matrix due to max-ortho bases can be evaluated with relative root mean square error lower than 10–11 for expansion orders up to 128, if the matrix elements are expressed as a linear combination of impedance integrals due to Legendre polynomials and their derivatives. Particularly, Legendre polynomials and their derivatives within impedance integrals should be calculated by the use of the well known recurrent formulas. In this way, ultra high expansion orders (tested up to order 128) can be applied, and the number of unknowns is reduced approximately 10 times with respect to modified basis functions. The possibilities of the method are further extended by the use of singular basis functions which take into account the edge effect. Namely, at the junction of two vi elements whose generatrices form an angle smaller than 180o, as well as at the open edge of the structure, the surface charge density is theoretically infinite. Using polynomial bases, like modified and max-ortho, to approximate infinite charge density is inefficient. The implementation of singular basis functions with infinite derivative at the location of the observed edge effect increases the efficiency of the method. However, the singular terms in the basis functions additionally complicate the evaluation of the impedance integrals. By selecting new optimal variable transforms for the singularity cancellation technique, precise evaluation of the impedance integrals due to singular bases can be performed up to machine precision. The advantages of implementation of singular basis functions are illustrated on several examples of bi- conical antennas with arm lengths spanning from 0,25λ to 10λ. It is shown that for the same number of unknowns achieved accuracy of the results is 3–5 orders of magnitude higher compared to the case where regular bases are used. The new method enables very efficient EM analysis of axially symmetric metallic structures with high precision (7–10 significant digits) in a very wide frequency range, so that dimensions of the structure (radial and axial) can span from 10–6λ to 100λ. The experience and knowledge gained through implementation of maximally orthogonalized ultra high order basis functions, singular basis functions, and singularity cancellation techniques for numerical evaluation of impedance integrals can be extended to arbitrary metallic-dielectric structures.
true
Uticaj ekvivalentne temperature Maksvelovog spektra gasa slobodnih elektrona na proboj gasova pri malim vrednostima pritiska i međuelektrodnog rastojanja
The influence of equivalent temperature of maxwell spectrum of free electron gas spectrum of free electron for gas discharge at small pressure values and interelectrode gap.
Polazeći od kinetičke teorije gasova, odnosno od Maksvelove raspodele po brzinama uz pretpostavku da se svaki gas može posmatrati kao smeša gasova koja se sastoji od atoma (molekula) gasa, jona gasa i gasa slobodnih elektrona uz poznate izraze za efikasni presek elastičnih interakcija, u integralnom obliku, izveden je izraz za prvi Tauzendov jonizacioni koeficijent. U tako izvedenim izrazima pod pretpostavkom Taunsendovog mehanizma proboja, koji je karakterističan za male vrednosti pritiska i međuelektrodnog rastojanja, figuriše ekvivalentna temperatura (energija) gasa slobodnih elektrona, kao direktna promenljiva u jonizacionim procesima. Iz tako dobijene zavisnosti izveden je izraz za proboj gasa. U cilju ispitivanja zavisnosti probojnog napona od probojne temperature (energije) odgovarajućeg Maksvelovog gasa slobodnih elektrona, izvršena je teoretska analiza pomenute zavisnosti, što uključuje izvođenje izraza za zavisnosti probojnog napona od probojne temperature (energije slobodnih elektrona), pod pretpostavkom Taunsendovog mehanizma proboja i homogenog električnog polja. Provera adekvatnosti dobijenih izraza izvršena je poređenjem, pod usvojenim pretpostavkama, dobijenih rezultata sa odgovarajućim eksperimentalno dobijenim rezultatima. Eksperimentalni rezultati su dobijeni merenjem zavisnosti DC probojnog napona plemenitih gasova u okviru Pašenovog minimuma za homogeno električno polje. Pod istim uslovima vršena su i merenja vrednosti impulsnih probojnih napona (izuzetno sporim impulsnim naponom). Eksperimentalni postupak je vršen pod dobro kontrolisanim laboratorijskim uslovima uz kombinovanu mernu nesigurnost manju od 5%. Dobijeni rezultati se odnose na: 1-analizu uslova pod kojim se na gas slobodnih elektrona, posmatran kao komponenta gasne smeše, može primeniti Maksvelova raspodela, odnosno Drajveštajnove raspodele (veoma slična Maksvelovoj); 2-izvođenje izraza za koeficijent jonizacije, pod pretpostavkom primenljivosti Maksvelove odnosno Drajveštajnove raspodele, u zavisnosti od proizvoda pritiska i međuelektrodnog rastojanja; 3-izvođenje uslova za proboj gasa prema Taunsendovom mehanizmu, na osnovu tako dobijenih koeficijenata jonizacije i određivanje koeficijenta u tako izvedenim uslovima za proboj; 4-poređenje dobijenih izraza za jonizacione koefiicjente sa odgovarajućim izrazima prema Taunsendu i prema Takašiju; 5-prikaz zavisnosti parametara Maksvelovog spektra gasa slobodnih elektrona od parametara izolacionog sistema izolovanog plemenitim gasom pri malim vrednostima pritiska pd.
Starting from the kinetic theory of gases, or more exactly, the Maxwell’s distribution according to velocity on the assumption that each gas can be observed as a gas mixture consisting of gas atom (molecule), gas ion and free electrons gas with the known expressions for efficient cross-section of elastic interaction, in integral form, the first Townsend ionization coefficient was derived. In expressions derived in such way, and under assumption of the Townsend breakdown mechanism, which is typical of small pressure values and interelectrode distance, equivalent temperature (energy) of free electrons’ gas, as well as direct variable in ionization processes. The expression for gas breakdown was derived from dependency obtained in such a way. With the aim of investigating the breakdown voltage dependence on breakdown temperature (energy) of the corresponding Maxwell’s gas of free electrons, the theoretical analysis of the mentioned dependence was performed, which includes deriving expressions for breakdown voltage dependence on breakdown temperature (free electrons’ energy), on the assumption of the Townsend breakdown mechanism and homogeneous electric field. Taking into account the adopted assumptions the adequacy check out of the obtained expressions was completed by comparing the obtained results with the corresponding experimentally obtained results. The experimental results were obtained by measuring the DC breakdown voltage dependence of noble gases within the Paschen’s minimum for a homogeneous electric field. A series of measurements was carried out for impulse breakdown voltage values (by exceptionally slow impulse voltage) under the same conditions. The experimental procedure was performed under well-controlled laboratory conditions combined with measurement uncertainty that was less than 5%. The obtained results refer to: 1. The analysis of conditions under which the Maxwell’s distribution can be applied the free electrons’ gas, observed as the gas mixture component, can be applied, or under which Druyvenstein’s distribution (which is very similar to Maxwell’s) can be applied; 2. Deriving the ionization coefficient expression, on the assumption of applicability of Maxwell’s that is Druyvenstein’s distribution, depending on the product of pressure and interelectrode distance; 2. Deriving the conditions for gas breakdown according to the Townsend mechanism, on the basis of ionization coefficients obtained in such a way and determining the coefficient under breakdown conditions derived in such manner; 4. The comparison of the obtained expressions for ionization coefficients with the corresponding expressions according to Townsend and Takashi; 5. The review of parameter dependence of the Maxwell’s gas spectrum of free electrons in relation to the parameters of insulating system isolated by a noble gas at small values of the pd pressure.
true
Primena neuralnih mreža u rešavanju transportnih problema
Application of neural network in solving of transportation problems
Dvadeseti vek je obeležen ekspanzijom u razvoju, priozvodnji i primeni računara i sofisticiranih elektronskih uređaja najrazličitijih namena. Niska cena računara, dostupnost Interneta, stalni razvoj i poboljšanja na nivou pristupne mreže, razvoj servisa koji omogućavaju najširi spektar mogućnosti i obavezno obrazovanje u domenu računarske pismenosti, stvorili su milijarde korisnika na celom svetu. Danas na tržištu postoji izuzetno veliki broj telekomunikacionih korisničkih uređaja i usluga koji se permanentno povećavaju, usložnjavaju i modifikuju. Među velikim brojem korisničkih servisa, korisnicima su na raspolaganju i vrlo sofisticirani multimedijalni servisi koji rade u realnom vremenu i u interakciji sa jednim ili većim brojem drugih korisnika (mogućnost kreiranja i distribucije multimedijalnog sadržaja, streaming audio i video fajlova, video konferencije, četovi, rad sa distribuiranim bazama podataka, itd.). Mobilni uređaji danas omogućavaju realizaciju skoro svih integrisanih korisničkih funkcionalnosti i servisa, koje su do skoro realizovali posebni uređaji. Danas imamo mogućnost prijema TV signala na mobilnim uređajima, pristup web i mail serverima i servisima, pristup specijalizovanim korisničkim servisima i informacijama i sl. Obzirom da je prisutna vrlo aktivna tržišna konkurencija, i da davaoci usluga uvek žele da ponude novi i bolji priozvod, to neminovno dovodi do stalnog razvoja, usavršavanja i primene novih tehnoloških rešenja kod proizvođača. Svi navedeni servisi svoj način rada baziraju na prenosu većeg broja informacija, čime se gustina saobraćaja u komunikacionim mrežama stalno povećava. Savremeni servisi zahtevaju strukturu mreža sa većim propusnim opsezima, većim brojem alternativnih putanja, bržim prenosom, malim kašnjenjem itd. Sa druge strane, korisnici očekuju kvalitetne servise, dostupnost mreže, garancije određenog kvaliteta i brz odziv sistema. Imajući sve ovo u vidu, jasno je da struktura i kvalitet komunikacione mreže moraju permenentno da se povećavaju i poboljšavaju. Pored toga, neophodna je stalna evaluacija uređaja i algoritama za rutiranje, koji mogu da obezbede pouzdanu i kvalitetnu komunikaciju. Danas postoji veliki broj stručnih radova koji na najrazličitije načine pokušavaju da doprinesu unapređenju kvaliteta prenosa signala. Kako su zahtevi sve složeniji, nameće se potreba za sve sofisticiranijim tehnologijama i alatima koji mogu da ponude bolje performanse. Jadna ovakva grupa alata je bazirana na primeni modela i mehanizama koji postoji kod visoko inteligentnih živih bića, i naziva se veštačka inteligencija (Artificial intelligence). Posebnu podgrupu čine alati bazirani na principima rada ljudskog mozga, u smislu detekcije nadražaja, prenosa informacija, memorisanja i načina odlučivanja. Ova naučna disciplina obrađena je kroz Veštačke neuralne mreže (Artificial neural networks). U ovoj disertaciji su prikazane pojedine mogućnosti veštačkih neuralnih mreža, date njihove komparativne prednosti u odnosu na standardne računarske uređaje i izvršena njihova klasifikacija u odnosu na veći broj kriterijuma. Posebno je obrađena jedna grupa veštačkih neuralnih mreža, pod nazivom rekurentne neuralne mreže, iz razloga što pokazuju vrlo dobre rezultate u rešavanju složenih optimizacionih problema. Kako je istaknuti predstavnik ove grupe mreža Hopfield-ova neuralna mreža, to je ona izabrana za rešavanje transportnih problema kojima se bavi ova disertacija. U disertaciji je detaljno analizirano kreiranje matematičkog modela Hopfield- ove neuralne mreže i mogućnosti hardverske realizacije istog. Posebna pažnja posvećena je definisanju termina energijske funkcije, koji ima ključnu ulogu u radu Hopfield-ove neuralne mreže. Obzirom na problem ispravnog izbora energijske funkcije, obrađene su relacije koje definišu opštu konvergenciju energijske funkcije i analizirani su uslovi za stabilnost rada ove mreže. Obzirom da se ova mreža može koristiti za rešavanje složenih optimizacionih problema, ukazano je na mogućnost pojave lažnih stanja koja se mogu javiti u njenoj implementaciji. Polazeći od opšteg modela Hopfield-ove neuralne mreže, njene primene na rešavanje problema Trgovačkog putnika, ali i modifikacija koje su drugi autori realizovali u cilju pronalaženja najkraće putanje, objašnjene su sve relacije koji opisuju datu mrežu, a koje su kasnije korišćene kao polazna osnova za dalje modifikacije koje uvodi disertacija. Shodno opisanim potrebama za unapređenjem kvaliteta protokola za rutiranje, u ovoj disertaciji je razvijen modifikovani model Hopfield-ove neuralne mreže namenjen dinamičkom multicast rutiranju u paketskom saobraćaju. Ukazano je na probleme u distribuciji i rutiranju multimedijalnog sadržaja, i na bazi definisanih uslova za njihov prenos, opisani su postojeći algoritmi i metode koje se trenutno koriste. Kako je primarni problem u multicast rutiranju formiranje multicast stabla, u ovoj disertaciji je predloženo rešenje bazirano na Hopfield-ovoj neuralnoj mreži. Da bi predloženo rešenje odgovaralo realnim uslovima rada, analizirana je višekriterijumska optimizacija pet različitih mrežnih parametara i opisan način za dobijanje multicast putanje u mreži proizvoljne topologije. Dobijeni rezultati su poređeni sa rezultatima dostupnim u stručnoj literaturi. Analizirajući veći broj različitih kriterijuma, dobijeni rezultati pokazuju bolji ili bar isti kvalitet. Pored toga, analiziran je veći broj performansi predloženog rešenja (brzina rada, potreban broj iteracija, verovatnoća blokade, skalabilnost rešenja, itd.). U disertaciji je predložen i novi način određivanja optimalnog broja iteracija u radu Hopfield-ove neuralne mreže, i na taj način kreirana logika u cilju minimizacije potrebnog broja iteracija i vremena rada mreže. Pokazano je da dobijeni rezultati imaju daleko bolje performanse od sličnih u dostupnoj literaturi. Imajući u vidu da je predloženo rešenje vezano za rutiranje informacija u računarskim mrežama, analizirani su i algoritmi za rutiranje u optičkim mrežama. Obzirom da sve veći broj komunikacionih mreža prelazi na tehnologiju optičkog vlakna, posebno zahtevni uslovi se nameću kod algoritama za rutiranje svetlosnog signala. U disertaciji je predložen novi algoritam za rutiranje i dodelu talasnih dužina u potpuno optičkim mrežama. Ovde su prikazane sve faze u kreiranju modela, modifikacije originalne topologije optičke mreže, kreiranje modifikovane energijske funkcije Hopfield-ove neuralne mreže i primena iste na problem rutiranja svetlosnog signala. Predloženo rešenje omogućava veliki stepen skalabilnosti i prilagodljivosti na realna ograničenja u pojedinim optičkim mrežama. Pored ovoga,analizirani su i rezultati za dodatni predloženi model koji rešava problem faze rutiranja (unutar algoritma rutiranja i dodela talasnih dužina), dok su za potrebe dodele talasnih dužina korišćeni, opisani, standardizovani algoritmi. Kako je Hopfield-ova neuralna mreža pokazala dobre rezultate u domenu višekriterijumske optimizacije u domenu rutiranja signala, analizirana je i moguća primena njenog modifikovanog modela na rešavanje opštijih transportnih problema. U disertaciji je predloženo rešenje za pronalaženje nove putne trase, koja je sastavni deo analize i planiranja saobraćaja i prostora. Obzirom da je potrebno analizirati uticaj geomorfoloških i društveno-ekonomskih karakteristika terena na izgradnju trase, ali i uticaj trase na političke, ekonomske i ekološke faktore, predloženo je rešenje na bazi višekriterijumske analize većeg broja nekonzistentnih faktora sa različitim uticajima. Predloženo rešenje je testirano u realnim uslovima na terenu Republike Srbije i može da posluži kao pomoćno sredstvo prilikom izrade studija o novim putnim pravcima. Sva predložena rešenja, ili modeli kojima se dodatno mogu poboljšati postojeći algoritmi, izloženi u ovoj disertaciji, ukazuju da je primena Hopfield-ove neuralne mreže opravdana u većem broju različitih tipova problema koji zahtevaju rešavanje višekriterijumskih optimizacionih problema.
The twentieth century has been marked by the expansion and development of new computer applications as well as sophisticated electronic devices for various purposes. Low cost of computer components, Internet availability, continuous development and improvement of the network access, development of services that enable the widest range of options and compulsory education in the field of computer literacy, have generated billions of users worldwide. Nowadays, the market of telecommunications users’ devices and services, is extremely rich and it is constantly increasing, and being modified by more complex applicationes and devices. Among the large number of services, users are offered by highly sophisticated multimedia services that work in real time, interacting with one or more users. They have the oportunity to create and distribute multimedia content, to stream audio and video files, to create video conferences, chats, working with distributed databases, etc. Mobile devices now allow realization of almost all of the integrated users’ service and functionalities, which until recently have been realized by special devices. Today we have the possibility of receiving TV signals on mobile devices, access to web and mail servers and services, as well as access to specialized services and information. Concerning very active market competition, and the fact that providers always want to offer new and better products, continuous development is inevitable, combined with constant improvement and application of new technological solutions. All above mentioned services are based on transfer of large amounts of information, and on that way they are constantly increasing the density of traffic in communication networks. Modern services require network structure with higher bandwidth ranges, larger number of alternative paths, fast transfer, low delays etc. On the other hand, users expect high quality services, fast access to networks, the guarantee of certain quality and quick response systems. Bearing this in mind, it is clear that the structure and quality of communication network permanently needs to grow and improve. In addition, ongoing evaluation of devices and algorithms for routing providing reliable and solid communication is needed.. Today, there are a large number of scientific papers aimed to improve the quality of signal transmission. With the increase of complexity of demands, the need for more sophisticated tools, that can offer better performances exists. One such group of tools is based on application of models and mechanisms that exist in highly intelligent entities, is artificial intelligence. Its special sub-group consists of tools based on the principles of the human brain, in terms of stimulus detection, transmission, storage, and decision-making process. This scientific discipline is realized through artificial neural networks. This dissertation presents capabilities of artificial neural networks, their comparative advantages over standard computing devices, and clasifies them in relation to a number of criteria. Special emphasis is given on recurrent neural networks as a very important artificial neural networks. They show very good results in solving of complex optimization problems. As a prominent representative of this group of networks Hopfield neural network is considered here. This type of network has been chosen for solving of the transport problems, dealt with in this dissertation. The dissertation examines in detail the creation of a mathematical model of Hopfield's neural network and the capabilities of its hardware implementation. Special attention has been paid to defining of the Energy function, which plays a key role in the Hopfield neural network implementation. Given the proper choice of energy function, the relations that define the general convergence of energy functions were handled and analyzed the conditions for the stability of the network. Since this network can be used for solving complex optimization problems, we point out the possibility of false situations that may arise in its implementation. Starting from the general model of Hopfield's neural network, its application in solving of the Traveling salesman problem, as well as modifications that were implemented by other authors with the goal of finding of the shortest path, all relationships that describe a given network are explained. These have been used later on as a basis for further modifications introduced by this thesis. In connection to the requirements described for improving of the quality of a routing protocol, in this dissertation, a modified model of Hopfield's neural network designed for dynamic multicast routing in packet traffic is shown. The problems in the distribution and routing of multimedia content were defined, and based on defined transfer preconditions the existing algorithms and methods that are currently used are described here. The primary problem in multicast routing is multicast tree establishment, and the solution based on the Hopfield's neural network is proposed here. In order to fit the proposed solution with the real conditions, the multicriteria optimization of five different network parameters with the description of obtainig of the multicast path in a network of arbitrary topology is explained. The results were compared with results available in the literature. By analyzing a number of different criteria, the obtained results show a better or at least the same quality of the final solution then others. In addition, we analyzed the performance of a number of solutions (operation speed, the required number of iterations, the probability of blocking, scalable solutions, etc.). The dissertation proposes a new way of determining of the optimal number of iterations in the Hopfield neural network implementation. In this way a logic of minimizing of the required number of iterations and network operation time is derived. It is shown that the results are significantly better than similar ones in literature. Since the proposed solution is given regarding the routing information in computer networks, the algorithms for routing in optical networks were analyzed, too. Having in mind the existence of increased number of communication networks and that it goes towards the optical fiber technology implementation, particularly demanding conditions are imposed to algorithms for routing of the optical signal. The dissertation proposes a new algorithm for routing and wavelength assignment in all-optical networks. This solution shows all of the stages in the creation of models, modifications of the original topology of optical networks, creating a modified energy function of Hopfield's neural network and usage of the same one in optical routing problem. The proposed solution allows a high degree of scalability and adaptability to the real limitations in certain optical networks. We have further analyzed the results of the additional proposed model that solves the problem of routing phase (within the routing algorithm and assignment of wavelengths), while for the allocation of wavelengths we have used the described standardized algorithms. Since Hopfield's neural network showed good results in the field of multicriteria optimization in the domain of routing signal, we analyzed its possible application of the modified model to solve more general transport problems. The dissertation proposed solution for finding optimal alignment of the new road junction, which is an integral part of business analysis and transportation planning. Whereas it is necessary to analyze the influence of geomorphologic features of the ground on construction of the route, and track the impact of political, economic and environmental factors, the proposed solution is based on multicriteria analysis of a number of factors inconsistent with the effects differed. The proposed solution is tested in real conditions on the ground of the Republic of Serbia, and it can be used as an aid in conducting studies on new roads. All proposed solutions, or models that can further improve the existing algorithms, are presented here. They indicate that the use of Hopfield's neural network is justified in a number of different types of problems that need solving multicriteria optimization problems.
true
Novi algoritmi za povećanje verovatnoće formiranja sigurnih senzornih barijera korišćenjem mobilnih čvorova
New algorithms for improving the probability of strong barrier coverage using mobile nodes
U ovoj tezi se razmatra problem pokrivenosti bežičnim senzorskim mrežama u slučaju stohastičkog razmeštaja čvorova u određenom regionu od interesa. Stohastički razmeštaj čvorova je jedini način instalacije mreže u situacijama kada je pristup terenu otežan ili nemoguć. Tipični primeri ovog načina instalacije nalaze se u oblasti vojnih primena. U vojnim aplikacijama, najčešće je potrebno obezbediti dobru pokrivenost regiona od interesa senzorskim poljem, bez obzira na činjenicu što se senzori ne mogu manuelno smestiti u optimalnim pozicijama gde je pokrivenost maksimalna pri minimalnom broju senzorskih čvorova i pri određenom nivou robusnosti i redundatnosti mreže. Štaviše, ovakav vid instalacije se može realizovati uglavnom samo upotrebom artiljerije ili avijacije te je, u ovim uslovima, pre nego što se pristupi optimizaciji mreže po pitanju potrošnje energije, sigurnosti i pouzdanosti, potrebno odrediti fundamentalne parametre pri kojima se može postići osnovna funkcionalnost mreže, odnosno određeni stepen pokrivenosti područja od interesa i potrebna konektivnost. Područje od interesa može biti region bilo kojeg oblika. U vojnim aplikacijama, to je najčešće geografsko područje površine od 5 do 20 . Ako je cilj očitavanje veličina iz čitavog regiona, govorimo o pokrivenosti regiona. Drugi čest slučaj primene senzorskih mreža u vojnim aplikacijama, odnosi se na obezbeđivanje određene linije između dva regiona po pitanju prelaska neprijateljskih vojnih trupa sa jednog regiona na drugi. U ovom slučaju se govori o pokrivenosti barijere. U ovoj tezi se vrši modeliranje i analiza stohastičke implementacije bežičnih senzorskih mreža da bi se, u okviru modela, izvršila evaluacija pokrivenosti regiona kao i evaluacija i poboljšanje pokrivenosti senzorskom barijerom korišćenjem mobilnih čvorova ili robota. U ovom pravcu se, pod određenim početnim pretpostavkama i aproksimacijama, daju uslovi za postizanje očekivanog nivoa određenog tipa pokrivenosti kao i određivanje uticaja fundamentalnih parametara mreže na kvalitet pokrivenosti za one tipove pokrivenosti gde se kvalitet može izraziti u pogledu očekivane verovatnoće detekcije objekata ili događaja. Međutim, kada je u pitanju sigurna senzorska barijera, ona može postojati ili ne postojati. U tom smislu se ne može govoriti o stepenu kvaliteta pokrivenosti već samo o verovatnoći postojanja jakih (ili sigurnih) senzorskih barijera. Na osnovu teorijskih izvedbi iz literature, može se zaključiti da je sigurno postojanje jake senzorske barijere moguće ostvariti samo pod određenim strogim uslovima instalacije. Međutim, kako su ti uslovi praktično teško ostvarivi, u ovom radu se predlažu dva algoritma za dopunjavanje mreže koristeći mobilne čvorove ili robote u uslovima kada nisu zadovoljeni strogi uslovi postojanja jakih senzorskih barijera. Za implementaciju algoritma se razvija simulator kojim se određuje pozicija i broj dodatnih mobilnih čvorova potrebnih da bi se u svakoj situaciji sa velikom verovatnoćom formirala jaka senzorska barijera. Funkcionalnost simulatora se verifikuje korišćenjem teorijski izvedenih rezultata za pokrivenost regiona i pokrivenost slabom senzorskom barijerom. Umesto korišćenja potpuno mobilnih mreža (koje su manje fleksibilne u pogledu troškova realizacije mreže), ovde se koriste hibridne mreže, gde se predloženi algoritmi za određivanje pozicija i broja dodatnih čvorova mogu praktički realizovati na centralizovan ili distribuiran način. Za dopunjavanje mreže se mogu koristiti mobilni čvorovi (koji se kreću do određene tačke, a zatim izvršavaju istu funkciju kao i svi ostali statički čvorovi) ili roboti (koji nose određeni broj fiksnih čvorova i pozicioniraju ih u određenim tačkama mreže). U okviru teze se takođe predlažu algoritmi za pomeranje ovih uređaja u cilju ostvarivanja sigurnih senzorskih barijera. Uzevši u obzir neke trendove razvoja u sistemskom i softverskom inženjerstvu, koji se odnose na povećanje efikasnosti razvoja rešenja koristeći odgovarajuće modele unapređivanja procesa i na metodologiju mašinksog prepoznavanja podatkovnih struktura, u ovoj tezi se daje opis razvoja rešenja u kontekstu integracionog modela sazrevanja procesa a najvažniji aspekti procesa automatskog prepoznavanja pozicija i osnovnih funkcionalnih karakteristika čvorova u zavisnosti od okolnosti instalacije se opisuje kroz metode konceptualnog modeliranja.
This thesis treats the problem of the network coverage in randomly deployed wireless sensor networks. Placing the nodes quasi-randomly is often the only way of deploying the network in geographically inaccessible regions. This way of deployment is often present in military applications. In military applications, high quality of the network coverage is usually a primary goal, despite the fact that the nodes cannot be placed manually in the optimal positions where they would all together provide the highest coverage, a certain degree of robustness, and redundancy, by using the minimum number of nodes. Bearing in mind that this kind of installation is usually realized by using artillery or the aircrafts, before approaching the energy, security, and reliability optimization, it is important for the application to meet the basic functionality, i. e., it is important for the values of the basic parameters that provide a certain degree of the network coverage and connectivity for a given area of interest to be assessed. The area of interest can be a region of any shape. In military applications, it is usually the region of size 5-20 . If the aim is for the whole area to be covered, we deal with the area coverage. Another usual example of the appliance of the wireless sensor networks in military applications is related to the detection of intruders while attempting to cross the line between two regions. In this case, we deal with the barrier coverage. This thesis covers the modeling of the stochastic implementation of the wireless sensor networks, the analysis of the area and the barrier coverage, as well as the methods for improving the barrier coverage by using mobile nodes or robots. Under the initial assumptions and approximations, the conditions for an expected degree of the specific types of coverage are given for those types of coverage whose quality can be expressed in terms of the expected detection probability. But, regarding the strong barrier coverage, instead of assessing the coverage quality, only the information on barrier coverage creation can be derived, i.e., a strong barrier coverage can either be accomplished or not accomplished. According to the theoretical results from literature, it can be concluded that the full barrier coverage can be achieved almost always if the very strict conditions are met. These conditions, however, are hard to be provided. Therefore, in this thesis, two algorithms for creating the full barrier coverage, by filling the network breaks with mobile nodes or by using robots, in the situations where these conditions are not met, are proposed. In order for the algorithms to be implemented, a simulator is developed. According to the rules defined in algorithms, the simulator determines the number of breaks, the position and the appropriate number of additionally needed nodes for the strong barrier coverage to be provided. The simulator functionality is verified by using the theoretically derived results for the area coverage and the weak barrier coverage. Instead of using the fully mobile network, in order to make the price of the network more flexible, the hybrid networks are considered in this thesis. The proposed algorithms for improving the strong barrier coverage can be executed in centralized or distributed manner. To mend the barrier gaps, either mobile nodes (that fill the gaps and become part of the network just as same as the other stationary nodes) or mobile robots (that place the stationary nodes into the appropriate positions) can be used. Algorithms for moving the mobile devices in hybrid networks are also proposed. Bearing in mind some of the development trends in system and software engineering, related to improving the efficiency of the development process by using the appropriate methodologies and machine-based identification of the data structures, in this thesis, the description of the technical solution is given in the context of CMMI (Capability Maturity Model Integration) while the most important aspects of the automatic identification of the position and the sensing characteristics of the nodes are presented by using the methodologies of the conceptual modeling.
true
Modelovanje uticaja vremenskog profila solarnog zračenja na efekte rada fotonaponskih sistema u eletroenergetskom sistemu
Modeling the influence of the solar radiation time profile on the photovoltaic systems operation effets in a power system
Predmet istraživanja ove doktorske disertacije je modelovanje i analiza uticaja fotonaponskih sistema na elektroenergetski sistem. Naučni doprinosi doktorske disertacije su: • Razvijena je unapređena metodologija za valorizaciju električne energije proizvedene u fotonaponskim elektranama u uslovima slobodnog tržišta i na taj način je omogućena bolja procena ekonomičnosti rada fotonaponskih elektrana u uslovima slobodnog tržišta električne energije. Matematički model za procenu kvantifikatora za valorizaciju električne energije proizvedene u fotonaponskim panelima u uslovima slobodnog tržišta je dat kroz korelacionu analizu tipičnih dijagrama proizvodnje fotonaponskih sistema i dijagrama cene električne energije na tržištu električne energije. • Razvijen je i unapređen stohastički model ekvivalenta mreže i na taj način je omogućeno sagledavanje uticaja disperzovanih fotonaponskih sistema na gubitke aktivne snage u distributivnoj mreži. Ova metodologija omogućava procenu efekata rada krovno integrisanih fotonaponskih sistema na smanjenje gubitaka aktivne snage u distributivnoj mreži. • Razvijena je inovirana metodologija za optimizaciju prostorne orijentacije fotonaponskih panela u pogledu profita koji vlasnik fotonaponske elektrane stiče na slobodnom tržištu električne energije.
The subject of investigation of this doctor dissertation is modeling and analysis of the influence of photovoltaic systems upon a power system. The scientific contributions of the doctoral dissertation are: • An advanced methodology of the electrical energy valorization produced by photovoltaic plants under free market conditions has been developed thus obtaining a better estimate of the economy of operation of photovoltaic plants under free market conditions was enabled. Mathematical model for estimating the quantifiers of valorization of electrical energy produced by photovoltaic panels under free market conditions is presented through a correlation analysis of typical diagrams of photovoltaic plants production and diagrams of electrical energy prices at the electrical energy market. • An advanced stochastic model of the network equivalent has been developed thereby it became possible to perceive the influence of dispersed photovoltaic systems upon active power losses in a distribution network. This methodology permits estimation of the effects of operation of the building integrated photovoltaic systems on reduction of active power losses in a distribution network. • An innovative methodology for optimization of spatial orientation of photovoltaic panels with respect to the profit gained by the owner of a photovoltaic plant at the free market of electrical energy has been developed.
true
Plasma channel evolution in the triggered lightning discharges
Еволуција плазма канала код тригерованих атмосферских пражњена
Evolucija kanala je od bazičnog značaja za proučavanje fizike kanala munje, elektrodinamike atmosfere kao i za elektroinženjersku praksu. Cilj ove disertacije je razvoj detaljnih analitičkih i numeričkih algoritama za primenu u proučavanju evolucije kanala munje. Kao prvo, klasične numeričke metode izračunavanje trodimenzionalnih integrala su korišćene za izračunavanje aksijalnog električnog polja duž ose kanala munje. Rezultati merenja kod negativnih trigerovanih atmosferskih pražnjenja su korišćeni kao ulazni parametri tokom proračuna. U prorčunu su korišćeni različiti inženjerski modeli - TL, MTLE, MTLP, MTLE i GTCS model. Za izračunavanje profila podužne električne provodnosti u kanalu munje korišćeni su dobijeni rezultati za električno polje kao i Omov zakon u lokalnom obliku. Izvršeno je poređenje sa eksperimentalnim rezultatima. Pokazano je da su srednje vrednosti navedenih fizičkih veličina u skladu sa eksperimentalnim rezultatima. Na osnovu eksperimentalnih rezu- ltata i poluempirijskih formula, mogu se proceniti ostali bitni parametri kanala munje (koncentracija i temperatura nosilaca). Posle detaljnog pregleda svih metoda koji se koriste u literaturi za prou- čavanje dinamike kanala munje, odlučeno je da se koristi GTCS model. Za tu svrhu potrebno je precizno izračunati funkciju pražnjenja iz Volterine integralne jednačine prve vrste. Jednačina je rešena analitički, metodom Laplasove transformacije, kao i metodom konvolucionih kvadratura (CQ me- thod (engl. convolution quadrature method)). Jednačina je takođe rešena numeri- čkom metodom modifikovane kompozitne trapezne formule (MCTF method), koja je jedna od numeričkih metoda korišćena da obezbedi veoma visok stepen tačnosti uz minimalne aproksimacije. Dobijeni analitički rezultati su pokazali odlično slaganje sa numeričkom metodom. Dobijena funkcija pražnje- nja kanala se koristi za proračun drugih fizičkih parametara duž radijalnog i aksijalnog pravca. Na osnovu prethodno navedenih rezultata, date su smernice za dalje prouča- vanje dinamike kanala munje.
Plasma channel evolution is of fundamental interest for the study of the lightning physics, the electrodynamics of the atmosphere and for electrical engi- neering practice. The subject of this dissertation is the development of detailed and efficient numerical algorithms for application in the study of the lightning channel evolution. First off all, the classical numerical methods for the calculating the three - dimensional integrals were used to calculate the axial electric field along the lightning channel axis. The results measurement of the negative triggered lightning discharge have been used as input parameters during calculation. Different engineering models were used in the calculation - the TL, the MTLE, the MTLP, the MTLE and the GTCS model. Obtained results for electric field as well as the point form of Ohm’s law have been used to calculate the profile of longitudinal electrical conductivity on the lightning channel. The comparison with the experimental results was performed. It has been shown that the mean values of the mentioned physical quantities are in accordance with the experimental results. On the basis of the experimental results and semi-empirical formulas, other essential parameters of the lightning channel (the concentration and the temperature of the carriers) can be estimated. After a detailed overview of all methods used in the literature to study the lightning channel dynamics, the decision was to use the GTCS model. For this purpose, it is necessary to precisely calculate the channel discharge function from Volterra integral equation of the first kind. This equation is solved analytically by Laplace transformation, as well as by the convolution quadrature method (CQ method). The Volterra equation is also solved by the modified composite trapezoidal formula method (MCTF method), which is one of the numerical methods used to provide a very high degree of accuracy with minimal approximations. Results showed excellent agreement with the analytical method. Obtained channel discharge function is used in the calculations of other physical parameters along radial and axial directions of the channel. Based on the afore stated results, the guidelines for further study of the lightning channel dynamics are given.
true
Algoritmi za distantne zaštite na bazi digitalne fazne komparacije
Distance relay algorithms based on digital phase comparator
U disertaciji su prezentovani novi algoritmi distantnih zaštita bazirani na faznoj komparaciji u vremenskom domenu. Predstavljeni algoritmi su poređeni sa algoritmima na bazi diskretne Furijeove transformacije (DFT) koji se najčešće koriste u praksi i predstavljaju klasično rješenje u ovoj oblasti. Osnovna razlika između naših i klasičnih rješenja je u tome što ne računamo fazore, nego signale ostavljamo u vremenskom domenu gdje uz mali broj računskih operacija određujemo da li je kvar lociran unutar podešene zone reagovanja. Osnovna prednost ovog pristupa je u tome što značajno redukuje broj računskih operacija koje algoritam izvršava u svakom periodu odabiranja. Zavisno od oblika karakteristike reagovanja, predloženi algoritmi zahtijevaju između 30 i 70% manje računskih operacija u odnosu na klasično rješenje bazirano na DFT. Predloženi pristup ima dvije potencijalno slabe tačke. Prva je da istovremeno prisustvo viših harmonika istog reda u ulaznim signalima može uzrokovati greške u detekciji kvara. Detaljnom analizom i testiranjem smo pokazali da je i pri uslovima koji su mnogo strožiji od realnih uslova u mreži, pomenuta greška neznatna, tako da praktično nema uticaja na tačnost detekcije kvara. Drugi potencijalni nedostatak predloženih algoritama je neophodnost unošenja dodatnih vremenskih kašnjenja u signale kako bi bile realizovane željene karakteristike reagovanja. To bi teoretski trebalo dovesti do toga da predloženi algoritmi budu u svakoj situaciji sporiji od kontrolnih. Testiranja pokazuju da to nije tačno i da su neki predloženi algoritmi u prosjeku čak i brži od klasičnih. Specifičnost fazne komparacije u vremenskom domenu je i u tome što nije moguće direktno odrediti frekventnu karakteristiku algoritma, koja je značajan pokazatelj pouzdanosti i sigurnosti prilikom detekcije kvara. Zbog toga smo predložili indirektan način za procjenu frekventne karakteristike i uradili poređenje sa frekventnom karakteristikom algoritama na bazi DFT. Pokazali smo da su frekventne karakteristike praktično identične. Algoritmi sadrže filter koji značajno redukuje opadajuću jednosmjernu komponentu, kao i prelazne procese kod kapacitivno spregnutih naponskih transformatora. Algoritmi su testirani na klasičan način, determinističkim pristupom kroz promjenu pojedinih promjenljivih u simulaciji u širokom opsegu. Time su dobijeni pokazatelji brzine, pouzdanosti i sigurnosti, ali nije omogućena njihova stvarna procjena niti poređenje algoritama. To je i glavni nedostatak ovog načina testiranja koji je prisutan u većini radova. Zbog toga smo formirali stohastički model mreže baziran na stvarnim funkcijama raspodjele najvažnijih promjenljivih u sistemu. Gustine raspodjela promjenljivih su određene na osnovu prikupljenih podataka iz prenosnih mreža Bosne i Hercegovine i Srbije. Na osnovu više hiljada simulacija kvarova i reagovanja algoritama distantnih zaštita, određeni su pouzdanost, sigurnost i raspodjela vremena reagovanja svakog algoritma. Time je omogućeno mnogo objektivnije poređenje algoritama. Pokazalo se da u prosjeku najbrže reaguje i ima najbolje opšte rezultate jedan od predloženih algoritama. Prikupljeni podaci iz prenosnih mreža, odnosno formirani stohastički model za procjenu pouzdanosti i sigurnosti, mogu biti korisni i drugim autorima prilikom testiranja algoritama raznih tipova zaštita prenosnih vodova.
In this doctoral dissertation new distance relay algorithms, based on the time-domain phase comparison principle, are presented. The algorithms are compared to the algorithms based on discrete Fourier transformation (DFT), which are the most commonly used in practice and considered as a classical solution in this area. The main difference between the proposed and the classical solution is that we do not calculate phasors. Instead we use signals in the time-domain, where using a smaller number of mathematical operations we can detect the fault. The main advantage of this approach is a significant reduction of mathematical operation that the algorithm executes in each sampling period. Depending on the shape of the operation characteristic, our algorithms demand 30 to 70% fewer mathematical operations in comparison to a classical solution based on DFT. The proposed approach potentially has two weak spots. The first one is that simultaneous presence of the same order higher harmonics in input signals can cause an error in fault detection. We showed, through detailed analysis and testing, that even in harder conditions than in a real power system, the error is insignificant, so practically there is no impact on the fault detection. The second problem is the necessity of adding an additional time delay in input signals, in order to get a desired operating characteristic. That should result in the proposed algorithms being slower than the algorithms based on DFT, but tests show that this is not true. Some of our algorithms are, on average, even faster than the classical ones. A specific property of the time-domain phase comparison is that it is not possible to determine the frequency response, which is a significant indicator of security and dependability. We proposed an indirect way to evaluate the frequency response and compare it with the frequency response of DFT. We proved that the frequency responses are practically identical. The algorithms include a filter which mitigates decaying DC-offset in the current as well as the capacitive coupled voltage transformer transients. Algorithms are tested in a classical way, through changing some variables in simulation over a wide range. That provides indicators of speed, dependability and security, but it is not enough for their real assessment nor for comparison of the algorithms. That is the main weakness of this approach, which is common in most papers. For that reason we made a stochastic power system model based on probability density functions of the main variables in a power system. In order to get necessary probability density functions, we were collecting data from transmission systems in Bosnia and Herzegovina, as well as Serbia. We carried out a few thousand fault simulations and determined dependability, security and speed for all tested algorithms. That enables a more objective comparison of the algorithms. The final result is that the highest speed, and the highest mark has one of the proposed algorithms. The data we collected, and stochastic model for dependability and security assessment, could be useful for other authors to test different kinds of transmission line protection algorithms.
true