Datasets:

jerteh

/

PaSaz

like 0

Metod detekcije objekata sa malom refleksijom i velikom fluktuacijom na radarskoj slici

The method of image processing for detection of targets with small reflection and big fluctuation

Postoje brojne metode za detekciju objekata koji liˇce na taˇcke u bitmapiranimslikama sa nijansama sive boje. Problem definisanja praga prihvatanja ili odbijanja rezultata detekcije je najˇceš´ce zanemaren ili je prepušten ekspertima. Detekcija malih objekata sa malom radarskom refleksnom površinom i sa velikom fluktuacijomreflektovanog signala pomo´cu radarskog senzora predstavlja izazovan inženjerski problem. Takvi objekti su problematiˇcni za detekciju i lociranje ako se koriste klasiˇcne metode obrade radarske slike. Tokomposlednjih petnaest godina rad¯ena je serija eksperimenata sa raznim tipovima radarskih senzora i nekoliko tipova procesora signala. Pokušano procesiranje slike sa ciljem uspešne detekcije svih tipova objekata na moru. Kao posebna klasa problema identifikovan je problem detekcije malih objekata na morskoj površini - malih plovila. Na globalnom nivou problem detekcije i identifikacije objekata namoru donekle je rešen uvod¯enjemAutomatskog Identifikacionog Sistema - AIS. Nažalost veliki brojmalih plovnih objekata nije opremljen AIS primopredajnicima. Subjekti koji nameravaju da se bave sumnjivimaktivnostima namerno izbegavaju posedovanje AIS primopredajnika. Ratni brodovi nisu u obavezi da koriste sistem AIS. Prikaz originalnog pristupa u pokušaju da se reši problem detekcije malih objekata iznet je u ovoj disertaciji. Pored uˇcesnika u saobra´caju na moru klasa pomorskih navigacionih radara detektuje i elemente reljefa obale. ˇCesto delovi obale liˇce na male pokretne objekte i operateri na sistemima su u nedoumici kako da klasifikuju uoˇceni odraz. U ovom radu izneta je ideja kako da se osmisle prostorni filteri kojima se vrši odbacivanje neinteresantnih detekcija koje potiˇcu od kopna. Postoje brojne metode za detekciju objekata koji liˇce na taˇcke u bitmapiranimslikama sa nijansama sive boje. Problem definisanja praga prihvatanja ili odbijanja rezultata detekcije je najˇceš´ce zanemaren ili je prepušten ekspertima. Detekcija malih objekata sa malom radarskom refleksnom površinom i sa velikom fluktuacijomreflektovanog signala pomo´cu radarskog senzora predstavlja izazovan inženjerski problem. Takvi objekti su problematiˇcni za detekciju i lociranje ako se koriste klasiˇcne metode obrade radarske slike. Tokomposlednjih petnaest godina rad¯ena je serija eksperimenata sa raznim tipovima radarskih senzora i nekoliko tipova procesora signala. Pokušano procesiranje slike sa ciljem uspešne detekcije svih tipova objekata na moru. Kao posebna klasa problema identifikovan je problem detekcije malih objekata na morskoj površini - malih plovila. Na globalnom nivou problem detekcije i identifikacije objekata namoru donekle je rešen uvod¯enjemAutomatskog Identifikacionog Sistema - AIS. Nažalost veliki brojmalih plovnih objekata nije opremljen AIS primopredajnicima. Subjekti koji nameravaju da se bave sumnjivimaktivnostima namerno izbegavaju posedovanje AIS primopredajnika. Ratni brodovi nisu u obavezi da koriste sistem AIS. Prikaz originalnog pristupa u pokušaju da se reši problem detekcije malih objekata iznet je u ovoj disertaciji. Pored uˇcesnika u saobra´caju na moru klasa pomorskih navigacionih radara detektuje i elemente reljefa obale. ˇCesto delovi obale liˇce na male pokretne objekte i operateri na sistemima su u nedoumici kako da klasifikuju uoˇceni odraz. U ovom radu izneta je ideja kako da se osmisle prostorni filteri kojima se vrši odbacivanje neinteresantnih detekcija koje potiˇcu od kopna. Poslednji deo ukratko prikazuje neke od implementacija u kojima je koriš´cen predloženi model kao i predlog objektnog modela za implementaciju.

There are numerous methods of detection of objects seen like points in 8 bit gray leveled bitmap image. The problem of defining the threshold for accepting or rejecting the results of detection is often neglected or left to the experts. The localization and detection of small objects at sea having small radar reflective surfaces, and with large fluctuations in the reflected signal, which is usual when using a radar sensor, represent a challenging engineering problem. Such objects are difficult to detect and locate when using classical methods of processing radar images. During the last fifteen years, series of experiments with various types of radar sensors and several types of signal processors have been made. The aimof the experiments was to performsuccessful detection of all types of vessels at sea. As a special class of identified problems, there is the problem of detection of small objects on the sea surface - small boats. At the global level, the problem of detection and identification of objects at sea is solved by introducing the Automatic Identification System - AIS. Unfortunately, a large number of small vessels are not equipped with AIS transponders. Entities intending to deal with some kind of illegal activity deliberately avoid introduction of AIS transceivers. Warships are not required to use AIS. This dissertation offers an original approach to the problem of detection of small objects at sea. An attempt to solve the problem of detection of small objects is presented in this dissertation. In addition to the participants in traffic at sea, the class of marine navigation radar detects elements of the coast. Often parts of the coast look like small objects, thus creating a problemto systems’ operators with classification of the observed reflection. This paper summarizes the idea of devising new spatial filters that are used for discarding uninteresting detection reflected from the mainland. The last section summarizes some of the implementations in which the proposed model is used, as well as a proposal for implementation of the object oriented programing model.

Kondenzatori vrlo velike kapacitivnosti na bazi sulfida bakra

Supercapacitors based on copper sulfides

Ova disertacija predstavlja doprinos razvoju superkondenzatora novog tipa zasnovanog na prirodnim sulfidima bakra. Ispitivanje superkondenzatora (zbog velikih kapacitivnosti, odnosno vremenskih konstanti) zahteva znatnu modifikaciju i prilagođavanje klasičnih elektrohemijskih metoda i instrumentalne tehnike. U okviru disertacije urađena je matematička analiza kojom su definisani parametri eksperimentalnih metoda, kako sedam modifikovanih standardnih, tako i dve nove – Dirakove naponske i nagibne strujne. Hardver i softver nephodan za njihovu realizaciju predstavlja značajan deo ove disertacije. Metode su najpre proverene na posebno konstruisanom fizičkom modelu – električnom kolu sa komercijalnim superkondenzatorom od 1,5F, kao i simulacijom na računaru. Sve metode su primenjene pri razvoju novog tipa superkondenzatora na bazi prirodnih minerala bakra – kovelina (CuS) i halkozina (Cu2S). Maksimalna dobijena kapacitivnost na halkozinu bila je oko 100F/cm2. Na kraju je realizovan i ispitan protip superkondenzatora sa halkozinom kao aktivnim elektrodnim materijalom.

In this dissertation contribution is given in the development of a new kind of supercapacitors based on copper sulfide minerals. Examination of supercapacitors needs considerable modification of standard electrochemical methods and techniques because of the big time constants as a consequence of enormous high capacitancies. Parameters for modified electrochemical methods were obtained by mathematical analysis of such systems. Also, two new methods have been established – Dirack voltage pulse and linear sweep amperometry. Hardware and software for seven modified standard and two newly developed methods are an important part of the dissertation. The methods are firstly verified on specially constructed physical model of electrical circuit containing one 1,5F supercapacitor, and by simulation on PC. All methods and instrumental techniques were applied to develop a new kind of supercapacitors based on natural copper minerals covellite (CuS) and chalcocite (Cu2S). Maximum capacitance obtained on chalcocite was about 100 F/cm2. Finally, a prototype of supercapacitor with chalcocite as an active electrode material was realized.

Srbija, Bugarska i Tatari u drugoj polovini XIII veka

Serbia, Bulgaria and Tatars in the second half of the thirteenth century

Doktorska disertacija Srbija, Bugarska i Tatari u drugoj polovini XIII veka paralelno posmatra dva niza istoriografskih pitanja. Prvi se odnosi na srpsko-bugarske odnose tokom pomenute epohe, a drugi na mongolski (tatarski) uticaj na jugoistočnu Evropu. Ove dve teme stoje u neraskidivoj vezi: prvu je nemoguće obuhvatiti bez druge, a druga je osvetljena iz više perspektiva zahvaljujući prvoj. Gornju hronološku granicu istraživanja predstavlja prodor Mongola na Zapad (1241-1242), koji je ostavio duboke tragove razaranja i izazvao niz migratornih kretanja. Donju čine unutrašnji sukobi epskih proporcija među Tatarima u crnomorskim stepama, koji su doveli do promene karaktera njihovog uticaja, kao i do novog masovnog naseljavanja nomadskih populacija južno od Dunava. Ova epoha je u balkanskim okvirima započela smrću bugarskog vladara Ivana Asena II (1218-1241), a završila se pokušajem stvaranja personalne unije dveju država Južnih Slovena i sporazumom kralja Milutina sa Vizantijom 1299. godine, čime je najavljena hegemonija države Nemanjića u regionu. Teza je zasnovana na širokom spektru narativnih (vizantijski, latinski, orijentalni, slovenski i drugi pisci), dokumentarnih (povelje papa, ugarskih, srpskih i bugarskih vladara, građa iz anžujskih arhiva) i arheoloških izvora. Okosnicu istraživanja čine politički aspekti, ali su oni neraskidivo isprepletani sa vojnim, geografskim i društvenim. Teza je strukturalno podeljena na Uvodne primedbe, šest poglavlja i Zaključna razmatranja. U Uvodnim primedbama dat je pregled izvora i literature, osvrt na geografiju regiona i terminološke napomene. U prvom poglavlju je predstavljen tok mongolske invazije i njene posledice (1241-1242). Drugo se bavi obrazovanjem mongolske države u Crnomorskim stepama (tzv. Zlatna horda) i njenim uticajem na balkanske zemlje sredinom XIII veka. Sledeća celina je posvećena analizi srpsko-bugarskih odnosa, kao i unutrašnjim i spoljnim činiocima koji su oblikovali njihov međunarodni položaj. Četvrto poglavlje je okrenuto usponu i aktivnostima mongolskog velikaša Nogaja, a peto se bavi njegovim uticajem na prilike u Bugarskoj i Srbiji. U poslednjem poglavlju je analizirano nasleđe Nogajeve epohe u jugoistočnoj Evropi, kroz emigraciju nomadskih Alana i Tatara, bliske odnose srpske države i severozapadnih bugarskih zemalja, kao i zajedničko srpsko-bugarsko-tatarsko delovanje protiv Ugarske početkom XIV veka. Sprovedena istraživanja rešavaju više hronoloških nepoznanica (vreme ustoličenja Nogaja na donjem Dunavu, datovanje braka kralja Milutina i ugarske princeze Jelisavete, pripajanje Braničeva državi kralja Dragutina, itd.) i otkrivaju neosnovanost nekih ustaljenih hipoteza u nauci (npr. talaštvo Teodora Svetoslava u Nogajevim zemljama). Osim toga, istraživanja pokazuju da su do tada skladni bugarsko-srpski odnosi 1284. godine prerasli u faktički vojni i politički savez, zatim da je Srbija deceniju kasnije na kratko postala u punom smislu vazal tatarske državne tvorevine obrazovane na donjem Dunavu, kao i da su bliske veze između Srbije i severozapadnih bugarskih zemalja, obrazovane krajem XIII veka pod Nogajevim pokroviteljstvom, nadživele ovog mongolskog velikaša pune dve decenije.

Doctoral Dissertation Serbia, Bulgaria and Tatars in the Second Half of the Thirteenth Century deals with two parallel historiographical issues. The former is Serbian-Bulgarian relations during this epoch; the latter is Mongol (Tatar) impact on Southeastern Europe in the proposed timeframe. These two topics are mutually inseparable. The first is impossible to explain without taking the latter into account, while the second is, in turn, illuminated from multiple perspectives by the first. Chronologically, the research starts with the Mongol Invasion of Europe (1241- 1242), which left deep traces of destruction and caused a series of migrations. It finishes with inner conflicts among the Tatars in the Black Sea steppes, which brought changes in the character of their influence and led to a new mass settlement of nomadic populations south of the Danube. In the Balkan frameworks, this epoch began with the death of Bulgarian ruler John Asen II (1218-1241); it ended with an attempt to create a personal union between the two South Slavic states and with king Milutin’s agreement with Byzantium in 1299, thereby announcing the hegemony of the Nemanjić state in the region. The dissertation is based upon a wide range of narrative (Byzantine, Latin, Oriental, Slavic and other writers), documentary (сharters of Popes, Hungarian, Serbian and Bulgarian rulers, material from the Angevin archives), and archaeological sources. Main focus is devoted to political aspects; they are, however, closely intertwined with the military, social and geographical aspects. Structurally, the dissertation is divided into Introductory remarks, six chapters and Conclusion. Introductory remarks provide an overview of the sources and literature, review of the regional geography and terminology used in the text. In the first chapter, the course of events during Mongol invasion of Europe (1241-1242) is presented, as well as its consequences. The second section deals with the establishment of Mongol state in the Black Sea steppes (so-called Golden Horde), and its influence on the Balkan countries in mid Thirteenth century. The next chapter is devoted to the analysis of the relations between Serbia and Bulgaria, internal and external factors that shaped their position. The fourth part is dedicated to the rise of Nogai and his career, while the next one is focused upon his impact оn Bulgaria and Serbia. The final chapter deals with his legacy, envisaged through emigration of nomadic Alans and Tatars, close relations between Serbian state and northwestern Bulgarian lands, as well as joint Serbian, Bulgarian and Tatar military activities against Hungary, at the beginning of the Fourteenth Century. The conducted researches solve several chronological issues (such as the time of Nogai’s establishment on the Lower Danube, date of marriage between king Milutin and Hungarian princess Elizabeth, inclusion of Braničevo into the state of king Dragutin, etc.). They reveal groundlesness of some of the established hypotheses in historiography (e.g. the hostage status of Bulgarian Prince Theodore Svetoslav in the Nogai’s lands). Also, it has been shown that: harmonious Bulgarian-Serbian relations grew into a de facto military and political alliance in 1284; a decade later, Serbia became a dependant of the Tatar state formed on the Lower Danube, although for a short time; the ties between Serbia and northwestern Bulgarian lands, formed in the late Thirteenth century under Nogai’s patronage, outlived him for another two decades.

Iterativno rešavanje integralnih jednačina elektromagnetskog polja korišćenjem metoda fizičke optike

Iterative solution of electromagnetic field integral equations using physical optics method

Numerički egzaktno rešavanje integralnih jednačina elektromagnetskog polja moguće je metodom momenata (MoM). Kada se primeni na električno velike probleme (npr. letelice), MoM rezultuje velikim sistemima linearnih jednačina, čije rešavanje predstavlja usko grlo metode. Problem je dvojak: računarska memorija potrebna za smeštanje matrice sistema je velika (srazmerna kvadratu broju jednačina), a računarsko vreme rešavanja dugo (srazmerno trećem stepenu broja jednačina, u slučaju rešavanja LU dekompozicijom). Jedan način (delimičnog) prevazilaženja ovog problema jeste paralelizacija rešavanja sistema linearnih jednačina. Korisnicima običnih („personalnih“) računara na raspolaganju su rešenja zasnovana na korišćenju višejezgarnih centralnih procesora ili grafičkih procesora. Na ovaj način može se značajno skratiti vreme rešavanja sistema linearnih jednačina, ali ne i memorijsko zauzeće. U svakom slučaju, primena MoM-a na personalnim računarima ima jasne granice. Razvijene su i brojne aproksimativne metode, između ostalog zasnovane i na MoM-u, za rešavanje električno velikih problema. Svaka od njih ima svoje prednosti i nedostatke, ali se ni za jednu ne može reći da je univerzalna u pogledu rešavanja proizvoljnih problema. Cilj našeg istraživanja bio je razvijanje nove aproksimativne metode, zasnovane na MoM-u, za rešavanje integralnih jednačina elektromagnetskog polja. Nova metoda trebalo je da, na personalnom računaru, omogući rešavanje problema za jedan red veće električne veličine od onih koji se mogu rešiti primenom MoM-a. Nova metoda realizovana je kao iterativni postupak u kojem se pravac korekcije određuje modifikovanom PO tehnikom, a rezultujući sistem linearnih jednačina je, za električno velike probleme, za više redova veličina manji od odgovarajućeg MoM sistema. Numerički eksperimenti su pokazali da je vreme potrebno za dobijanje rešenja (zadovoljavajuće tačnosti) značajno kraće od onog potrebnog za dobijanje MoM rešenja, dok je memorijsko zauzeće zanemarljivo u odnosu na ono kod primene MoM-a. Električna veličina problema koji se mogu analizirati personalnim računarom povećana je za najmanje red veličine u odnosu na MoM.

Numerically exact solution of electromagnetic field integral equations is possible using Method of Moments (MoM). When applied to electrically large problems (e.g. airplanes), MoM results in large system of linear equations. Solving it is a bottleneck of the method, for two reasons: occupied computer memory is large (proportional to number of equations squared) and computer solution time is long (proportional to third power of number of equations, when using LU decomposition solver). A way to (partially) circumvent this problem is parallelization of solution process. Personal computer (PC) users have at their disposal software solutions based on multi-core CPUs and GPUs, which are able to significantly decrease solution time, but memory occupation stays intact. Anyway, application of MoM on PCs has clear limitations. Different approximate methods, many of them based on MoM, are developed for treating electrically large problems. Each of them has its own advantages and shortcomings, but neither is universal in respect of treating an arbitrary electromagnetic problem. The goal of our research is development of new approximate method, based on MoM, for solution of electromagnetic field integral equations. The new method should enable solution of the order of magnitude electrically larger problems than MoM. The new method is formulated as iterative procedure in which the direction of correction is determined using modified PO technique. For electrically large problems the resulting system of linear equations is much smaller than the corresponding MoM system. Numerical experiments show that (accurate enough) results can be obtained in much shorter time than with MoM, whereas memory occupancy can be neglected when compared to those of MoM. Electrical size of problems that can be analyzed using PC is increased by (at least) an order of magnitude.

Tunelovanje elektromagnetskih talasa kroz kompleksne optičke sredine

Tunneling of electromagnetic waves through complex optical media

Predmet ove disertacije je analiza karakteristika tunelovanja elektromagnetskih talasa kroz različite kompleksne optičke sredine, sa posebnim osvrtom na metamaterijale sa negativnim indeksom prelamanja. Teorijski su razmatrane različite strukture, te su za njih, korišdenjem brojnih matematičkih metoda, izvedeni izrazi za vremena tunelovanja u: (1) nelinearnim disperzivnim anizotropnim strukturama, (2) linearnim anizotropnim strukturama (poluprovodničkim metamaterijalima), a proračunat je i uticaj Gus-Henkenovog pomeraja na ova vremena. Vremena tunelovanja za različite, postojede, strukture su izračunate koridenjem numeričke metode pod nazivom metoda pogađanja (shooting method), koja je odgovarajuda za ovu problematiku. Takođe, teorijski je razmatrana propagacija elektromagnetnog zračenja u anizotropnim strukturama sastavljenim od naizmenično postavljenih slojeva dva različita poluprovodnika, te je, na osnovu ove analize, dizajniran poluprovodnički metamaterijal sastavljen od naizmenično postavljenih slojeva različito dopiranih InGaAs. Mogudnost negativnog prelamanja u ovim materijalima potvrđena je primenom FDTD (Finite-Difference Time-Domain) numeričke metode, koja omogudava simulaciju realnog prostiranja elektromagnetnog zračenja kroz različite strukture.

Subject of this dissertation is the analysis of electromagnetic waves tunneling properties through different complex optical media, with emphasis on metamaterials with negative index of refraction. Different structures are theoretically considered, i.e., using numerous mathematical methods, expressions for tunneling times are derived for: (1) nonlinear dispersive anisotropic structures, (2) linear anisotropic structures (semiconductor metamaterials). Also, influence of Goos- Hänchen shift on these times is calculated. Tunneling times for existing structures are calculated by using suitable numerical method: shooting method. Besides these, propagation of electromagnetic radiation through anisotropic structures composed of alernately placed layers of two different semiconductors is analyzed. On the grounds of this analysis, semiconductor metamaterial composed of alternately placed layers of differently doped InGaAs is designed. The possibility of obtaining negative refraction in these structures is confirmed by using FDTD (Finite-Difference-Time-Domain) numerical method, which enables simulation of electromagnetic radiation propagation through various structures.

Evaluacija uspeha učenika oštećenog sluha u redovnoj školi

Evaluation of success in children with hearing impairment in regular schools

Cilj istraživanja je bio da se utvrdi školski uspeh gluvih učenika koji su uključeni u sistem redovnog školovanja, dok su posebni ciljevi bili: utvrditi školski uspeh gluvih učenika koji su uključeni u redovne razrede iz pojedinačnih nastavnih predmeta; utvrditi povezanost dužine ranog surdološkog tretmana i školskog uspeha gluvih učenika uključenih u redovne razrede; utvrditi oblike ponašanja gluvih učenika uključenih u redovne razrede; utvrditi povezanost ranog surdološkog tretmana i integracije gluvih učenika u redovne škole...

Aim of this research was to determine the academic achievement of students with hearing impairment who are involved in regular schooling system, while specific objectives were: to identify school achievement of students with hearing impairment included in regular classes from individual subjects; to determine correlation between early audiology treatment and school achievement of students with hearing impairment included in regular classes; to determine behaviors of students with hearing impairment included in regular classes; to determine the relationship between early audiology treatment and integration of students with hearing impairment into regular schools...

Pridruživanje podataka Singerovim memorijskim filtrom za praćenje pokretnih ciljeva u gustom klateru

Singer memory filter data association for moving target tracking in heavy clutter.

Predmet istraživanja ove disertacije je postupak pridruživanja podataka u procesu praćenja pokretnih ciljeva. U osnovi, poznato je da senzori obezbeđuju merenja, na osnovu kojih se može izvršiti procena (estimacija) parametara i stanja sistema. U većini slučajeva je tačno poznat izvor merenja tako da se procena vrši standardnim metodama (Kalman-Bucy filter). Asocijacija podataka je neophodna u situacijama kada nije poznat izvor merenja, kao što je osmatranje prostora radarom ili sonarom. Radar šalje impuls elektromagnetske energije i procesira primljene odjeke. Kada snaga primljenog signala pređe određeni prag, dolazi do detekcije. Detekcije se projektuju u osmatračke koordinate koje postaju ulazna merenja za blok praćenja ciljeva. Ta merenja mogu poticati od ciljeva, ali isto tako i od slučajnih objekata i fenomena. Izvor svakog merenja je nepoznat. Merenja koja potiču od ciljeva (”tačna” merenja) nisu pouzdana, jer su u svakom interval merenja prisutna samo sa određenom verovatnoćom detekcije. Neželjena merenja se obično zovu “klater”, ona se pojavljuju slučajno, i obično se modeliraju kao Poasonov proces opisan gustinom broja klater merenja u prostoru osmatranja. Broj ciljeva i njihovo postojanje su tipično ''a priori'' nepoznati, tako da je primarni zadatak algoritama asocijacije podataka određivanje postojanja i broja ciljeva u prostoru osmatranja. To podrazumeva primenu teorijskih metoda koje iz skupa pristiglih merenja u toku jednog vremenskog ciklusa okretanja antene izdvajaju one podskupove koje potiču od postojećih ciljeva ili koje potiču od nekih novih ciljeva. Ostale detekcije mogu se smatrati ''lažnim alarmom'' i posledica su uticja klatera i šumova u kanalu. Primena ovih metoda podrazumeva pripremne radnje, kao što je formiranje tzv. ''prozora praćenja'' odnosno sužavanje celokupnog prostora osmatranja na oblast od interesa, itd. Zbog a priori neznanja broja i položaja ciljeva, tragovi se inicijalizuju koristeći nepouzdana merenja. Pridruživanje podataka Singerovim memorijskim filtrom za praćenje ciljeva u gustom klateru ____________________________________________________________________________________________ Z. Radosavljević: Doktorska disertacija iv Pored tačnih tragova (koji prate ciljeve), inicijalizuju se i lažni tragovi (koji ne prate ciljeve). Neophodni deo algoritama za praćenje pokretnih ciljeva jeste diskriminacija lažnih tragova; tj. procedura koja treba da potvrdi broj tačnih tragova, i izbriše (većinu) lažnih tragova. Predložena metodologija praćenja pokretnih ciljeva (Integrated Track Splitting-ITS), predmet je intenzivnog istraživanja. ITS filter rekurzivno računa verovatnoću postojanja traga u svakom skenu, koja se koristi za diskriminaciju lažnih tragova. Jedan rekurzivni ciklus ITS algoritma sastoji se iz tri koraka. Prvi korak je propagacija stanja cilja. U ovom koraku, komponente traga, propagiraju kao Markov proces (korak predikcije Kalmanovog filtra). Sledi korak selekcije merenja komponentama traga iz prtethodnog skena. Svaka propagirana komponenta traga selektuje merenja pristigla u tekućem skenu. Na kraju, korak estimacija stanja traga gde se pdf estimacije stanja traga dobija se kao kombinacija međusobno isključivih komponenti estimacije stanja pdf-a, koja je uslovljena postojanjem cilja i istorijom pripadnosti komponente merenjima pristiglim u N prethodnih skenova. Ha taj način, svako novo merenje, upareno sa komponentom iz prethodnog skena, daje novu komponentu. Istorijat nastanka komponenti traga utiče na tačnost praćenja, naročito pri niskim verovatnoćama detekcije cilja. Nedostatak teorijskog i praktičnog poređenja između različitih dubina istorije komponenti traga, jedan je od problema koji se javljaju u praktičnim primenama praćenja pokretnih ciljeva. Drugi problem realizacije je metodologija ''iskustva'' kod određivanja parametara filtra. S druge strane treba smanjiti numeričku kompleksnost. Kompleksnost optimalna implementacija ITS algoritama raste eksponencijalno sa vremenom, pa je potrebno izabrati pod-optimalnu implementaciju. Sistematska teorijska obrada i praktična uputstva primene ITS filtera nisu objavljeni. Praćenje u okruženju klatera inicira stvarne (koji prate cilj) i lažne (koji ne prate cilj) tragove. Odbacivanje lažnih tragova (FTD-False Track Discrimination) je veoma važna osobina sistema za praćenje. ITS filter je efikasna metoda za automatsko praćenje ciljeva u okruženju klatera. Praćenje traga u ITS metodi takođe uključuje estimaciju trajektorije cilja u obliku skupa međusobno Pridruživanje podataka Singerovim memorijskim filtrom za praćenje ciljeva u gustom klateru ____________________________________________________________________________________________ Z. Radosavljević: Doktorska disertacija v isključivih komponenti (Gausiana). Broj komponente traga raste eksponencijalno u svakom skenu kao rezultat njihove interakcije sa pristiglim merenjima (merenja potiču od cilja ili šuma). Kontrola broja komponenti se ostvaruje 'otsecanjem ' slabih komponenti (prune) ili 'sjedinjavanjem' komponenti sa bliskim trajektorijama cilja (merge). Odnos između FTD i kontrole komonenata je deo istraživanja koje treba da rezultuje praktičnim uputstvom za korišćenje ITS filtra. U disertaciji biće date teorijske i praktične osnove ITS algoritama pri korišćenju Singerove metode dubine istorije merenja traga. Biće dat teorijski prikaz traga kao unije komponenti. Estimacija stanja svake komponente je izlaz filtra koji daje jedno merenje u svakom skenu. Reprezent svake komponente je istorija pripadnosti merenja cilju, iskazana preko odgovarajuće verovatnoće. Zato što su one međusobno isključive, zbir verovatnoća svih komponenata traga jednak je jedinici. Svako stanje komponente se sastoji od verovatnoće postojanja komponente i funkcije gustine verovatnoće stanja komponente pdf, pod uslovom da komponenta postoji. Standardnm pretpostavkama praćenja cilja senzorom konačne rezolucije (merenje potiče od jednog izvora) i tačkastog cilja (cilj može da kreira jedno merenje po skenu sa verovatnoćom detekcije). U svakom skenu, senzor daje određeni slučajni broj klater merenja, definisan gustinom klatera. Tragovi se inicijalizuju u svakom skenu korišćenjem merenja od dva konsekventna skena u skladu sa zahtevom maksimalne brzine cilja. Formirani ITS trag se rekurivno ažurira u svakom skenu. Verovatnoća postojanja traga se koristi za potvrđivanje ili gašenje svakog traga nastalog u susednim skenovima. Uticaj klatera dovodi do određene greske pri potvrđivanju tragova. Neki od potvrđenih tragova mogu biti i lažni. Rezultat FTD procesa je potvrđivanje ili gašenje nekih tragova. Funkcija gustine verovatnoće trajektorije stanja traga (pdf) aproksimira se određenim brojem komponenti traga (Gausiana). Ova aproksimacija dospeva u kontrolu komponenata traga. Poboljšanje aproksimacije trajektorije rezultuje poboljšanjem FTD performansi kao i povećanjem broja 'zadržanih' komponenti sa povećanjem računarskih zahteva. Kompromis između FTD i raspoloživih Pridruživanje podataka Singerovim memorijskim filtrom za praćenje ciljeva u gustom klateru ____________________________________________________________________________________________ Z. Radosavljević: Doktorska disertacija vi računarskih resursa je dodatni cilj disertacije. Teorijsko i praktično poređenje različitih dubina istorije ITS algoritama je neophodno da bi se utvrdilo njihovo nastajanje i uzajmna ograničenja. Disertacija definiše probleme koje treba da reši asocijacija podataka, dok se u nastavku unapređuje ITS predlaganjem optimalna metode asocijacije podataka podoptimalnom implementacijom.

The subject of the research topic presented in this doctoral dissertation is data association of moving target tracking. Basically, it is known that sensors provide the measurement, which results can effectively be used for evaluation (estimation) of the parameters and state of the system. In most cases, the exact source of measurement is known, thus estimation is performed by using standard methods (Kalman-Bucy filter). Data association is necessary in scenarios where source of measurements is not known, such as the observation of radar or sonar. Radar sends a pulse of electromagnetic energy and processes the received echoes. When the power of the received signal exceeds a certain threshold, there is a detection. Detections are projected in the observation coordinates which then become input measurements for the target tracking system. This measurement may derive from the targets, but also from random objects and phenomena. Source of each measurement is unknown. Measurements derived from targets ("true" measurements) are not reliable, because in each measurement interval they are present only with a certain probability of detection. Adverse measurements are commonly called "clutter", they appear randomly, and are usually modeled as a Poisson process described Clutter number density measurements in space observation. Number of targets and their existence are typically ‘a priori’ unknown, thus primary task of data association algorithms is to determine the existence and the number of targets in the space of observation. This includes the application of theoretical methods which, from a set of measurements received during a cycle time of turning the antenna, allocate subsets that derive from existing targets or from some new targets. Other detection can be considered as a 'false' alarm and is consequence of clutter and noise influence in the channel. Application of these methods involves preparatory activities, such as the formation of the so-called 'monitoring window ' or narrowing of the entire Придруживање података Сингеровим меморијским филтром за праћење циљева у густом клатеру ____________________________________________________________________________________________ З. Радосављевић: Докторска дисертација viii observation space on area of interest only, and so on. Because the number and positions of targets are not a priori known, tracks are initialized using unreliable measurements. In addition to accurate tracks (which follow the target), false tracks are also initialized (which do not follow targets). A necessary part of moving targets tracking algorithms is discrimination of false tracks; procedure which should confirm (most of) accurate tracks and delete (most of) false tracks. The proposed methodology for moving targets tracking -Integrated Track Splitting (ITS) is the subject of intensive research. ITS filter recursively calculates probability of the existence for each track in each scan, and this is used to discriminate false tracks. A recursive cycle ITS algorithm consists of three steps. The first step is the propagation of the target state. In this step, tracks components, are propagated as a Markov process (step prediction of Kalman filter). Next step is selection of track components measurements from previous scan. Each component tracks from previous scan, selects the number of measurements in the current scan. Finally, the state estimation step follow, where tracks pdf estimation is obtained as a combination of mutually exclusive components of state estimation pdf's, which is conditioned by the existence of target history and belonging component measurements arrived in N previous scans. Thus, each new measurement, paired with the component from the previous scan, provides a new component. History of the tracks components affects the accuracy of tracking, especially at low probabilities of targets detection. The lack of theoretical and practical comparisons between various depth history of tracks components, is one of the problems that arise in practical applications of moving targets tracking. Another problem is the implementation of the 'experience' methodology in determining the parameters of the filter. On the other hand, numerical complexity should be reduced. The complexity of optimal implementation of ITS algorithms grows exponentially with time, so it is necessary to select a sub-optimal implementation. The theoretical analysis and practical application instructions for ITS filters were not disclosed yet. Tracking the surrounding clutter initiate true (which follow the target) and false (which do not follow the target) tracks. The rejection of the false track (FTD-False Track Discrimination) is a very important feature of targets tracking systems. ITS filter is an efficient method for automatic targets tracking in clutter environment. Tracks control Придруживање података Сингеровим меморијским филтром за праћење циљева у густом клатеру ____________________________________________________________________________________________ З. Радосављевић: Докторска дисертација ix in the ITS method also includes the estimation of target trajectories in the form of a set of mutually exclusive components (Gaussian). Number of tracks components grows exponentially in each scan as a result of their interaction with the incoming measurements (measurements originate from a target or clutter). Control of the number of components is achieved by excision of weak components (prune) or conjoining component with close target trajectories (merge). The relationship between FTD and control of components is part of the research, which should result in practical guidelines for the use of ITS filter. The dissertation provides theoretical and practical foundation in the use of ITS algorithms with Singer method of depth of history. Also, dissertation provides theoretical view of track as Union of components. Estimation of each component is the output of the filter that gives one measurement at each scan. Representative of each component is the history of belonging of the measured track to target, expressed through appropriate probabilities. Because they are mutually exclusive, the sum of the probability of all track components is equal ‘one’. Each state of component consists of the probability of existence and probability density functions of component state estimation pdf, provided that the component exists. The usual assumptions of target tracking is a final resolution sensor (measurement comes from a single source) and the ‘point’ target (target can create one measurement per scan with a probability of detection). In each scan, the sensor provides a certain number of random measurements clutter, defined by its density. Tracks are initialized in each scan using measurements from the two consequential scans, in accordance with the requirement of ‘maximum target velocity’. Each ITS track recursively updates every scan. The probability of the track existence is used to confirm or decline all tracks generated in the consequently scans. The effect of clutter leads to some errors when acknowledging the tracks. Some of the confirmed tracks can be false. The result of the FTD process is to ‘confirm’ оr ‘terminate’ some tracks. Track trajectory pdf is approximated by a number of tracks components (Gaussian). This approximation is due to the control of tracks components. Improving approximation trajectories results with improvement of FTD performance and increases the number of 'retained' components with increasing computational demands. Придруживање података Сингеровим меморијским филтром за праћење циљева у густом клатеру ____________________________________________________________________________________________ З. Радосављевић: Докторска дисертација x Compromise between FTD and available computing resources is an additional aim of the dissertation. Theoretical and practical comparison of different depths of history ITS algorithms it is necessary to determine their formation and contract, the mutual constraints. The dissertation defines the problems that should be addressed by the data association, while below improves ITS proposing optimal methods of sub-optimal implementation.

Uticaj akustičkog i biološkog šuma na kvalitet merenja signala otoakustičke emisije

The influence of acoustic and biological noise on signal quality measurements of otoacoustic emissions

Biološka osnova nastanka signala otoakustičke emisije (OAE), akustička priroda njegovog nastanka i izuzetna maskiranost raznim oblicima smetnji jesu polazna osnova razvoja raznih metoda detekcije, obrade i interpretacije signala OAE, odnosno samog efekta otoakustičke emisije u svetlu karakterizacije funkcionisanja kohlee. Metode merenja OAE se razlikuju po načinu generisanja stimulusa, mehanizmima generisanja odziva kao i po načinima obrade signala i statističkog tumačenja dobijenih rezultata. Postojeće metode merenja i obrade signala OAE jesu osnova savremenih sistema za merenje OAE, koji su našli široku primenu u kliničkoj praksi. Međutim, nisu iscrpljene sve mogućnosti, pre svega teorijske a zatim i metodološke prirode, prvenstveno u domenu ekstrakcije signala OAE iz šuma. Poboljšanjem uslova merenja smanjuje se uticaj spoljašnje buke (šumova) na mogućnost ekstrakcije signala otoakustičke emisije i biološki šum postaje dominantan izvor smetnji. Istraživanje karakteristika bioloških šumova je jedan od preduslova za uspešno rešavanje problema detekcije OAE u prisustvu biološkog šuma. Na pažljivo odabranom uzorku, u skladu sa potrebama istraživanja, analizirani su biološki šumovi snimljeni u ušnom kanalu. Pokazalo se da postoji značajna korelacija bioloških šumova u levom i desnom uhu i da koeficijent korelacije raste sa porastom intenziteta bioloških šumova. Ovakvo ponašanje bioloških šumova predstavlja povoljnu okolnost sa stanovišta primene adaptivnih metoda filtriranja i adaptivnih metoda potiskivanja šuma. U disertaciji su analizirane karakteristike najznačajnijih bioloških smetnji koje su prisutne kod merenja signala OAE. Ova saznanja su omogućila izbor optimalane adaptivne metode za detekciju TEOAE u prisustvu biološkog šuma. Osnovna ideja za potiskivanje biološkog šuma sastojala se u korišćenju biološkog šuma iz drugog ušnog c kanala za adaptivno potiskivanje biološkog šuma u ušnom kanalu u kome se meri OAE signal. Pošto je signal OAE duboko uronjen u šum, neophodno je bilo kombinovati nekoliko postupaka u toku predobrade signala kako bi se postiglo dobro potiskivanje šuma. Primenjeni su postupci adaptivnog filtriranja, ponderisanog usrednjavanja i oduzimanja procenjene vrednosti signala OAE od signala odziva snimljenog u ušnom kanalu. Da bi se na objektivan način procenio doprinos postupka za potiskivanje šuma zasnovanog na korišćenju dve sonde, postavljene u oba ušna kanala, posebna pažnja je posvećena eksperimentalnim uslovima snimanja TEOAE, biološkog šuma i formiranja zašumljenog signala TEOAE. Kao mera uspešnosti predloženih postupaka korišćeni su SNR, reproduktivnost i trajanje merenja TEOAE. Parametri dobijeni u toku detekcije OAE, kada su korišćeni postupci za potiskivanje šuma (predobrada), poređeni su sa parametrima dobijenim u toku detekcije OAE bez predobrade signala. Najveće potiskivanje šuma ostvaruje se u najnižim frekvencijskim opsezima i ono opada sa porastom frekvencije. Potiskivanje šuma na frekvencijama iznad 3500 Hz je zanemarljivo. Ovo je u skladu sa teorijskim očekivanjima kada se ima u vidu spektralna gustina snage bioloških šumova. Dobijeni rezultati pokazuju da postoji statistički značajna razlika između rezultata dobijenih postupkom predobrada + detekcija signala OAE i rezultata dobijenih postupkom detekcija signala OAE bez predobrade. Ova značajnost je za SNR na nivou p = 0.009, a za reproduktivnost na nivou p = 0.002. Pri tome se trajanje merenja TEOAE smanjuje za 9.72 %. Zbog statističke prirode postupka detekcije signala OAE neophodno je bilo proveriti da li predloženi postupci predobrade signala dovode do detekcije nepostojećih TEOAE. U namenski izvedenom eksperimentu izvršena su merenja OAE u pasivnim komorama zapremine 1cm3 i 2cm3. Analiza dobijenih rezultata je pokazala da postupci predobrade signala u toku detekcije TEOAE ne dovode do detekcije nepostojećih otoakustičkih emisija.

Biological basis of the otoacoustic emission signal (OAE), acoustic nature of its generation and exceptional masking with various forms of interference are the starting point for the development of various methods of detection, signal processing and interpretation of OAE signal, or the effect of otoacoustic emission in light of the characterization of the cochlea functioning. OAE measurement methods differ in the way of stimulus generation, response generation mechanisms and the methods of signal processing and statistical interpretation of results. Existing methods of measurement and OAE signal processing are the basis of modern systems for OAE measurement, which have found wide application in clinical practice. However, all possibilities are not exhausted, especially the theoretical and methodological nature aproaches, primarily in the field of OAE signal extraction from the noise. Improvement of the measuring conditions reduces the influence of external noise on the possibility of extraction of otoacoustic emissions, and biological noise becomes the dominant source of interference. The study of characteristics of biological noise is one of the prerequisites for successful troubleshooting OAE detection in the presence of biological noise. In carefully selected sample, in accordance with the needs of research, we analyzed the biological noise recorded in the ear canal. It turned out that there was a significant correlation of biological noise in the left and right ear and the correlation coefficient increases with the intensity of biological noise. This behavior of biological noise is a favorable circumstance from the point of application of adaptive filtering methods and adaptive noise suppression method. The dissertation analyzes the characteristics of the most important form of biological interference that are present during OAE measurement. These findings have enabled the selection of optimal adaptive methods for the detection of TEOAE in the presence of biological noise. The basic idea of suppressing biological noise consisted in e the use of biological noise from the other ear canal for adaptive suppression of biological noise in the ear canal in which OAE signal is measured. Since the OAE signal is deeply immersed in the noise it was necessary to combine several steps during preprocessing the signal to achieve good noise suppression. The applied methods are adaptive filtering, weighted averaging and estimated value of OAE signal subtraction from the signal response recorded in the ear canal. In order to objectively assess the contribution of the procedure for noise suppression based on the use of two probes, placed in both ear canals, particular attention was paid to experimental conditions for recording TEOAE and biological noise and the formation of noisy TEOAE signals. As a measure of the goodness of the proposed procedures SNR, reproducibility, and duration of TEOAE measurements were used. The parameters obtained during the OAE detection methods when used to suppress noise (preprocessing), were compared with parameters obtained during the detection of OAE signals without preprocessing. The greatest noise suppression is achieved in the lowest frequency bands, and it decreases with increasing frequency. The suppression of noise at frequencies above 3500 Hz is negligible. This is consistent with theoretical expectations when we consider the power spectral density of biological noise. The results show a statistically significant difference between the results obtained with preprocessing + OAE signal detection and the results obtained with OAE signal detection procedure without preprocessing. The statisticaly significant diference of the SNR is at the level of p = 0.009, and for the reproducibility is at the level of p = 0.002. At the same time it reduces the TEOAE measurement duration for 9.72%. Due to the statistical nature of OAE signal detection procedure it was necessary to verify that the proposed methods of preprocessing don't produce false OAE signal detection. In a dedicated experiment OAE measurements were performed in passive chambers with volumes 1cm3 and 2cm3. Analysis of the results showed that signal preprocessing procedure in the detection of TEOAE not lead to detection of non- existent otoacoustic emissions.

Spektrometrija fotonskog zračenja u radiološkom opsegu energija pomoću CdTe detektora

Photon spectrometry in diagnostic radiology energy range with the cdte detector.

U ovoj disertaciji je prikazan metod za korišćenje detektora od kadmijum-telurida (CdTe) u spektrometriji fotona u medicinskoj dijagnostičkoj radiologiji. Detekcija fotonskog zračenja i određivanje njegovog spektra CdTe poluprovodničkim detektorom podrazumeva i analizu deformacija spektra. Ove deformacije se javljaju usled neželjenih posledica interakcije upadnih fotona i elektrona detektora, kao i usled nekompletnog transporta nosilaca naelektrisanja unutar detektora. Da bi se te posledice uklonile potrebno je upotrebiti metod anfoldinga (unfolding). Anfolding je jedna vrsta dekonvolucuje i njime se izračunava pravi energetski spektar upadnog zračenja, a prema deformisanom, izmerenom spektru. Ovaj metod je podrazumevao rad u više etapa: (1) eksperimentalna merenja CdTe spektrometrom; (2) simulaciju realnog eksperimenta primenom Monte Karlo metoda, (3) modelovanje uticaja električnog polja unutar poluprovodničkog detektora na nosioce naelektrisanja, i (4) numeričko rešavanje sistema jednačina u slučaju loše uslovljene matrice sistema. Simulacija eksperimenta je izvršena programom MCNP5 i PTRAC izlaznom datotekom. Datoteka PTRAC je omogućila uključivanje modela transporta nosilaca u Monte Karlo simulaciju i algoritam je predstavljen dalje u tekstu. Izabrana numerička metoda za rešavanje sistema jednačina u slučaju loše uslovljene matrice je metod dekompozicije na singularne vrednosti sa odsecanjem. Pokazano je da se rezultati primenjenog anfolding metoda dobro slažu sa referentnim spektrom. Anfolding metod je onda uspešno iskorišćen za postprocesiranje nekolicine spektara X-zračenja koji se koristi u medicinskoj radiologiji snimljenih pomoću CdTe detektora.

The method for using a cadmium-telluride (CdTe) radiation detector for the photon spectrometry in the diagnostic medical radiology is presented here. The detection of photon radiation and the determination of its spectrum with the CdTe semiconductor, also involves the analysis of the deformation of the spectrum. These distortions are due to the unwanted consequence of the interaction of photons with detector electrons, as well as due to the incomplete charge carrier transport within the detector. In order to eliminate these consequences, it is necessary to use the unfolding method. The unfolding method is a deconvolution method that calculates true energy spectrum of incident radiation according to convoluted, measured spectrum. This method involved working in several phases: (1) real spectrum measurement; (2) simulation of the real experiment using the Monte Carlo method, (3) modeling of the influence of the electric field within the semiconductor detector to the charge carriers transport, and (4) the numerical solution of systems of equations in the case of poorly conditioned system matrix. The experiment was simulated with the program MCNP5 and a PTRAC output file. The PTRAC file enabled the inclusion of carrier transport model into Monte Carlo simulation and the algorithm is presented in the text. A truncated single value decomposition method was used as the numerical method for solving systems of equations in the case of poorly conditioned system matrix. The results showed that the results of applied unfolding method agree well with the referential spectrum. The unfolding method was than succesfuly used for the post-processing of several X-spectra used in the field of diagnostic radiology recorded using the CdTe detectors.

Adsorpcija jona teških metala iz vodenih rastvora korišćenjem pšenične slame kao adsorbensa

Adsorption of heavy metal ions from aqueousb solution using wheat straw as an adsorbent.

U disertaciji su prikazani rezultati adsorpcije Cu2+, Ni2+, Zn2+ i Pb2+ jona iz sintetičkih rastvora korišćenjem pšenične slame kao adsorbensa. Eksperimenti adsorpcije su izvođeni u reaktoru sa mešanjem i u adsorpcionoj koloni. Izvršena je karakterizacija pšenične slame; određene su fizičke, hemijske i fizičko-hemijske osobine. Određen je sadržaj organskih materija u vodi od ispiranja slame destilisanom vodom na osnovu hemijske potrošnje kiseonika (HPK). Dobijena vrednost za HPK je iznosila 18 mgO2/dm3. Određen je sadržaj alkalnih i zemnoalkalnih metala, pre i nakon ispiranja slame destilisanom vodom, kao i nakon zasićenja datim jonima metala. Određen je i kapacitet izmene katjona (KIK) između slame i vodene faze, standardnom metodom jonske izmene sa NH4Cl. Ukupni KIK iznosio je 1,86 mmol Mez+/g slame (Mez+ ovde predstavlja jone alkalnih i zemnoalkalnih metala). Na osnovu dobijenih rezultata za KIK dominantan jon u izmenjivom položaju je Ca2+ jon. Određena je tačka nultog naelektrisanja (pHzpc) pšenične slame, pri čemu je dobijena vrednost od 6,5 za koncentraciju 0,1 M KNO3, odnosno 6,7 za koncentraciju 0,01 M KNO3. SEM analiza pšenične slame, pre i posle adsorpcije datih jona metala, ukazuje na jasne morfološke promene površine pšenične slame nakon adsorpcije. FTIR analiza pšenične slame pre i posle adsorpcije jona bakra, pokazala je da su O-H i C-O funkcionalne grupe u strukturi pšenične slame moguća mesta vezivanja Cu2+ jona. Do porasta pH vrednosti rastvora, tokom ispiranja slame destilisanom vodom, dolazi usled prelaska H+ jona iz vodene faze u strukturu slame, gde bivaju izmenjeni sa Ca2+ jonima. Porast provodljivosti rastvora, tokom ispiranja slame destilisanom vodom posledica je porasta koncentracije Na+, K+, Ca2+ i Mg2+ jona u rastvoru koji se ispiranjem samoizlužuju i prelaze iz strukture slame u vodenu fazu. Maksimalni kapacitet adsorpcije (9,3 mg g-1 slame) dobijen je za adsorpciju Pb2+ jona, dok je kapacitet adsorpcije Cu2+, Ni2+ i Zn2+ jona iznosio 5 mg g-1, 2,5 mg g−1 i 3,25 mg g−1 slame, redom. Do smanjenja pH vrednosti tokom adsorpcije Cu2+, Zn2+ i Pb2+ jona dolazi usled deprotonizacije funkcionalnih grupa u slami i prelaska H+ jona u rastvor. Porast pH vrednosti rastvora tokom adsorpcije Ni2+ jona govori o tome da se paralelno sa adsorpcijom Ni2+ jona na slami odvija i adsorpcija H+ jona. Za opisivanje kinetike adsorpcije datih jona na slami, najbolje slaganje sa eksperimentalnim podacima dao je kinetički model pseudo-drugog reda. Za opisivanje ravnoteže adsorpcije Cu2+, Ni2+ i Zn2+ jona na pšeničnoj slami najbolje slaganje sa eksperimentalnim podacima dali su Langmuir-ov i Temkin-ov model adsorpcione izoterme. Za opisivanje ravnoteže adsorpcije Pb2+ jona na slami, najbolje slaganje sa eksperimentalnim podacima dao je Temkin-ov model adsorpcione izoterme. Za eksperimente adsorpcije u koloni, stepen adsorpcije Cu2+, Ni2+ i Zn2+ jona, pre dostizanja probojne tačke, je iznosio 99,4 %, 99,1 % i 99,65 %, redom. Stepen adsorpcije Cu2+, Ni2+ i Zn2+ jona iz ternarnog rastvora datih jona metala, pre dostizanja probojne tačke, je iznosio 99,2 %, 95 % i 90 %, redom. Pretpostavka je da se zemnoalkalni metali (pre svega kalcijum), prisutni u molekulskoj strukturi pšenične slame, mehanizmom jonske izmene izmenjuju sa datim jonima teških metala i protonima iz rastvora tokom procesa adsorpcije. Afinitet adsorbensa prema datom jonu metala je u korelaciji sa fizičko- hemijskim osobinama (jonski radijus, elektronegativnost, elektronska konfiguracija) datog jona metala. Dobijeni rezultati ukazuju da se pšenična slama može uspešno koristiti kao jeftin, prirodni adsorbens za adsorpciju Cu2+, Ni2+, Zn2+ i Pb2+ jona iz vodenih rastvora.

The results of adsorption of Cu2+, Ni2+, Zn2+ and Pb2+ ions from synthetic solutions by using wheat straw are presented in this disertation. Physical, chemical and physico-chemical characterization of wheat straw was carried out. Chemical oxygen demand (COD) of rinsed water was determined prior to the adsorption experiments. The consumption of oxygen was 18 mg O2/dm3. The content of alkali and alkaline-earth metals, before and after rinsing of wheat straw with water, and after loading the adsorbent with Cu2+, Zn2+, Ni2+ and Pb2+ ions, was given. The cation exchange capacity (CEC) of alkali and alkaline-earth metals, exchanged in the reaction with NH4Cl, was also determined. The determined total CEC of the wheat straw was 1.86 mmol Mez+ g-1 (Mez+ denotes exchanged alkali and alkaline earth metal ions). Based on the CEC results, the dominant ion in the exchangeable position is Ca2+ ion. The point of zero charge (pHzpc) was determined. The obtained value was 6.5 for . 0.1 M KNO3, i.e., 6.7 for 0.01 M KNO3. SEM analysis of wheat straw before and after adsorption of considered metal ions indicates, the obvious morphological changes of the wheat straw surface after adsorption of the metal ions. The FTIR spectra analysis of wheat straw before and after adsorption of Cu2+ ions revealed that O-H and C-O functional groups are possible sites for Cu2+ ions bonding. The increase in the initial pH value of the solution, during rinsing wheat straw with distilled water, occurs due to a transfer of H+ ions from the aqueous phase to the molecular structure of straw, to be exchanged there with Ca2+ ions. The increase in the conductivity of the solution, during rinsing wheat straw with distilled water, was due to the increase in the concentration of Na+, K+, Ca2+ i Mg2+ ions in the solution, which were leached from the straw via a leaching mechanism. The maximum adsorption capacity of 9.3 mg g-1 was achieved for Pb2+ ions, while the adsorption capacity for Cu2+, Ni2+ and Zn2+ ions was 5 mg g-1, 2.5 mg g-1 and 3.25 mg g-1, respectively. Kinetics of adsorption was modelled by a pseudo-second order kinetic model showing a very good fitting with experimental data for all the considered metal ions. The adsorption equilibrium data for Cu2+, Ni2+ and Zn2+ ions can be well modelled by the Langmuir and Temkin adsorption isotherm model. The adsorption equilibrium data for Pb2+ ions can be well modelled by the Temkin adsorption isotherm model. In the column experiments, the adsorption degree of Cu2+, Ni2+ and Zn2+ ions, before reaching the breakthrough point, was 99.4 %, 99.1 % and 99.65 %, respectively. The adsorption degree of Cu2+, Ni2+ and Zn2+ ions from ternary solution, before reaching the breakthrough point, was 99.2 %, 95 % and 90 %, respectively. It is assumed that the alkaline-earth metals (mostly calcium), from wheat straw, are substituted by the considered heavy metal ions and protons during the adsorption process. The affinity of the adsorbent to a considered metal ion is correlated with the physico-chemical properties (ionic radius, electronegativity, electronic configuration) of a given metal ion. The obtained results show that wheat straw can be successfully used as cheap, natural adsorbent for the adsorption of Cu2+, Zn2+, Ni2+ and Pb2+ ions from aqueous solutions, contaning few ppm of heavy metal ions.

Ispitivanje kompozitnih materijala pomoću ugrađenih fiberoptičkih senzora prilikom udara kontrolisanom energijom

Composite materials testing using embedded fiberoptic sensors by controlled energy impact

Fiberoptički senzori (FOS) koriste se poslednjih 30 godina za praćenje stanja struktura. Većina savremenih istraživanja odnose se na rešenja na bazi Bragovih rešetki, rešetki sa dugom periodom, interferometarske senzore ili optičku reflektometriju u vremenskom domenu, imajući u vidu značajno bolju osetljivost ovih rešenja u poređenju sa intenzitetskim. Sa druge strane, intenzitetski senzori su jeftiniji i njihovo ispitivanje je tehnički manje zahtevno. U okviru ove teze ispitivana je upotreba unutrašnjeg intenzitetskog FOS za praćenje stanja strukture prilikom udara kontrolisanom energijom. Utvrđena je korelacija između dubine modulacije intenziteta svetlosti i udarne sile na različitim temperaturama i za različite tipove materijala. Kompozitni materijal sa polietilenskom matricom, ojačanjem u vidu staklenog mata i ugrađenim telekomunikacionim vlaknima podvrgnut je udarima, korišćenjem uređaja za testiranje udarima kontrolisane energije, a koji je služio kao kalibracioni instrument. Prenosna funkcija senzora generisana je na osnovu povezivanja dubine modulacije intenziteta svetlosti, udarne sile i rastojanja optičkog vlakna od mesta udara. Prezentovani su uslovi za validnost prenosne funkcije. Predložena je ortogonalna mreža optičkih vlakana za detekciju opasnih udara i izračunato rastojanje između vlakana u toj mreži za posmatrani materijal. Takođe je predložena kalibraciona procedura i algoritam za određivanje mesta udara. Metod proračuna udarne sile, promene impulsa udarača i mesta udara proveren je prilikom ponovljenih udara na prethodno kalibrisanom uzorku. Senzor ima mogućnost pružanja informacija o vremenu udara, a može i generisati alarm kada je sila veća od granične vrednosti sile nakon koje nastaju oštećenja u materijalu. Utvrđeno je da u procesu izrade uzorka sa polietilenskom matricom i staklenim matom kao ojačanjem, dolazi do homogenizacije optičkog vlakna sa rešetkom sa dugom periodom i da nema promene spektra pri primenjenom statičkom opterećenju. Pokazano je da plastično optičko vlakno ugrađeno u vlaknima ojačan uzorak sa epoksidnom matricom, kao unutrašnji intenzitetni FOS, poseduje inferiornu osetljivost u odnosu na telekomunikaciono vlakno.

Fiberoptic sensors (FOS) have been used for structural health monitoring for decades. Most of the recent research is dedicated to fiber Bragg grating, long period grating (LPG), interferometric and OTDR based solutions due to their sensitivity superiority when compared to intensity FOS. On the other hand, intensity FOS is less expensive and sensor interrogation is significantly less demanding. In this thesis, the use of intrinsic intensity fiber-optic sensors for structural health monitoring was investigated. The correlation between the light intensity modulation depth and the impact force was determined under different testing temperatures and for different types of samples. Polyethylene-based composite samples with glass mat reinforcement and embedded telecommunication fibers were subjected to impact using high-speed puncture impact tester as a calibration device. Transfer function of the sensor was generated by linking the light intensity modulation depth with the impact force and the impact point to fiber distance. Conditions for transfer function validity were presented. To detect any possible dangerous impact, a net of crossed fibers was proposed and the net spacing for the given material was calculated. Appropriate non- damaging calibration procedure was proposed, as well as the algorithm for locating the impact point. The method for calculating the impact force, the projectile momentum change, and the place of impact was verified by consecutive strikes on the previously calibrated sample. The sensor can also provide information on the time of the impact and an alarm signal if the force exceeds the damaging threshold force. It was determined that the optical fiber with LPG homogenizes when embedded during material processing in polyethylene-based composite sample with glass mat reinforcement and that there is no difference in sensor response under static loading. When used as an intrinsic intensity FOS, the plastic optical fiber, embedded in epoxy- based composite sample with glass fiber reinforcement, was shown to provide inferior sensitivity compared to telecommunication fibers.

Software system for multi-criteria planning and management of hybrid microgrids

Softverski sistem za više-kriterijumsko planiranje i upravljanje hibridnom mikro-mrežom

Efikasno upravljanje energetskom infrastrukturom u kontekstu sve većeg udela distribuiranih resursa i obnovljivih izvora energije u kombinaciji sa skladištima energije postaje sve veći izazov. Kada se još ima u vidu da su ovakvi hibridni sistemi sve češće povezani na distributivnu mrežu i da funkcionišu u kontekstu liberalnog tržišta energije, postaje jasno zašto postojeći sistemi za upravljanje imaju sve više problema da ostvare efikasan rad. Ova složenost upravljanja takođe otežava i bržu penetraciju obnovljivih izvora i doprinosi nepredvidivosti po pitanju njihove ekonomičnosti. Ova teza predlaže inovativni softverski sistem koji donosi napredne algoritme za planiranje i upravljanje hibridnim mikro-mrežama, pri čemu se oslanja na postojeće sisteme nadzora upravljanja za potrebe prikupljanja podataka i neophodnih informacija sa fizičkog sloja, ali i za prosleđivanje i sprovođenje donesene kontrolne akcije. Imajući u vidu ovu vrstu zavisnosti od postojećih sistema, jedan od doprinosa teze je bio upravo razvoj inovativnog pristupa za jednostavnu integraciju sa postojećim IKT sistemima koji često koristite posebne, nestandardne, protokole i mehanizme za komunikaciju. S tim u vezi, predloženo rešenje prevazilazi potrebe za komunikacionom i semantičkom interoperabilnosti i otvara prostor za napredno rezonovanje. U nastavku je dat kratak pregled konkretnih rezultata kao i osnovnih doprinosa u odnosu na postojeća rešenja. Pre svega, razvijen je alat za planiranje, odnosno dizajn, mikro-mreže koji slično postojećim pristupima razmatra hibridne energetske infrastrukture sa konvencionalnim i obnovljivim izvorima energije, različitim elementima za konverziju kao i skladištima energije putem odgovarajućih matematičkih modela različite kompleksnosti. Međutim, predložena metodologija unapređuje postojeće alate i pristupe u planiranju kroz integralno razmatranje domena električne i toplotne energije, razmatranja kako izolovanih sistema tako i onih povezanih na distributivnu mrežu, ali pre svega, primenom fleksibilnosti na strani potrošnje za optimalno dimenzionisanje elemenata mikro-mreže. Pored toga, za potrebe rangiranja višestrukog broja prihvatljivih alternativa, korišćeni su posebni algoritmi za višekriterijumsko odlučivanje. Takođe, razvijen je inovativni model potrošnje koji ima zadatak da utvrdi očekivani profila potrošnje energije. Predloženi model se ne naslanja na uobičajene modele koji se zasnivaju na fizičkom modelu zgrade već se uvodi metodologiju koja polazi od navika i zahteva samog korisnika kao i postojećih potrošača. U okviru teze, razvijen je i alat za upravljanje mikro-mrežom koji ima za cilj nadogradnju postojećih sistema za upravljanje. Naime, razvijeni alat je testiran u realnom okruženju gde je ostvario prosečne uštede troškova reda 10% kroz jednostavnu, softversku, nadgradnju postojeće upravljačke infrastrukture. Dobijeni rezultati su potpuno uporedivi sa onima koji se dobijaju prilikom neuporedivo većih investicija koje se tiču zamene pojedinih uređaja za neke efikasnije. Predložena metodologija, koja je implementirana u okviru ovog alata, se konceptualno oslanja na neke od postojećih pristupa ali pravi kvalitativni pomak u odnosu na njih jer uvodi u razmatranje upravljanje potrošnjom, i to na nivou pojedinačnih uređaja, što dalje omogućuje direktnu primenu rezultata optimizacije na odgovarajuće kontrolne akcije, što nije slučaj sa postojećim metodama. Konačno, u okviru teze je razvijen i model kompleksnih objekata i njihove odgovarajuće energetske infrastrukture kao i unificirani model za komunikaciju kako bi se omogućila semantička interoperabilnost postojećih nadzornih i kontrolnih sistema koji često koriste posebne, nestandardne, komunikacijske protokole. Korišćenjem tehnologija semantičkog veba i polazeći od već postojećih modela i standarda, razvijeni model služi kao baza znanje o čitavom sistemu, a ne samo energetskoj infrastrukturi, čime se omogućava unapređenje algoritama za upravljanje energijom i omogućava primena dinamičkih mera energetske efikasnosti.

Efficient operation and management of energy systems in the context of increasing penetration of distributed resources and renewable energy sources combined with energy storages, gird-tied configurations and deregulated energy market context pose a complex energy-dispatching problem, the one conventional energy management systems are not able to cope with. This complexity also hinders faster adoption of renewables and aggravates their cost effectiveness. This thesis proposes an innovative software system, which brings advanced microgrid planning, and operation algorithms while leveraging existing supervision and control systems to collect data and information from the field level but also to forward desired control action. Moreover, the thesis also proposes innovative approach for seamless integration with diverse, legacy, ICT systems thus contributing to system large-scale applicability. Following is a breakdown of concrete results and outputs with emphasis on their specific advancements comparted to existing approaches. Similarly, to existing approaches the underlying methodology of microgrid planning tool considers hybrid energy infrastructures with both conventional and renewable energy sources, converters and storages, featured by their mathematical representation. However, proposed methodology advances existing tools and planning approaches by simultaneously considering both electric and thermal energy domain, by considering both grid-tied and isolated deployment scenarios and, foremost, by leveraging demand side flexibility for optimized dimensioning of energy assets. Moreover, multi-criteria decision-making algorithms are used to rank multiple feasible planning alternatives. For the purpose of determining expected energy demand profile, an approach to leverage user habits and behaviour together with characteristics of existing energy infrastructure, rather than building physics, was developed. The developed microgrid management tool aimed at providing additional intelligence to existing supervision and control systems resulting in the cost savings typically exceeding 10%, as demonstrated in a real world use case, at the cost of ICT system retrofit. The achieved level of cost savings is comparable to those of retrofit of energy assets, which require considerable higher investments. The underlying microgrid management methodology complements existing approaches that optimize energy imports for a desired demand and available storages, by introducing appliance-level demand response actions, which can be immediately translated into control actions. Lastly, an ontology-based facility model of complex facilities and corresponding energy infrastructure was developed together with unified messaging format to enable semantic interoperability of existing diverse, legacy, supervision and control systems, which use proprietary communication protocols. By employing semantic web technologies and leveraging existing models and standards, the developed model stores contextual knowledge about the entire facility, and not just energy infrastructure, thus enables enhancement of energy management algorithms and allowing for application of dynamic energy efficiency measures.

Analiza ugaone raspodele incidentne energije spoljašnje buke primenom mikrofonskog niza

Analysis of the incident energy anular distribution of ambiental noise by microphone array

Problem buke u savremenom svetu postao je danas jedan od nekoliko faktora koji dominatno utiču na život ljudi, bilo da se radi o životnom ili radnom prostoru. Problem buke naročito je izražen u urbanim uslovima, gde je struktura zvučnog polja vrlo složena jer buka potiče od velikog broja različitih i dislociranih zvučnih izvora koji emituju zvuk različitog nivoa i spektralnog sadržaja. Procena karakteristika zvučnog polja, odnosno vrste i prostorne distribucije izvora zvuka, kao i mogućnost klasifikacije pojedinih tipova izvora buke predstavljaju osnov za planiranje mogućih intervencija koje će dovesti do smanjenja nivoa buke u okruženju. Monitoring spoljašnje buke standardno se vrši na bazi merenja jednim neusmerenim mikrofonom. Osnovna ideja rada je da se analizira mogućnost primene mikrofonskog niza za detaljnu vremensku, frekvencijku, i prostornu karakterizaciju zvučnog polja, lokalizaciju izvora i kategorizaciju urbanih zona. U okviru rada osmišljen je i realizovan mikrofonski niz sa malim brojem mikrofona čija je konfiguracija optimizovana u skladu sa zadatim frekvencijskim opsegom zvučnog polja u urbanim sredinama. Primenom više algoritama za prostorno-vremensku obradu signala određena je struktura zvučnog polja u urbanim uslovima. Takođe, izvršeno je poređenje performansi algoritama u skladu sa postavljenim zahtevima i namenom rezultata. U okviru sprovedenih istraživanja izvršena su dugovremena merenja spoljašnje buke mikrofonskim nizom na različitim lokacijama i određena je struktura zvučnog polja u prostoru ispred fasade. Izabrane su lokacije koje se značajno razlikuju po konfiguraciji terena. Karakteristike polja na svim posmatranim lokacijama detaljno su analiziranae u pogledu lokalizacije i karakterizacije izvora zvuka. Razvijena je metodologija za eksperimentalno određivanje ugaone raspodele incidentne energije spoljašnje buke na fasadi na osnovu rezultata lokalizacije. Dobijene raspodele korišćene su u daljoj analizi kao osnov za stastitičku kategorizaciju urbanih zona. Izvršena je višedimenziona statistička analiza ugaonih raspodela spoljašnje buke za različite lokacije u urbanim uslovima sa ciljem da se utvrde kategorije sa sličnim v karakteristikama zvučnog polja. Eksperimentalno određene ugaone raspodele incidentne energije spoljašnje buke korišćene su za predikciju izolacione moći fasadnih pregrada. Proračun terenske izolacione moći fasadnih pregrada izvršen je na osnovu poznatih građevinskih parametara i eksperimentalno utvrđene ugaone raspodele incidentne energije. Time je stvorena mogućnost za sagledavanje razlika u ostvarenim vrednostima izolacione moći fasadne pregrade za različite kategorije urbanih zona.

The noise problem in today’s modern world has become one of the key factors that dominantly influence the lives of people in both the living and working areas. The noise problem is particularly expressed in the urban environments, where the structure of the sound field is quite complex due to a large number of different and spatially dislocated sound sources, which emit sounds of different levels and spectral content. The assessment of the sound field characteristics, i.e. the type and the spatial distribution of the sound sources, as well as the possibility of classifying individual types of noise sources, form the basis for planning potential interventions which would lead to a decrease in the environment’s noise levels. Standard outdoor noise monitoring is based on the measurements with one omnidirectional microphone. The basic idea of this work is to analyze the possibilities of microphone array application for a detailed temporal, spectral and spatial characterization of the sound field, source localization and urban area categorization. A microphone array with a small number of microphones, whose configuration is in accordance with the frequency range of the sound field in urban environments, has been designed and realized within the scope of this research. The sound field structure in urban areas has been determined using several algorithms for space-time signal analysis. Furthermore, the comparison of algorithms’ performance has been conducted in accordance with given demands and results’ aims. Within the research, long time interval measurements of the outdoor noise were performed in various locations and the sound field structure in front of the façade was determined. The selected locations have significant differences in terms of terrain configuration. The characteristics of the sound field in all the observed locations were analyzed in detail in terms of localization and characterization of sources. Based on the localization results, a methodology was developed for the experimental determination of the angular distribution of the outdoor noise incidence energy. Obtained distributions were used in the detailed analysis as the basis for statistical vii categorization of the urban zones. Multidimensional statistical analysis of the outdoor noise angular distribution was performed for different locations within the urban area, with the goal of determining the categories with the similar characteristics of the sound field. Experimentally determined outdoor noise incidence energy angular distributions were used for predicting the façade’s sound reduction index values. The calculation of the in-situ sound reduction index of façade partitions was based on known physical parameters of the partitions and the experimentally obtained incident energy angular distributions. This procedure allowed for the comparison of differences in the achieved values of the façade partition sound reduction index for different categories of urban zones.

Modelovanje i analiza uticaja prostornog i vremenskog profila snage vetra u projektovanju i eksploataciji vetroelektrana u elektroenergetskom sistemu

Modeling and analysis of influence of spatial and temporal wind power profile on the design and exploitation of wind farms in electric power system

Predmet istraživanja doktorske disertacije je modelovanje i analiza uticaja prostornog i vremenskog profila snage vetra u projektovanju i eksploataciji vetroelektrana u elektroenergetskom sistemu (EES). Osnovni doprinosi radu su: 1. Razvijen je model za visinsku ekstrapolaciju merenih podataka o brzini vetra za procenu proizvodnje vetroelektrana. Model se bazira na metodi minimuma sume kvadrata odstupanja. Primenom ovog modela na set merenih podataka o brzini vetra sa najmanje tri visine dobija se sintetički set mernih podataka o brzini vetra na željenoj visini postavljanja vetroturbine. Estimirana baza mernih podataka se može koristiti kao ulazni set podataka za profesionalne softvere za regionalnu analizu resursa vetra, kao i procenu proizvodnje vetroelektrana. Algoritam je testiran korišćenjem realnih merenih podataka o brzini vetra sa tri lokacije koje karakterišu različiti topografski i klimatski uslovi. Sprovedene analize pokazuju da visinska ekstrapolacija merenih podataka o brzini vetra, prema predloženom modelu, daje bolje rezultate u proceni resursa energije vetra i smanjuje grešku u proceni proizvodnje vetroelektrana u odnosu na klasičnu primenu programa WAsP, koji standardno koristi podatke samo sa jedne merne visine. 2. Razvijen je model dinamičkog koeficijenta smicanja vetra. Visinski profil brzine vetra se u vetroenergetici opisuje sa dva modela i to stepenim zakonom i logaritamskim zakonom. Oba modela se praktično podjednako koriste u inženjerskoj praksi. Nedostatak oba modela je što u sebi sadrže fiksne parametre, koji ne omogućavaju dinamičko sagledavanje profila brzine vetra. Kombinujući ova dva pristupa u tezi je analitički raspregnut uticaj dominantnih faktora koji utiču na visinski profil brzine vetra. U razvijenom modelu visinski profil brzine vetra se modeluje sa koeficijentom smicanja vetra koji ima dve komponente: jednu statičku, koja je posledica hrapavosti terena, i drugu, vremenski promenljivu komponentu, koja uvažava dijabatske promene visinskog profila brzine vetra. Razvijeni model ima značajnu praktičnu 5 primenu u regionalnoj analizi potencijala energije vetra, kao i u proceni prosečne višegodišnje proizvodnje vetroelektrane na osnovu kratkoročnih merenja. Praktična primenljivost modela je testirana na realnim projektima vetroelektrana u regionu Banata i Pešterske visoravni. 3. Uveden je indeks korelisanosti profila proizvodnje vetroelektrane i profila cene na berzi električne energije. Za sagledavanje ekonomičnosti projekta vetroelektrane u uslovima slobodnog tržišta, pored procene godišnje proizvodnje, u tezi je analizirana korelisanost između dijagrama proizvodnje i dijagrama cena na berzi električne energije. Definisan je indeks korelacije između vremenskog profila snage proizvodnje vetroelektrane i vremenskog profila cene na berzi, koji omogućava valorizaciju električne energije iz vetroelektrana na slobodnom tržištu električne energije. Za procenu ovog indeksa potrebno je imati podatke o prosečnom dnevnom i sezonskom profilu brzine vetra na lokaciji vetroelektrane, kao i prosečnom dnevnom i sezonskom profilu cena na analiziranoj berzi. Definisanje indeksa korelisanosti omogućava investitoru da preciznije sagleda ekonomičnost projekta, kao i vrednost presečnog proizvedenog MWh na berzi električne energije. Za svaku berzu električne energije mogu se definisati teorijske granične vrednosti indeksa korelisanosti, koje su samo funkcije faktora kapaciteta vetroelektrane. Uporednom analizom procenjenog indeksa korelisanosti za neku vetroelektranu i odgovarajućih graničnih vrednosti može se proceniti kvalitet vetra u pogledu dnevnog i sezonskog profila. 4. Definisana je metodologija za procenu uticaja vremenskog profila proizvodnje vetroelektrana na gubitke u distributivnoj mreži. U radu je razvijen model za procenu uticaja vetroelektrane na gubitke snage u priključnoj distributivnoj mreži. Uveden je indeks korelisanosti profila proizvodnje i profila potrošnje, koji predstavlja meru kvaliteta vremenskog dijagrama proizvodnje vetroelektrane u pogledu gubitaka. Uvedeni indeks ima opšti karakter i za njegov proračun potrebno je poznavati profil potrošnje i profil proizvodnje odabranog vetroagregata. Njegova brojna vrednost predstavlja očekivanu relativnu promenu gubitaka na distributivnom vodu nakon priključenja vetroelektrane. Sprovedene analize omogućavaju distributivnim kompanijama da sagledaju pozitivne efekte rada vetroelektrana na distributivnu mrežu u pogledu smanjenja gubitaka. 6

The research subject in this dissertation is the modeling and analysis of the effects of spatial and temporal profiles of wind power in the design and operation of wind farms in power systems. The main contributions of this work are: 1. A new mathematical model for vertical extrapolation of the measurement data for wind speed taken at several measurement heights. The model is based on the method of least squares (LES). By applying the proposed model on the sets of measured data taken at least at three measurement heights, one can obtain a synthetic set of data at a desired height where the wind power potential is analyzed. The basic idea is that during the process of estimation of the wind power potential, the measurement data are first extrapolated by the proposed method, and then by using program WAsP the spatial extrapolation is carried out. The algorithm is tested by one year wind speed measurement data taken at three locations characterized by different topographies of the terrain and different climatic conditions. The performed analyses show that pre- processing of measurement data by the proposed method results in a better estimate of the wind power potential at a height which is greater than the measurement heights compared to that obtained by the standard application of WAsP program which makes use of measurement data taken at one measurement height. 2. A new model for dynamic wind shear coefficient. In wind energy, the height profile of the wind speed is usually described by two models based on exponential and logarithmic law. Both models are almost equally used in engineering practice. The principle lack of both models is that they contain fixed parameters, which do not allow the dynamic consideration of wind speed profiles. Combining these two approaches, the analytical decoupling of the dominant factors affecting the height profile of the wind speed is provided. In the developed model, height profile of the wind speed is modelled with two parameters, one static, which is a consequence of the terrain roughness, and other time-varying parameter, which is the consequence of atmospheric stability. The developed model has important practical applications in 8 regional analysis of wind energy potential, as well as in long term wind resource assessment. The practical applicability of the model was tested on real projects of wind farms in the region of Banat and the Pester plateau. 3. Introduction of correlation index between the production profile and prices profile at an electricity market. For the purpose of assessing the economy of the project of a wind power plant (WPP) under open market conditions, the analysis of the correlation between the production diagram and diagram of the market prices has been performed and, in addition, the annual electricity production has been estimated. In this dissertation the correlation index between the production profile and prices profile at an electricity market has been introduced. This index represents a measure of wind quality regarding the daily and seasonal variations. For the purpose of this index estimation, one should have measurement data concerning average daily and seasonal profiles of the wind at the WPP location, as well as the average daily and seasonal profile of prices at the considered auction. For each electricity market it is possible to define theoretical limiting values of the correlation index for a WPP which are functions of the capacity factor of the WPP only. By a comparative analysis of the estimated correlation index of a WPP and the corresponding limiting values, the quality of the wind regarding daily and seasonal profile can be estimated. 4. A methodology for estimating the influence of wind power plant production time profile on distribution network losses is defined. In this thesis, a model for estimating the optimal power of wind power plant, connected to a distribution line, providing the minimal energy losses, is developed. The correlation index between the production and load profiles is introduced which represents the measure of quality of the wind power production time diagram with respect to losses. The introduced index has a global character and it is necessary to know the load profile and production profile of the chosen wind power plant. The numerical value of the index represents the expected relative change of the distribution line losses, after the wind power plant is connected. The performed analysis introduces the necessity and provides the basis for revision of technical recommendations for connection of small-scale embedded power plants to a distribution system. This revision is necessary to ensure, in an optimal way, positive effect of operation of wind power plants and of other distributive sources in a distribution system. 9

Formalizacija i automatsko dokazivanje teorema euklidske geometrije

Formalization and automation of Euclidean geometry

Napredak geometrije kroz vekove se može razmatrati kroz razvoj različitih aksiomatskih sistema koji je opisuju. Upotreba aksiomatskih sistema započinje sa Hilbertom i Tarskim ali se tu ne završava. Čak i danas se razvijaju novi aksiomatski sitemi za rad sa euklidskom geometrijom...

The advance of geometry over the centuries can be observed through the development of dierent axiomatic systems that describe it. The use of axiomatic systems begins with Euclid, continues with Hilbert and Tarski, but it doesn't end there. Even today, new axiomatic systems for Euclidean geometry are developed...

Pojačavači snage u klasi A sa istovremenim konjugovanim prilagođenjem i prilagođenjem po snazi pri velikim signalima na izlaznom pristupu

Power amplifiers in class a with simultaneous conjugate and large signal power matching at the output port

Disertacija se bavi razvojem metodologije projektovanja i unapređenjem topo- logija pojačavača snage u klasi A sa istovremenim konjugovanim prilagođenjem i prilagođenjem po snazi pri velikim signalima na izlaznom pristupu. Jedan od osnovnih zahteva za svaki pojačavač snage jeste efikasnost. U slučaju pojačavača snage u klasi A, tranzistor treba prilagoditi po snazi pri velikim signalima na izlaznom pristupu da bi se opterećenju predala maksimalna moguća snaga. Osim prilagođenja po snazi, pojačavači male i srednje snage često moraju biti i konjugovano-kompleksno prilagođeni na izlaznom pristupu. Ovaj zahtev je bi- tan pošto je uobičajeno da je sledeći stepen projektovan tako da optimalno radi kada je konjugovano-kompleksno prilagođen na ulaznom pristupu. Pojačavači male i srednje snage su obično implementirani u standardnoj complementary metal-oxide semiconductor (CMOS) tehnologiji, dok naredni stepen može biti filtar za potiskiva- nje neželjenih harmonika ili pojačavač snage napravljen u nekoj od III-V generacija poluprovodničke tehnologije sa znatno većom izlaznom snagom. Standardni pojačavač sa zajedničkim sorsom može biti prilagođen po snazi pri velikim signalima ili konjugovano prilagođen na izlaznom pristupu, ali se oba pri- lagođenja ne mogu, u opštem slučaju, ostvariti istovremeno. Da bi se prevazišlo pomenuto ograničenje, predložena je topologija pojačavača sa drejn-gejt povratnom spregom kao i kompletna metodologija projektovanja. U svrhu verifikacije razvijene teorije, projektovan je širokopojasni pojačavač snage u 130 nm RFCMOS tehnologiji. Postignuta je maksimalna izlazna snaga od 7 dBm i efikasnost od 20%, u opsegu učestanosti od 6 do 9 GHz. Koeficijent refleksije na ulaznom i izlaznom pristupu je manji od -8,5 dB i -9,5 dB, redom. Rezime

The dissertation deals with the development of design methodology and the enhancement of topology of power amplifiers in class A with simultaneous conjugate and power matching for large signals at the output port. One of the major demands for any power amplifier is power efficiency. In the case of a power amplifier in Class A, the transistor needs to be power matched for large signals at the output port in order to deliver maximum power to the load. This criterion is particularly important in the case of an amplifier with an output power of the order of 30 dBm or more. In addition to power matching, small and medium power amplifiers often have to be conjugately matched at the output port as well. This requirement is impor- tant because it is quite common that the next stage is designed so that it works optimally when it is conjugately matched at the input port. Small and medium power amplifiers are usually implemented in standard complementary metal-oxide semiconductor CMOS technology, while the next stage can be a filter for suppressing unwanted harmonics or a power amplifier made in one of the III-V generations of semiconductor technology with significantly higher output power. A standard common-source amplifier can be designed to be either large signal power matched or conjugately matched at the output port, but not both of them at the same time. In order to overcome the mentioned limit, we propose a simple topology based on a drain-gate feedback amplifier, together with a complete design methodology. For the purpose of verifying the developed theory, a broadband power amplifier is designed in 130 nm RFCMOS technology. The amplifier achieves 7 dBm peak output power and maximum power efficiency of about 20% in the frequency range from 6 to 9 GHz. Input and output return loss are better than 8.5 and 9.5 dB, respectively. Abstract

Procene gradijenata funkcija i normi operatora u teoriji harmonijskih funkcija

Estimates of gradients and operator norm estimates in harmonic function theory.

U ovoj tezi razmatraju se ocene normi operatora u teoriji harmonijskih funkcija i procene gradijenata funkcija. Prvi rezultati navedeni u tezi su nejednakosti Švarcovog tipa za ho- lomorfna preslikavanja jedinične lopte Bn u jediničnu loptu Bm, zatim analogne nejednakosti za holomorfne funkcije na disku D bez nula, kao i pluriharmonijske funkcije na jediničnoj lopti Bn sa kodomenom u (−1, 1). Ovim su data poboljšanja tvrđenja dobijenih u [32] i [18]. Takođe, dat je nov dokaz činjenice da su pozitivne harmonijske funkcije na poluravni kontrak- tivne, ako se na poluravni H i na R+ posmatraju hiperboličke metrike([47]). Osim toga, druga glava sadrži i primere koji pokazuju da isto ne važi ako se posmatraju poluprostori viših dimenzija. Navedeni rezultati objavljeni su u [55]. Rezultati navedeni u trećoj glavi vezani su za određivanje tačnih polu- normi jedne težinske varijante Berezinove transformacije, posmatrane na L∞(Bn) sa kodomenom u glatkom Blohovom prostoru ([57]). Četvrta glava sadrži brojne rezultate vezane za Bergmanovu projekci- ju. Rešen je problem Markovića i Kalaja postavljen u [28], koji se odnosi na određivanje tačne polunorme Bergmanove projekcije iz L∞(Bn) na B(Bn). Činjenica da se na Blohovom prostoru posmatra polunorma zadata pomoću M−invarijantnog gradijenta predstavljala je srž problema. Takođe, date su i optimalne procene gradijenta Bergmanove projekcije Lp funkcije na jediničnoj lopti Bn, kao i posledice koje se odnose na Košijevu projekciju i odgovarajuće ocene gradijenata u Hardijevim i Bergmanovim prostorima. Do- bijene su i tačne vrednosti Blohovih polunormi i normi Košijeve projekcije na L∞(Sn). Dobijeni rezultati bazirani su na [56] i [58]. U poslednjoj glavi dat je dokaz dela hipoteze Holenbeka i Verbickog iz [26]. Tačnije, određene su tačne norme operatora � |P+|s + |P−|s� 1 s za 0 < s ≤ 2 na Lp(T), gde je P+ Risova projekcija i P− = I − P+. Osim toga, date su i odgovarajuće dualne nejednakosti i dokazana je njihova optimalnost. Rad [45] inspirisan je rezultatima iz [25] i [33]. •

УНИВЕРЗИТЕТ У БЕОГРАДУ МАТЕМАТИЧКИ ФАКУЛТЕТ Петар Мелентијевић Процене градијената функција и норми оператора у теорији хармонијских функција ДОКТОРСКА ДИСЕРТАЦИЈА Београд, 2018. UNIVERSITY OF BELGRADE FACULTY OF MATHEMATICS Petar Melentijević Estimates of gradients and operator norm estimates in harmonic function theory DOCTORAL DISSERTATION Belgrade, 2018. МЕНТОР • проф. др Милош Арсеновић, ЧЛАНОВИ КОМИСИЈЕ • проф. др Милош Арсеновић, редовни професор Математичког факултета • проф. др Оливера Михић, редовни професор Факултета организационих наука • проф. др Данко Јоцић, редовни професор Математичког факултета Процена градијената функција и норми оператора у теорији хармонијских функција • РЕЗИМЕ. У овој тези разматрају се оцене норми оператора у теорији хармонијских функција и процене градијената функција. Први резултати наведени у тези су неједнакости Шварцовог типа за хо- ломорфна пресликавања јединичне лопте Bn у јединичну лопту Bm, затим аналогне неједнакости за холоморфне функције на диску D без нула, као и плурихармонијске функције на јединичној лопти Bn са кодоменом у (−1, 1). Овим су дата побољшања тврђења добијених у [32] и [18]. Такође, дат је нов доказ чињенице да су позитивне хармонијске функције на полуравни контрак- тивне, ако се на полуравни H и на R+ посматрају хиперболичке метрике([47]). Осим тога, друга глава садржи и примере који показују да исто не важи ако се посматрају полупростори виших димензија. Наведени резултати објављени су у [55]. Резултати наведени у трећој глави везани су за одређивање тачних полу- норми једне тежинске варијанте Березинове трансформације, посматране на L∞(Bn) са кодоменом у глатком Блоховом простору ([57]). Четврта глава садржи бројне резултате везане за Бергманову пројекци- ју. Решен је проблем Марковића и Калаја постављен у [28], који се односи на одређивање тачне полунорме Бергманове пројекције из L∞(Bn) на B(Bn). Чињеница да се на Блоховом простору посматра полунорма задата помоћу M−инваријантног градијента представљала је срж проблема. Такође, дате су и оптималне процене градијента Бергманове пројекције Lp функције на јединичној лопти Bn, као и последице које се односе на Кошијеву пројекцију и одговарајуће оцене градијената у Хардијевим и Бергмановим просторима. До- бијене су и тачне вредности Блохових полунорми и норми Кошијеве пројекције на L∞(Sn). Добијени резултати базирани су на [56] и [58]. У последњој глави дат је доказ дела хипотезе Холенбека и Вербицког из [26]. Тачније, одређене су тачне норме оператора � |P+|s + |P−|s� 1 s за 0 < s ≤ 2 на Lp(T), где је P+ Рисова пројекција и P− = I − P+. Осим тога, дате су и одговарајуће дуалне неједнакости и доказана је њихова оптималност. Рад [45] инспирисан је резултатима из [25] и [33]. • КЉУЧНЕ РЕЧИ. Шварцова лема, Бергманова пројекција, Березино- ва трансформација, Рисова пројекција, јединична лопта, Хардијеви простори, Бергманови простори, Блохови простори • НАУЧНА ОБЛАСТ. Математика • УЖА НАУЧНА ОБЛАСТ. Комплексна анализа • УДК. 517.544(043.3) • AMS КЛАСИФИКАЦИЈА. 47B35, 31B10, 30H05, 31A05 i Estimates of gradients and operator norm estimates in harmonic function theory • ABSTRACT. In this thesis we study sharp estimates of gradients and operator norm estimates in harmonic function theory. First, we obtain Schwarz-type inequalities for holomorphic mappings from the unit ball Bn to the unit ball Bm, and then analoguous inequalities for holomorphic functions on the disk D without zeros and pluriharmonic functions from the unit ball Bn to (−1, 1). These extend results from [32] and [18]. Also, we give a new proof of the fact that positive harmonic function in the upper-half plane is a contraction with resprect to hyperbolic metrics on both H and R+ ([47]). Besides that, in the second chapter, we construct the examples to show that the analoguous does not hold for the higher-dimensional upper-half spaces. All mentioned results are from the authors’ paper [55]. In the third chapter we intend to calculate the exact seminorm of the weighted Berezin transform considered as an operator from L∞(Bn) to the ”smooth” Bloch space ([57]). The fourth chapter contains results concerning Bergman projection. We solve the problem posed by Kalaj and Marković in [28] on determining the exact seminorm of the Bergman projections from L∞(Bn) to the B(Bn). The crucial obstacle is the fact that B(Bn) is equipped with M−invariant gradient seminorm. Also, we provide the sharp gradient estimates of the Bergman projection of an Lp function in the unit ball Bn, as well as its consequences on Cauchy projection and certain gradient estimates for the functions from the Hardy and Bergman spaces.We obtain the exact values of the Bloch’s seminorms and norms for the Cauchy projection on L∞(Sn). These results are based on the papers [56] and [58]. The last chapter contains the proof of the one part of Hollenbeck-Verbitsky conjecture from [26]. Exactly, we find the exact norms of � |P+|s + |P−|s� 1 s for 0 < s ≤ 2 on Lp(T), where P+ is the Riesz projection and P− = I − P+. Also we give the appropriate dual estimates and prove that they are sharp. The paper [45] is motivated by the results from [25] and [33]. •

Određivač pravca zasnovan na pasivnoj višekanalnoj detekciji elektromagnetskog signala

The direction finder based on the passive multichannel detection of the electromagnetic signal

Pasivno određivanje pravca dolazećeg elektromagnetskog talasa (passive direction finding - DF) je jedna od tehničkih disciplina standardno primenjivana u različitim civilnim i vojnim oblastima. Tipičan primer su pasivni sistemi za ranu detekciju borbene tehnike koja emituje elektromagnetsko zračenje, kao što su letelice ili plovila sa aktiviranim radarima. U oblasti civilne primene, upotreba DF-a je mnogo raznovrsnija i masovnija, pa se takvi sistemi rutinski koriste u svrhu detekcije pokretnih i stacionarnih izvora zračenja, prepoznavanja i otklanjanja interferencije u radio-vezama, lociranju neautorizovanih predajnika, u sigurnosnim i bezbednosnim servisima, itd. Zbog široke oblasti primene, DF sistemi mogu biti različitih karakteristika, kao što su: a) portabilnost – mogućnost integracije uređaja u male prenosive sisteme, ili male pokretne sisteme (bespilotna letelica, lako terensko vozilo, patrolni čamac, i dr.); b) rad u realnom vremenu; v) detekcija zračenja na više frekvencijskih kanala, itd. Da bi se ispunili tako specifični zahtevi, potrebno je sistemski pristupiti dizajnu, što je glavna tema doktorske disertacije. Kako je za akviziciju bilo koje fizičke veličine potreban odgovarajući senzor, akvizicija elektromagnetskih talasa, postiže se upotrebom odgovarajućih antena, koje su u DF tehnikama tipično integrisane u odgovarajući antenski niz. Za primenu definisanu u disertaciji, postoji više tipova pogodnih antena. Sa ciljem selekcije odgovarajuće antene napravljen je pregled najpogodnije klase antena – antene horn tipa, i zaključeno je da je piramidalni horn najbolje rešenje. Cilj metoda procesiranja signala dobijenih sa antenskog niza kao senzora je određivanje pravca dolazećeg elektromagnetskog signala. Mogu da budu realizovane različitim tipovima algoritama, pa se shodno tome metodi međusobno i razlikuju. Iako postoje super-rezolucione tehnike koje daju najbolje rezultate po pitanju rezolucije, one nisu efikasne i pogodne za rad u portabilnim sistemima, tako da je za dizajn odabrana jedna od klasičnih tehnika amplitudskog tipa. Analizirane su osnovne tehnike amplitudskog tipa i odabrana je najpogodnija sa stanovišta efikasnosti hardverske i softverske realizacije, kao i sa stanovišta uvedenog kriterijuma – amplitudskog dinamičkog opsega. Određivanje pravca elektromagnetskog signala iz jednog izvora, ukoliko nema dodatnih narušavajućih faktora, relativno je lako rešiv problem. Ozbiljan problem nastaje kada se radi o više izvora i koji su pri tom fizički bliski, pa određivanja pravca za svaki od njih može da bude u opštem slučaju nerešiv. Super-rezolucioni metodi, sudeći prema 6 literaturi, mogu da reše taj problem u velikom broju slučajeva, ali po cenu brzine, veličine opreme, potrošnje energije, itd. Ako je potrebno portabilno rešenje onda je taj problem u osnovi nerešiv. U praksi, fizički bliski izvori emituju zračenje na različitim frekvencijskim kanalima, pa je sukcesivnom detekcijom po svakom kanalu moguće detektovati i sve takve izvore. Takav postupak je nazvan višekanalna detekcija. Pošto se antenski niz kao senzor koristi na više frekvencija, njegove karakteristike nisu iste za svaki frekvencijski kanal. Uz to elementi antenskog niza, u opštem slučaju, nemaju identične karakteristike, bez obzira na istu geometriju. Zbog toga je u disertaciji predložen postupak kalibracije po svim antenama, i po svim frekvencijskim kanalima, kao i formiranje odgovarajuće tabele pretraživanja koja omogućuje kasnije brzo procesiranje podataka. Teoretski, za definisanu primenu, svaki cirkularni antenski niz sa usmerenim antenama prilagođenim za definisani frekvencijski opseg, mogao bi da posluži svrsi. Međutim kako je u pitanju portabilna primena, cilj je uklapanje u zadate dimenzije, odnosno minimizacija gabarita i potrebnog broja elementa antenskog niza. Zbog toga je definisana originalna procedura projektovanja piramidalnih horn antena, koja redefiniše standardni kriterijum optimalnosti, i uvodi dodatne kriterijume koji optimizuju same antene po pitanju električnih i mehaničkih karakteristika, kao i kompletan antenski niz po pitanju broja upotrebljenih elementa. Kompletan DF sistem je realizovan primenom savremene analogno-digitalne metodologije projektovanja elektronskih sistema, kao i primenom robusnih softverskih rešenja. Svi blokovi sistema, njihova funkcija i realizacija detaljno su opisani u disertaciji. Za realizovani sistem i procenjenu verovatnoću grešaka koje se generišu tokom rada, izračunata je verovatnoća ispravne detekcije. Određene su metrološke karakteristike realizovanog DF sistema, merenjem karakteristika pojedinih sklopova sistema i celog sistema u celini. Izračunate su merne nesigurnosti kako za pojedinačne sklopove, tako i za ceo sistem, čime je dato viđenje upotrebne vrednosti sistema.

Passive radio direction finding DF is a technical discipline commonly used in many civil and military applications. Typical applications are passive early warning systems, such are aircrafts and vessels equipped with radars. In area of civil applications, DF usage is much more manifold and widespread: such systems are used for detection of moving or stationary radiation sources, in reconnaissance and cleaning interference in radio communications, localization of non-authorized transmitters, in intelligent and security systems, etc. According to widespread application area, DF systems might have different characteristic, such are: a) portability – ability of integration into small transferable or mobile systems (unmanned aircraft, light infantry vehicle, patrol boat, etc.); b) real time operation; c) different channel signals detection, etc. With an aim to be able to achieve all specific DF features, it is mandatory to perform a systematic design, what is the main topic of this doctoral dissertation. For the acquisition of any physical signal, appropriate sensor is required. For the case of acquisition of the electromagnetic waves in DF techniques, the sensor is based on convenient antenna, typically organized in corresponding antenna array. For the application given in the dissertation, there are several suitable types of antennas. In order to select appropriate antenna, an overview of all suitable antennas types – horn antennas, are given, and the conclusion is that the most suitable selection is the pyramidal horn antenna. The purpose of processing signals obtained from the antenna array as a sensor, is to determine the direction of the incoming electromagnetic signal. The processing can be achieved by several techniques, and accordingly, each differ one to another. Although, there are super-resolution techniques that provides the best results in terms of resolution, they are not efficient and suitable for portable systems applications, thus, the classical techniques with amplitude detection are selected for consideration. Basic amplitude detection techniques are analysed, and one is selected as the most suitable from the point of the software and hardware efficiency and implementation, same as from the point of the introduced criterion - amplitude dynamic range. Determination of an electromagnetic wave Direction, from unique radiation source, with no interfering conditions, is a quite easy task. The problem arises when it comes to presence of multiple, physically very close, radiation sources, so in general, determination of direction, per each of them, might be unsolvable. Super-resolution methods, according to literature, are able to solve such tasks in most of cases, for the price of speed, equipment size, energy consumption, etc. If a portable solution is requested, then this problem is essentially unsolvable. In practice, physically close sources radiate at different frequency channels, so by successive detection on each 8 channel it is possible to detect all possible sources. Such a procedure is called multichannel detection. Since the antenna array, considered as a sensor, is used on multiple frequencies, its characteristics are not the same for each frequency channel. In addition, the elements of an antenna array do not have identical characteristics in general, regardless of the same geometry. Therefore, in the dissertation, a calibration procedure is proposed for all antennas, and at all frequency channels. Accordingly, constitution of lookup table that allows, efficient data processing is proposed. Theoretically, for a defined application, any circular antenna array with directional antennas adapted for the defined frequency range can be applied. However, the portable application is concerned, so the goal is to fit it into the given dimension, that is, minimize dimension and number of antenna array elements. For this reason, an original pyramidal horn antenna design procedure is defined, which redefines the standard optimization criterion, and introduces additional criteria for antenna optimization, in terms of electrical and mechanical characteristics, as well as the antenna array optimization in term of number of elements. The complete DF system was realized using the modern mix-signal electronic systems design methodology, and by applying the robust software solutions. Detail description of all system blocks, their function and realization is given in the dissertation. Using the estimated probability of errors generated during operation, the probability of correct detection is calculated. For the case of designed system, and the estimated probability of errors, the probability of correct detection is evaluated. The metrology characteristics of the realized DF system are determined, by measuring the characteristics of individual system components and the entire system itself. Measurement uncertainty for individual components and entre system are calculated, demonstrating the system quality value.

Direktno upravljanje asinhronim motorom u režimu slabljenja polja

Direct torque control of induction motor in field weakening regime

Predmet naučne rasprave u ovom radu je razvoj algoritma za direktno upravljanje asinhronim motorom u režimu slabljenja polja, koji treba da obezbedi maksimalne performanse pri radu sa velikim brzinama. U prvom poglavlju navedeni su motivi za istraživanje, detaljno je analizirana stručna literatura i identifikovani su nerešeni problemi vezani za predmetno istraživanje. Istaknute su manjkavosti dosadašnjih rezultata u upravljanju asinhronim motorom u režimu slabljenja polja, kao i smernice za razvoj nove strukture za direktno upravljanje koja će da u potpunosti iskoristi raspoložive resurse motora i pridruženog pogonskog pretvarača, te je dat sadržaj daljeg izlaganja. Drugo poglavlje disertacije bavi se formiranjem matematičkih modela asinhronog motora i pridruženog pogonskog pretvarača prilagođenih direktnom upravljanju. Matematički modeli izvedeni su uz uobičajene idealizacije koje se koriste u opštoj teoriji električnih mašina. U ovom delu opisan je i postupak upravljanja pogonskim pretvaračem zasnovan na metodi modulacije vektora napona. Predmet trećeg poglavlja je analiza tranzijentne momentne karakteristike asinhrone mašine u režimu slabljenja polja. Analizirana je mogućnost dobijanja maksimalnog momenta u zavisnosti od karakteristika pogonskog pretvarača. Istražen je uticaj strujnog, naponskog, te istovremenog strujnog i naponskog limita na maksimalni moment u režimu slabljenja polja. Analizirana je problematika kolapsa fluksa rotora, koja je posledica naponskog limita u slabljenju polja. U četvrtom poglavlju detaljno su analizirani karakteristični pristupi upravljanja asinhronim motorom u slabljenju polja. Poglavlje se sastoji od četiri dela. Prvo je izveden opšti model vektorski upravljanog asinhronog motora, na osnovu kojeg su analizirane varijante vektorskog upravljanja prema fluksu rotora i fluksu statora u slabljenju polja. Nakon toga je analiziran jedan tipičan algoritam direktnog upravljanja u slabljenju polja. Na kraju poglavlja analiza je zaokružena međusobnim poređenjem karakteristika ova tri reprezentativna postojeća koncepta upravljanja asinhronom mašinom. U petom poglavlju istražena je promena momenta, flukseva statora i rotora, klizanja, sinhrone brzine i struje u limitu napona tokom jedne upravljačke periode, a u limitu napona. Izveden je izraz za promenu momenta u limitu napona tokom upravljačke periode u opštem slučaju. Na osnovu sprovedene analize, na originalan način je u kompleksnoj ravni prikazano kretanje odgovarajućih polifazora asinhronog motora u režimu slabljenja polja. Šesto poglavlje sadrži sintezu originalne upravljačke strukture za direktno upravljanje momentom u slabljenju polja. Polazna osnova za formiranje strukture je zahtev da se raspoloživi napon invertora u potpunosti iskoristi. Pošto su u naponskom limitu komponente fluksa i momenta spregnute, kako u stacionarnom stanju, tako i u prelaznim režimima, jedina fizički nezavisna upravljačka veličina tada je fazni stav (ugao) napona statora. U cilju formiranja algoritma za direktno upravljanje u slabljenju polja, prvo je izveden dinamički matematički model asinhrone mašine u limitu napona. Usvojena je modifikacija generalizovanog matematičkog modela u sinhrono rotirajućem koordinatnom sistemu, te je tako modifikovani model linearizovan i dobijena je odgovarajuća funkcija prenosa. Usvojene se odgovarajuće aproksimacije koje znatno olakšavaju izvođenje funkcije prenosa i data je detaljna analiza opravdanosti tih aproksimacija, poređenjem ponašanja aproksimiranog i kompletnog modela. Na osnovu usvojene funkcije prenosa asinhronog motora pri radu sa punim naponom u slabljenju polja, predložena je originalna upravljačka struktura za regulaciju momenta, a zatim i postupak za određivanje potrebnih parametara regulatora za dobijanje željenog odziva. Kao projektni zahtevi za performansama regulatora momenta postavljeni su brz aperiodski odziv i nulta greška stacionarnog stanja. Nakon sinteze, dobijene performanse predloženog regulatora momenta ilustrovane su putem računarskih simulacija. Putem računarskih simulacija analiziran je i uticaj varijacije parametara regulatora i napona napajanja. U nastavku je formirana i nadređena struktura za upravljanje brzinom u slabljenju polja. Predložena je struktura regulatora i usvojen je postupak za izbor vrednosti parametara. Kao projektni zahtev za performanse regulatora brzine postavljeni su striktno aperiodski odziv bez preskoka, kao i nulta greška stacionarnog stanja. Prilikom projektovanja brzinske petlje dinamički je uvažena i promena limita po prevalnom momentu. Na kraju poglavlja, performanse predloženog regulatora brzine proverene su računarskom simulacijom. Eksperimentalna verifikacija predloženog rešenja opisana je u sedmom poglavlju. Kao reprezentativni, odabrana su tri karakteristična radna režima: prvi, u kome je ispitivan odziv na pravougaonu referentnu vrednost momenta pri konstantnoj brzini obrtanja, zatim drugi, u kome je ispitivano ubrzavanje pogona do trostruke nominalne brzine pri pravougaonoj referentnoj vrednosti momenta, te treći, gdje je ispitan rad pogona sa zatvorenom povratnom spregom po brzini. Rezultati eksperimenta prikazani su u obliku osciloskopskih snimaka i detaljno su komentarisani. U osmom poglavlju (Zaključak) istaknuti su glavni doprinosi pojedinih poglavlja i doktorske disertacije u celini. Takođe, ukazano je i na moguće pravce daljeg razvoja u razmatranoj oblasti. Na kraju rada nalazi se spisak korišćene literature, te prilog koji sadrži matematička izvođenja koja bi opteretila osnovni tekst, podatke korištene u računarskim simulacijama, kao i tehnički opis konkretno realizovanog prototipa.

Тhe subject of the dissertation is development of an algorithm for direct control of induction motor in field weakening regime, which should provide improved performance when working at high speeds. In the first Chapter motivation for the research is given, available references are analyzed in detail, and unsolved problems in the field of research are identified. Limitations of the existing solutions for induction motor control in field weakening mode are pointed out. Guidelines for the development of new structure for direct control which will fully utilize motor and power converter capabilities are given. The outline of the rest of thesis is presented. The second Chapter of the dissertation starts with the development of mathematical models of induction motor and matched power converter suitable for direct control. Mathematical models are developed with usual approximations used in the general theory of electric machines. The space vector modulation technique for power inverter control is also described in this section. The subject of the third Chapter is the analysis of torque transient characteristicsof an induction machine in field weakening regime. The ability of obtaining the maximum torque dependening on voltage and current limits of a drive is investigated. The influence of current limit, voltage limit, and combined current and voltage limits on torque characteristics in field weakening regime and ability of achieving maximal torque is investigated. The problem of rotor flux collapse, as a consequence of voltage limit in the field weakening mode, is explained in detail. The fourth Chapter analyzes in detail typical approaches for control of induction motors in field weakening regime. The Chapter consists of four parts. In the first part a general model of vector controlled induction motor drive is derived. On the proposed model, various stator and rotor flux vector controls in the field weakening regime are analyzed in the second and the third part of the Chapter. After that, a typical algorithm for direct torque control applicable in the field weakening regime is presented. At the end of the Chapter, all three presented methods are compared to each other. In the fifth Chapter the changes in torque, stator and rotor fluxes, slip and synchronous speed during a single control period at the voltage limit are analyzed. The expression for torque change at voltage limit is derived for the general case. Based on the presented equations, the original representation of induction motor polyphasor trajectories in the field weakening regime is given. The sixth Chapter contains a synthesis of the original control structure for direct torque control in field weakening regime. The proposed structure is based on the requirement to fully utilize available inverter voltage. Since both flux and torque producing components are coupled at the voltage limit, in both steady state and during transients, the only independent control variable is stator voltage phase angle. In order to create an algorithm for direct torque control in field weakening mode, a dynamic model of induction machine at voltage limit is developed first. The model is developed as a modification of generalized mathematical model in the synchronously rotating reference frame. This model is linearized around the operating point, and the induction motor transfer function in field weakening mode is obtained. The adopted approximations are discussed, by comparing behaviors of approximate and full order models using computer simulations. Based on the adopted transfer function, the original control structure of torque regulation is presented. Design performance requirements for torque regulator are zero torque error in steady state and fast transient response. The procedure for creating the control structure and calculating the regulator gains is explained in detail. The key part of the structure is torque regulator with variable gains, i.e., Gain Scheduling torque regulator is used. Performance of the proposed torque controller is illustrated through computer simulations. Impacts of motor parameter changes and supply voltage variations are analyzed through computer simulations too. At the end of the Chapter, the outer speed control loop synthesis procedure is presented. Performance requirements are strictly aperiodic speed response and zero error in steady state. Control structure and parameter tunings are explained in detail. In the proposed structure, the control limits are dynamically changed following the break-down torque changes. Performance of the proposed speed controller is validated by the computer simulation. Experimental verification of the proposed solution is described in seventh Chapter. As a representative, three typical operating modes are verified: first, the response to a rectangular change of torque reference at constant speed, the second mode, in which the motor accelerates and decelerates to the triple rated speed, and the third, where the closed speed loop performance is tested. Experimental results are presented in a form of oscilloscope readings and are discussed in detail. In the eighth Chapter (Conclusion) the major contributions from all chapters and the dissertation in general are highlighted. Possible directions for further work in the field of research are pointed out. Bibliography, mathematical derivations excluded from the main text for clarity, data used in computer simulations, as well as the technical description of the realized prototype are given at the and of the dissertation.

Algoritmi distribuirane detekcije i estimacije zasnovani na konsenzusu

Consensus based distributed detection and estimation algorithms

Ova disertacija razmatra nekoliko važnih problema u distribuiranoj obradi signala, to jest, u distribuiranoj detekciji i estimaciji signala...

This work deals with several important problems of distributed signal processing, i.e., distributed signal detection and estimation. Novel algorithms for distributed change detection and target tracking using sensor networks, as well as algorithm for fault detection and isolation using overlapping system decomposition are proposed. All the proposed algorithms are based on the introduction of a consensus strategy, whose dynamics is combined in parallel with detection or estimation dynamics, enforcing in such a way agreement between the sensors of the used sensor networks or between overlapping subsystems of the monitored systems, and obtaining robust and ecient solutions of the considered problems. In the rst part a novel distributed algorithm derived from the Generalized Likelihood Ratio is proposed for real time change detection using sensor networks. The algorithm is based on a combination of recursively generated local statistics and a global consensus strategy, and does not require any fusion center. The problem of detection of an unknown change in the mean of an observed random process is discussed and the performance of the algorithm is analyzed in the sense of a measure of the error with respect to the corresponding centralized algorithm. The analysis encompasses asymmetric constant and randomly time varying matrices describing communications in the network, as well as constant and time varying forgetting factors in the underlying recursions. An analogous algorithm for detection of an unknown change in the variance is also proposed. Simulation results illustrate characteristic properties of the algorithms including detection performance in terms of detection delay and false alarm rate. They also show that the theoretical analysis connected to the problem of detecting change in the mean can be extended to the problem of detecting change in the variance. A new distributed fault detection and isolation (FDI) methodology is proposed in the second part, in the form of a multi-agent network representing a combination of a consensus based FDI observer for residual generation and a consensus based decision making strategy for change detection, applicable in real time. The proposed observer is based on overlapping system decomposition and a combination between the local optimal stochastic FDI observers and a dynamic consensus strategy. It is shown how the proposed algorithm can generate residuals which provide, under general conditions concerning local models and the network topology, high eciency, scalability and robustness. The proposed decision making strategy provides solutions for two particular cases: a) local detection for non-overlapping parts of the identied subsystems; b) a consensus based strategy for FDI in the overlapping parts. Selected examples illustrate the applicability of the proposed methodology in practice...

Razvoj kontekstno-osetljivih korisničkih interfejsa

Development of context-sensitive user interfaces

Dobro dizajniran, intuitivan i privlačan za korišćenje korisnički interfejs predstavlja ključni faktor uspeha računarskih proizvoda i sistema. Radi unapređenja razvoja i upotrebljivosti korisničkih interfejsa potrebno je uzeti u obzir karakteristike korisnika. Ovo zahteva interdisciplinaran pristup i korišćenje znanja iz različitih oblasti kao što su računarske, saznajne i biološke nauke. Pored toga, potrebno je uzeti u obzir karakteristike medija i fizičkog okruženja u kojem se odvija interakcija čoveka i računara. Razvoj korisničkog interfejsa treba da uvaži i karakteristike hardverskih uređaja koji se koriste u komunikaciji sa korisnikom, dostupne softverske resurse, kao i karakteristike programskih sistema koji treba da koriste korisnički interfejs. U skladu sa tim, uvodi se pojam kontekstno-osetljivog interfejsa koji se definiše kao korisnički interfejs koji je prilagodljiv kontekstu interakcije sa konkretnim korisnikom. Kontekst interakcije čine tri klase entiteta: korisnik računarskog sistema (čovek); hardverska i softverska platforma pomoću kojih korisnici interaguju sa sistemom i fizičko okruženje u kojem se odigrava interakcija sa sistemom. Posmatrajući evoluciju razvoja softvera uočavamo povećanje nivoa apstrakcije na kojem se softver opisuje. Dostignuti nivo razvoja omogućava platformski nezavisnu specifikaciju softvera koja se postepeno ili automatizovano prevodi u izvršne aplikacije za različite softverske i hardverske platforme. Arhitektura upravljana modelima, koja se koristi za razvoj složenih programskih rešenja, hijerarhijski organizuje koncepte i modele u više nivoa apstrakcije. Ovo je posebno bitno imajući u vidu da je razvoj kontekstno-osetljivih korisničkih interfejsa složen proces koji uključuje modelovanje velikog broja elemenata na različitim nivoima apstrakcije. U ovoj tezi smo istraživali problem unapređenja razvoja kontekstno-osetljivih korisničkih interfejsa. Predloženo je rešenje koje omogućava automatizaciju razvoja korisničkog interfejsa prilagođenog kontekstu interakcije čoveka i računara. Rešenje se ogleda u proširenju jezika za modelovanje, standardnog procesa razvoja softverskih sistema (Unified proces) i razvojnih alata elementima specifičnim za interakciju čoveka i računara. U skladu sa prethodnim, razvijen je model kontekstno-osetljive interakcije čoveka i računara i predloženi su modeli korisničkih interfejsa na različitim nivoima apstrakcije. Zbog standardizacije, široke prihvaćenosti, i dostupnosti razvojnih alata, odlučili smo se za proširenje UML (Unified Modeling Language) jezika za modelovanje i ATL (Atlas Transformation Language) jezika za transformacije modela. Primena predloženog pristupa je demonstrirana na primerima dve studije slučaja iz različitih domena.

Well-designed, intuitive and catchy-to-use user interface represents key issue of success of computer products and systems. In order to improve development and usability of user interfaces it is needed to take into account user’s charasteristics. This entails interdisciplinary approach and use of knowledge from different fields such as computing, cognitive and biological sciences. In addition, it is needed to consider features of the physical environment and the medium in which interaction between human and computer takes place. Development of user interface must include characteristics of hardware devices employed in interaction with the user, availabale software resources, as well as characteristics of software systems using the interface. According to stated, concept of context-sensitive user interface is introduced, defined as a user interface adaptable to context of interaction with concrete user. Context of interaction is decomposed into three classes of entitites: user of a computer system (human); hardware and software platforms by which users interact with the system and physical environment in which interaction with system happens. Looking at the evolution of software development, we can notice that the abstraction level on which software is described is increasing all the time. The latest trend is to specify software using platform-independent models, which are then gradually and (semi-) automatically transformed into executable applications for different platforms and target devices. Model- driven architecture, used for development of complex software solutions, hierarchically organizes concepts and models into multiple abstraction levels. This is especially important regarding development of context-sensitive user interfaces which appears to be a complex process involved with modeling of a large number of elements on different abstraction levels. In this thesis, we have been exploring problem concerned with the improvement of development of context-sensitive user interfaces. Solution enabling automation of development of user interface adaptable to context of interaction between human and computer is proposed. Solution includes extensions of modelling language, standard software development process (Unified process) and development tools with the elements specific for human-computer interaction. Based on previous, model of context-sensitive human-computer interaction has been developed and user interface models on different abstraction levels have been proposed. For reasons of standardization, wide acceptance and availability of development tools, we have decided to extend UML (Unified Modeling Language) modeling language and ATL (Atlas Transformation Language) language for model transformations. Application of the proposed approach is demonstrated with examples of two case studies from different domains.

Ekspertski sistem za praćenje i prepoznavanje ciljeva u okruženju izviđačkih radara

Expert system for tracking and recoissance targets in the environment of ground surveillance radar

Jedan od najznačajnijih procesa u vojnim primenama je procena situacije. Procena situacije predstavlja kontinuirani proces izdvajanja informacija prikupljenih senzorima u cilju procene mogućih pretnji i opasnosti tokom nekog perioda vremena. Za potrebe procene situacije koriste se senzori koji rade u različitim delovima elektromagnetnog spektra (radar, televizijska i termovizijska kamera, akustički senzori, seizmički senzori i drugi), dok se ostvaruje i fuzija podataka dobijenih različitim senzorima. Svaka od ovih vrsta senzora ima svojih prednosti i nedostataka u odnosu na druge vrste. Prednosti korišćenja radara se ogledaju u povećanoj daljini detekcije prisustva cilja i mogućnosti rada u složenim meteorološkim uslovima, dok su nedostaci manja rezolucija u odnosu na optoelektronske senzore i mogućnost ometanja usled aktivnog zračenja energije. U procesu osmatranja i kontrole zemljišta upotreba radarskih senzora u savremenim borbenim dejstvima je naročito izražena. Za razliku od radara za osmatranje vazdušnog prostora, radari za izviđanje zemljišta (izviđački radari) se susreću sa nizom dodatnih problema usled jakog klaterskog okruženja i naizmeničnog pojavljivanja i nestajanja ciljeva. Cilj istraživanja predstavljenog u ovoj disertaciji predstavlja projektovanje i testiranje ekspertskog sistema za praćenje i prepoznavanje ciljeva u okruženju izviđačkih radara. Ovaj ekspertski sistem je zasnovan na skrivenim Markovljevim lancima kao modelu složenog stohastičkog procesa. Kao vektori obeležja korišćeni su parametri autoregresivnih modela realnih radarskih signala koji su prikupljeni za potrebe ovog istraživanja. U radu je predložena nova mera za izbor reda ovih modela zasnovana na maksimalnoj separabilnosti različitih klasa. Rezultati primene nove mere poređeni su sa merama za izbor reda autoregresivnih modela predloženim u literaturi. Pored ovoga, izvršena je analiza koncentracije energije spektrograma realnih radarskih signala na osnovu mera predloženih u literaturi. Ova analiza je sprovedena radi određivanja optimalnih vrednosti vrste i dužine prozora spektrograma koji obezbeđuju maksimalnu koncentraciju energije u vremensko – frekvencijskoj ravni. Na osnovu izabranih modela autoregresivnih procesa iv izvršen je izbor dimenzija kodne knjige kvantizacije skrivenih Markovljevih lanaca. Ovaj izbor je izvršen na osnovu određivanja broja klastera metodom oduzimanja potencijala na takav način da se minimizuje rastojanje podataka od centara klastera. Nakon ovoga, izabrani su elementi skrivenih Markovljevih lanaca i projektovan i testiran ekspertski sistem zasnovan na ovim modelima. Na osnovu testova na ilustrativnim sekvencama sagledani su i neki nedostaci predloženog rešenja ekspertskog sistema za praćenje i prepoznavanje ciljeva u okruženju izviđačkih radara. Istraživanje sprovedeno u okviru ove doktorske disertacije nije samo u funkciji modernizacije savremenih vojnih sistema, već rezultati predstavljeni u ovoj disertaciji mogu biti iskorišćeni i za projektovanje različitih civilnih sistema za nadgledanje saobraćaja, javnih mesta, autobuskih i železničkih stanica, trgova, kao i za obezbeđenje objekata od nacionalnog značaja kao što su kompleksi za snabdevanje vodom, aerodromi, elektroenergetska postrojenja i slično.

One of the most important processes in the military to change the assessment of the situation. Assessment is a continuous process of extracting information collected by sensors in order to assess possible threats and risks over a period of time. For the purposes of assessing the situation using sensors that work in different parts of the electromagnetic spectrum (radar, television and thermal imaging cameras, acoustic sensors, seismic sensors, etc.), while the realized and fusion of data obtained by different sensors. Each of these sensors has its advantages and disadvantages compared to other species. The advantages of using radar is reflected in the increased distance detecting the target and the possibility of working in complex meteorological conditions, while the disadvantages of lower resolutions in relation to the optoelectronic sensors and the possibility of disruption due to active energy radiation. In the process of monitoring and control of land use radar sensors in modern combat is particularly pronounced. Unlike radar surveillance of airspace, ground surveillance radars (GSR) are faced with a number of additional problems due to strong clutter environment and intermittent appearance and disappearance goals. The aim of the research presented in this thesis is the design and testing of an expert system for tracking and identifying targets in an environment of ground surveillance radar. This expert system is based on the hidden Markov chain model as a complex stochastic process. As feature vectors were used autoregressive model parameters in real radar signals that are collected for the purpose of this research. In this thesis is proposed a new measure for selecting the order of these models based on the maximum separability of different classes. Results of the new measures were compared with measures for the selection order autoregressive model proposed in the literature. In addition, an analysis of the real concentration of energy spectra of radar signals based on the measures proposed in the literature. This analysis was conducted to determine the optimal value of the type and length of the window spectrogram that provide the maximum concentration of power in the time - frequency level. Dimension of quantization codebook for using hidden Markov chains is choosen based on selected parameters of radar echo signal models. This selection is performed based on the method of determining the number of clusters seizure potential in such a way as to minimize the distance between the data centers of the clusters. After this, the vi elements of hidden Markov chains are selected and expert system for tracking and recognizing radar targets is designed and tested based on these chains. On the basis of tests on the illustrative sequences analyzed are some disadvantages of the proposed solutions expert system for tracking and identifying targets in an environment reconnaissance radar. Research conducted in the framework of this doctoral thesis is not only a function of modernization of modern military systems, but the results presented in this thesis can be used for the design of various civil systems for traffic monitoring, public places, bus and railway stations, squares and the like, as well as for securing objects of national importance such as complexes for water supply, airports, power plants and the like.

Određivanje razlike dinamičkog i svetskog vremena i predviđanja promena u brzini zemljine rotacije

Determination of the difference between dynamical time and universal time and predictions of variations in the earth's rotation.

Problem preciznog predvi]anja Zemljine prostorne orijentacije je u direktnoj vezi sa poznavanjem budu'ih vrednosti parametara Zemljine rotacije...

The problem of precise prediction of the Earth's spetial orientation is directly related to the knowledge of the forthcoming values of the Eart's rotation parametars...

Uticaj loših električnih kontakata na nastanak početnog požara i metoda za njihovo otkrivanje u niskonaponskim električnim instalacijama

The influence of poor electrical contacts on the occurrence of initial fire and a method for their detection in low-voltage electrical installations

Pouzdano funkcionisanje niskonaponskih električnih instalacija predstavlja važan preduslov za normalno odvijanje tehnoloških procesa u industriji, normalan rad javnih ustanova (škola, pozorišta, muzeja, bolnica...), neometano funkcionisanje domaćinstava itd. Usled lošeg izbora ili neadekvatne ugradnje, korišćenja ili održavanja električnih instalacija, javljaju se kvarovi koji mogu da dovedu do požara, električnog udara ili prekida u napajanju (dela) objekta električnom energijom. Posebna pažnja mora da se posveti kvarovima koji mogu da izazovu požar, a koje klasične zaštitne komponente (automatski prekidač, osigurač sa topljivim umetkom, zaštitni uređaj diferencijalne struje itd.) ne mogu da detektuju. Takve kvarove najčešće izazivaju loši električni kontakti. Loš električni kontakt u niskonaponskoj električnoj instalaciji može se pojaviti u razvodnom ormanu (razvodnoj tabli), ali i bilo gde u strujnom kolu (na oštećenom provodniku, u utikaču ili utičnici, u produžnom kablu, na prijemniku električne energije itd.). Ukoliko se ne otkloni, vremenom se povećava njegova temperatura, što može dovesti do topljenja ili paljenja izolacije električne komponente i okolnog zapaljivog materijala. Lošim električnim kontaktom u električnoj instalaciji može se nazvati onaj kontakt kod koga je: - smanjena sila pritiska naleganja kontaktnih površina bakarnog provodnika i zavrtnja na priključnom elementu elektroinstalacione komponente (redukcija momenta zatezanja), - ostvareno nepotpuno naleganja (preklapanje) površine bakarnog provodnika i kontaktne površine na priključnom elementu elektroinstalacione komponente (smanjenje efektivne površine kontakta), ili - povećan sloj oksida na mestu električnog kontakta (star i/ili korozijom oštećen kontakt). Niti domaći niti međunarodni pravilnici i standardi kojima su propisane periodične verifikacije kvaliteta niskonaponskih električnih instalacija ne sadrže procedure pomoću kojih bi se u ranoj fazi detektovao loš električni kontakt. Međutim, jedna od tehnika koje se danas koriste u tu svrhu je vršenje preventivnih termovizijskih pregleda (samo u SAD postoji standard koji pokriva ovu oblast). Ovakvim pregledom mogu jednostavno i efikasno da se otkriju loši električni kontakti na kojima su stvoreni preduslovi za nastanak početnog požara na niskonaponskoj električnoj instalaciji. Detektovani kvarovi i neregularnosti najčešće se mogu otkloniti jednostavnim intervencijama električara ili tehničkih službi korisnika objekta. Glavni nedostatak pomenute tehnike termovizijskog pregleda je nemogućnost da se pregledaju (snime) električni kontakti koji nisu direktno (vizuelno) dostupni infracrvenom senzoru kamere. Cilj doktorske disertacije je analiza uticaja loših električnih kontakata na nastanak početnog požara i razvoj metode za njihovo rano otkrivanje u niskonaponskim električnim instalacijama. Na osnovu velikog broja izvršenih eksperimenata u kojima su korišćene razne kombinacije materijala od kojih se izrađuju kontakti (bakar–bakar, bakar–mesing i bakar–nerđajući čelik), uz pomoć posebno konstruisane aparature simuliran je loš električni kontakt koji nastaje usled nepotpunog naleganja (preklapanja) bakarnog provodnika i površine elektroinstalacionog elementa na priključku elektroinstalacione komponente. Utvrđeno je da bakar–nerđajući čelik predstavlja kritičnu kombinaciju materijala za koje kontakt pri istim uslovima (istom strujnom opterećenju, preseku provodnika i stepenu pogoršanja kontakta u smislu istog procentualnog smanjenja efektivne površine preklapanja) najbrže dostiže maksimalnu trajno dozvoljenu temperaturu. Na osnovu velikog broja eksperimenata utvrđen je i uticaj redukcije momenta zatezanja zavrtnja u elektroinstalacionoj komponenti na nastajanje „vrele“ tačke u šuko-utikačima i šuko-utičnicama. Zaključeno je da su za nastanak „vrele“ tačke nepovoljniji loši električni kontakti nastali u ugrađenoj električnoj instalaciji objekta (gde je uobičajen bakarni provodnik punog preseka) u odnosu na loše električne kontakte nastale u napojnim kablovima prijemnika (gde se uobičajeno koristi licnasti bakarni provodnik). Na osnovu velikog broja izvedenih eksperimenata u kojima je analiziran uticaj loših električnih kontakata na nastanak „vrelih“ tačaka na šuko- utičnicama, kao i na osnovu podataka dobijenih iz velikog broja stručnih nalaza o verifikaciji kvaliteta niskonaponskih električnih instalacija u industrijskim i administrativnim objektima (koje je izradila Laboratorija za ispitivanje niskonaponskih električnih i gromobranskih zaštitnih instalacija Elektrotehničkog fakulteta u Beogradu), razvijena je nova metoda za otkrivanje loših električnih kontakata. Ona je zasnovana na uspostavljanju korelacije između izmerenih otpornosti petlje kvara i petlje kratkog spoja i stepena pogoršanja kontakta, odnosno povećanja njegove električne otpornosti. Urađen je niz eksperimenata sa utičnicom koja je bila korišćena dugi niz godina (30 godina u analiziranom slučaju), kao i niz eksperimenata sa novim utičnicama, ali sa električnim kontaktima koji su prethodno bili namerno izloženi koroziji. Da bi se ubrzao proces korozije, u jednoj od novih utičnica priključak faznog provodnika je tretiran carskom vodom (koncentrovana HNO3 : koncentrovana HCl = 1:3 (v / v)), a u drugoj novoj utičnici priključak neutralnog provodnika tretiran je koncentrovanom sumpornom kiselinom (H2SO4). Korišćenjem ovih hemikalija simuliran je efekat dugotrajnog izlaganja električnih kontakata elektroinstalacionih komponenti kiseloj atmosferi. Upoređivanjem rezultata merenja koja su za cilj imala da se analiziraju efekti redukcije momenta zatezanja zavrtnja na priključcima testiranih utičnica ((stare (korišćene) utičnice i dve nove utičnice (sa namerno korodiranim spojevima)) na zagrevanje i električnu otpornost kontakta, pokazano je da one imaju slične karakteristike. U dostupnim domaćim i stranim standardima i naučnostručnoj literaturi, uz izuzetak nekoliko patenata koji su namenjeni za industrijske i specijalne primene, nisu predstavljene metode koje mogu da posluže za pravovremeno otkrivanje loših električnih kontakata u stambenim, poslovnim i javnim objektima. Primenom metode razvijene u predmetnoj disertaciji, koja je namenjena otkrivanju loših električnih kontakata u strujnim kolima, redukovala bi se verovatnoća nastajanja loših električnih kontakata i opasnih „vrelih“ tačaka na njima koje izazivaju požar (metoda bi se primenjivala u okviru redovne procedure za proveru efikasnosti sistema zaštite zasnovanog na automatskom isključenju napajanja u slučaju kvara – električnim merenjem otpornosti petlje kvara i otpornosti petlje kratkog spoja). Time bi se rešio problem nemogućnosti da se termovizijskim pregledom otkriju „vrele“ tačke na lošim kontaktima sa kojima senzor kamere nema direktan vizuelni kontakt. Predložena metoda za otkrivanje lošeg električnog kontakta primenjena je prilikom verifikacije kvaliteta niskonaponskih električnih instalacija u stambenim jedinicama različite starosti (10–60 godina). Proverena je i mogućnost uspostavljanja veze između starosti stambenih jedinica (tačnije starosti njihovih električnih instalacija) i verovatnoće nastanka opasne „vrele“ tačke na lošem električnom kontaktu u strujnim kolima monofaznih i trofaznih prijemnika. Nažalost, nikakva veza nije uspostavljena. Za njeno eventualno uspostavljanje bilo bi neophodno primeniti novu metodu na više hiljada utičnica u stanovima širokog opsega starosti. Neotkriven loš električni kontakt u električnoj instalaciji može da prouzrokuje redni električni luk koji je čest uzrok požara. Za razliku od lošeg električnog kontakta, redni električni luk može biti otkriven primenom savremenih zaštitnih komponenti. U laboratorijskim uslovima, eksperimentalno je izvršena provera funkcionalnosti rada AFDD (Arc-Fault Detection Device) komponente namenjene evropskom tržištu, čija je osnovna namena zaštita od neželjenog rednog električnog luka. Provera funkcionalnosti je počivala na velikom broju laboratorijskih eksperimenata pomoću kojih su simulirane skoro sve relevantne situacije uspostavljanja rednog električnog luka, i to za različite kombinacije materijala elektroda (bakar–bakar i bakar–mesing). Registrovano je pravovremeno reagovanje AFDD komponente u svakom od tih slučajeva. Takođe, realizovani su eksperimenti pomoću kojih je konstatovano da AFDD komponenta nije reagovala pri normalnom radu nelinearnih prijemnika koji sadrže energetsku elektroniku (npr. personalni računar, štampač, mikrotalasna pećnica, klima-uređaj itd.), kao i pri normalnom radu prijemnika koji sadrže kolektorske motore (npr. usisivač, bušilica, fen itd.), iako je za njihov rad normalno postojanje električnih lukova. Na osnovu rezultata tih eksperimenata izrađena je preporuka za instaliranje AFDD komponente u objektima u kojima se okuplja veliki broj ljudi, odnosno dato je mišljenje o adekvatnosti njene upotrebe u Republici Srbiji.

Reliable functioning of low-voltage electrical installations represents an important prerequisite for normal functioning of technological processes in the industry, normal operation of public institutions (schools, theaters, museums, hospitals,...), proper operation of households, etc. Due to poor selection or inadequate installation, improper use or maintenance of electrical installations, failures can occur causing fire, electric shock or interruption in power supply of a facility (or its part). Special attention must be given to fire-causing failures, which cannot be detected by conventional protecting devices (such as: miniature circuit breakers, fuse-links, residual current devices, etc.). Such failures are most often caused by poor electrical contacts. Poor electrical contacts in low-voltage electrical installations can occur in the distribution cabinet (switchboard) or anywhere in the electrical circuit (on damaged conductors, in plugs or receptacles, in extension cords, in electrical loads, etc.). If such a failure is not resolved, the temperature of the poor contact increases gradually, which can lead to melting or ignition of the insulation of an electrical component and the surrounding flammable material. In an electrical installation, a poor electrical contact can be considered as characterized by: - a reduced pressure force between the contact surfaces of the copper conductor and the screw of the electrical component terminal (reduction of the torque), - an incomplete overlap of the surface of the copper conductor and the contact surface at the electrical component terminal (reduction of the contact surface), or - an increased oxide layer at the point of electrical contact (old and/or corrosion- damaged contact). Neither local nor international regulations and standards, which define periodic verification of the quality of low-voltage electrical installations, contain procedures by which a poor electrical contact would be detected at an early stage. However, one of the techniques used for this purpose today is performing preventive thermographic inspections (a standard covering this area is available only in the USA). By such inspections a poor electrical contact which creates prerequisites for the occurrence of an initial fire in a low-voltage electrical installation can easily and effectively be detected. Detected failures and irregularities can most frequently be eliminated by simple interventions of electricians or facility technical services. The main disadvantage of the thermographic inspection is the inability to inspect electrical contacts which are not directly (visually) accessible to the camera sensor. The aim of the doctoral dissertation is the analysis of the impact of poor electrical contacts on the occurrence of an initial fire and the development of a method for their early detection in low-voltage electrical installations. Based on a large number of experiments carried out with various combinations of materials from which contacts are made (copper–copper, copper–brass and copper–stainless steel), by a specially designed apparatus a poor electrical contact occurrence due to an incomplete overlap of a copper conductor and a corresponding element of an electrical device was simulated. It was concluded that copper–stainless steel represents a critical combination of materials for which a contact most rapidly reaches the maximum permissible temperature under the same conditions (the same load, conductor cross-section and contact deterioration degree in terms of the same percentage reduction of the overlap surface). Based on a large number of experiments, the impact of the reduction of the torque of tightening of the screw in the electrical component on the occurence of a „hot“ spot in plugs and receptacles was also determined. It was concluded that poor electrical contacts developed in built-in electrical installations of buildings (where the copper conductor of the full cross-section is common) are more unfavourable for the occurrence of a „hot“ spot compared to poor electrical contacts developed in power cables of electric loads (where the stranded copper conductor is commonly used). Based on a large number of experiments in which the impact of poor electrical contacts on the occurrence of „hot“ spots in receptacles was analysed, as well as on the data obtained from a large number of documents related to periodic verifications of the quality of low-voltage electrical installations in industrial and administrative facilities (issued by the Laboratory for testing low-voltage electrical and lightning protection installations at the Faculty of Electrical Engineering in Belgrade), a new method for detecting poor electrical contacts has been developed. It was based on the establishment of a correlation between the measured line to earth and line to neutral short circuit loop resistances and the degree of the contact deterioration, i.e. the increase of its electrical resistance. A number of experiments were performed on a receptacle that has been used for many years (30 years in the analysed case), as well as on new receptacles, the electrical contacts of which were intentionally exposed to corrosion prior to the experiments. In order to accelerate the corrosion process, in one of the new receptacles the connection of the phase conductor was treated with aqua regia (concentrated HNO3 : concentrated HCl = 1: 3 (v / v)), and in the other the connection of the neutral conductor was treated with concentrated sulfuric acid (H2SO4). Using these chemicals the effect of long-term exposure of electrical contacts of electrical components to an acidic atmosphere was simulated. Comparing the results of the measurements aimed to analyze the effects of reducing the screw torque tightening at the receptacle terminals of the old (used) receptacle and two new receptacles (with deliberately corroded connections) on heating and electrical resistance of electrical contacts, it was shown that they possess similar characteristics. In the available domestic and foreign standards and scientific literature, with the exception of several patents intended for industrial and special applications, methods and techniques that can be applied for the timely detection of poor electrical contacts in residential, commercial and public buildings have not been presented. Applying the method developed in this dissertation, intended for the detection of poor electrical contacts in electrical circuits, the probability of maintaining poor electrical contacts and dangerous „hot“ spots that may cause fire would be reduced (the method would be applied as part of the regular procedure for checking the efficiency of the protective measure based on the automatic switching off of the power supply in case of failure - by measuring both the line to earth and the line to neutral short circuit loop resistances). This would overcome the problem of the impossibility of a thermal imager to detect „hot“ spots in poor contacts that do not have direct visual contact with the camera sensor. The proposed method for detecting poor electrical contacts was applied in performing the verification of the quality of low-voltage electrical installations in old (10–60 years) residential buildings. The possibility of establishing a relation between the age of the residential buildings (more precisely, the age of their electrical installations) and the probability of the occurrence of a dangerous „hot“ spot in a poor electrical contact in the electrical circuits of single-phase and three-phase loads has also been explored. Unfortunately, this relation has not been established. Its establishment would require to apply the new method for the detection of poor electrical contacts in thousands of receptacles in apartments of different ages. An undetected poor electrical contact in an electrical installation may cause a series electric arc, which is a common cause of fire. Opposite to poor electrical contacts, series electric arcs can be detected using modern protecting devices. The functionality of the AFDD (Arc-Fault Detection Device), intended for the European market, the primary purpose of which is the protection against unwanted series arc, has been experimentally tested in laboratory conditions. The functionality test was based on a large number of laboratory experiments, by which almost all relevant situations of establishing a series electric arc for various combinations of materials of electrodes (copper–copper and copper–brass) have been simulated. In each of these cases, a timely response of the AFDD component has been registered. In addition, experiments aimed to check if the AFDD component does not operate during the normal operation of non- linear electrical loads containing power electronics (for example, personal computers, printers, microwave ovens, air conditioners, etc.), as well as loads containing collector motors (e.g., vacuum cleaners, drills, fans, etc.), were also successful. Based on the results of these experiments, a recommendation has been made for the installation of AFDD components in buildings in which a large number of people gather, i.e. an opinion on the adequacy of their application in the Republic of Serbia has been given.

Promene društvenih odnosa u poljoprivredi Vojvodine i njen razvoj 1950-1957. godine

Promene društvenih odnosa u poljoprivredi Vojvodine i njen razvoj 1950-1957. godine

Agrarna politika vlasti u periodu od 1945-1950. godine je shodno ideološkim određenjima, bila usmerena na promenu društvenih odnosa na selu. Međutim, osim ideoloških razloga ne treba zanemariti činjenicu da je usitnjen zemljišni posed već ispoljio mnoge slabosti i da su u uslovima uvećane potražnje za poljoprivrednim proizvodima ti problemi dodatno ubedili vlast u neophodnost koncentracije zemlje i formiranja krupnih, pre svega, zadružnih gazdinstava. U cilju realizacije tog cilja, primenjene su rigidne ekonomske i političke mere, naročito na prostoru Vojvodine među kojima je bilo i brutalnih metoda izrazito oprečnih tradicionalnim vrednostima i etičkom kodeksu očuvanom u seoskim sredinama. Ovakva agrarna politika izazvala je veliko nezadovoljstvo seljaštva i veliki pad poljoprivredne proizvodnje. Razni vidovi opstrukcije velikog dela stanovništva prema državi i strah od gladi u jeku sukoba sa SSSRom, ubedili su partijsko rukovodstvo zemlje u neophodnost promene dotadašnjeg odnosa prema poljoprivredi i seljaštvu. Mada se nije lako odustajalo od politike kolektivizacije, gore navedeni razlozi doveli su do njenog napuštanja već 1950. godine, a zadružni sistem od 1953. godine transformisan je u ekonomski racionalnije organizacije. Ove promene, nisu značile i kraj primene revolucionarnih mera na selu. One su posebno naglašene 1953. godine ograničenjem privatnog poseda na 10 hektara po domaćinstvu. Ipak, agrarna politika KPJ od tada se razvija na principima većeg uvažavanja seljaštva i uspostavljanja ekonomske saradnje sa privatnim posedom. Novu politiku karakteriše dobrovoljnost kooperacije i značajnije ulaganje u prerađivačku industriju i poljoprivredu proizvodnju.

Agrarian policy of the Government in the period from 1945 to 1950, according to ideological determinations, was pointed to the changes of social relations in a village. However, beside ideological reasons, it should not be neglected the fact that fragmented land property had already shown many weaknesses and that in the circumstances of higher requests for agricultural products, those issues persuaded the Government in the necessity of gathering land and forming large collective economies. In order to realize that goal, rigid economic and political measures were taken, especially in the territory of Vojvodina, some of them very brutal, totally opposite to traditional values and ethical codex preserved in rural areas. This agrarian policy caused great discontent among peasantry and huge decline of agricultural production. Different ways of obstruction by people towards the state and the fear of hunger in the middle of the conflict with Soviet Union (USSR), persuaded the Party leader-ship of the State in the necessity of changing the relation towards agriculture and peasantry. Even though it was not easy to give up the policy of collectivization, reasons mentioned above brought to it's abandonment in 1950, and collective economy sys-tem was transformed in 1953 into economically more rational organizations. Those changes did not led to the end of use revolutionary measures in a village. They were even emphasized in 1953 by limiting the private properties to 10 hectares per house-holds. However, agrarian policy since that period was developing on the principles of higher respect of the peasantry and establishing cooperation with private property. New policy was characterized by willingly cooperation and significant investment in manufacturing industry and agrarian production.

Sinteza optimalne strukture reverzibilnog mehanizma laminatora primenom aksiomatskog projektovanja

Optimal structure synthesis of reversible mechanism of laminator using axiomatic design

Slojevanje ili laminacija testane trake je jedna specifična operacija koja postoji kod linija za proizvodnju tvrdog keksa i krekera što je drugačije u odnosu na ostalu konditorsku opremu. Tom operacijom formira se višeslojna struktura testa koja ima višestruko pozitivan efekat na krajnji proizvod. Umiruju se naponi u testanoj traci izazvani stanjivanjem trake, dobija se vazdušasta struktura od mehurića zarobljenih između susednih slojeva (povećan volumen) i pretapaju se slojevi testa dajući uniformnu traku. Kvalitet laminiranja testa prvenstveno zavisi od kvaliteta same mašine – laminatora i njenog uređaja za slojevanje, odnosno reverzibilnog mehanizma koji treba da obradi ulazne sirovine (brašno, ulje, masnoća, aditivi), bez obzira na odstupanja njihovog kvaliteta. Cilj ove disertacije je da se metodološkim pristupom kreira takva klasa i struktura mehanizma sa integrisanim pogonom koja će ostvariti zahtevano kretanje, a kao krajnji rezultat dati dobro formiranu, slojevitu testanu traku.Uprkos činjenici da testo može biti promenljivog viskoziteta i adhezivnosti proces kontinualnog slaganja i presavijanja mora kontinualno da se odvija uz istovremeno oslobađanje indukovanih napona. Ovi zahtevi mogu da se ostvare uspešno sintetizovanom strukturom i pogonom reverzibilnog mehanizma nove generacije. Dosadašnja istraživanja i studije o kretanju, oblikovanju i slojevanju testane trake, a sa aspekta optimalne sinteze strukture i pogona izvršnih mehanizama ne pružaju mogućnost neposredne sinteze novog rešenja. Radi pronalaženja kvalitetnijeg rešenja potrebno je bilo primeniti alate savremenih metoda, kao na primer aksiomatsko projektovanje. Primenom metodologije aksiomatskog projektovanja, veoma jasno, parametarski, i kvantifikovanim kriterijumima uspešno se definišu tehnološki zahtevi i projektni prostor novog rešenja. V Pošto se radi o oblasti konstruisanja prehrambeno - konditorskih mašina što se istražuje unutar fabričkih razvojno - istraživačkih centara, to je dostupan veoma mali broj publikovanih ostvarenja. Najznačajnija dostupna dokumentacija vezana je za patentna rešenja.Tu je definisan primarni prostor postojećih rešenja i slobodni prostor za sintezu novog rešenja koje neće zadirati u postojeće patentne zahteve. U tom smislu, razmatranjem već primenjenih rešenja uređaja za slojevanje, izabran je cilj istraživanja čijim realizovanjem se omogućava iznalaženje originalnog rešenja koje treba da predstavlja tehnološki napredak u odnosu na postojeće stanje u ovoj tehnici. Na svetskom tržištu postoje linije za proizvodnju keksa i krekera kod kojih se koriste različita rešenja mehanizama za laminaciju testane trake. Višegodišnjim ispitivanjem pokazalo se da nijedno od tih rešenja ne može da da izlazni proizvod visokog kvaliteta za ulazne sirovine koje variraju u širokom dijapazonu. Za razliku od tih poznatih rešenja novokonstruisani reverzibilni mehanizam za laminiranje pokazao je izuzetnu fleksibilnost i primenjivost kod sirovina slabijeg kvaliteta. Tim je otvoren širok prostor za razvoj, primenu i plasman ove klase reverzibilnog mehanizma za slojevanje. Zaključuje se da je potpuno originalno postavljena metodologija za sintezu optimalne strukture reverzibilnog mehanizma, da je razvijeno primenjivo rešenje, da je definisan 3D model i izvršena virtualna verifikacija funkcionalnosti, da je zatim prema originalnoj radioničkoj i sklopnoj dokumentaciji izrađeno prototipsko rešenje i da je kroz operativno ispitivanje dokazana potpuna validnost postavljenog rešenja.

Layering or lamination of a dough strip is a specific operation that appears in product lines for hard biscuits and crackers, which is different compared to other conditory equipment. This operation forms a multilayered structure of the dough that has multiple positive effects to the final product. Thus, the dough strip stresses induced by the strip thinning are soothed, the airy structure, created by the bubbles trapped between adjacent layers, is obtained (increased volume) and the dough layers are blended giving a uniform dough strip. The quality of a laminated dough depends primarily on the quality of the machine – the laminator and its layering device, i.e. reversible mechanism which processes the incoming raw materials (flour, oil, grease, additives), regardless of the quality differences. The aim of this thesis is the creation of such a class and structure of mechanism with integrated drive by a methodological approach that will achieve the required movement and give the well formed layered dough structure. Despite the fact that dough viscosity and adhesiveness can vary, the stacking and folding process must proceed continually with the releasing the induced stresses at the same time. These requirements can be achieved successfully by the synthesized structure and reverse drive mechanism of the new generation. Previous research and studies of the motion, forming and layering of a dough strip do not provide the possibility of direct synthesis of a new solution in terms of optimal synthesis of the structure and drive of the executive mechanisms. In order to find better solutions, it was necessary to apply the tools of modern methods, such as axiomatic design. By the applying the axiomatic design methodology, the technological requirements and the design space of a new solution are defined successfully, clearly, parametric and by the quantifiable criteria. Since this is a field of constructing food - conditory machines, which are explored within the factory development and research centers, a very small number of published VII works, are available. The most important accessible documents are related to the patent solutions. The primary space of the existing solutions is defined here, as well as the free space for the synthesis of a new solution that will not interfere with existing patent requirements. Regard this, by the consideration of implemented solutions for layering device, the research goal was chosen, which realization enables the finding the original solution that would present a technological improvement over the current situation in this technique. There are lines for the production of cookies and crackers on the world market based on the use of different mechanism solutions for the dough strip lamination. Long term study showed that none of these solutions can deliver the high quality product output for the input raw materials which vary in a wide range. In contrast to these known solutions, the newly-designed reversible mechanism for lamination demonstrated exceptional flexibility and applicability for lower quality raw materials. On this way, the wide space for the development, implementation and marketing of this class of layering reversible mechanism is opened. It is concluded that the set of methodologies for the synthesis of optimal structure of the reversible mechanism is founded completely original, the applicable solutions has been developed, 3D model has been defined and virtual functionality verification has been realized, then, the prototype solution is accomplished after original technical documentation, and the complete validity of the solutions is proved by the operational test.

Optimizacija problema upravljanja odnosima koristi i troškova pri raspoređivanju projekata primenom metaheurističkih algoritama

The optimization of the benefit/costs tradeoff during the scheduling of the projects by applying metaheuristic algorithms.

U ovoj doktorskoj disertaciji razmatrano je modeliranje procesa odlučivanja u prisustvu neizvesnosti. Analizirana su dva tipa problema jedan je problem optimizacije odnosa korist/trošak pri raspoređivanju projekata, a drugi je klasifikacija podataka opisanih atributima među kojima neki nedostaju. Osnovni problemi pri modeliranju procesa odlučivanja u prisustvu neizvesnosti su izbor adekvatnog tretmana neizvesnosti i izbor metode za donošenje odluke. Jedan od ciljeva disertacije je ispitivanje pogodnosti primene metaheurističkih algoritama na razmatrane optimizacione probleme. Glavna mera evaluacije njihovih performansi se sastoji od vrednosti funkcije cilja (bilo da je u pitanju odnos korist/trošak pri raspoređivanju projekata ili klasterovanje). Kod problema raspoređivanja projekata može se razmatrati i nivo satisfakcije za ograničenja koja karakterišu problem (tzv. nivo praga). Druga metoda evaluacije primenjenih metaheurističkih metoda je vreme potrebno da bi se pronašlo rešenje. Diskutuje se i o parametrima koji kontrolišu algoritme metaheurističkih metoda i o pravilnom izboru njihovih vrednosti kako bi se postigle maksimalne performanse implementacija na testiranim primerima razmatranih problema. Kod problema optimizacije odnosa korist/trošak neizvesnost je modelirana primenom trougaonih fazi brojeva, a zatim su primenjene metaheurističke metode simulirano kaljenje i genetski algoritam za rešavanje tako dobijenog fazi optimizacionog problema. Testirani problemi formulisani su metodom fazifikacije koja je predložena od strane (Ribeiro et al. 1999.). Predstavljeni eksperimentalni rezultati za niz fazi problema pokazuju efikasnost primenjenih metoda simuliranog kaljenja i genetskog algoritma. Neznatno kvalitetnija rešenja dobijena su primenom genetskog algoritma, a poboljšanje u odnosu na prethodne rezultate iz literature u oba slučaja je oko 20%. Drugi deo disertacije se bavi problemom klasterovanja podataka sa nedostajućim vrednostima atributa i donošenja odluka u takvim okolnostima. Glavne faze u rešavanju postavljenog problema su nalaženje najpogodnijeg rastojanja koje će se koristiti u slučajevima kada podaci iz nekog razloga nedostaju i izbor metode za rešavanje samog problema klasterovanja. U cilju teorijskog i praktičnog doprinosa predložena je metrika koja se zasniva na logici. Primenom verovatnoće dokazana je teorema kojom se određuju vrednosti težinskih koeficijenata atributa koji opisuju objekte za klasterovanje. Sve predloženo je implementirano kroz metaheurističku metodu promenljivih okolina, kao i neke njene modifikacije, i primenjeno na realne i probleme iz literature. Klasifikacijom pacijenata koji boluju od autoimunih bolesti, na osnovu baze podataka Kliničkog centra Srbije, postignuta je preciznost klasterovanja od 93.33%. Kao drugi realan primer, analizirano je sedam baza evropske komisije koje sadrže podatke vezane za poštanski saobraćaj. Postignuta je preciznost klasterovanja od 90%-96.96%. Da bi se uporedila efikasnost pristupa zasnovanih na metodi promenljivih okolina korišćeno je i devet baza dostupnih na internetu i dobijeni rezultati poređeni su sa postojećima iz literature. Eksperimenti su pokazali veliku stabilnost metode promenljivih okolina, u osam od devet slučajeva najbolje rešenje postignuto je u svih sto izvršavanja. Osim toga, kvalitet dobijenih rešenja znatno je premašio rezultate iz literature.

In this doctoral dissertation the modelling process has been taken into consideration in the presence of uncertainty. Two types of problems were analyzed: one is the optimization of the benefit/costs tradeoff during the distribution of the projects and the other is the classification of data described by the attributes among which some are missing. The basic problems during the modelling of the decision making in the presence of uncertainty are the choice of the adequate treatment of uncertainty and the choice of the method for making a decision. One of the aims of the work is investigating the benefits of applying the metaheuristic algorithms on the considered optimization problems. The main measure for the evaluation of their performances is the value of objective function (for both problems: optimization of benefit/costs tradeoff during the project scheduling and clustering of incomplete data). Considering the project scheduling problem the level of satisfaction related to the problem constraints could also be taken into account. The other evaluation criteria of the applied metaheuristic methods is the time required for finding the solution. The influence of the parameters which control the algorithms of the metaheuristic methods is examined, as well as their appropriate values leading to the maximum performances of the implementation could be reached on the tested examples of the considered problems. As for the optimization problem of the profit/costs tradeoff, the uncertainty is modelled by applying the triangle fuzzy problems and then the metaheuristic methods, simulated annealing and genetic algorithm were applied for solving the obtained fuzzy optimization problem. The tested problems are formulated by the fuzzification method which was suggested by (Ribeiro et al. 1999). The represented experimental results for the set of fuzzy problems show the efficiency of the applied methods: simulated annealing and genetic algorithm. Genetic algorithm seems to produce slightly better solution than the simulated annealing. However, both methods out performed the existing form the literature for about 20%. The secund part of the work deals with the clustering data problem with the missing values of the attributes and making decisions in such circumstances. The main phases in solving the considered problem are finding the most appropriate distance, which will be used in the cases when the data are missing for some reasons and choosing the method for solving the clustering problem. As the theoretical and practical contribution, the metric, based on the logic principles, was proposed. By applying the probability, the theorem was proved defining the values of the weighting coefficients related to attributes that describe the objects for clustering. The proposed metric was implemented in the variable neighborhood search metaheuristic method as well as in some of its modifications. The implemented methods have been applied on the real life problems from the literature. Classifying the patients who suffer from some auto-immune diseases, stored in the database of Clinical Centre of Serbia, the precision of the clustering of 93.33% was achieved. As another real life example, seven databases of the European Commision (Board), which contain the data for the mail service, have been analyzed. The clustering efficiency of 90% - 96.96% was achieved. In order to compare the efficiency of the approach based on the variable neighborhood search method, nine databases available on the internet were used and the obtained results were compared with the existing ones from the literature. The experiments showed large stability of variable neighborhood search method: in eight out of nine cases the best solution was reached in all hundred repetitions. Besides that, the quality of the obtained solutons have considerably surpassed the results from the literature.

Statistička obrada signala za dijagnostiku grešaka i prediktivno održavanje sistema

Statistical signal processing for fault diagnosis and predictive system maintenance.

Cilj istraživanja predstavljenog u ovoj tezi je razvoj novih metoda za de- tekciju i izolaciju otkaza u sistemima koji se mogu opisati opštim, neline- arnim, stohastičkim modelima u prostoru stanja. Literatura obiluje rezul- tatima vezanim za linearne sisteme. Sa druge strane, broj dostupnih rešenja primenjivih na nelinearne sisteme znatno je manji, uprkos tome što se oni često sreću u oblastima hemijske procesne industrije, avionike, aeronautike, robotike, biomedicine itd. Među ovim rešenjima su najbrojniji pristupi zasnovani na nelinearnim opserverima. Osim što je prateća matematička teorija veoma komplikovana, rezultati do kojih je moguće doći najčešće se od- nose samo na određene uske klase nelinearnih sistema. Alternativni, opštiji pristupi koriste proširene i unscented Kalmanove filtre. Međutim, ove tehnike podrazumevaju linearizaciju originalnog modela i/ili pretpostavku o normalnosti raspodela, i ne garantuju konvergentnost niti optimalnost u opštem slučaju. U okviru ove teze su kao osnovni alati korišćene metode za sekvencijalno testiranje hipoteza, i rekurzivna Monte Karlo tehnika za estimaciju stanja nelinearnih sistema, u literaturi poznata pod nazivom čestični filtar. Prednosti u odnosu na prošireni i unscented Kalmanov filtar su: 1) za- garantovana asimptotska konvergentnost, kao i 2) optimalnost u smislu mi- nimalne varijanse estimacije stanja koju je moguće ostvariti pod određenim uslovima. Osim toga, čestični filtar ne koristi linearizaciju, niti raspo- dele od interesa moraju biti normalne, pa je primenjiv na stohastičke modele potpuno proizvoljnog oblika. Osnovni naučni doprinos čine tri nova algoritma za dijagnozu otkaza. U njima se kombinuju neki dobro poznati koncepti prvobitno razvijeni za li- nearne sisteme, i savremena teorija čestičnog filtriranja. Osim toga, ra- zvijena je nova, robusna metoda za sekvencijalno testiranje višestrukih hipo- teza. Predložena rešenja su temeljno verifikovana kroz iscrpne računarske simulacije. Kao test-primeri su korišćeni modeli standardnih nelinearnih procesa često sretanih u literaturi. Posebna pažnja je posvećena analizi mogućnosti primene razvijenih algoritama u praksi. iv Dobijeni rezultati svedoče o efikasnosti i pouzdanosti predloženih me- toda. Osnovne prednosti u odnosu na slična postojeća rešenja su: 1. intuitivnost, koja proizlazi iz korišćenja standardnih, dobro pozna- tih i opšte prihvaćenih koncepata originalno razvijenih za linearne sisteme, 2. numerička jednostavnost, omogućena svođenjem opšteg problema detek- cije i izolacije na relativno malu grupu otkaza od najvećeg praktičnog interesa, ili upotrebom tehnika poput marginalizacije, koja omogućava smanjenje broja potrebnih računskih operacija, i 3. praktična upotrebljivost, koja je ostvarena tako što su od najranijih etapa razvoja razmatrane neizbežne pojave poput dejstva poremećaja i grešaka modeliranja. Ovime je unapređena oblast dijagnoze otkaza u sistemima opisanim opštim nelinearnim stohastičkim modelima u prostoru stanja. Osim toga, razvi- jene metode generišu pouzdane procene stvarnih stanja sistema, bez obzira na eventualno prisustvo otkaza u senzorima ili aktuatorima. Time su omogu- ćeni upravljanje u zatvorenoj sprezi tolerantno na otkaze, kao i prediktivno održavanje sistema.

The purpose of the research presented in this thesis is to develop new methods of detecting and isolating faults in systems which can be described by general, nonlinear, stochastic state-space models. Numerous results pertaining to linear systems are available in the literature. On the other hand, the number of solutions applicable to nonlinear systems is significantly smaller, in spite of their frequent occurrence in the fields of chemical process industry, avionics, aeronautics, robotics, biomedicine etc. The most common of these solutions are based on the use of nonlinear observers. However, the underlying mathematical theory in such ap- proaches is rather complex. Moreover, the resulting algorithms are usually applicable only to certain narrow classes of nonlinear systems. Alternative, more general ap- proaches employ extended and unscented Kalman filters. Their main drawbacks are the use of linearisation of the original model and/or the required assumption of the normality of probability distributions. Consequently, no convergence or optimality results are available in the general case. The main tools used within this thesis are sequential hypothesis testing methods and a recursive Monte Carlo technique for state estimation, known in the literature as the particle filter. The advantages with respect to the extended and unscented Kalman filters are: 1) guaranteed asymptotic convergence and 2) optimality in the sense of minimum state estimation variance, which can be achieved under certain conditions. In addition, the particle filter does not rely on linearisation, nor does it require the probability distributions of interest to be Gaussian. As a results, it is applicable to completely general stochastic state-space models. The main scientific contribution consists of three new fault diagnosis algorithms. These combine some well known concepts originally developed for linear systems, and modern particle filtering theory. Moreover, a new, robust method for multiple hypothesis testing was developed. The proposed solutions were thoroughly verified through exhaustive computer simulations. Models of benchmark nonlinear plants, frequently encountered in the literature, were used as test cases. Special attention was devoted to the analysis of the possibilities for the practical implementation of the developed algorithms. The obtained results testify to the efficiency and reliability of the proposed vi methods. The key advantages with respect to similar existing solutions are: 1. intuitiveness, which stems from the use of standard, well known and generally accepted concepts, originally developed for linear systems; 2. numerical simplicity, enabled by the reduction of the general fault diagnosis problem to a small group of faults of greatest practical interest, or by the use of techniques such as marginalisation, which enables the number of required computations to be reduced; 3. practical usability, which was achieved by considering from the earliest stages the unavoidable phenomena such as disturbances or plant-model mismatches. Our findings advance the filed of fault diagnosis in systems described by general stochastic state-space models. In addition, the developed algorithms generate reli- able estimates of the true model states, even in the presence of possible sensor or actuator faults. This paves the way for fault-tolerant feedback control and predictive system maintenance.

Prostiranje svetlosti u kompleksnim fotonskim rešetkama sa zasićujućom nelinearnošću

Propagation of light in complex photonic lattices with saturable nonlinearity

U ovoj disertaciji teorijski i eksperimentalno su analizirani linearni i nelinearni efekti koji prate prostiranje vidljive laserske svetlosti kroz različite jednodimenzionalne (1D) fotonske rešetke napravljene od materijala sa zasićujućom (saturacionom) nelinearnošću, kao što je, na primer, litijum niobat (LiNbO3). Zbog nelinearnog odziva kakvim se odlikuju ovakvi materijali, moguće je formiranje stabilnih prostornih lokalizovanih struktura na granicama između dve uniformne rešetke istih, odnosno različitih perioda, kao i unutar binarnih superrešetki, kako u nelinearnom režimu, tako i u linearnom kada do lokalizacije dolazi u okolini defekta. U doktorskoj tezi razvijeni su odgovarajući matematički modeli i primenjene različite numeričke metode za dobijanje odgovarajućih rezultata, dok su neki od rezultata potvrđeni i eksperimentalno u nelinearnim 1D fotonskim rešetkama proizvedenim u LiNbO3. Korišćeni matematički modeli zasnovani su na sistemima spregnutih diferencno-diferencijalnih jednačina, tačnije 1D diskretnim nelinearnim Šredingerovim (Schrödinger) jednačinama sa nelinearnostima Kerovog (Kerr) i zasićujućeg tipa, dok je stabilnost dobijenih rešenja ispitivana primenom metoda linearne analize stabilnosti. Za rešavanje stacionarnih i dinamičkih jednačina izloženih u tezi, korišćene su numeričke metode pod nazivom Gaus-Njutnova metoda (Gauss-Newton) i Runge-Kuta (Runge-Kutta) metoda, respektivno. Dobijeni numerički i eksperimentalni rezultati pokazuju da prisustvo defekta utiče na formiranje lokalizovanih modova kako u nelinearnom, tako i u linearnom režimu, tj. pri malim upadnim snagama nedovoljnim za ispoljavanje nelinearnog odziva sredine. Rezultati pokazuju da pored pojave linearnih lokalizovanih stanja, u nelinearnom režimu dolazi do narušavanja simetrije lokalizovanih rešenja kada su ispunjeni odgovarajući uslovi. U slučaju prostiranja svetlosti kroz binarne superrešetke, rezultati ukazuju na postojanje novih tipova lokalizovanih rešenja, kao i otvaranje dodatnog procepa u zonskoj strukturi posmatrane rešetke. Pored ovoga, prikazani su i numerički rezultati dobijeni analizom međusobnih interakcija solitonskih rešenja, kao i ispitivanjem osobina modova lokalizovanih na površini rešetke.

In this dissertation the linear and nonlinear effects accompanying the propagation of visible laser light through a different one-dimensional (1D) photonic lattices made of material with a saturable nonlinearity, such as, for example, lithium niobate (LiNbO3) are both investigated, theoretically and experimentally. Due to the nonlinear response of these materials to the intensity of incident radiation, the light passing through the photonic lattice causes local change in the refractive index, providing necessary conditions for the formation of stable localized spatial structures at the interfaces between two uniform lattices of the same or different periods, and also within the binary superlattices, in the nonlinear regime, as well as in the linear regime when the localization occurs in the vicinity of the defect. In this dissertation, appropriate mathematical models are developed and various numerical methods for obtaining the relevant results are applied. Some of the results have been confirmed experimentally in 1D nonlinear photonic lattices fabricated in LiNbO3. Mathematical models are based on systems of coupled difference- differential equations, namely the 1D discrete nonlinear Schrödinger equation with the nonlinear term of Kerr and saturable type, while the stability of the obtained solutions was investigated using the method of linear stability analysis. To solve stationary and dynamic equations presented in this dissertation, numerical methods such as the Gauss-Newton and Runge-Kutta methods are performed, respectively. The obtained numerical and experimental results show that the presence of the defect affects the formation of localized modes in the nonlinear, as well as in the linear regime, i.e. for intensities insufficient for the manifestation of the nonlinear response of the medium. Furthermore, the results indicate to the possible appearance of the symmetry breaking of the mode profiles of certain nonlinear localized solutions when appropriate conditions are met. In the case of light propagation through binary superlattices, results show the emergence of new types of localized solutions and the opening of an additional gap in the observed energy spectrum of the lattice. Additionally, numerical results obtained by analyzing the mutual interaction of soliton solutions, as well as examining the properties of localized modes at the surface of the lattice, are presented as well.

Povećanje efikasnosti metoda za merenje intenziteta električnog polja u okolini baznih stanica javnih mobilnih sistema

Improving the efficiency of methods for measurement of electric field strength in the vicinity of public mobile system base stations

Rastući zahtevi korisnika za pristupačnim i visokokvalitetnim telekomunikacionim servisima doveli su do intenzivnog razvoja sistema i uređaja za bežične komunikacije. Zbog prostorne rasprostranjenosti naročito se ističu javni mobilni sistemi (GSM - Global System for Mobile Communications, DCS - Digital Communication System, UMTS - Universal Mobile Telecommunication System, LTE - Long-Term Evolution). Sa obzirom da ovi sistemi kao osnov za međusobnu komunikaciju koriste elektromagnetne talase, javlja se potreba za objektivnim sagledavanjem nivoa elektromagnetnog zračenja u životnom okruženju. Da bi se uticaj elektromagntetnog zračenja na ljude kontrolisao, usvojene su norme kojima se ograničava izlaganje, iskazane kroz bazična ograničenja i izvedene referentne granične nivoe. U okviru ove disertacije dat je pregled bazičnih ograničenja i referentnih graničnih nivoa definisanih Pravilnikom o granicama izlaganja nejonizujućim zračenjima, ICNIRP (International Commission on Non-Ionizing Radiation Protection) preporukom, FCC (Federal Communications Commission) preporukom, IEEE (Institute of Electrical and Electronics Engineers) standardom i ARPANSA (Australian Radiation Protection and Nuclear Safety Agency) standardom. Da bi se ispitala usklađenost instalacije bazne stanice sa referentnim graničnim nivoima, sprovode se merenja intenziteta elektromagnetnog polja u okolini baznih stanica, korišćenjem metoda definisanih međunarodnim standardima. Ove metode se razvijaju i nadograđuju u skladu sa razvojem telekomunikacih sistema, kao i poboljšanjima mernih uređaja. Važan segment u razvoju ovih metoda, koji je posebno aktuelan u poslednje vreme, je određivanje merne nesigurnosti rezultata merenja. U ovoj disertaciji, predložena je praktična metoda za procenu merne nesigurnosti rezultata merenja intenziteta električnog polja, za slučajeve širokopojasnog mernog sistema i frekvencijski selektivnog mernog sistema. Ova metoda bazirana je na zakonu propagacije Doktorska disertacija Elektrotehnički fakultet Univerziteta u Beogradu iii nesigurnosti i omogućava procenu kombinovane standardne merne nesigurnosti rezultata merenja, kao i proširene merne nesigurnosti sa intervalom poverenja od 95%. Cilj ove disertacije bio je povećanje efikasnosti metoda za merenje intenziteta elektromagnetnog polja u okolini baznih stanica koje je ostvareno unapređenjima u nekoliko segmenata. U okviru disertacije, data je procedura za povećanje efikasnosti skraćivanjem intervala merenja (u odnosu na interval od 6min propisan standardima). Na osnovu rezultata merenja sprovedenih u okolini multisistemske bazne stanice, pokazano je da se usrednjavanjem u kraćim vremenskim intervalima od 1min, 30s i 15s mogu uspešno proceniti vrednosti dobijene standardnim postupkom, pri čemu se merni proces ubrzava približno 6, 12 i 24 puta, respektivno. Naravno, u ovim slučajevima u obzir mora biti uzeta dodatna standardna merna nesigurnost vrednosti 3.07%, 3.80% i 4.42% za intervale usrednjavanja 1min, 30s i 15s, respektivno. Kao što je pokazano, uprkos činjenici da su korišćeni intervali usrednjavanja kraći od zahtevanih 6min, predložena merna procedura obezbeđuje rezultate sa prihvatljivim gubitkom kvaliteta merenja. U slučaju sistematskih merenja, često se koristi veliki broj pojednostavljenih mernih sistema koji koriste monoaksijalne antene. U disertaciji su sprovedena istovremena merenja intenziteta sve tri prostorne komponente električnog polja, za sedam različitih tipova okruženja, i na bazi ovih rezultata razvijen je model za konverziju rezultata merenja intenziteta jedne prostorne komponente električnog polja, u intenzitet koji se dobija izotropnim merenjima. Po ovom modelu, pri korišćenju monoaksijalne merne sonde za odredivanje ukupnog intenziteta električnog polja, potrebno je primeniti dodatni mulitplikativni faktor konverzije vrednosti 1.95, i potrebno je uzeti u obzir dodatnu standardnu mernu nesigurnost vrednosti 33.07%. Pored dobro poznatog kratkotrajnog fedinga, na promenljivost intenziteta elektromagnetnog polja baznih stanica, direktno utiču saobraćajno opterećenje, automatska kontrola snage i diskontinualna predaja. U okviru disertacije, sprovedena je analiza dugoročne promenljivosti intenziteta električnog polja, u vremenskom intervalu od 7 dana. Za potrebe analize, sprovedena su band-selektivna merenja u realnom mrežnom okruženju na 7 urbanih lokacija. Pored toga, određena je dodatna merna nesigurnost koja potiče od promenljivosti saobraćajnog opterećenja i funkcionalnosti predajnika, za raličite intervale usrednjavanja: 10s, 30s, 1min, 6min, 15min, 30min, 1h, Doktorska disertacija Elektrotehnički fakultet Univerziteta u Beogradu iv 3h, 5h i 10h. Rezultati pokazuju da se nesigurnost smanjuje sa povećanjem intervala usrednjavanja. Poređenjem opsega standardne merne nesigurnosti za različite sisteme, zaključeno je da su najveće vrednosti za UMTS sistem (4,69-14.69%), zatim GSM sistem (4.04-12.11%) i na kraju DCS sistem (2.71-7.92%). Važan deo istraživanja u oblasti merenja intenziteta elektromagnetnog polja jesu i rezultati merenja sprovedenih u realnim okruženjima baznih stanica. U okviru disertacije, sprovedena je statistička analiza rezultata merenja intenziteta električnog polja na 664 različite lokacije baznih klasifikovane u tri osnovne kategorije: stubovi (316 lokacija), lokacije sa instalacijom na objektu (276 lokacija) i lokacije sa instalacijom unutar objekta (72 lokacije). Rezultati merenja za kategoriju stubovi pokazuju da ni u jednom slučaju ukupni intenzitet električnog polja ne prelazi vrednost od 10V/m, što je znatno ispod referentnih graničnih nivoa propisanih ICNIRP preporukom, kao i referentnih graničnih nivoa propisanih Pravilnikom o granicama izlaganja nejonizujućim zračenjima. Za kategoriju instalacije na objektu, u slučaju outdoor merenja, referentne granične vrednosti propisane ICNIRP preporukama prevaziđene u 2.5% lokacija, dok su vrednosti propisane Pravilnikom o granicama izlaganja nejonizujućim zračenjima, prevaziđene su u 15.6% lokacija. Treba istaći da su ove vrednosti izmerene na krovovima objekata, u zonama gde ljudi ne borave, i čak u većini slučajeva nemaju pristup. Sa druge strane, indoor ukupni intenzitet električnog polja nije prevazišao ICNIRP i Srpske referentne granične nivoe ni u jednom slučaju. Na određenom broju lokacija indoor mikro baznih stanica, maksimalna vrednost intenziteta električnog polja, izmerena na horizontalnoj udaljenosti od 1m prevazilazi referentne granične vrednosti propisane ICNIRP preporukama i propisane Pravilnikom o granicama izlaganja nejonizujućim zračenjima. Apsolutna maksimalna vrednost prevazilazi Srpske referentne granične nivoe za približno 38% lokacija, i prevazilazi ICNIRP referentne granične za približno 7% lokacija. Savremeni pristup u analizi izloženosti ljudi radio-frekvencijskim elektromagnetnim poljima podrazumeva određivanje ukupnog izlaganja ljudi, i to kombinovanjem izlaganja u downlink-u (koje potiče od baznih stanica i access point-a) i izlaganja u uplink-u (koje potiče od mobilnih korisničkih uređaja). U okviru disertacije razvijena je metoda za određivanje merne nesigurnosti nove metrike nazvane indeks Doktorska disertacija Elektrotehnički fakultet Univerziteta u Beogradu v izloženosti, koja je definisana u okviru EU FP7 projekta LEXNET (Low-EMF Exposure Future Networks). Ova metoda daje analitički izraz za izračunavanje relativne standardne merne nesigurnosti indeksa izloženosti, na bazi poznatih standardnih mernih nesigurnosti brojnih ulaznih parametara. Pored toga, pokazano je da je funkcija gustine verovatnoće indeksa izloženosti približno normalna, za slučaj kada su ulazni parametri promenljive sa normalnom raspodelom i malim relativnim standardnim mernim nesigurnostima, što je najčešće slučaj u praksi.

Increasing user demand for accessible and high quality telecommunication services has led to intensive development of systems and devices for wireless communications. Because of their wide distribution, public mobile systems (GSM - Global System for Mobile Communications, DCS - Digital Communication System, UMTS - Universal Mobile Telecommunication System, LTE - Long-Term Evolution) are the most important. Having in mind that these systems use electromagnetic waves as basis for the communication, there is a need for an objective assessment of the electromagnetic radiation level in the living environment. To control the influence of electromagnetic radiation on people, norms that limit human exposure were adopted and stated as basic restrictions and reference levels. In this dissertation, an overview of basic restrictions and reference levels defined with Regulation on the limits of exposure to non-ionizing radiation, ICNIRP (International Commission on Non-Ionizing Radiation Protection) guidelines, FCC (Federal Communications Commission) recommendation, IEEE (Institute of Electrical and Electronics Engineers) standard and ARPANSA (Australian Radiation Protection and Nuclear Safety Agency) standard, is given. In order to evaluate compliance of base station with reference levels, measurement of electromagnetic field in the vicinity of base station should be conducted, using methods defined by the international standards. These methods are developed and upgraded in line with the development of telecommunication systems as well as with the improvements to the measuring devices. An important part in the development of these methods, which is especially important in recent time, is evaluation of measurement result uncertainty. In the dissertation, practical procedure for evaluation of uncertainty of electric field strength measurements, for the cases of broadband and frequency selective measurement system, is developed. This procedure is based on the low of propagation of uncertainty, and Doktorska disertacija Elektrotehnički fakultet Univerziteta u Beogradu vii enables evaluation of combined standard measurement uncertainty as well as expanded measurement uncertainty with confidence interval of 95%. The aim of this dissertation was improving the efficiency of methods for measurement of electric field strength in the vicinity of public mobile system base stations, which was achieved with improvements in several areas. In this dissertation, the procedure for improving the efficiency by using shorter time intervals of averaging (comparing to 6min that standards stipulate), is given. Based on the results of measurements conducted in the vicinity of multisystem base station, it was shown that the results can be estimated by using shorter time intervals of averaging 15s, 30s and 1min, and measurement process can be accelerated approximately by 6, 12 and 24 times, respectively. In these cases, additional standard measurement uncertainty of 3.07%, 3.80% and 4.42%, for the intervals of averaging 15s, 30s and 1min, should be taken into account. Shorter time-averaging intervals would reduce the total duration of the exposure assessment survey, while not compromising too much on measurement quality. A number of simplified measuring systems with mono-axial probe is frequently used in the case of systematic measurements. In the dissertation, simultaneous measurements of all three spatial components of electric field strength were conducted, for seven different environment types. Based on these results, model for the conversion from mono-axial measurements to isotropic case, is developed. On the basis of values averaged for all seven chosen environments, it was concluded that the additional multiplicative conversion factor value of 1.95 should be applied and additional uncertainty in measurement results of 33.07% should be taken into account. Besides the well-known short-term fading, variability of electric field strength of base stations directly depends on traffic load, automatic transmitter power control and discontinuous transmission. In this dissertation, long-term variability of electric field strength, in time interval of seven days, was conducted. For this analysis, band-selective measurements were carried out at 7 different locations in urban area. Furthermore, an additional traffic uncertainty caused by telecommunications traffic and transmitter functionalities for different time intervals of averaging: 10s, 30s, 1min, 6min, 15min, 30min, 1h, 3h, 5h and 10h, was calculated. The results show that uncertainty decreases with increasing of averaging interval. Comparing the uncertainty ranges corresponding Doktorska disertacija Elektrotehnički fakultet Univerziteta u Beogradu viii to different systems it can be concluded that the highest values are for the UMTS (4,69- 14.69%), than GSM (4.04-12.11%) and finally DCS (2.71-7.92%). An important part of research in the field of measurement of electromagnetic radiation, are the results of measurements carried out in realistic base stations environments. In this dissertation, statistical analysis of electric field strength in the vicinity of base stations for 664 different location was performed. Depending on how the antenna systems were installed, the base station locations were classified into three basic categories: masts (316 locations), locations with installations on buildings (276 locations) and indoor (72 locations). The measurement results showed that the electromagnetic field strength in the vicinity of base station antenna systems installed on the masts did not exceed in any case the ICNIRP or Serbian national reference levels. Precisely, the results, with a maximum value lower than 10V/m for electric field strength, were significantly below the reference levels. Measurement results for the locations with antenna installations on buildings showed that the maximum recorded value of total electric field strength has exceeded ICNIRP general public exposure reference levels at 2.5% of all locations and Serbian national reference levels at 15.6% of all locations. It should be emphasized that the values exceeding the reference levels were observed only outdoor, while in indoor the total electric field strength in no case exceeded defined reference levels. The results of measurement campaign for indoor microcell base station locations showed that at certain number of analyzed locations the intensity of electromagnetic exposure exceeded both Serbian national (at 38% of locations) and ICNIRP (7% of locations) reference levels. At these locations, the maximum values of electric field strength were recorded within the horizontal distance of 1 m from antenna installations. Modern approach in the assessment of human exposure to radio-frequency electromagnetic fields is based on the total human exposure, determined by combining downlink exposure (exposure originating from base stations and access points) and uplink exposure (exposure originating from wireless user devices). In this dissertation, the procedure for evaluation of uncertainty of a novel metric named Exposure Index (defined by EU FP7 project LEXNET - Low-EMF Exposure Future Networks), was developed. The developed procedure gives analytical expression for the calculation of Doktorska disertacija Elektrotehnički fakultet Univerziteta u Beogradu ix Exposure Index uncertainty, based on the known uncertainties of a number of input parameters. Moreover, it was shown that the probability distribution function of the Exposure Index can be considered as normal, in the case that all input parameters of the Exposure Index are normally distributed with small uncertainties, which is usually the case in reality.

Unapređenja sistema za detekciju plagijarizma u izvornom programskom kodu

Improving source code plagiarism detection systems

Obrazovanje u oblasti računarstva uključuje praktičan rad kroz programske zadatke koji su česta meta plagijarizma. U ovom radu su diskutovani različiti aspekti plagijarizma u programskom kodu u akademskom okruženju, izvršena je uporedna analiza softverskih sistema za detekciju sličnosti i predložena njihova unapređenja. Izabrani sistemi su evaluirani korišćenjem tri različita programa nad kojima je korišćeno više od 20 tipova leksičkih i strukturalnih izmena koje su primenjivane na kod tokom 1, 2, 4, i 8 sati rada. Primenjeno je i realno opterećenje koje je uključivalo zadatke obima od 50 do 1000 linija programskog koda sa tri različita predmeta koje je pohađalo od 100 do 300 studenata. Rezultati su pokazali da 5-10% studenata, shodno metrici i kriterijumima ove teze, plagira svoja rešenja. Analizirana su tri pravca za unapređenje ovih sistema: paralelizacija, vizuelizacija rezultata u vidu grafa i analiza rezultata metodama za analizu socijalnih mreža. Paralelizacija nad sekvencijalnom verzijom baziranom na algoritmu RKR-GST sprovedena je na centralnom procesoru korišćenjem Pthreads niti i na grafičkom procesoru korišćenjem CUDA tehnologije. Evaluacija korišćenjem dva skupa test primera pokazuje ubrzanja do 6 puta na centralnom procesoru i do 4 puta na grafičkom procesoru. iv U detekciji plagijarizma se mogu koristiti metrike i metode za analizu socijalnih mreža: centralnost po stepenu, relaciona i beta centralnost, stepen klasterizacije, analiza klika i sl. Pomoću njih su karakterisane četiri najčešće šeme kolaboracije, a potvrda osmišljenog pristupa je verifikovana na dva realna skupa primera kroz koje je pokazan i značaj pravilnog određivanja praga sličnosti. Radi vizuelnog prikaza mreže, izvršena je integracija sa alatom Gephi i razmatrani različiti načini za raspoređivanje čvorova.

Computing education involves practical training through programming assignments which are frequent targets for plagiarism. In this thesis, different aspects of source code plagiarism in academic environment are discussed. Comparative analysis of source code similarity detection systems was performed and several improvеments were proposed. Selected systems were evaluated using simulated plagiarism based on three programming assignments produced after 1, 2, 4, and 8 hours of work on baseline version using more than 20 types of lexical and structural modifications. Real-life student codes from three different courses were also used for evaluation. The courses were attended by 100 to 300 students, and the solutions varied from 50 to 1000 lines of code. The results show that 5-10% of students plagiarized their solutions, according to the criteria used in this thesis. Three improvements of such systems were proposed: paralellization, similarity network visualisation, and analysis of results using social network analysis. Sequentual version based on RKR-GST algorithm was parallelized on the CPU using Pthreads, while GPU implementation is based on CUDA. Observed speedups over the sequential version of the code were between 4 on the GPU and 6 on the CPU. vi Social network analysis measures and methods, such as degree, betweeness, and beta centrality, clustering coefficient, clique analysis were used for plagiarism detection in this study. The most frequent four collaboration patterns were characterized using aforementioned measures. The proposed approach was verified using two real-life test sets. The importance of proper choice of similarity threshold was also discussed. Visualization of the similarity network was performed using Gephi and different node layouts were considered.

Mere nekompaktnosti na Hilbertovim S-modulima

Measures of noncompactness over hilbert C-modules

U prvom poglavlju izlaž se teorija o uniformnim prostorima i merama nekompaktnosti na metričkim i uniformnim prostorima...

In the first section we present the theory on uniform spaces and measures of noncompactness in metric and uniform spaces...

Modelovanje gradiranih fotonskih plazmonskih kristala koji rade u režimu metamaterijala

Modelling of graded photonic and plasmonic crystals in metamaterial regime

U ovoj tezi su istraˇzeni elektromagnetski kompozitni materijali u cilju realizacije sredina sa elektromagnetskim osobinama koje ne postoje u prirodnim materijalima. Istraˇzivani kompoziti su jednodimenzionalni i dvodimenzionalni fotonski i plazmonski kristali. Jednodimenzionalne strukture se sastoje od slojeva odgovaraju´ceg oblika, dok se dvodimenzionalne strukture sastoje od ˇstapi´ca u vazduhu ili cilindriˇcnih rupa u dielektriku. Pored struktura sa periodiˇcnim ponavljanjem jediniˇcnih ´celija, razmatraju se i gradirane strukture dobijene prostornom promenom geometrije ili dielektriˇcne permitivnosti jediniˇcnih ´celija. Fotonski kristali se sastoje iskljuˇcivo od dielektrika, dok su plazmonski kristali kombinacija dielektriˇcnih i plazmonskih materijala, metala ili poluprovodnika. Posebna klasa razmotrenih kompozitnih struktura su planarni plazmonski kristali koji se sastoje od paralelnih traka od plazmonskih materijala na dielektriˇcnom substratu. Sve kompozitne strukture se razmatraju u reˇzimu metamaterijala gde veliˇcina jedini ˇcne ´celije nije zanemarljiva u odnosu na talasnu duˇzinu svetlosti, ali se efektivni parametri ipak mogu definisati. I jednodimenzionalne i dvodimenzionalne strukture su homogenizovane pomo´cu Maksvel Garnetove teorije. Planarni plazmonski kristali nisu opisani efektivnim parametrima, nego se razmatraju kao nizovi istih i podtalasnih rezonatora sa definisanim plazmonskim rezonancijama tako da rezonantne karakteristike plazmonskih kristala predstavljaju kolektivni odziv svih rezonatora. Posebna paˇznja je posve´cena izboru odgovaraju´cih materijala u kompozitima. Strukture namenjene vodenju elektromagnetskog polja treba da imaju ˇsto manje gubitke tako da u ovom sluˇcaju strukture treba da budu iskljuˇcivo dielektriˇcne. Plazmonski materijali kao ˇsto su metali i poluprovodnici se moraju koristiti u slede´cim sluˇcajevima: u realizaciji visoke anizotropije dielektriˇcne permitivnosti i u realizaciji rezonantnih struktura. Prvi metod koji je istraˇzivan za vodenje elektromagnetskog polja je transformaciona optika. U ovom metodu, prave linije polja u slobodnom prostoru se na odgovaraju´ci naˇcin transformiˇsu u linije polja sa ˇzeljenim oblikom. Maksvelove jednaˇcine su invarijantne prilikom koordinatnih transformacija dok se materijalni parametri menjaju u skladu sa primenjenom transformacijom. Dobijeni materijalni parametri su opisani anizotropnom i prostorno nehomogenom dielektriˇcnom permitivno ˇs´cu i magnetskom peremabilnoˇs´cu u opˇstem sluˇcaju. Ovde je data procedura kojom se nalaze jednostavniji materijalni parametri pomo´cu koncepta redukovanih parametara pri ˇcemu se zadrˇzava ista disperzija ili pomo´cu koncepta transformacija sa jediniˇcnim Jakobijanom. Kao rezultat primenjenih procedura, pokazuje se da je mogu´ce realizovati uredaje na bazi transformacione optike pomo´cu anizotropne i prostorno nehomogene dielektriˇcne permitivnosti. Anizotropija je onda realizovana slojevima ploˇca, dok se nehomogenost moˇze realizovati odgovaraju´com gradacijom debljine ili permitivnosti ploˇca.parametri menjaju u skladu sa primenjenom transformacijom. Dobijeni materijalni parametri su opisani anizotropnom i prostorno nehomogenom dielektriˇcnom permitivno ˇs´cu i magnetskom peremabilnoˇs´cu u opˇstem sluˇcaju. Ovde je data procedura kojom se nalaze jednostavniji materijalni parametri pomo´cu koncepta redukovanih parametara pri ˇcemu se zadrˇzava ista disperzija ili pomo´cu koncepta transformacija sa jediniˇcnim Jakobijanom. Kao rezultat primenjenih procedura, pokazuje se da je mogu´ce realizovati uredaje na bazi transformacione optike pomo´cu anizotropne i prostorno nehomogene dielektriˇcne permitivnosti. Anizotropija je onda realizovana slojevima ploˇca, dok se nehomogenost moˇze realizovati odgovaraju´com gradacijom debljine ili permitivnosti ploˇca.parametri menjaju u skladu sa primenjenom transformacijom. Dobijeni materijalni parametri su opisani anizotropnom i prostorno nehomogenom dielektriˇcnom permitivno ˇs´cu i magnetskom peremabilnoˇs´cu u opˇstem sluˇcaju. Ovde je data procedura kojom se nalaze jednostavniji materijalni parametri pomo´cu koncepta redukovanih parametara pri ˇcemu se zadrˇzava ista disperzija ili pomo´cu koncepta transformacija sa jediniˇcnim Jakobijanom. Kao rezultat primenjenih procedura, pokazuje se da je mogu´ce realizovati uredaje na bazi transformacione optike pomo´cu anizotrparametri menjaju u skladu sa primenjenom transformacijom. Dobijeni materijalni parametri su opisani anizotropnom i prostorno nehomogenom dielektriˇcnom permitivnoˇs´cu i magnetskom peremabilnoˇs´cu u opˇstem sluˇcaju...

In this thesis we investigate electromagnetic composite materials in order to realize media with the electromagnetic properties not achievable in the nature. The investigated composites are one dimensional and two dimensional photonic and plasmonic crystals. One dimensional structures consist of appropriately shaped slabs, whereas two dimensional structures consist of rods in air or cylindrical holes in a dielectric host. Beside the structures with the periodic arrangement of the unit cells, we consider graded structures obtained by a spatial variation of either cells’ geometry or dielectric permittivity. Photonic crystals are all dielectric structures whereas the plasmonic crystals are combination of dielectric and plasmonic materials, metals or semiconductors. Special class of the composite structures are the planar plasmonic crystals which consist of parallel ribbons made from plasmonic material on a dielectric substrate. All composites are considered in the metamaterial regime where unit cell size of the composites are not negligible in comparison to light wavelength, but the effective parameters can be still well defined. Homogenization of both one and two dimensional structures are done within the framework of Maxwell Garnett theory. Planar plasmonic crystals are not described by effective parameters, but they are considered as arrays of the same and subwavelength resonators with well defined plasmonic resonances so the resonant behaviour of the plasmonic crystals is the collective response of all resonators. Special attention is devoted to the choice of appropriate materials in the composites. The structures aimed for guiding of electromagnetic field should have as low as possible losses so the structures built form dielectrics only are the most preferably in this case. The plasmonic materials such as metals or semiconductors have to be used in the following cases: in the realization of extreme anisotropy in the dielectric permittivity and in the realization of resonant structures. The first method investigated for guiding of electromagnetic field is transformation optics. In this method, straight field trajectories in free space are appropriately transformed into the desired field trajectories. Maxwell equations are invariant under the applied coordinate transformations while the material parameters are scaled accordingly. The obtained material parameters are described with anisotropic and spatially inhomogeneous dielectric permittivity and magnetic permeability in a general case...

Razvoj multifunkcionalne bioinformatičke platforme zasnovane na potencijalu elektron-jon interakcije bioloških molekula

Development of multifunctional bioinformatics platform based on electron-ion interaction potential of biological molecules

Iako dalekodoseţne meĎumolekulske interakcije (interakcije na rastojanjima >5Å) igraju značajnu ulogu u prepoznavanju i selektivnom meĎusobnom privlačenju molekula u biološkim sistemima, do sada nije razvijen odgovarajući softverski paket koji bi omogućio da se ova vaţna osobina uključi u izučavanje biološki aktivnih molekula. U ovom radu je razvijena multifunkcionalna softverska EIIP/ISM platforma zasnovana na fizičkim parametrima organskih molekula koji definišu njihovu dalekodoseţnu interakciju. Ova platforma omogućava (i) izučavanje protein-protein interakcije i interakcije izmeĎu proteina i malih molekula, (ii) izučavanje veze izmeĎu strukture i funkcije proteina, (iii) procenu uticaja mutacija na biološku funkciju proteina, (iv) praćenje funkcionalne evolucije proteina, (v) dizajniranje molekula ţeljene biološke aktivnosti i (vi) selekciju potencijalnih terapeutskih molekula. Rezultati primene EIIP/ISM platforme na različite konkretne probleme kao što su evolucija virusa gripa, analiza mutacija na LPL proteinu koje predstavljaju faktor rizika za nastanak kardiovaskularnih bolesti, identifikacija terapeutskih targeta za viruse HIV-1 i viruse gripa, selekcija kandidata za antibiotike i lekove za SIDU virtuelnim pretraţivanjem molekulskih biblioteka, koji su prikazani u ovom radu, potvrdili su primenjivost ove platforme u rešavanju širokog spektra problema u molekularnoj biologiji, biomedicini i farmaciji.

Although long-range intermolecular interactions (interactions acting on distances >5Å) play an important role in recognition and targeting between molecules in biological systems, there is no one appropriate software package allowing use of this important property in investigation of biologically active molecules. The multifunctional EIIP/ISM software, which is based on physical parameters determining long-range molecular properties, was developed in this thesis. This novel and unique platform allows (i) investigation of protein-protein and protein-small molecule interactions, (ii) analysis of structure/function relationship of proteins, (iii) assessment of biological effects of mutations in proteins, (iv) monitoring of the functional evolution of proteins, (v) ―de novo‖ design of molecules with desired biological function and (vi) selection of candidate therapeutic molecules. Results of application of the EIIP/ISM platform on diverse problems (e.g. the evolution of influenza A viruses, assessment of biological effects of mutations on the LPL protein, representing a risk factor for cardiovascular diseases, identification of therapeutic targets for HIV and influenza viruses, virtual screening of molecular libraries for candidate antibiotics and anti-HIV drugs) which are presented in this thesis, confirm the applicability of this platform on broad spectrum of problems in molecular biology, biomedicine and pharmacology.

Primena neuralnih mreža u prepoznavanju šapata

Application of neural networks in whispered speech recognition.

Nedavno postignuti uspesi dubinskih neuralnih mreža u različitim zadacima mašinskog učenja su doprineli da vestačke neuralne mreze ponovo zauzmu bitnu ulogu u automatskom prepoznavanju govora. U ovom doktoratu je ispitana primena vestačkih neuralnih mreza u prepoznavanju šapata. Šapat kao specifičan oblik govora predstavlja ozbiljan problem za aktuelne sisteme automatskog prepoznavanja govora. Naime, neusaglašeni scenariji u kojima je sistem obučen sa neutralnim govorom a testiran sa šapatom daju veoma loše rezultate prepoznavanja. U eksperimentima ovog doktorata su demonstrirani različiti pristupi sa kojima se ovaj degradirani uspeh u automatskom prepoznavanju izolovanih reči može poboljšati primenom neuralnih mreža. Predlozena su dva nova sistema bazirana na neuralnim mrezama  jedan na višeslojnim perceptronima, (MLP-IF sistem) a drugi na dubinskim neuralnim mrežama (DNN- HMM sistem). Oba sistema u svojim front-end delovima imaju zadatak da reše ključni problem smanjenja akustičkih razlika između neutralnog govora i šapata. MLP-IF sistem u tu svrhu koristi inverzno filtriranje čime se potiskuje zvučnost iz govora koji nakon obrade po svojim spektralnim karakteristikama postaje sličniji šapatu. Drugi sistem, koji je u formi tandem DNN-HMM sistema, isti zadatak rešava na suprotan način tako što iz šapata rekonstruiše karakteristike neutralnog govora. Naime, u osnovi front-end dela DNN-HMM sistema je poseban tip dubinskih neuralnih mreža, poznat kao dubinski denoising autoenkoder koji zahvaljujući svojoj dubinskoj strukturi na efikasan način vrši rekonstrukciju kepstralnih karakteristika neutralnog govora iz šapata. Oba sistema su testirana u različitim obuka/test scenarijima sa tri različita tipa kepstralnih koeficijenata: MFCC (Mel-Frequency Cepstral Coefficients), TECC (Teager-Energy Cepstral Coefficients) i TEMFCC (Teager-based Mel-Frequency Cepstral Coefficients), pri čemu su se TECC govorna obeležja pokazala kao najpodesnija za prepoznavanje šapata. MLP-IF sistem zahvaljujući inverznom filtriranju i TECC obeležjima ostvaruje poboljšanje uspeha prepoznavanja izolovanih reči u šapatu za 12,5% u odnosu na klasični MLP sistem sa MFCC obeležjima. Najbolje performanse poseduje tandem DNN-HMM sistem koji sa TECC obeležjima poboljšava prepoznavanje šapata za 31% u poređenju sa tradicionalnim HMM sistemom i ostvaruje tačnost od 92,81% u prepoznavanju reči izgovorenih u šapatu.

The recent success of Deep Neural Networks (DNN) in different machine learning tasks has significantly contributed to the rise in the popularity of artificial neural networks (ANN) and their today’s role in Automatic Speech Recognition (ASR). This thesis examines how artificial neural networks can benefit in automatic whispered speech recognition. Whisper, as a specific form of verbal communication, represents one of the most challenging problems in current automatic speech recognition systems. Namely, the performance of traditional ASR systems trained on neutral speech degrades significantly when a whisper is applied. The experiments of this thesis present different approaches based on the application of neural networks that improve isolated word recognition under whispered speech condition. Two systems based on neural networks are proposed  one that is based on multilayer perceptrons (MLP-IF system), and another that is based on a deep neural network (DNN-HMM system). Both systems have the same task to solve in their front-ends and that is to alleviate the acoustic mismatch between neutral speech and whisper. For that purpose, MLP-IF system uses inverse filtering which suppresses sonority from a neutral speech which after such treatment becomes more similar to whisper in terms of their spectral characteristics. The second system, which is in the form of tandem DNN-HMM system, solves the same task in an opposite way by reconstructing characteristics of neutral speech from whispered speech. Namely, the basis of the front-end part of the DNN-HMM system is a special type of deep neural network, known as deep denoising autoencoder, which thanks to its deep structure effectively performs reconstruction of cepstral characteristics of neutral speech from whispered samples. Both systems were tested in different train/test scenarios with three types of cepstral features: MFCC (Mel- Frequency Cepstral Coefficients), TECC (Teager-Energy Cepstral Coefficients) and TEMFCC (Teager-based Mel-Frequency Cepstral Coefficients). The TECC features proved to be the most suitable for whisper recognition. Due to inverse filtering and the application of TECC features, MLP-IF system improves isolated word recognition in whispered speech by 12.5% compared to the traditional MLP system with MFCC features. The best performance has the tandem DNN-HMM system which with TECC features improves whisper recognition by 31% in comparison with the traditional HMM system, and achieves 92.81% accuracy in isolated word recognition under whisper conditions.

Statistics of Langmuir waves associated with type III solar radio bursts / Sonja Vidojević.

Statistika Lengmirovih talasa povezanih sa Sunčevim radio erupcijama tipa III Sažetak: Međuplanetski snopovi elektrona proizvedeni u Sunčevim erupcijama nesta- bilni su u Sunčevom vetru i proizvode Lengmirove talase na plazmenoj frekvenciji, fp. Ovi talasi se zatim konvertuju u takozvane radio-erupcije tipa III – elektromagnetno zračenje koje se slobodno prostire na fp ili na prvom harmoniku. Erupcije tipa III se posmatraju kao zračenje koje se pomera od visokih ka niskim frekvencijama u kilometarskom opsegu. Posle prvog teorijskog objašnjenja koje su dali Ginzburg i Železnjakov 1958. godine (Ginzburg and Zhelezniakov, 1958a), sa nekoliko poboljšanih modela izvršen je pokušaj da se detaljno opišu fizički procesi od kojih potiču erupcije tipa III. Nestabilnost usled “izbočine na repu” (“Bump-on-tail”, eng.) raspodele po brzinama čestica, generisanje Lengmirovih talasa, kon- verzija Lengmirovih talasa u radio-zračenje putem nelinearnih talasno-talasnih interakcija itd, su mehanizmi koji su detaljno proučavani. Sa posmatračkog stanovišta, od posebnog interesa su takozvane erupcije tipa III za koje je snop elektrona, koji se kreće duž linija sila magnetnog polja, posmatran pomoću satelita na mestu izvora emitovanja direktno–in situ zajedno sa Lengmirovim talasima i rezultujućom radio-emisijom. Samo nekoliko erupcija tipa III posmatranih in situ je do sada proučeno i opisano u literaturi. Prvi zadatak u ovoj disertaciji je bio da se ispitaju posmatranja čestica, radio i elektro- statičkog zračenja Sunčevog vetra zabeleženih pomoću satelita Wind in situ tokom prvih 16 godina i da se izdvoje erupcije tipa III. Primenjujući pažljivo odabrane stroge kriteri- jume, dobijen je skup od 36 događaja visokog kvaliteta. Sa ovako brojnim skupom posma- tranja koji je statistički reprezentativan za posmatranu pojavu, sada je sa većim poverenjem moguće postaviti posmatračka ograničenja na modele generisanja erupcija tipa III. Nakon izdvajanja statističkog skupa, za svaki od dogaljaja je precizno određen oblik raspodele snage Lengmirovih talasa posmatrane u spektralnom domenu. Ova posmatranja su mode- lovana sistemom Pirsonovih raspodela verovatnoće i pokazano je da raspodela verovatnoće logaritama spektralne gustine snage Lengmirovih talasa pripada trima “glavnim” tipovima Pirsonovih raspodela verovatnoće: tipu I, tipu IV i tipu VI. Dodatno su modelovani instru- mentski efekti vremena integracije radio-prijemnika satelita Wind na posmatrane raspodele snage Lengmirovih talasa. Kombinujući posmatranja i naše modele, pokazali smo da nije moguće izvesti definitivan zaključak da je raspodela energije Lengmirovih talasa u real- nom vremenu log-normalna što predviđaju neke teorije kao što je Robinsonova teorija sto- hastičkog rasta (Robinson, 1992). U poslednjem delu disertacije, pokazali smo kako naš skup in situ posmatranja visokog kvaliteta od 36 događaja tipa III može da se koristi u budućim studijama kojima bi se još bolje postavila ograničenja na teorijske modele. Mi smo, na primer, ispitali korelaciju snage Lengmirovih talasa sa energijom brzih elektrona i sa snagom radio-zračenja.

Statistics of Langmuir waves associated with type III solar radio bursts Abstract: Interplanetary electron beams, produced by solar flares, are unstable in the solar wind and generate Langmuir waves at the local plasma frequency, fp. These waves are then converted into the so-called type III radio bursts which are freely propagating electromagnetic emissions at fp or its harmonic. The type IIIs are therefore observed as drifting emissions from high to low frequencies, in the kilometric wavelengths range. Since the first theoretical explanation by Ginzburg and Zhelezniakov (1958a), several refined models have attempted to describe in details the physical processes at the origin of type III bursts. The mechanisms of “Bump-on-tail” instabilities, Langmuir waves generations, conversion of these Langmuir waves into radio emissions throughout nonlinear wave-wave interactions etc, have been studied in detail. Of particular interest from the observational point of view are the so called in situ type III bursts for which the electron beam, at the origin of the emission and traveling along open interplanetary magnetic field lines, is observed directly in situ by a spacecraft, together with the local Langmuir waves and the resulting radio emissions. Until now only a few of these in situ type IIIs have been reported in the literature. The first research study performed in this thesis was to examine the first 16 years of radio, waves and particles data recorded by the Wind spacecraft in the Solar Wind and to look for in situ type IIIs. Applying rigorous and careful criteria, this examination has yielded to a data set of 36 high-quality events. With such a numerous data set, which is statistically representative of the studied phenomenon, it is now possible to constrain obser- vationally and with a better confidence the type III generation models. After having built our statistical dataset, we have studied, for each of the events, the precise shapes of the Langmuir wave power distributions, observed in the spectral domain. We have fitted these observed distributions by a Pearson’s system of probability distributions and have shown that the probability distributions of the logarithm of the Langmuir waves power spectral density belong to three “main” types of Pearson’s probability distributions: type I, type IV and type VI. In addition we have modeled the effects of the instrumental integration time of the Wind radio receivers on the observed Langmuir wave power distributions. By combining our observations with our models we have shown that it was not possible to con- clude definitively, that the distribution of the Langmuir waves energy in the real temporal domain is lognormal, as it is predicted in some theories as the Stochastic Growth Theory by Robinson (1992). In the last part of the thesis, we have shown how our high-quality data set of 36 in situ type III events can be used for further studies that could allow to constrain the theoretical models even better. For instance we have investigated the correlation between the Langmuir waves power and the energy of impulsive electron or with the power of the radio emissions themselves.

Prilog teoriji apstraktnih metričkih prostora

Contribution to the theory of abstract metric spaces

Ova disertacija je prilog Metričkoj teoriji fiksne tačke, oblasti koja se u poslednje vreme intenzivno razvija. Sastoji se iz pet poglavlja. U prvom poglavlju dokazana je jedna poznata lema, koja se primenjuje u dokazu Banach-ovog principa kontrakcije, iskazanog za orbitalno kompletne metričke prostore. U drugom poglavlju dokazano je osam teorema koje generališu neke poznate rezultate o fiksnim tačkama u metričkim prostorima (Boyd-Wong-ov, Ćirićev, Pant-ov i druge). Neke od tih teorema su modifikacije poznatih, a tri su potpuno nove. U trećem poglavlju dokazane su tri teoreme koje generališu rezultat Nemыckiй-a o fiksnim tačkama preslikavanja u kompaktnim metričkim prostorima, kao i jedno uopštenje teoreme Edelstein-a. Kod nekih teorema su originalni dokaz i iskazane posledice. U četvrtom poglavlju razmatraju se b-metrički prostori, kao uopštenja metričkih. Izvedena je generalizacija teoreme Zamfirescu-a u b-metričkom prostoru i neke njene primene. Dat je još jedan novi rezultat koji se odnosi na slaba skoro kontraktivna preslikavanja. U petom poglavlju razmatraju se konusni metrički prostori. Dokazane su dve teoreme kao analogoni odgovarajućih rezultata iz običnih metričkih prostora. Izvedena je i jedna potpuno nova teorema koja za posledicu ima Banach-ovu teoremu u konusnim metričkim prostorima nad konusom koji ne mora biti normalan. Dokazana je i generalizacija Fischer-ove teoreme u konusnim metričkim prostorima sa regularnim konusom. Većina rezultata snabdevena je odgovarajućim primerima koji pokazuju u čemu se ti rezultati razlikuju od već poznatih.

This dissertation is the contribution to the Metric fixed point theory, the area that has recently been rapidly developing. It contains five chapters. The first chapter gives the proof of one already known lemma. This lemma is used in the proof of Banach’s theorem for orbital complete metric spaces. The second chapter contains the proofs of eight theorems, which generalize some known results from the theory of fixed points in metric spaces (Boyd-Wong’s, ´ Ciri´c ’ s, Pant’ s, and other). Some of these theorems are modifications of the known ones, while three are completely new. Three theorems are proven in the third chapter. They generalize the result of fixed point of mapping defined in compact metric space given by Nemytzki, as well as one generalization of Edelstein’s theorem. The proofs and stated corollaries of some theorems are original. Chapter four discusses b-metric spaces as a generalization of metric spaces. The generalization of Zamfirescu’s theorem of b-metric spaces is presented as well as some of its applications. A new result concerning weakly almost contractive mappings is also determined. Chapter five contains some new results in cone metric spaces. Two theorems are presented as the analogue of the same theorems in the setting of standard metric spaces. A completely new theorem is established which results in the Banach’s theorem in cone metric spaces whereby the cone does not need to be normal. A generalization of Fisher’s theorem in cone metric spaces over a regular cone is also proven. Almost all results in this dissertation are confirmed by corresponding examples, which explain how these results differ from the already known results. Key Words: Cone metric space, normal cone, b-complete metric space, Banach’s contraction, generalized contraction, fixed point, weakly compatible, weakly semi compatible, weakly commutative, the sequence of bounded variation

Matematički modeli i metode rešavanja novog problema raspoređivanja vozila pri optimizaciji transpoorta poljoprivrednih sirovina

Mathematical models and solution methods for the novel vehicle scheduling problem for optimizing the transport of agricultural raw materials.

Problemi optimizacije nalaze primenu u svim sferama realnog ži- vota. Adekvatno modeliranje problema optimizacije i razvoj adekvatnih mate- matičkih metoda za njihovo rešavanje od velikog su značaja za funkcionisanje mnogih realnih sistema. Predmet ove doktorske disertacije je nov problem raspoređivanja vozila (engl. Vehicle Scheduling Problem - VSP) pri optimi- zaciji transporta poljoprivrednih sirovina. Optimizacija transporta siro- vina koje se koriste za preradu do finalnog proizvoda je od velikog značaja u početnoj fazi proizvodnje. To je posebno izraženo kod prerade poljoprivred- nih sirovina, jer je njihova cena na tržištu niska, pa troškovi transporta imaju najveći udeo u ukupnim troškovima proizvodnje finalnog prozvoda. Iz tog razloga, smanjenje neophodnog vremena i uloženih finansijskih sredstava u fazi transporta sirovina, direktno utiče na povećanje profitabilnosti preduzeća. Varijanta problema VSP koja je razmatrana u disertaciji nastala je iz po- trebe optimizacije transporta šećerne repe u jednoj fabrici za proizvodnju šećera u Srbiji, ali se može primeniti i u širem kontekstu, za optimiza- ciju transporta sirovina ili robe koji se organizuje u velikim kompanijama pri istim ili sličnim uslovima. Problem uključuje niz specifičnosti po kojima se razlikuje od postojećih varijanti problema raspoređivanja vozila, te matematički modeli koji su do sada predloženi u literaturi ne opisuju adekvatno razmatrani problem. U disertaciji je analizirana složenost razmatranog problema. Relaksa- cija predloženog problema poređena je sa problemom paralelnog raspoređi- vanja mašina (engl. Parallel Machine Scheduling Problem - PMSP). Na osnovu ekvivalentnosti sa problemom particije (engl. Partitioning problem), poznato je da PMSP NP-težak. U disertaciji je dokazana analogija između relaksa- cije razmatranog problema i PMSP i zaključeno je da je predloženi problem NP-težak. Razvijeni su novi adekvatni matematički modeli koji obuhvataju sve karak- teristike razmatranog problema. Predloženi matematički modeli su upore- đeni u smislu efikasnosti korišćenjem egzaktnih rešavača Lingo 17 i CPLEX MIP 12.6.2. Eksperimentalni rezultati su pokazali da oba egzaktna rešavača pronalaze optimalna ili dopustiva rešenja samo na realnim instancama pro- blema malih dimenzija, što je i očekivano, imajući u vidu složenost pro- blema. Iz tog razloga, korišćenje heurističkih i metaheurističkih metoda predstavlja adekvatan pristup za rešavanje instanci problema većih dimen- zija. S obzirom na specifičnosti razmatranog problema, postojeće imple- mentacije heurističkih i metaheurističkih metoda za probleme raspoređi- vanja ili rutiranja vozila iz literature ne mogu biti direktno primenjene na ovaj konkretan problem. Iz tog razloga su, u ovoj disertaciji, dizajnirane i implementirane različite varijante metaheuristike pod nazivom Metoda promenljivih okolina (engl. Variable Neighborhood Search - VNS), kao i jedna varijanta Pohlepne stohastičko-adaptivne procedure pretrage (engl. Greedy Randomized Adaptive Search Procedure - GRASP), čiji su elementi prilagođeni karakteristikama razmatranog problema raspoređivanja vozila. Prvo je razmatran problem VSP-P koji predstavlja potproblem predložene varijante problema raspoređivanja vozila. VSP-P je dobijen isključivanjem ograničenja koja se odnose na usklađivanje vremena stizanja kamiona na svaku od lokacija i u krug fabrike. Za rešavanje VSP-P dizajnirana su dva algo- ritma: varijanta Osnovne metode promenljivih okolina (engl. Basic Variable Neighborhood Search - BVNS) i Pohlepna stohastičko-adaptivna procedura pre- trage (engl. Greedy Randomized Adaptive Search Procedure - GRASP). Obe metode su testirane na instancama dobijenim na osnovu podataka iz prakse i na skupu generisanih instanci većih dimenzija. Eksperimentalni rezultati pokazuju da i BVNS i GRASP metoda na realnim instancama malih dimenzija dostižu sva poznata optimalna rešenja, dobijena egzaktnim rešavačima. Na skupu re- alnih instanci srednjih dimenzija, BVNS metoda je dostigla ili poboljšala gornje granice vrednosti funkcije cilja dobijenih CPLEX rešavačem za ogra- ničeno vreme izvršavanja od 5h. Takođe, BVNS daje rešenja boljeg kvaliteta u odnosu na GRASP metodu na skupu realnih instanci srednjih dimenzija i ge- nerisanih instanci većih dimenzija. Analizom svih eksperimentalnih rezul- tata, zaključeno je da obe metode predstavljaju adekvatne pristupe za rešavanje razmatranog problema. Ipak, uočava se da je BVNS bolja kada je u pitanju kva- litet dobijenih rešenja, dok je GRASP efikasniji od BVNS u pogledu vremena izvršavanja. Za polazni problem raspoređivanja vozila (VSP) koji uključuje sve karak- teristike realnog problema, dizajnirane su i implementirane tri varijante VNS metode: Osnovna metoda promenljivih okolina (engl. Basic Variable Neigh- borhood Search - BVNS), Adaptivna metoda promenljivih okolina (engl. Skewed Variable Neighborhood Search - SVNS) i Poboljšana osnovna metoda promenlji- vih okolina (engl. Improved Basic Variable Neighborhood Search - BVNSi). Prve dve metode koriste iste strukture okolina, pri čemu BVNS koristi strate- giju najboljeg poboljšanja (engl. Best improvement strategy), a SVNS strategiju prvog poboljšanja (engl. First improvement strategy) u fazi lokalne pretrage. Sve tri metode su testirane na realnim i generisanim instancama. Eksperi- mentalni rezultati na skupu realnih i generisanih instanci su pokazali da BVNS daje rešenja boljeg kvaliteta u odnosu na SVNS, ali je vreme rada SVNS metode značajno kraće. Treći dizajnirani algoritam (BVNSi) predstavlja po- boljšanu varijantu BVNS metode, koja koristi opštije strukture okolina u odnosu na BVNS i SVNS, što je doprinelo jednostavnosti BVNSi algoritma i brzini izvršavanja. Testirane su dve varijante BVNSi metode koje koriste različite strategije u okviru procedure lokalnog pretraživanja: BVNSiF sa strategijom prvog i BVNSiB sa strategijom najboljeg poboljšanja. Analizom i poređenjem rezultata zaključeno je da u pogledu kvaliteta dobijenih rešenja kao i stabilnosti, najbolje performanse ima BVNS metoda, ali je SVNS naj- brža. Varijante poboljšane metode BVNSi uspevaju da pronađu nova najbolja rešenja za dve realne instance srednjih dimenzija i da za značajno kraće vreme reše instance većih dimenzija, u poređenju sa BVNS i SVNS, redom. Sa druge strane, varijante BVNSi metode nisu tako stabilne kao BVNS i SVNS metode na skupu realnih i generisanih instanci. U slučaju jedne generisane instance većih dimenzija, obe BVNSi metode su imale značajno lošije perfoermanse u odnosu na ostale metode, što je negativno uticalo na prosečne vrednosti funkcije cilja i vremena izvršvanja. Predloženi problem raspoređivanja vozila je od velikog praktičnog zna- čaja za optimizaciju transporta poljoprivrednih sirovina. Planirano je da dobijeni rezultati budu primenjeni u praksi, delimično ili u potpunosti. Sa teorijskog aspekta, razvijeni matematički modeli predstavljaju naučni doprinos oblastima optimizacije i matematičkog modeliranja. Razvijene va- rijante VNS metoda prilagođenih problemu i poređenja njihovih performansi predstavljaju naučni doprinos oblasti metaheurističkih metoda za rešavanje problema optimizacije.

Optimization problems arise from many real-life situations. The devel- opment of adequate mathematical models of optimization problems and appropriate solution methods are of great importance for performance of real systems. The sub- ject of this doctoral dissertation is a novel vehicle scheduling problem (VSP) that arises from optimizing the transport of agricultural raw materials. The organization of transport of raw materials is of great importance in the initial phase of produc- tion. This is particularly evident in the case of agricultural raw materials, because their price in the market is very low, and therefore, the costs of their transport represent the largest part of the total production cost. For this reason, any reduc- tion of time and money spent in this early production stage directly increases the company’s profitability. The considered variant of VSP arises from optimizing the transport of sugar beet in a factory for sugar production in Serbia, but it can also be applied in a wider context, i.e., to optimize the transport of raw materials or goods in large companies under the same or similar conditions. The considered problem involves a number of specific constraints that distinguish it from existing variants of the vehicle scheduling problem. Therefore, mathematical models proposed in the literature for other variants of VSP do not describe adequately the considered problem. The complexity of the newly introduced VSP is analyzed. It is proven that the introduced VSP belongs to the class of NP-hard problems by comparing its relaxation with the Parallel Machine Scheduling Problem (PMSP). PMSP is known to be NP-hard, as it is equivalent to the Partitioning problem. From the established analogy between the relaxation of the considered VSP and PMSP, it is concluded that the VSP introduced in this dissertation is NP-hard. New mathematical models of the considered problem that involve all problem specific properties, are developed. The proposed mathematical models are compared in sense of efficiency by using Lingo 17 and CPLEX MIP 12.6.2 solvers. Experi- mental results showed that both exact solvers provided optimal or feasible solutions only for small-size real-life problem instances. However, this was expectable, having in mind the NP-hardness of the considered problem. Therefore, heuristic and meta- heuristic method seem to be appropriate approaches for solving problem instances of larger dimension. Due to specific properties of the considered problem, the ex- isting implementations of heuristic and metaheuristic methods for vehicle routing and scheduling problems can not be directly applied. For this reason, different vari- ants of well-known Variable Neighborhood Search (VNS) metaheuristic, as well as Greedy Randomized Adaptive Search Procedure (GRASP), are designed. The con- structive elements of the proposed VNS and GRASP implementations are adapted to the characteristics of the considered vehicle scheduling problem. A subproblem of the proposed variant of vehicle scheduling problem, denoted as VSP-P is considered first. VSP-P is obtained from the initial VSP by excluding problem specific constraints regarding vehicle arriving times to each location and to the factory area. Two metaheuristic algorithms are designed as solution meth- ods for this subproblem: Basic Variable Neighborhood Search - BVNS, and Greedy Randomized Adaptive Search Procedure - GRASP. Both proposed approaches were tested on instances based on real-life data and on the set of generated instances of lager dimensions. Experimental results show that BVNS and GRASP reached all optimal solutions obtained by exact solvers on small-size real-life problem instances. On medium-size real-life instances, BVNS reached or improved upper bounds ob- tained by CPLEX solver under time limit of 5 hours. BVNS showed to be superior compared to GRASP in the sense of solution quality on medium size real-life in- stances, as well as on generated large-size problem instances. However, general conclusion is that both proposed methods represent adequate solution approaches for the subproblem VSP-P. BVNS provides solutions of better quality compared to GRASP, while GRASP outperforms BVNS regarding the average CPU time required to produce its best solutions. For the initial vehicle scheduling problem (VSP) that includes all problem specific constraints, three VNS-based metaheuristic methods are designed and implemented: Basic Variable Neighborhood Search - BVNS, Skewed Variable Neighborhood Search - SVNS, and Improved Basic Variable Neighborhood Search - BVNSi. BVNS and SVNS use the same neighborhood structures, but different search strategies in lo- cal search phase: BVNS uses Best improvement strategy, while SVNS uses First improvement strategy. All three VNS-based methods are tested on real-life and generated problem instances. As it was expected, experimental results showed that BVNS outperformed SVNS regarding solution quality, while SVNS running time was significantly shorter compared to BVNS. The third designed algorithm BVNSi represents a variant of BVNS that uses more general neighborhood structures com- pared to the ones used in BVNS and SVNS. The use of such neighborhood structures lead to the simplicity of BVNSi and shorter running times compared to BVNS. Two variants of BVNSi method that exploit different strategies in Local search phase are designed: BVNSiB with best improvement strategy and BVNSiF with First improvement strategy. The results of computational experiments for all proposed VNS-based methods for VSP are analyzed and compared. Regarding the quality of the obtained solutions, BVNS method shows the best performance, while SVNS needed the shortest average running times to produce its best solutions. Two vari- ants of BVNSi method succeeded to find new best solutions on two medium size real life instances and to solve large size instances in shorter running time compared to BVNS and SVNS, respectively. However, both BVNSiB and BVNSiF turn out to be less stabile than BVNS and SVNS on real-life and generated inatances. In the case of one large-size generated instance, both BVNSi variants had significantly worse performance compared to BVNS and SVNS, which had negative impact on their average objective values and average running times. The proposed vehicle scheduling problem is of great practical importance for optimizing the transport of agricultural raw materials. It is planned to use the ob- tained results in practice (partially or completely), as a support to decision makers who organize transportation of in the early production phase. From the theoretical point of view, the developed mathematical models represent a scientific contribu- tion to the fields of optimization and mathematical modeling. The variants of VNS methods that are developed and adapted to the problem, as well as comparison of their performances, represent a scientific contribution to the field of metaheuristic methods for solving NP-hard optimization problems.

Primena metoda kombinatorne optimizacije za rešavanje problema formiranja grupa u nastavi

Combinatorial optimization methods for solving the problems of group formation in classes

Predmet ovog rada je predstavljanje novih matematičkih modela i algoritama kombinatorne optimizacije, koji se mogu primeniti za rešavanje problema formiranja grupa u nastavi. Naime, postoje razni problemi koji podrazumevaju izdvajanje određenih grupa individua iz konačnog skupa, po unapred utvrđenim kriterijumima grupisanja. Neki od njih su NP-teški problemi kombinatorne optimizacije, razmatrani u ovoj disertaciji: problem maksimalne podele skupa na dva dela (engl. Maximum Set Splitting Problem - MSSP), problem formiranja dobro balansirane eksperimentalne i kontrolne grupe (engl. Well-Balanced Experimental and Control Group Formation Problem - WBECGFP), problem formiranja k dobro balansiranih grupa (engl. Balanced Multi-Weighted Attribute Set Partitioning Problem - BMWASP), problem formiranja grupa za kolaborativno učenje (engl. Collaborative Learning Groups Formation Problem - CLGFP) i problem minimalnog reprezentativnog skupa (engl. Minimum Hitting Set Problem - MHSP). Sam proces formiranja grupa predstavlja kompleksan i vremenski veoma zahtevan zadatak, pa je iz tih razloga neophodna i softverska podrška kako bi se on efikasno i uspešno izvršio. Neki od problema na koje se može naići u nastavi su ekvivalentni navedenim NP-teškim problemima i njihovim specijalnim slučajevima, posebno kada je potrebno uzeti u obzir veliki broj individua, karakteristika i kriterijuma za njihovo raspoređivanje u grupe. Cilj istraživanja prikazanog u ovom radu je rešavanje problema kombinatorne optimizacije: MSSP, WBECGFP, BMWASP, CLGFP i MHSP. Dobijena rešenja razmatranih problema se mogu primeniti za unapređivanje procesa organizovanja i izvođenja nastave, procesa deljenja i usvajanja novih znanja, za uspešnije izvođenje obrazovnih eksperimentalnih istraživanja, za povećanje motivacije kod studenata kroz grupni i timski rad, i sve to u cilju što kvalitetnije nastave matematike i računarstva. S obzirom na raznovrsne zahteve Rezime iv koji se postavljaju od strane organizatora nastave u pogledu broja, veličina i sastava grupa koje treba formirati, kao i kriterijuma koje pri tome treba uzeti u obzir, ova disertacija daje praktični doprinos metodologiji što boljeg raspoređivanja individua u grupe primenom matematičkih modela i algoritama kombinatorne optimizacije. Predloženi algoritami su implementirani u javno dostupne aplikacije, kako bi se omogućila što jednostavnija primena od strane korisnika svih obrazovnih profila. Za razmatrane probleme optimizacije, neophodno je najpre definisati adekvatne matematičke modele. Imajući u vidu složenost problema, dizajnirane su metaheurističke metode za njihovo rešavanje. Različite varijante genetskog algoritma (engl. Genetic Algorithm - GA) prikazane su za rešavanje MSSP i MHSP problema, dok je za rešavanje problema WBECGFP, BMWASP i CLGFP prikazana metoda promenljivih okolina (engl. Variable Neighbourhood Search - VNS), posebno prilagođena svakom od navedenih problema. Iako se predstavljena istraživanja odnose na rešavanje konkretnih NP-teških problema i na raspoređivanje studenata u specifične grupe, razmatrani problemi imaju širu primenu. Predloženi modeli i metode mogu se primeniti za rešavanje drugih problema, kao što su problemi raspoređivanja: zaposlenih, zadataka, mašina, paketa, putovanja, časova, lekcija, kao i drugih objekata i subjekata. U cilju potvrde adekvatnosti formulisanih matematičkih modela i efikasnosti predloženih metaheurističkih metoda, sproveden je veliki broj eksperimenata. Predloženi algoritmi su testirani na odgovarajućim test primerima iz literature kao i na novogenerisanim (javno dostupnim) test primerima. Dobijeni rezultati su poređeni sa rezultatima egzaktnih ili već postojećih metoda, u odnosu na unapred definisane indikatore poređenja. Eksperimentalni rezultati su pokazali da predloženi pristupi imaju bolje performanse u poređenju sa drugim metodama i da mogu da obezbede visoko kvalitetna rešenja za različite dimenzije problema, u razumnom vremenskom roku. Rezime v

The subject of this thesis is to present new mathematical methods and algorithms of combinatorial optimization, which could be applied for solving the problems of group formation in classes. Namely, there are various problems that require the selection of certain groups forming of individuals from the finite set, based on previously determined grouping criteria. Some of these are NP-hard problems of combinatorial optimization, which are taken into consideration in this thesis: Maximum Set Splitting Problem - MSSP, Well-Balanced Experimental and Control Group Formation Problem - WBECGFP, Balanced Multi-Weighted Attribute Set Partitioning problem – BMWASP, Collaborative Learning Groups Formation Problem - CLGFP and Minimum Hitting Set Problem - MHSP. The process of group formation represents a complex and time- consuming task, thus requiring a necessary software support for the efficient and successful completion of the task. Some of the problems that could come up in the course of teaching are equivalent to the aforementioned NP-hard problems and their special cases, especially when there is a need to take into consideration a large nunber of individuals, charactersitics and criteria for their assigning into groups. The objective of the research presented in this thesis is solving combinatorial optimization problems: MSSP, WBECGFP, BMWASP, CLGFP и MHSP. The obtained results of the considered problems can be applied for: upgrading the process of the organization and performance of teaching, the process of splitting and the adoption of new knowledge, as well as to achive more successful performance of educational experimental researches, and to increase student’s motivation through group and team work. The objective is to achieve higher quality teaching of mathematics and computing. Taking into consideration various requests put forward by the organizers of teaching, in terms of number, size and group composition needed to be formed, as well as the criteria needed to be taken into account, this thesis provides a practical contribution to the methodology of optimal distribution of individuals into groups by applying mathematical Резиме vii models and combinatorial optimization algorithms. The proposed algorithms are implemented in publicly available applications, such that users of all educational profiles are able to use them. The first step in studying the considered optimization problems is to define adequate mathematical models. As the considered problems are very complex, metaheuristic methods are designed in order to find high quality solutions. Different versions of Genetic Algorithm - GA were presented for solving MSSP and MHSP problems, while a Variable Neighbourhood Search - VNS method was presented for solving WBECGFP, BMWASP and CLGFP problems, especially adapted to all of the aforementioned problems. Although the presented researches refer to solving concrete NP-hard problems, as well as to the distribution of students into specific groups, the considered problems have shown to have a wider application. The proposed models and methods can be applied for solving other problems, such as the problems of arrangement of employees, tasks, machines, packets, travel, classes, lectures, as well as organizing other objects and subjects. In order to confirm the adequacy of the formulated mathematical models and efficiency of the proposed metaheuristic methods, a large number of experiments was conducted. The proposed algorithms have been tested on the relevant instances from literature as well as on newly generated (publically available) instances. The obtained results have been compared with the results of the exact or already existing methods, with reference to the previously defined comparison indicators. Experimental results have shown that the proposed approaches have a better performance compared to other methods and that they can provide high quality solutions for different problem dimensions, within a reasonable execution time. Резиме viii

Unapređenje postupka zaštićene tople ploče za određivanje toplotne provodnosti termoizolacionih materijala

Improvements to the guarded hot plate method for thermalk insulating materials thermal conductivity measurement.

Ova doktorska teza obuhvata istraživački rad na unapređenju merenja toplotne provodnosti termoizolacionih čvrstih materijala u Metrološkoj laboratoriji za temperaturu i termofizičke veličine (MLTV) Instituta za nuklearne nauke VINČA. Toplotna provodnost se, kao jedna od najznačajnijih termofizičkih veličina čvrstih materijala, meri različitim eksperimentalnim metodama. Metoda koja kod slabo provodnih materijala može obezbediti rezultate sa najmanjom mogućom mernom nesigurnošću je metoda zaštićene tople ploče i ona se kao referentna koristi u svim najznačajnijim nacionalnim metrološkim laboratorijama u svetu. Eksperimentalna postavka kojom se primenjuje ova metoda je ustanovljena u Institutu VINČA još 60-ih godina prošlog veka i bila je funkcionalna do sredine 80-ih. Toplotna provodnost termoizolacionih materijala se u tom periodu merila sa relativnom mernom nesigurnošću od 3 % do 10 %. U skladu sa ubrzanim razvojem MLTV početkom ovog veka, kao i njenim kasnijim priznavanjem kao imenovane nacionalne laboratorije za termofizičke veličine, pristupilo se revitalizaciji eksperimentalne postavke i unapređenju postupka određivanja toplotne provodnosti. U okviru rada na ovoj disertaciji je unapređena kako složena merna aparatura kroz modelovanje, izradu i testiranje novih elemenata i sistema, tako i kalibracioni i merni postupak i način obrade mernih signala. Kao rezultat širokog unapređenja, danas se u Institutu VINČA toplotna provodnost termoizolacionih čvrstih materijala može meriti sa visokom tačnošću i preciznošću, tj. malom relativnom mernom nesigurnošću koja dostiže nivo od 2 %. Tim rezultatom, koji je potvrđen preko međunarodnog međulaboratorijskog poređenja izvršenog u okviru Eura-Thermal projekta, Metrološka laboratorija za temperaturu i termofizičke veličine Instituta VINČA se svrstala u red značajnih svetskih laboratorija te vrste.

This PhD thesis implies research on improving the thermal conductivity measurement of thermal insulations in the Metrological Laboratory for Temperature and Thermophysical Properties (MLTV) of the Institute of Nuclear Sciences VINČA. Thermal conductivity, as one of the most important thermophysical properties of solid materials, is measured by various experimental methods. A method that may provide results with a lowest possible measurement uncertainty for poorly conductive materials is the guarded hot plate method and it is used as a reference technique in all most important national metrology laboratories in the world. A related experimental setup was established at the Institute VINČA in the 1960s and was functional until the mid-80s. Thermal conductivity of thermal insulation materials was measured in this period with a relative uncertainty of 3 % to 10 %. In accordance with a fast development of MLTV at the beginning of this century, as well as its subsequent recognition as the designated laboratory for thermophysical properties, a revitalization of the experimental setup and an improvement of the corresponding measurement process began. Within the scope of this dissertation, the measurement apparatus was improved through modelling, designing and testing of new elements and systems, as well as through new calibration, measuring and data reduction procedures. As a result of these improvements, today, the Institute VINČA is capable to measure the thermal conductivity of insulation materials with a high accuracy and precision, i.e. with small relative uncertainties down to a level of 2 %. With this result, which was confirmed through an international inter-laboratory comparison performed within the framework of the Eura-Thermal project, the MLTV of the Institute VINČA has been ranked among the world's leading laboratories of this kind.

Functional and imperative reactive programming based on a generalization of the continuation monad in the C++ programming language

Функционално и императивно реактивно програмирање употребом генерализације монаде наставка у програмском језику C++.

Postoji veliki broj problema koji zahtevaju pisanje programskih sistema koji imaju komponente koje se izvršavaju međusobno asinhrono jedne od drugih. Na primer, kod distribuiranih izračunavanja postoje distribu- irani čvorovi koji obrađuju podatke asinhrono jedni od drugih, sistemi sa servisno-orijentisanim arhitekturama moraju da obrađuju zahteve koje do- bijaju od klijenata asinhrono, a u više-jezgarnim i heterogenim sistemima postoji više odvojenih zadataka koji se izvršavaju istovremeno radi što boljeg iskorišćenja hardverskih resursa. Čak i obični grafički programi imaju asinhrone komponente – grafičko korisničko okruženje ne sme da bude zaglavljeno ni u jednom trenutku, što znači da bez obzira na to šta program trenutno radi, ne sme da prestane da obrađuje i reaguje na ulazne događaje koji dolaze od korisnika. Neophodnost asinhronog izvršavanja donosi do- datno usložnjavanje programa zbog umetanja izvrtanja kontrole izvršavanja (IoC – Inversion of Control), a zbog upotrebe sistema za asinhron prenos po- ruka ili zbog upotrebe sistema za obradu događaja. Zbog izvrtanja kontrole izvršavanja, program postaje težak za razvoj i teško razumljiv. U ovoj disertaciji prikazan je jedan metod za rešavanje problema koji nastaju zbog izvrtanja kontrole izvršavanja. Predstavljen je način za mo- delovanje i sinhronih i različitih tipova asinhronih zadataka upotrebom monade nastavka. Monada nastavka služi kao osnovna primitiva za izgrad- nju složenijih struktura kontrole toka izvršavanja programa koje simuli- raju strukture kontrole toka izvršavanja programskog jezika C++. Takođe je predstavljeno da monada nastavka dozvoljava izgradnju naprednijih struk- tura kontrole toka programa koje su specijalizovane za paralelizaciju ili za transakciono izvršavanje. Predstavljen je i metod za simuliranje kon- cepata funkcionalnog programiranja kroz generalizaciju monade nastavka koja dozvoljava asinhronim zadacima da generišu više rezultata nezavisno. Pokazano je da ovi metodi omogućavaju implementaciju programskih sistema sa asinhrono izvršenim komponentama pisanjem naizgled sinhronog impera- tivnog ili funkcionalnog koda, dok je prevodiocu prepušteno da napravi asinhrone zadatke i prevede napisani kod u program kod kog se delovi izvr- šavaju asinhrono. Dodatna prednost predstavljenog metoda je što omogućava automatizaciju rukovanja životnim vekom podataka u programu bez potrebe za skupljačem otpadaka. U uvodnom poglavlju (poglavlje 1) je ukratko opisana motivacija, dopri- nosi i organizacija disertacije. Poglavlje 2 sadrži detaljniji opis pro- blema koji metodi razvijeni u okviru ove disertacije rešavaju, postojeće načine implementacije programskih sistema sa asinhronim komponentama, kao osnovne pojmove teorije kategorija. Upotreba monade nastavka u kon- tekstu asinhronog programiranja i koncepta buućih vrednosti je opisana u poglavlju 3. Ovo poglavlje definiše i reaktivne tokove (generalizaciju monade nastavka, kao i transformacije nad njima) koji omogućavaju imple- mentaciju programskih sistema kao toka podataka i niza transformacija nad tim tokom. Poglavlje 4 definiše strukture kontrole toka programa bazirane na monadi nastavka, dok poglavlje 5 prikazuje implementaciju. Po- slednje poglavlje sadrži diskusiju i poređenje uobičajenih metoda za im- plementaciju programskih sistema sa asinhrono izvršavanim komponenama sa metodima predstavljenim u ovoj disertaciji. Dodaci disertaciji sadrže istoriju ideja prezentovanih u disertaciji (dodatak A), anketu zajednice profesionalnih programera o čitljivosti koda implementiranog metodima predstavljenih u disertaciji u odnosu na alternative (dodatak B), kao i primenu funkcionalno reaktivne tehnike na biblioteku Qt (dodatak C). Metodi predstavljeni u ovoj disertaciji su uspešno primenjeni i testi- rani u nekoliko slobodnih i vlasničkih softverskih projekata koje koriste stotine miliona ljudi širom sveta. U ovoj disertaciji opisan je primer sistema upravljanja projektima koji se zasniva na sličnom sistemu koji je deo projekta KDE Plasma. Ova disertacija takođe sadrži važne delove implementacije biblioteke AsynQt koja proširuje biblioteku Qt, i njenu primitivu za konkurentno izvršavanje – klasu QFuture – funkcionalnim reaktivnim programskim obrascima zasnovanim na metodu predloženom u ovoj disertaciji.

There is a big class of problems that require software systems with asynchronously executed components. For example, distributed computations have the distributed nodes that process the data asynchronously to one anot- her, service-oriented architectures need to process separate requests asynchrono- usly, and multi-core and heterogeneous systems need to have multiple separa- te tasks running concurrently to best utilize the hardware. Even ordinary GUI applications need asynchronous components – the user interface needs to be re- sponsive at all times which means that no matter in what state the program is in, it needs to process and react to the input events coming from the user. The necessity of concurrency and asynchronous execution brings in the added com- plexity of the Inversion of Control (IoC) into the system, either through mes- sage passing or through event processing. IoC makes code difficult to develop and reason about, it increases component coupling and inhibits clean functional or object-oriented software design. In this dissertation, a method for solving the problems that IoC introduces is presented. It presents a way to model both synchronous and different types of asynchronous tasks with the continuation monad. The continuation monad serves as a primitive to build more complex control flow structures that mimic the control flow structures of the host programming language. It also allows for building more complex control structures specialized for parallelism, transactional execution, and for simulating functional programming idioms with asynchronous tasks through a generalization of the continuation monad that allows the asynchronous tasks to generate results one at a time. This allows for writing programming systems with asynchronously executed components by writing seemingly synchronous imperati- ve or functional code while leaving it up to the compiler to do all the heavy lifting and convert the written code to asynchronously executed set of tasks. Another benefit of the presented method is that it allows for easier automatic handling of the data lifetime without the need for garbage collection. This method has been successfully applied and tested in several Free/Libre Open Source Software and proprietary real-world software projects used by hun- dreds of millions of people around the world. In this dissertation, an example of a secure project management system is described which is based on a similar system implemented as a part of the KDE Plasma project. This dissertation also contains the important parts of the implementation of the AsynQt library which extends the Qt library, and its concurrency primitive – QFuture class – with functional reactive programming patterns based on the method proposed in this dissertation.

Metod za detekciju ivica defekata u procesu proizvodnje kartona primenom wavelet transformacije

Wavelet based method for edge detection of defect on coated board during production proccess

Disertacija razmatra problem detekcije ivica defekata u slikama kartona upotrebom wavelet transformacije. Osobine wavelet transformacije, pre svega predstavljanje singulariteta u signalu malim brojem koeficijenata omogućava realizaciju efikasnog detektora ivica. U okviru ove doktorske disertacije dat je detaljan pregled postojećih detektora ivica baziran na diferenciranju i upotrebi Gausovog filtra. Veoma iscrpno su opisani i postupci zasnovani na wavelet transformaciju. Analizirani su načini poređenja predloženih sa drugim detektorima ivica i veoma su retki slučajevi u kojima autori primenjuju objektivne metode poređenja detektora, a i tada su rezultati komparacije zasnovani na sintetičkoj slici, zbog čega se izvedeni zaključci ne mogu generalizovati na realne slike. Ilustrovani su nedostaci klasičnog detektora ivica kada se primene na slike defekata. Predstavljene su najvažnije karakteristike wavelet transformacije i data je teorijska osnova detekcije singulariteta u signalu upotrebom wavelet transformacije. Kako karakteristike signala imaju uticaj na performanse detektora ivica, ispitivanjem slika defekata na kartonu izveden je matematički model ivice defekata. Na osnovu matematičkog modela i karakteristika slika defekata na kartonu ustanovljeno je da se bolji rezultat ostvaruje kada se u proizvod uključe tri skale wavelet transformacije, a ne samo dve kako je prvobitno predložno, jer je, pre svega, izraženije potiskivanje šuma. Zatim, pokazano je da se predloženi algoritam može primeniti na proizvoljan skup slika, pri čemu se na osnovu karakteristika slika može utvrditi početna skala za formiranje proizvoda koeficijenata wavelet transformacije. U nastavku, prikazane su i analizirane dostupne metodologije komparacije detektora ivica. Ustanovljeno je da se mora primeniti objektivna metoda zasnovana na korišćenju istinite mape ivica. Da bi se ostvarilo adekvatno poređenje, realizovana je baza od 50 slika defekata na kartonu sa odgovarajućim istinitim mapama ivica. Komparacija detektora ivica je izvršena i na osnovu postojeće baze svakodnevnih slika i slika iz vazduha (slike iz ptičje perspektive), kao i odgovarajućih mapa ivica. Za poređenje su izabrani klasični i najčešće korišćeni detektori ivica: Sobel, Canny i Marr-Hildreth, zatim dva detektora bazirana na wavelet transformaciji i noviji, često korišćen detektor ivica – SUSAN (eng. Smallest Univalue Segment Assimilating Nucleus) detektor...

This thesis considers edge detection of defects on a coated board based on the wavelet transform. Properties of the wavelet transform, above all the possibility to represent singularity in the signal with a few coefficients, gives opportunity to realize the efficient edge detector. This thesis gives a detailed description of existing edge detection methods based on differentiation and Gaussian filtering with in-depth review of the wavelet transform techniques. It is analyzed how authors compare suggested methods with other edge detectors. If was found that only few authors use objective evaluation and those comparisons are based on a synthetic image and cannot be applied to real images. Shortcomings of classical edge detector when it is used with coated board images are shown. The most important properties of the wavelet transform are presented with theoretical background of singularity detection. Characteristics of the signal have influence on the edge detector performances and the model of an edge on coated board is developed. Using this model it is shown that better result are obtain when three, instead of originally suggested two scale of wavelet transform are multiplied. Additionally, it is shown that proposed algorithm can be applied on arbitrary image set and only initially scale for multiplication must be determined. Subsequently, the edge detector evaluation methods are shown and analyzed. It is concluded that objective methodology based on the ground truth images must be used. For the evaluation set of 50 defect images on coated board and 50 corresponding ground truth images are created. The comparison is also performed with on available sets of 50 object and 10 aerial images. Testing of proposed algorithm is done by comparing it with the classical and frequently used edge detector: Sobel, Canny and Marr-Hildreth; with two edge detectors based on the wavelet transform and one newly and commonly used edge detector – SUSAN (Smallest Univalue Segment Assimilating Nucleus) detector...

Primena Grebnerovih baza na probleme popločavanja

Application of the Grobner bases theory to tiling problems.

Predmet ove doktorske disertacije je primena algebarskih tehnika na jednu od centralnih tema kombinatorike i diskretne geometrije - poliomino poplocavanja. Poliomino poplocavanja su interesantna ne samo matematicarima nego i zicarima i biolozima, a imaju i primenu u racunarskim naukama. U ovoj disertaciji akcenat je stavljen na mogucnost da se posebna klasa problema poplocavanja koja su invarijantna u odnosu na delovanje konacne grupe resi primenom teorije Grebnerovih baza za prstene polinoma nad prstenom celih brojeva Z. Metoda koja se koristi odrazava duboku povezanost izmedu algebre, geometrije i kombinatorike...

Subject of this doctoral thesis is the application of algebraic techniques on one of the central topics of combinatorics and discrete geometry - polyomino tiling. Polyomino tilings are interesting not only to mathematicians, but also to physicists and biologists, and they can also be applied in computer science. In this thesis we put some emphasis on possibility to solve special class of tiling problems, that are invariant under the action of nite group, by using theory of Gr obner basis for polynomial rings with integer coecients. Method used here re ects deep connection between algebra, geometry and combinatorics...

Primena pravila pridruživanja i metoda podržavajućih vektora za predviđanje T - ćelijskih epitopa

Application of association rule and support vector machine technique for T - cell epitope prediction

Istraživanje podataka (eng. Data mining)je interdisciplinarno polje informatike koje se bavi automatskim ili polu-automatskim otkrivanjem znanja u podacima. Osnovni zadatak istraživanja podataka je izdvajanje netrivijalnih, prethodno nepoznatih i potencijalno korisnih obrazaca, odnosa i veza u podacima i statistički značajnih struktura iz velikih kolekcija podataka. Imperativ je da dobijeni rezultati budu novi, valjani, korisni i razumljivi. Tehnike za istraživanje podataka uključuju statističke modele, matematičke algoritme i metode mašinskog učenja...

Data mining is an interdisciplinary subfield of computer science, including various scientific disciplines such as: database systems, statistics, machine learning, artificial intelligence and the others. The main task of data mining is automatic and semi-automatic analysis of large quantities of data to extract previously unknown, nontrivial and interesting patterns. Rapid development in the fields of immunology, genomics, proteomics, molecular biology and other related areas has caused a large increase in biological data. Drawing conclusions from these data requires sophisticated computational analyses. Without automatic methods to extract data it is almost impossible to investigate and analyze this data. Currently, one of the most active problems in immunoinformatics is T - cell epitope identification. Identification of T - cell epitopes, especially dominant T - cell epitopes widely represented in population, is of the immense relevance in vaccine development and detecting immunological patterns characteristic for autoimmune diseases. Epitope-based vaccines are of great importance in combating infectious and chronic diseases and various types of cancer. Experimental methods for identification of T - cell epitopes are expensive, time consuming, and are not applicable for large scale research (especially not for the choice of the optimal group of epitopes for vaccine development which will cover the whole population or personalized vaccines). Computational and mathematical models for T - cell epitope prediction, based on MHC-peptide binding, are crucial to enable the systematic investigation and identification of T - cell epitopes on a large dataset and to complement expensive and time consuming experimentation [16]. T - cells (T - lymphocytes) recognize protein antigen(s) only when degradated to peptide fragments and complexed with Major Histocompatibility Complex (MHC) molecules on the surface of antigen-presenting cells [1]. The binding of these peptides (potential epitopes) to MHC molecules and presentation to T - cells is a crucial (and the most selective) step in both cellular and humoral adoptive immunity. Currently exist numerous of methodologies that provide identification of these epitopes. In this PhD thesis, discussed methods are exclusively based on peptide sequence binding to MHC molecules. It describes existing methodologies for T - cell epitope prediction, the shortcomings of existing methods and some of the available databases of experimentally determined linear T - cell epitopes. The new models for T - cell epitope prediction using data mining techniques are developed and extensive analyses concerning to whether disorder and hydropathy prediction methods could help understanding epitope processing and presentation is done. Accurate computational prediction of T cell epitope, which is the aim of this thesis, can greatly expedite epitope screening by reducing costs and experimental effort. These theses deals with predictive data mining tasks: classification and regression, and descriptive data mining tasks: clustering, association rules and sequence analysis. The new-developed models, which are main contribution of the dissertation are comparable in performance with the best currently existing methods, and even better in some cases. Developed models are based on the support vector machine technique for classification and regression problems. A new approach of extracting the most important physicochemical properties that influence the classification of MHC-binding ligands is also presented. For that purpose are developed new clustering-based classification models. The models are based on k-means clustering technique. The second part of the thesis concerns the establishment of rules and associations of T - cell epitopes that belong to different protein structures. The task of this part of research was to find out whether disorder and hydropathy prediction methods could help in understanding epitope processing and presentation. The results of the application of an association rule technique and thorough analysis over large protein dataset where T cell epitopes, protein structure and hydropathy has been determined computationally, using publicly available tools, are presented. During the research on this theses new extendable open source software system that support bioinformatic research and have wide applications in prediction of various proteins characteristics is developed. A part of this thesis is described in the works [71][82][45][42][43][44][72][73] that are published or submitted for publications in several journals. The dissertation is organized as follows: In section1 is illustrated introduction to the problem of identifying T - cell epitopes, the importance of mathematical and computational methods in this area, as well as the importance of T - cell epitopes to the immune system and basis for functioning of the immune system. In section 2 are described in details data mining techniques that are used in the thesis for development of new models. Section 3 provides an overview of existing methods for predicting the T - cell epitopes and explains the work methodologies of existing models and methods. It pointed out the shortcomings of existing methods which have been the motivation for the development of new models for the T - cell epitope prediction. Some of the publicly available databases with the experimentally determined MHC binding peptides and T - cell epitope are described. In section 4 are presented new developed models for epitopes prediction. The developed models include three new encoding schemes for peptide sequences representation in the form of a vector which is more suitable as input to models based on the data mining techniques. Section 5 reports results of presented new classification and regression models. The new models are compared with each other as well as with currently existing methods for T cell epitope prediction. Section 6 presents the research results of the T - cell epitopes relationship with ordered and disordered regions in proteins. In the context of this chapter summary results are presented which are shown in more detail in the published works [71][82][45][44]. Section 7 concludes the dissertation with some discussion of the potential significance of obtained results and some directions for future work.

Spektroskopska i fotometrijska analiza bliskih galaksija različitih morfoloških tipova

Spectroscopical and photometrical analysis of nearby galaxies of different morphological types

Cilj ove disertacije jeste svod-enje visedimenzionalnog prostora osobina galaksija na najmanji broj dimenzija, dovoljan da ih opise. U tu svrhu primenjena je statisticka analiza nad skupom parametara koji oslikavaju fundamentalne osobine galaksija, na morfoloski reprezentativnom uzorku od 2180 galaksija. Uzorak galaksija koriscen u ovoj disertaciji zasnovan je na Arecibo Legacy Fast ALFA (skraceno, Alfalfa) HI pregledu neba. Znacaj HI pregleda lezi u cinjenici da su galaksije birane iskljucivo na osnovu sadrzaja gasa (HI), tako da nema optickih selekcionih efekata. Iz ovog uzorka koji broji 10000 galaksija, izdvojeno je 2180 galaksija, koje raspolazu optickom spektroskopijom iz baze podataka Sloan Digital Sky Survey (skraceno, SDSS) i fotometrijom od ultraljubicastog (Galaxy Evolution Explorer, skraceno GALEX), preko optickog (SDSS) do bliskog infracrvenog dela spektra (Two Micron All Sky Survey, skraceno 2MASS). Skup analiziranih parametara odabran je uzimajuci u obzir poznate korelacije izmed-u fundamentalnih osobina galaksija, oslanjajuci se na prethodne radove iz ove oblasti. Parametri su detaljno testirani i med-usobno upored-eni, kako bi formirali konacan skup parametra iz veceg parametarskog prostora osobina galaksija. U svrhu izbora, tacnije odred-ivanja i testiranja parametara, prvo je izracunata kinematika zvezda pomocu javno dostupnog programa (pPXF), sa kojim su testirane empirijske i sinteticke biblioteke zvezdanih spektara. Konkretno, izracunata je disperzija brzina zvezda i visi momenti raspodele brzina duz pravca posmatranja. Ovo je do sada najveci uzorak galaksija za koji je izmerena detaljna kinematika zvezda ukljucujuci vise momente raspodele brzina. Velicina uzorka omogucava primenu statistickih testova na visi moment raspodele brzina (h4) za razlicite grupe morfoloskih tipova galaksija, koji su potvrdili ranije indikacije da elipticne i socivaste galaksije dele poreklo. Dalje su izvrsena nova merenja jacine odred-enih apsorpcionih linija (Likovi indeksi), od kojih su pojedine korisni indikatori starosti i metalicnosti, koje takod-e predstavljaju fundamentalne osobine galaksija. U krajnjoj statistickoj analizi, ispostavlja se da metali cnost nema znacajnu ulogu, dok se ukljucivanjem starosti galaksija u analizu bitno menjaju rezultati...

The goal of this thesis is to reduce multidimensional space of galactic properties to the smallest number of dimensions sucient to describe them. For this purpose, the statistical analysis is applied over the parameters that describe fundamental galactic properties on the morphologically representative sample of 2180 galaxies. The sample of galaxies used in this thesis is based on the Arecibo Legacy Fast ALFA (Alfalfa) blind HI survey. The importance of an HI blind survey lies in the fact that galaxies are chosen on the basis of their gas content (HI) solely, thus free of optical selection eects. From the initial sample counting 10000 galaxies, 2180 of them were chosen, since for this subsample the optical spectroscopy from the Sloan Digital Sky Survey (SDSS) was available and moreover the photometry in the UV (Galaxy Evolution Explorer, GALEX), and optical (SDSS) to the near-infrared (Two Micron All Sky Survey, 2MASS). Parameters are selected according to the previously established correlations between fundamental galactic properties, relying on the previous work. They are extensively tested and confronted between each other to be chosen from the larger parametric space. To select parameters, we rst measured stellar kinematics using publicly available code (pPXF), and tested both empirical and synthetic stellar libraries. In particular, we have measured the velocity dispersion and the higher moments of the line-of-sight velocity distribution function. This is the largest galaxy sample created so far with detailed stellar kinematics measured including higher moments of the line-of-sight velocity distribution function. The sample size allows statistical tests to be applied to the higher moment of the velocity distribution function (h4), with respect to the dierent groups of morphological galaxy types. Various tests agree with the previous indication that elliptical and lenticular galaxies have the same origin...

Elementarni operatori i transformacije tipa skalarnog proizvoda na idealima kompaktnih operatora generisanim p-modifikovanim normama i njihovim dualima

Elementary operators and inner product type transformers on ideals of compact operators generated by p-modified norms and their duals.

U ovom radu predstavljene su nejednakosti vezane za norme elementarnih operatora i transformacija tipa skalarnog proizvoda, sa posebnim osvr- tom na Xatenove norme ukoliko su familije operatora koje generixu date transformacije proizvo ne, odnosno Q norme ukoliko se barem jedna od datih familija sastoji od meusobno komutirajuih normalnih operatora. Između ostalog, pokazane su nejednakosti koje uopštavaju nejednakost...

In this paper we present some norm inequalities for certain elementary operators and inner product type transformers, specially for Schatten norms, if the families of operators generating those transforms consists of arbitrary operators, and Q norms if at least one of those families consists of mutually commuting normal operators. Among others, we present inequalities that are generalizing the inequality...

Prepoznavanje artikulaciono-akustičnih odstupanja glasova u patološkom govoru

Detection of articulatory-acoustic deviations in pathological speech

U kliničkoj praksi procena artikulacionih poremećaja uglavnom se vrši pomoću subjektivnih mera i metoda, audio vizuelnom procenim od strane obučenih eksperata- logopeda. U tu svrhu postoji određeni broj standardnih testova i procedura na osnovu kojih se daje kvalitativna i kvantitativna ocena patološkog izgovora. Ovakav pristup podložan je uticaju faktora koji potiču od ispitivača i uslova ispitivanja, a koji nisu posledica patološkog izgovora. Postojanje automatskog sistema za procenu patologije u govoru, u prvom redu bi dovelo do toga da ovakvi faktori postanu zanemarljivi, čime bi se višestruko unapredila klinička praksa. Ovo istraživanje fokusirano je na formiranje jedinstvenog sistema za detekciju patologije govora koji se bazira na aktuelnim dostignućima na polju obrade govora i standardnim metodama koje se u te svrhe koriste u logopedskoj praksi. U radu je prikazan razvoj sistema za procenu artikulacionih odstupanja govora koji je osmišljen tako da prati metodologiju logopedskog pristupa. Predstavljeni su rezultati modelovanja auditornog sistema čoveka prilikom detekcije akustičkih obeležja u govornom signalu koja učestvuju u manifestaciji patologije. Data su i istraživanja na temu segmentacije glasova pogodnih za dalju obradu radi utvrđivanja postojanja i stepena patologije. Prikazan je i rad na izradi automatskog sistema za klasifikaciju govora u smislu postojanja ili odsustva patologije u govornom signalu. Poseban deo posvećen je istraživanjima na temu prepoznavanja specifičnih artikulacionih odstupanja pojedinačnog glasa. Rezultati istraživanja pokazali su da se upotrebom algoritama koji se koriste u obradi govora može izvesti sistem za automatsku procenu patološkog izgovora. Rezultati koji se dobijaju automatskom procenom u visokom procentu se podudaraju sa ocenama dobijenim od strane eksperata.

In clinical practice, dyslalia is usually assessed by subjective measures and methods, using the audio-visual evaluation performed by the trained experts - speech therapists. Qualitative and quantitative evaluation of pathological pronunciation is based on a number of standard tests and procedures. This approach could be influenced by factors arising from the examiner and the test conditions, which do not occur as a consequence of pathological pronunciation. The existence of an automatic system for the evaluation of speech pathology would make these factors negligible, and further improve clinical practice. This research is focused on the system for speech pathology detection that is based on the current achievements in the field of speech processing and standard methods used for this purpose in speech therapy practice. In this dissertation, the development of the system for the assessment of articulation disorders is presented. The system is designed to follow the methods used in speech therapy. The model of human auditory system is presented, adjusted to the detection of the acoustic characteristics of the speech signal that participate in the manifestation of speech pathology. Also, research on automatic speech segmentation is described, which results in the segments suitable for further processing in order to determine the existence and degree of speech pathology. Subsequently, the system for classification of speech in terms of the existence or absence of speech pathology is introduced. A special section is devoted to the research on recognition of the specific articulatory deviations that occur during the pronunciation of individual phonemes. The results showed that it is possible to develop the system for automatic evaluation of speech pathology using the speech processing algorithms. The results obtained by the automatic evaluation match to a great extent to the ones given by the experts.

Asimptotska svojstva rešenja jednačina Emden-Faulera i njihovih uopštenja

Asymptotic properties of solutions of Emden-Fowler equations and their generalizations

Predmet disertacije jeste ispitivanje asimptotskih svojstava resenja jednaqina Emden-Faulera i njihovih uopstena. Oblast ko- joj ova disertacija pripada jeste Kvalitativna analiza obicnih diferencijalnih jednacina...

The subject of this dissertation is the investigation of asymptotic properties of solutions for dierential equations of Emden-Fowler type and their generalizations. The eld to which this dissertation belongs is a Qualitative theory of ordinary dife- rential equations...

Numerička analiza dinamike rezonantnih asteroida pod dejstvom fekta Jarkovskog

Numerical analysis of the dynamics of resonant asteroids under the influence of the Yarkovsky effect.

Predmet ispitivanja ove disertacije je interakcija izmedju rezonanci u srednjem kretanju i efekta Jarkovskog...

The subject of this dissertation is interaction between the mean motion resonan- ces and the Yarkovsky effect...

Algebarska svojstva spektralnih invarijanti u Florovoj homologiji

Algebraic properties of spectral invariants in Floer Homology.

U ovom radu definisemo i analiziramo spektralne invarijante u Florovoj homologiji za konormalno raslojenje i Florovoj homologiji za otvoren skup. Konstrukcija preslikavanja tipa Piunikin Salamon Xvarc, koju mi ovde izvodimo, neophodan je korak za dobru definisanost ovih invarijanti. Postojanje dodatnih algebarskih struktura, kao xto je proizvod na spomenutim Florovim homologijama, omoguava nam da izvedemo razne nejednakosti medju spektralnim invarijantama. Posma- tranjem perturbovano holomorfnih Rimanovih povrxi sa granicom mo- emo da uporedimo invarijante iz raznih Florovih homologija...

In this doctoral dissertation we dene and investigate spectral invariants in Floer homology for conormal bundle and Floer homology of an open sub- set. As a key step to well dened spectral invariants we give a construction of Piunikhin-Salamon-Schwarz isomorphism in both of these homologies. Ad- ditional algebraic structures, such as a product on Floer homology, give us various inequalities between spectral invariants. We can compare spectral in- variants from dierent Floer homologies by observing appropriate perturbed holomorphic Riemmanian surfaces with boundary...

Formiranje supermasivnih crnih rupa i uticaj sudara galaksija na njihovu evoluciju

Supermassive black hole formation and role of galaxy mergers in their evolution.

Mehanizam formiranja supermasivnih crnih rupa posmatranih u ranom Univerzumu nije u potpunosti razjasnjen...

Formation mechanism of supermassive black holes (SMBHs) observed in the early Universe is still not fully understood...

Formalizacija različitih modela geometrije i primene u verifikaciji automatskih dokazivača teorema

Formalization of various geometry models and applications in verification of automated theorem provers

U ovoj tezi predstavljena je interaktivna formalizacija modela raznih geometrija i algebarskih metoda automatskog dokazivanja geometrijskih te- orema. Predstavljen je naš rad na formalizaciji analitičke (Dekartove) pla- narne geometrije u okviru asistenta za dokazivanje teorema Isabelle/HOL. Dajemo nekoliko ekvivalentnih definicija Dekartove koordinatne ravni i dokazujemo da je ona model sintetičkih planarnih geometrija (korišćenjem aksiomatskog sistema Tarskog i aksiomatskog sistema Hilberta). Takođe, diskutujemo o nekoliko tehnika kojima se pojednostavljuju i automatizuju dokazi. Kako je jedan od naših ciljeva da podržimo korišćenje asistenata za dokazivanje teorema u matematičkom obrazovanju, naše izlaganje će biti blisko standardnim definicijama u udžbenicima, ali potpuno formalno i mašinski proverivo. Ova formalizacija predstavlja deo potrebne infra- strukture za implementaciju procedura odlučivanja baziranih na analitič- koj geometriji u okviru asistenata za dokazivanje teorema. Dodatno, formalno razmatramo i geometriju kompleksne ravni. Bliska povezanost između kompleksnih brojeva i geometrije je dobro poznata i pa- žljivo je izučavana još vekovima unazad. Osnovni objekti koji su izučavani su kompleksna ravan (obično proširena jednom beskonačnom tačkom), njeni objekti (tačke, prave i krugovi), i grupa transformacija koja na njih de- luje (na primer, inverzije i Mebijusove transformacije). U ovoj tezi, mi formalno posmatramo geometriju kompleksnih brojeva i predstavljamo pot- puno mehanički verifikovanu teoriju u okviru asistenta za dokazivanje te- orema Isabelle/HOL. Diskutujemo o različitim pristupima formalizaciji i glavnim prednostima pristupa koji je više algebarski orijentisan. Pored primena u formalizaciji matematike i u obrazovanju, ovaj rad je osnova za formalnu analizu neeuklidskih geometrija i njihove međusobne povezanosti. Takođe, predstavljamo i formalizaciju dela aksiomatskog sistema Tarskog u okviru Poenkareovog disk modela u sistemu Isabelle/HOL. Treće, analiziramo vezu između geometrije i polinoma, kao i primene koja ova veza daje. U euklidskoj geometriji objekti i relacije među njima mogu se izraziti polinomijalnim jednakostima. Dodatno, bilo koja geome- trijska konstrukcija može se izraziti skupom polinomijalnih jednakosti, a geometrijska tvrđenja se mogu dokazati korišćenjem algebarskih metoda (na primer, metod Grebnerovih baza ili Vuovom metodom) nad skupom polino- ma. Dajemo opis algoritma u sistemu Isabelle/HOL koji kao ulazni podatak prihvata geometrijsku konstrukciju zapisanu korišćenjem termova, a vraća odgovarajući skup polinoma. Naš dalji rad će biti primena metode Greb- nerovih baza u okviru sistema Isabelle/HOL nad generisanim polinomima u nameri da se dokaže ispravnost date konstrukcije. Dodatno, istražujemo kako se konstrukcije u trodimenzionalnom prostoru mogu prikazati kori- šćenjem polinoma. Istražujemo dva različita pristupa u izvođenju ovih polinoma i onda poredimo efikasnost algebarskih metoda u zavisnosti od korišćenog pristupa. Predstavljamo potpuno automatski sistem za prevođe- nje geometrijskih konstrukcija iz stereometrije u skup polinoma. Naš dalji rad će biti da povežemo predstavljeni sistem sa dinamičkim geometrijskim softverom i na taj način da omogućimo studentima lakše korišćenje ovog automatizovanog sistema za dokazivanje u stereometriji.

In this thesis is presented interactive formalization of various models of geom- etry and algebraic methods for automated proving geometry theorems. We present our current work on formalizing analytic (Cartesian) plane geome- tries within the proof assistant Isabelle/HOL. We give several equivalent defini- tions of the Cartesian plane and show that it models synthetic plane geometries (using both Tarski’s and Hilbert’s axiom systems). We also discuss several tech- niques used to simplify and automate the proofs. As one of our aims is to advocate the use of proof assistants in mathematical education, our exposure tries to re- main simple and close to standard textbook definitions, but completely formal and mechanically verifiable. This formalization presents the develop of the necessary infrastructure for implementing decision procedures based on analytic geometry within proof assistants. Furthermore, we investigate complex numbers. Deep connections between com- plex numbers and geometry had been well known and carefully studied centuries ago. Fundamental objects that are investigated are the complex plane (usually ex- tended by a single infinite point), its objects (points, lines and circles), and groups of transformations that act on them (e.g., inversions and Möbius transformations). In this thesis we treat the geometry of complex numbers formally and present a fully mechanically verified development within the theorem prover Isabelle/HOL. We discuss different approaches to formalization and discuss major advantages of the more algebraically oriented approach. Apart from applications in formalizing mathematics and in education, this work serves as a ground for formally inves- tigating various non-Euclidean geometries and their intimate connections. We also present a formalization of part of Tarski axiom system withing Poincare disk model in Isabelle/HOL. Further on, we analyze connections between geometry and polynomials and the use of these connections. In Euclidean geometry, objects and relations between them can be expressed as polynomials. Further, any geometry construction can be expressed by set of polynomials and geometry statements can be proved by using algebraic methods (e.g. the Gröbner bases method or Wu’s method) over that set of polynomials. We describe an implementation of an algorithm in Isabelle/HOL that accepts a term representation of a geometry construction and returns its cor- responding set of polynomials. Our further work will be to use the method of Gröbner bases within the Isabelle system on the generated polynomials, in order to prove correctness of the given construction. Furthermore, we investigate how spatial geometry constructions can be presented using polynomials. We inves- tigate two different approaches in deriving those polynomials and then compare efficiency of algebraic provers depending on the approach used. We present a fully automated system for transforming geometry constructions into set of polynomi- als. Our further work would be to relate these geometry provers with dynamic geometry software and thus make easier for students to use it.

Uticaj dinamičke polarizacije na interakciju naelektrisanih čestica sa ugljeničnim nancevima u dvofluidnom hidrodinamičkom modelu

The influence of dynamic polarization on charged particles interaction with carbon nanotubes in two - fluid hydrodynamic model

U ovoj disertaciji se razmatra uticaj efekata dinamičke polarizacije na kanalisanje naelektrisanih čestica kroz različite tipove jednoslojnih ugljeničnih nanocevi (SWNT). Na početku analize razmatra se interakcija naelektrisanih čestica sa četiri različita tipa nanocevi u okviru linearizovanog dvodimenzionog jednofluidnog i dvofluidnog hidrodinamičkog modela. Tipovi nanocevi koji se razmatraju su SWNT(6, 4), SWNT(8, 6), SWNT(11, 9) i SWNT(15, 10). Pomenuti hidrodinamički modeli se koriste za računanje potencijala lika gde se na taj način razmatra uticaj na kretanje naelektrisanih čestica duž putanje paralelne sa osom nanocevi. Brzine kretanja protona se uzimaju u opsegu od 1 do 10 a.u.. Proton srednje kinetičke energije (reda MeV) izaziva pojavu snažne dinamičke polarizacije valentnih elektrona na omotaču nanocevi što kao efekat ima indukovanje značajne sile lika na proton, odnosno pojavu gubitaka energije usled ekscitacije pomenutih elektrona. Pokazano je da dinamička sila lika izaziva snažan uticaj na ugaonu raspodelu kanalisanih protona kroz kratke nanocevi. Ustanovljeno je da su ove nove pojave posebno izražene kada se brzina kanalisanih čestica poklapa sa faznom brzinom kvaziakustičnog π plazmona. Analiza je uključila generisanje numeričkih rezultata kojima je prikazan uticaj faktora prigušenja, radijusa nanocevi i početne pozicije kanalisane čestice na potencijal lika unutar nanocevi. Izvršeno je poređenje dobijenih rezultata za potencijal lika za slučaj jednofluidnog i dvofluidnog modela za različite tipove nanocevi. Takođe je izvršena računarska simulacija kanalisanja čestica kao i poređenje prostorne i ugaone raspodele kanalisanih čestica u slučaju pomenutih hidrodinamičkih modela. U nastavku se analizira interakcija protona sa SWNT(6, 4) gde se uzimaju u obzir efekti dinamičke polarizacije u okviru dvodimenzionog proširenog hidrodinamičkog modela. Ovaj model se koristi za analitičko i numeričko računanje potencijala lika i zaustavne sile na proton koji se kreće paralelno sa osom nanocevi u slučajevima kada je trajektorija njegovog kretanja unutar i van nanocevi. Opseg brzina koji se razmatra je od 0.5 do 15 a.u.. Razmatra se uticaj efekata različitih ugaonih modova na zavisnost potencijala lika od brzine protona za različite tipove nanocevi. Takođe se računa vi prostorna i ugaona distribucija protona u slučaju proširenog dvofluidnog modela i poredi sa slučajem običnog dvofluidnog modela sa nultim faktorom prigušenja. Na kraju se analizira interakcija naelektrisanih čestica sa pravim i zakrivljenim jednoslojnim nanocevima pri uslovima kanalisanja kada je uračunata dinamička polarizacija valentnih elektrona ugljenika. Polarizacija je opisana linearizovanim dvofluidnim hidrodinamičkim modelom sa parametrima određenim iz nekoliko nezavisnih eksperimenata u vezi sa spektroskopijom gubitaka energije u ugljeničnim nanostrukturama. Hidrodinamički model se koristi za izračunavanje potencijala lika indukovanog kretanjem protona kroz četiri pomenuta tipa nanocevi pri brzini od 3 atomske jedinice. Potencijal lika se potom dodaje na Dojl-Tarneov atomski potencijal u cilju dobijanja ukupnog potencijala u pravim i zakrivljenim nanocevima. Na osnovu ovih proračuna simulira se proces kanalisanja protonskog snopa i određuje prostorna i ugaona raspodela kanalisanih protona kroz zakrivljene nanocevi i porede dobijeni rezultati sa slučajem kada se ne uzima u obzir dinamička polarizacija i potencijal lika.

In this dissertation the effects of dynamic polarization on charged particles channeling through various types of single – walled carbon nanotubes (SWNTs) are studied. At the very beginning of the analyze the interactions of charged particles with 4 different types of SWNTs by means of linearized two dimensional one and two fluid hydrodynamic models are studied. Types of SWNTs are (6, 4), (8, 6), (11, 9) and (15, 10). The models are used to calculate the image potential for a charged particle moving parallel to the axis of the SWNTs. Proton speeds between 1 and 10 a.u. are chosen. A proton that moves with average energy (MeV) will induce a strong dynamic polarization of valence electrons in the nanotubes which in turn will give rise to a sizeable image force on the proton, as well as a considerable energy loss due to the collective, or plasma, excitations of those electrons. The dynamic image force was shown to exert large influence in the angular distributions of protons channelled through short SWNTs. It is found that these quantities exhibit novel features when the particle speed matches the phase velocity of the quasiacoustic π plasmon. Numerical results are obtained to show the influence of the damping factor, the nanotube radius, and the particle position on the image potential inside the nanotube. Results for image potential in the one and two fluid hydrodynamic models are compared for different types of nanotubes. The spatial and angular distributions of protons are also computed and compared for the two models. After that, we study the interaction of charged particles with a SWNT(6, 4) under channelling conditions by means of the linearized, two dimensional (2D), two-fluid extended hydrodynamic model. We use the model to calculate analytically and numerically the image potential and the stopping force for a proton moving parallel to the axis of the SWNT, both inside and outside the nanotube at the speeds from 0.5 a.u. to 15 a.u.. The effects of different angular modes on the velocity dependence of the image potential are compared for a proton moving in different types of SWNTs. We also compute the spatial and angular distributions of protons in the 2D two-fluid extended hydrodynamic model and compare them with the 2D two- fluid hydrodynamic model with zero damping. At the end we investigate the interaction of charged particles with straight and bent single-walled carbon nanotubes under channelling conditions in the presence of dynamic polarization of the valence electrons in carbon nanotube wall. This polarization is described by a viii linearized, two-fluid hydrodynamic model with the parameters taken from recent modelling of several independent experiments on electron energy loss spectroscopy of carbon nanostructures. We use the hydrodynamic model to calculate the image potential for protons moving through four types of SWNTs at the speed of 3 atomic units. The image potential is then combined with the Doyle-Turner atomic potential to obtain the total potential in the bent carbon nanotubes. Based on that potential, we also compute the spatial and angular distributions of protons channeled through the bent carbon nanotubes, and compare the results with the distributions obtained without taking into account the image potential.

Analysis of optimal sizing and location of shunt capacitors in active distribution networks

Analiza optimalne snage i lokacije otočnih baterija kondenzatora u aktivnim distributivnim mrežama

Pove anje pot a nje za elekt čnom ene g jom zahteva pove anje b oja g ana m e , pove anje p o zvodnje, pot eb za p o enjem m e e, pove anje kapac teta t ansfo mato sk h stan ca kao ostal h elemenata m e e. Ove akt vnost s s t n veoma sk pe. Po ed toga, p otok eakt vne snage d st b t vn m s stem ma dodatno pove ava g b tke smanj je napone m e . Ovo je posebno z a eno kod vel k h opte e enja gde se napon značajno smanj j sa daljeno od t ansformatorske stan ce. P tome g b c m e se pove avaj . Kako b se smanj l g b c od ao profil napona u prihvatljivim granicama, koristi se kompenzacija reaktivne snage. Kompenzac ja eakt vne snage mo e b t ko sna ako se p av lno zabe e opt malna lokac ja snaga eđaja za kompenzac j . Ob čno se za kompenzac j eakt vne snage ko ste otočn kondenzato . Otočn kondenzato se ko ste za pobolj anje naponskog p of la, pove anje snage koja se mo e p enet k oz m e pobolj anje ef kasnosti d st b t vn h s stema. Postoje mnog nač n za nstal anje otočn h kondenzato a, al je veoma te ko od ed t najbolje lokac je snage otočn h kondenzato a. Po ed toga, stohast čka p oda opte e enja d st b t vnom s stem dodatno kompl k je ovaj zadatak. C lj ove dokto ske d se tac je je vođenje nove metode za p onala enje opt malne vel č ne lokac je otočn h kondenzato a p menom v ek te j mske opt m zac je. P edlo en metod, zasnovan na p men Monte Ca lo s m lac je, va ava neizvesnosti vezane za snage pot o nje čvo ov ma snage p o zvodnje obnovlj v h zvo a. Po ed toga, ovaj metod z ma obz ko elac j zmeđ sk pova sl čajn h promenljivih. Po t o nja akt vne snage d st b t vnoj m e modelovana je na osnov jednogod nj h me enja napojnoj t ansfo mato skoj stan c . zme en podac s konve tovan k v t ajanja opte e enja, nakon čega je k va t ajanja opte e enja modelovana v estepenom k vom opte e enja p menom tehn ke klaste ovanja. Za vetrogeneratore, podaci o brzini vet a dob jen s na osnov me enja na anal z anoj lokac j za pe od od jedne god ne. zme en podac s modelovan p menom We b ll- XIII ove aspodele g st ne ve ovatno e koja je standa dna aspodela za ove sv he. Sl čajne va jac je s nčevog z ačenja s modelovane Beta aspodelom g st ne ve ovatno e. Na osnov zme en h podataka na anal z anoj lokac j od eđen se pa amet ove aspodele. Optimizacioni problem je modelovan pomo dve k te j mske f nkc je s c ljem pobolj anja naponskog p of la m e optimalnim izborom lokacije i instalisane snage otočn h kondenzato a. C lj p ve k te j mske f nkc je je nap eđenje naponskog p of la m e e m n m zac jom s me kvad ata odst panja oček vane v ednost napona od referentne vrednosti. Cilj druge kriterijumske funkcije je minimizacija ukupne nstal sane snage otočn h kondenzato a. Jasno je da ako je nstal sana snaga otočn h kondenzato a mala, to e ez lt at manj m pobolj anjem p of la napona ob n to. Prema tome, ove dve kriterijumske funkcije su kontradikto ne. P edlo ene k te j mske funkcije su s m ltano opt m zovane p menom genet čkog algo tma sa so t anjem prema nedominaciji (algoritam NSGA-II). Algoritam NSGA- zab an je zato to p edstavlja jedn od standa dn h p oced a za e avanje p oblema v ek te j mske optimizacije. Pored toga, algoritam NSGA- se često ko st za e avanje p oblema opt malne loakc je otočn h kondenzato a. P edlo en metod je p menjen na ealnoj d st b t vnoj m e s ednjeg napona koja snabdeva nekoliko manjih mesta. Uprkos proizvodnji snage iz obnovljivih izvora, ovoj d st b t vnoj m e se maj značajn padov napona znatn g b c akt vne snage. P menom p edlo ene metode na d st b t vn m e za t azl č ta scena ja, dob jena s opt malna e enja za svak od scena ja. Svako e enje daje snage lokac je za g adnj otočn h kondenzato a anal z anoj m e . Set opt maln h e enja gene alno daje bolje nfo mac je o p oblema opt m zac je. Po ed toga, kada postoj v e komp om sn h opt maln h e enja, ma se ve a sloboda zbo konačnog e enja, je se najbolje e enje mo e zab at z pon đenog sk pa e enja. Kod prvog i drugog scenarija analizirana je instalacija jednog, odnosno dva otočna kondenzato a, sa bez z manja obz ko elac je zmeđ sl čajnih promenljivih. Kod t e eg scenarija analizirana je instalacija t otočna kondenzato a sa va avanjem ko elac je zmeđ sl čajn h va jabl . Uz odgova aj a po eđenja odab an je jedan scena o za dalj anal z . Po to s sva dob jena e enja opt malna, izbor konačnog e enja zahtevao je dodatnu analizu. Da bi se izabralo najbolje e enje l e enje koje najmanje na ava dozvoljene granice, u tezi je predlo eno po eđ vanje XIV m n maln h maks maln h oček van h v ednost napona kompletnog seta optimalnih e enja za najv e najn e n voe opte e enja. Za zab ano e enje z ač nat s oček van g b c akt vne snage i energije pre i nakon ugradnje otočn h kondenzatora kako bi se utvrdila oček vana god nja teda akt vne energije. Po ed teda akt vne energ je, za zab ano e enje s z ač nat t o kov g adnje kondenzato a. Po eđenjem teda t o kova pokazano je da g adnja kondenzato a ma p n ekonomsk sm sao.

The increase in power demand requires an increase in the number of feeder or feeder capacity, more generation and expand the network by increasing substation capacity as well as equipment capacity. These options are expensive. In addition, the flow of reactive power in distribution systems increases the losses and reduces the line voltages particularly at heavy loads where the voltage at buses reduces when moved away from the substation, and the losses are high. In order to minimise those losses and to maintain the voltage profile within acceptable limits, reactive power compensation is used. Reactive power compensation can be beneficial if they are correctly applied by choosing the correct location and size of the compensation. Commonly, the capacitors are used for the reactive power compensation. Shunt capacitors are used for a voltage profile improvement, maximise transmitted power flowing through the networks and improving the efficiency of distribution systems. There are many ways to install shunt capacitors, but it is very difficult to determine the best locations and size of shunt capacitors. In addition, the stochastic nature of loads in systems makes the placement becomes more complicated. The aim of this thesis is to introduce a new method for finding the optimal size and location of the shunt capacitors using multiobjective optimisation. The proposed method, based on the application of the Monte Carlo Simulation methods, respects the uncertainties associated with load demand and renewable sources power production. In addition, this method takes into account the correlation between sets of random variables. Active power consumption for the distribution network has been modeled by using one-year measurements at the feeding transformer station. The measured data converted to the load duration curve then the load duration curve was modeled by the multistep load model by applying clustering technique. For wind turbines, data on wind speed have been obtained by the measurements taken at the analyzed location for the period of one year. The Weibull PDF (Probability Density Function) adequately models the measured wind data thus the parameters of Weibull PDF were calculated. In the X same time, Beta PDF appropriately models the random variations of the solar radiation. On the basis of the measured data, the shape parameters of Beta PDF were obtained. The optimisation problem was modeled by two objective functions with the aim of improving voltage profile in a network by optimal selection of the locations and installed powers of the shunt capacitors. The first objective function aims to improve voltage profile of the network by minimising the sum of square deviations of the expected voltage value from the reference value. The goal of the second objective function is minimisation the sum of installed powers of shunt capacitors. Here, if the installed power of the shunt capacitors is small, this will result in a slight voltage profile improvement and vice versa. Therefore, the two objective functions are contradictory. The suggested objective functions were optimised simultaneously by applying the Non- dominated Sorting Genetic Algorithm II (NSGA-II). NSGA-II algorithm was chosen because it represents one of the standard procedures for solving multiobjective optimisation problems. In addition, NSGA-II algorithm has been widely used for solving the problem of optimum placement of shunt capacitors. The proposed method was applied to a real medium voltage distribution network supplying several small communities. Despite the production from renewable energy sources in this distribution network, there are significant voltage drops in the network and considerable active power losses. By applying this method to the distribution network in three scenarios, a set of optimal solutions was obtained for every scenario. Every solution gives specific powers and locations to install the shunt capacitors in the analyzed network. A set of optimal solutions generally gives more a complete description of the optimisation problem. In addition, because there are more compromise optimal solutions, more freedom in choosing the final solution will be given where the best suits solution can be chosen from the offered set of solutions. The first and the second scenarios contained the installation of only one and two shunt capacitors respectively with and without taking into account the correlation between the random variables. The third scenario analyzed the installation of three shunt capacitors with taking into account the correlation between the random variables. With some comparisons, resulting in one scenario has been taken for further analysis. All obtained solutions are optimal therefore an additional analysis is required. To choose the best solution or the solution that contains the least violations of the permitted XI limits, this thesis proposed comparing the minimum and maximum expected voltage values of the complete set of optimum solutions for the highest and the lowest load levels. Also, the expected active power and energy losses prior to and after building-in shunt capacitors were calculated to determine the expect annual active energy saving by the chosen solution. The cost of the expected annual energy saving and the cost of building-in shunt capacitors have been shown and compared to verify whether the building-in shunt capacitors from the chosen solution is a good economic investment.

Nelinearno valjanje broda pod dejstvom vetra i talasa

Nonlinear rolling of ships due to wind and waves

Proračun valjanja broda na uzburkanom moru predstavlja jedan od najznačajnijih problema dinamike broda. Uobičajeno je da se problem rešava linearnom jednačinom valjanja broda na talasima, pri čemu se uticaj vetra zanemaruje, a dobijena rešenja važe samo za male uglove nakretanja. Nasuprot tome, istraživanje predstavljeno u ovom radu zasniva se na nelinearnim diferencijalnim jednačinama kretanja broda izloženog stohastičkom dejstvu neregularnih bočnih talasa i udara bočnog vetra koje simuliraju ponašanje broda u realnoj bočnoj oluji. Statističkom analizom zakona valjanja određuje se verovatnoća gubitka stabiliteta broda: prevrtanja, velikih uglova nakretanja ili naplavljivanja kroz nezaštićene otvore. Ovaj postupak nazvan je probabilistička analiza sigurnosti broda u oluji. U radu su predstavljeni matematički modeli sa jednim i sa dva stepena slobode: nezavisne nelinearne jednačine valjanja i spregnute nelinearne nezavisne jednačine valjanja i zanošenja morskih brodova izloženih dejstvu promenljivog bočnog vetra i neregularnih talasa i rečnih brodova pod dejstvom promenljivog bočnog vetra. Svaki model sadrži više izvora nelinearnosti: moment stabiliteta aproksimiran polinomom visokog stepena, nelinearno prigušenje kretanja, moment vetra. Posebna pažnja poklonjena je ispravnom modelovanju haotičnog, udarnog vetra. Probabilistička analiza može se koristiti u raznovrsnim problemima sigurnosti broda. Razvoj novih propisa o stabilitetu broda jedna je od oblasti u kojima ovaj pristup nalazi primenu. U radu je predstavljen koncept mogućih probabilističkih propisa o stabilitetu morskih i rečnih brodova. Probabilistički postupak primenjen je i u istraživanju uzroka pomorskih nesreća (nesreća broda Kugar Ejs, iz 2006. godine). Primena probabilističke analize u postupku projektovanja na osnovu rizika omogućava uvođenje inovativnih tehničkih rešenja nastalih u cilju povećane efikasnosti i sigurnosti. Definicije opsega prihvatljivih metacentarskih visina i optimalne metacentarske visine morskih brodova predstavljaju verovatno najznačajniji doprinos rada projektovanju na osnovu rizika. Osim toga, rezultati analize sigurnosti brodova unutrašnje plovidbe, dobijeni probabilističkim postupkom, iskorišćeni su kako bi se dobile neke smernice za projektovanje rečnih kontejnerskih brodova.

Rolling of ships in severe weather is one of the most important problems of ship dynamics, which is typically solved by using a linear differential equation of rolling; the influence of wind is neglected and the obtained solutions are valid for small roll angles only. However, the research presented in this thesis makes use of a very dissimilar approach, based on dynamics of a ship subjected to stochastic action of irregular beam waves and gusting beam wind. Acquired time history is statistically analyzed in order to assess the probability of a ship stability failure: capsize, large amplitude rolling or flooding through unsecured openings. This approach is denoted as the probabilistic analysis of ship safety. The thesis introduces following mathematical models, developed in the course of the research: nonlinear differential equations of roll and coupled nonlinear differential equations of sway and roll of seagoing ships exposed to beam waves and wind gusts and inland vessels subjected to gusting wind. Probabilistic analysis may be used in a variety of ship safety problems. The thesis introduces concepts of possible future probabilistic ship stability regulations, both for the seagoing and inland vessels. The thesis also proposes options for the improvement of the present stability regulations for inland container vessels. The probabilistic approach was used to conduct analysis of one of the most significant maritime accidents in 2006, the car carrier Cougar Ace case. The application of the probabilistic analysis in the risk-based ship design emerges as a promising technique for introduction of innovative solutions that may provide increased efficiency and safety. Probably the most significant contribution of the thesis to the field of risk-based design is the definition of the range of acceptable metacentric heights of seagoing ships. Additionally, the probabilistic analysis was used to develop some guidelines for design of inland container vessels.

Modeli konačnih stanja u ekstrakciji informacija

Finite state models in information extraction

Disertacija je posvećena istraživanju naučne oblasti nazvane ekstrakcija informacija (engl. information extraction), koja predstavlja podoblast veštačke inteligencije, a u sebi kombinuje i koristi tehnike i dostignuća više različitih oblasti računarstva. Termin "ekstrakcija informacija" će biti korišćen u dva različita konteksta. U jednom od njih misli se na ekstrakciju informacije kao naučnu oblast i tada će se koristiti skraćenica IE, preuzeta iz anglosaksonske literature u značenju "Information Extraction". U drugom slučaju, kada se bude mislilo na sam proces i postupak izdvajanja informacija iz teksta, koristiće se oblik "ekstrakcija informacija". Ova disertacija predstavlja, pored pregleda postojećih metoda iz ove oblasti, i jedan originalni pristup i metod za ekstrakciju informacija baziran na konačnim transduktorima. Tokom istraživanja i rada na disertaciji, a primenom pomenutog metoda, kao rezultat formirana je baza podataka o mikroorganizmima koja sadrži fenotipske i genotipske karakteristike za 2412 vrsta i 873 rodova, namenjena za istraživanja iz oblasti bioinformatike i genetike. Baza i korišćeni metod su detaljno prikazani u nekoliko radova, publikovanih u časopisima ili izlaganih na međunarodnim konferencijama (Pajić, 2011; Pajić i sar. 2011a; Pajić i sar. 2011b) U glavi 1 dat je uvod u oblast ekstrakcije informacije, unutar koga je opisan istorijat i razvoj metoda ove oblasti. Dalje je opisana klasifikacija tekstualnih resursa nad kojima se vrši ekstrakcija informacija, kao i klasifikacija samih informacija. Na kraju glave 1 oblast ekstrakcije informacije je upoređena sa drugim srodnim disciplinama računarstva. Glava 2 je posvećena prikazu teorijskih osnova na kojima su zasnovana istraživanja ove disertacije. Razmatrana je teorija formalnih jezika i modela konačnih stanja, kao i njihova uzajamna veza i veza sa ekstrakcijom informacija. Akcenat je stavljen na konačne modele i metode koji su zasnovani na modelima konačnih stanja. Ovi metodi pokazuju veću preciznost od drugih metoda za ekstrakciju informacije, te su nezamenljivi u situacijama kada je tačnost izdvojenih podataka iz teksta od presudnog značaja. Pojedini pojmovi ekstrakcije informacija - jezik relevantnih informacija, jezik izdvojenih informacija, pravila ekstrakcije, definisani su iz ugla teorije formalnih jezika. Formulisano je i dokazano osnovno svojstvo relacije transdukcije za zadato pravilo ekstrakcije. Definisan je i pojam jezika konteksta informacija i dokazano je njegovo svojstvo regularnosti. U glavi 3 su prikazana dva različita pristupa projektovanju sistema za ekstrakciju informacija, (1) pristup baziran na mašinskom učenju i (2) pristup baziran na znanju, tzv. Knowledge Engineering pristup (u daljem tekstu KE). Sistemi koji koriste prvi pristup podrazumevaju postojanje anotiranih korpusa relevantnih tekstova, dok sistemi zasnovani na KE pristupu zahtevaju postojanje stručnjaka iz određene oblasti, koji su upoznati sa domenom u kome se vrši ekstrakcija informacija i koji definišu jasna pravila koja se koriste za ekstrakciju informacija iz teksta. Većina današnjih sistema za ekstrakciju informacija kombinuje ova dva pristupa, pa su tako pojedine komponente sistema bazirane na mašinskom učenju, a pojedine na pravilima za ekstrakciju. Detaljan prikaz arhitekture sistema za ekstrakciju informacija, kao i pregled postojećih sistema dat je na kraju glave 3, kao i prikaz programskog sistema UNITEX koji je i korišćen u okviru rada na ovoj disertaciji za potrebe ekstrakcije informacija. Glava 4 je posvećena prikazu originalnog metoda za ekstrakciju informacije koji je zasnovan na konačnim transduktorima i pripada grupi sistema koji su bazirani na KE pristupu. Ovaj metod jasno razlikuje dve faze procesa ekstrakcije. U prvoj fazi se vrši identifikacija i ekstrakcija delova teksta koji odgovaraju jednom entitetu, tj. jednom slogu podataka, dok se u drugoj fazi iz tih manjih delova teksta izdvajaju informacije o vrednostima pojedinih atributa tog entiteta. Prva faza jako zavisi od strukture teksta i ne mora da se izvrši pomoću transduktora. Algoritam prve faze se menja u zavisnosti od teksta iz koga se vrši ekstrakcija. Sa druge strane, druga faza uopšte ne zavisi od strukture teksta, već se u okviru nje primenjuju transduktori koji definišu pravila ekstrakcije. Ove transduktore konstruišu stručnjaci iz specifičnih oblasti. Imajući u vidu da se u okviru velikog broja domena na sličan način opisuju slični koncepti, ovi transduktori mogu biti ponovno korišćeni za neke druge procese ekstrakcije iz iste oblasti. U glavi 4 je detaljno opisan gore pomenuti metod, prikazan je primer njegove primene na enciklopedijski tekst iz oblasti mikrobiologije kao i rezultujuća baza podataka i data je evaluacija metoda. Glava 5 posvećena je problemu ekstrakcije informacije i obrade tekstova na srpskom jeziku, kao jednom od morfološki bogatih jezika, koji zbog te svoje osobine zahteva nešto drugačiji pristup u obradi nego što je to slučaj sa jezicima kakav je recimo engleski jezik. Ekstrakcija informacija iz tekstova na srpskom jeziku prikazana je na primeru korpusa tekstova o vremenskoj prognozi i vremenskim prilikama, preuzetih iz većeg broja elektronskih medija. U glavi 6 dati su neki od zaključaka, kao i smernice za dalja istraživanja iz ove oblasti.

This dissertation is on research and studying in scientific field called information extraction, which can be seen as a sub-area of artificial intelligence and which combines and uses techniques and achievements of several computer science areas. The term „information extraction“ will be used in two different contexts. In the first one, the term will refer to the scientific area and the acronym IE will be used in that case. In the second case, this term will refer to the very process of extracting information. Beside the IE state-of-the-art survey, an original approach and a method for information extraction based on finite state transducers are presented. A database with microbial phenotype and genotype characteristics, for 2412 species and 873 genera has been created, as a result of the research and the work on the dissertation. The database is intended for research, in bioinformatics and genetics. The method used for the creation of the database and the database itself are described in details and published in several journals and conference proceedings (Pajić, 2011; Pajić et al. 2011a; Pajić et al. 2011b). In the Section 1, the introduction to IE is given, together with the history of development of methods in this area. The classification of textual resources that are used for information extraction and classification of the information itself are described. At the end of the Section 1, IE is compared with other related disciplines of computer science. Section 2 contains some excerpts from formal language theory and abstract automata, on which the dissertation is based. The mutual relationship between these two areas and their connection with IE are described. The emphasis is put on the final state models and methods based on them. These methods show higher precision than other methods for extracting information, and are indispensable in situations where the accuracy of data extracted from the text is of crucial importance. Some specific terms of information extraction - the language of the relevant information, the language of extracted information and extraction rules, are defined from the perspective of formal language theory. The basic feature of the transduction relation for the given rule extraction is formulated and proved. The language of information context is defined and its regularilty is proven. Two different approaches to designing systems for information extraction, (1) an approach based on machine learning, and (2) knowledge-based approach (hereafter KE) are presented in Section 3. Systems using the first approach require the existence of annotated corpora, while the KE-based systems requires the existence of human experts in specific fields, who are familiar with the domain in which information extraction is to be performed. These experts define rules used to extract information from text. Most of the systems for the information extraction combine these two approaches, so certain components of the system are based on machine learning, and other are based on expert rules for extraction. An overview of an IE system’s architecture and of existing systems is given at the end of Section 3. Moreover, the software system UNITEX, which is used for the information extraction in this research, is described in more details. In the Section 4, an original method for information extraction, based on finite state transducers, is presented, as a KE based method. This method clearly distinguishes between two phases of the extraction. The first phase is carried out to identify and extract smaller parts of the text corresponding to one entity, i.e. one data record, while in the second phase, values of some entity attributes are extracted from those pieces of text. The first phase is highly dependent on the structure of the text and does not have to use transducers. The algorithm of the first phase is different for different texts, from which extraction is performed. On the other hand, the second phase does not depend on the structure of the text. Within the second phase, transducers are applied, defining the rules of the extraction. These transducers are constructed by experts in specific areas. Bearing in mind that similar concepts are often described in a similar way, in the large number of domains, these transducers can be re-used for other processes of extraction from the same area. In the Section 4, the above-mentioned method is described in detail and an example of its application to an encyclopedic text in the field of microbiology is given, as well as the resulting database and the data evaluation. The information extraction problem from texts in the Serbian language is presented in the Section 5, as an example of the approach needed when the natural language has a rich morphology. The data are extracted from the corpus of weather forecast reports in Serbian, collected from several Serbian electronic media. The whole process is done by UNITEX software, in which the electronic dictionaries for Serbian are applied. In the Section 6, conclusions and directives for future research in the area are given.

Mikromehanička analiza loma zavarenih spojeva čelika povišene čvrstoće

Micromechanical fracture analysis of high strength steel weldments

Industrijska primena čelika povišene čvrstoće u elementima zavarenih konstrukcija čini neophodnim poznavanje ponašanja spojeva ovih materijala pri žilavom lomu. Stoga, procena integriteta zavarenih struktura je potrebna da bi se obezbedio potrebni nivo sigurnsti i pouzdanosti, imajući u vidu uticaj ograničenog deformisanja i heterogenosti na ponašanje pri lomu kritičnih zona spoja: zone uticaja toplote (ZUT) i metala šava (MŠ), koji često imaju nižu žilavost i višu prelaznu temperaturu. Takođe, veoma je važno da procena integriteta bude urađena realno i da ne bude previše konzervativna, kako bi se sprečilo povećanje mase strukture i obezbedilo ekonomično korišćenje materijala. U ovoj disertaciji, mikromehanički pristup je korišćen za analizu uticaja mehaničke heterogenosti i ograničenog deformisanja na žilav lom zavarenih spojeva čelika povišene čvrstoće. Ovaj pristup je korišćen kao rešenje za problem prenosivosti parametara klasične mehanike loma. Takođe, motiv za njegovu primenu je i to što standardni parametri mehanike loma: faktor intenziteta napona, otvaranje prsline i konturni J- integral, ne mogu na odgovarajući način opisati odgovor materijala sa prslinom na dejstvo spoljnog opterećenja u svim uslovima, kao što su izraženo plastično tečenje (large scale yielding), različiti uticaji heterogenosti, obika i geometrije zavarenih konstrukcija u eksploataciji. Parametri mehanike loma, određeni laboratorijskim ispitivanjem epruveta, nisu direktno prenosivi na komponente i stoga se moraju uzeti u obzir dodatni faktori (kao što je uticaj ograničenog deformisanja). Takođe, cilj ove disertacije je određivanje mehaničkih osobina zona zavarenog spoja korišćenjem kombinovanog eksperimentalno- numeričkog postupka, što je naročito važno kod zona male širine u okviru ZUT koje su izložene opterećenju u transverzalnom pravcu. Rad na disertaciji je podrazumevao primenu metode konačnih elemenata i eksperimentalna ispitivanja. Eksperimentalna ispitivanja su urađena na zavarenim glatkim epruvetama (uz korišćenje ARAMIS stereometrijskog mernog sistema) za određivanje mehaničkih osobina, kao i epruvetama za savijanje u tri tačke i epruvetama za zatezanje sa početnom prslinom u metalu šava i zoni uticaja toplote za analizu ponašanja pri lomu. J-R krive i vrednosti parametra mehanike loma koje odgovaraju početku rasta prsline v određene su eksperimentalno i numerički na epruvetama sa početnom prslinom u ZUT i MŠ. Numerička analiza elasto-plastičnih modela konačnih elemenata (2D i 3D) je urađena u programskom paketu Abaqus, a mikromehanički kompletni Gursonov model (CGM) je primenjen preko korisničkog potprograma UMAT (autor Z.L. Zhang). Ograničeno deformisanje oko vrha prsline i promena troosnosti u ligamentu su numerički analizirani na epruvetama za savijanje u tri tačke i onim za zatezanje, da bi se analizirala prenosivost mikromehaničkih parametara oštećenja sa jedne epruvete na drugu. Takođe, mikromehaničkim pristupom je određen uticaj geometrije i rezultati su upoređeni sa eksperimentalnim podacima. Najveći deo numeričkih rezultata dobijenih korišćenjem CGM modela pokazuje dobro slaganje sa eksperimentalnim rezultatima. Analiza rezultata je pokazala da je mikromehanički pristup pouzdana tehnika za predviđanje žilavog loma u zavarenim spojevima čelika povišene čvrstoće, uključujući određivanje uticaja ograničenog deformisanja i heterogenosti. Pokazano je da mehanička heterogenost spoja i ograničeno deformisanje oko vrha prsline imaju veliki uticaj na otpornost prema nastanku i rastu prsline. Takođe, numerički određena otpornost prema nastanku i rastu prsline zavise od veličine konačnih elemenata. Štaviše, mikromehanički parametri oštećenja preneti su sa jedne na drugu komponentu, bez obzira na uticaj ograničenog deformisanja. Pokazano je da se određivanje mehaničkih osobina zona zavarenog spoja, naročito za delove ZUT male debljine, može koristiti kao alternativna tehnika standardnoj, koja se ne može primeniti za određivanje osobina zona spoja u slučaju dejstva opterećenja u transverzalnom pravcu (u odnosu na zavareni spoj).

The industrial application of high strength steel in structural welded components has increased the demand for understanding the ductile failure behavior of this type of welded materials. Therefore, integrity assessment of those welded structures is required in order to ensure a certain level of safety and reliability, having in mind the effects of constraints and heterogeneity on fracture behavior of crucial regions such as: heat affected zone (HAZ) and weld metal (WM) which usually have low toughness and higher transition temperature. It is also essential that the assessment is done in a realistic and not too conservative way in order not to increase the mass of the structure or impair the economic efficiency too much. In this thesis, micromechanical approach has been used to study the effect of mechanical heterogeneity and constraints on ductile fracture of high strength steel weldments. This approach has been used as a solution for the transferability problem of conventional fracture mechanics parameters. It has also been used on basis of that fracture mechanics parameters recommended by standard, such as: stress intensity factor, crack opening displacement and contour J-integral, cannot reliably describe the reaction of a pre-cracked material to the effects of external loading under all conditions such as: large scale yielding, various effects of heterogeneity, shape and geometry of real welded structures. The fracture mechanics parameters, determined from laboratory scale experiments are not also directly transferable to components and hence additional considerations (like constraint effects etc.) need to be taken care of. In addition, the aim of the thesis was to estimate precise mechanical properties using a combined experimental and numerical procedure for various welded joint regions, especially for narrow HAZ regions, when they are subjected to transversally applied load. The study was carried out using finite element method and experiments. Experimental analysis was carried out on: welded smooth tensile specimen with ARAMIS measuring system for estimation mechanical properties, welded single-edge notched bend and flat tensile specimens with pre-cracks in weld metal (WM) and heat-affected zone (HAZ) for studying the ductile fracture behavior. J-R curves and crack growth initiation iii values of fracture mechanics parameter were experimentally and numerically obtained for specimens with a pre-crack in HAZ and WM. Numerical analysis of elastic-plastic finite element models (2D and 3D) was performed in software package Abaqus, with micromechanical complete Gurson model (CGM) applied through user subroutine, UMAT (author: Z.L. Zhang). The crack tip constraint and variation of stress triaxiality in ligament were numerically analyzed on single-edge notched bend and tensile specimens to analyze the transferability of micromechanical damage parameters from one specimen to another. In addition, the geometry effects were also studied by the micromechanical approach and the results were compared with those of the experiments. Most of numerical results obtained with CGM model are in good agreement with the experimental results. Analysis of results shows that micromechanical approach is reliable technique to describe ductile fracture behavior in high strength steel weldments, including the effect of constraints and heterogeneity. It was also observed that the initiation toughness and the crack growth resistance are greatly affected by the mechanical heterogeneity of the weldment and the crack tip constraint. Additionally, the numerical values of the initiation toughness and the crack growth resistance are also affected greatly by the size of finite element. Moreover, micromechanical damage parameters could be transferred from one component to another without depending on constraint effect etc. The results also show that estimation of mechanical properties for various welded regions, especially for narrow HAZ regions, could be considered as alternative technique to the standard one, which may not be applied to estimate their mechanical properties when welded joint is transversally subjected to applied load.

Ocenjivanje indeksa repa raspodele korišćenjem nekompletnih uzoraka

On tail index estimation using samples with missing observations

Teza obrađuje ocenjivanje indeksa pravilne promenljivosti na nekompletnom uzorku zavisnih slučajnih veličina sa raspodelom teškog repa. Pod pretpostavkom ekstremalne zavisnosti dokazuje se konzistentnost ocene geometrijskog tipa, kao i konzistentnost i asimptotska normalnost Hilove ocene. Navode se primeri procesa koji zadovoljavaju zahteve dokazanih teorema i na kojima se mogu primeniti dobijeni rezultati.

For the sequence of heavy-tailed and possibly dependent random variables with the missing observations the estimation of the tail-index is considered. Under minimal but verifiable assumption of ''extremal dependence'' we proved the consistency of geometric-type estimator (Brito and Freitas, 2003). We extended results from Mladenovic and Piterbarg (2008) and proved the consistency and the asymptotic normality of the Hill estimator. Illustrative examples are provided.

Prilog dinamičkoj analizi višeslojnih anizotropnih ploča

Multilayered anisotropic plate dynamic analysis

Razmatranjem nelinearne teorije višeg reda istaknute su velike prednosti ovakve analize za višeslojne anizotropne strukture tipa ploča. Rešenja pomeranja ukazuju na prednost analize višeg reda, jer dobijena pomeranja opisuju kretanje ploče kao deformabilnog tela. Nelinearnim kinematskim relacijama određene su deformacije i na osnovu linearnih konstitutivnih veza naponi. Naponi sa uvedenim “međuslojevima” otvaraju mogućnost potpunijeg poznavanja naponskog stanja u višeslojnom laminatu. Posmatrano je slobodno oscilovanje ploče i kretanje ploče pod dejstvom transverzalne harmonijske prinude. Za uspostavljen režim oscilovanja određene su sve karakteristike neophodne za dimenzionisanje struktura ovog tipa.

Nonlinear higher order theory of plate is accepted to describe the advantage of this theory for multilayered anisotropic structure. Displacements solutions, as for three-dimensional body, are determinate via higher order theory. With nonlinear kinematic relations the strains are solved and the stresses with linear constitutive relations. Stacking sequence with “interface” has special purpose in delamination identification, because it gives more information about stress field in laminate. Free vibrations and forced transversal steady state oscillations without damping for annular plate are presented with displacements, strains and stresses, which are prescribed as function of time. Amplitude stresses are very important for structure failure improvements.

Spektralno prepoznavanje grafova i mreža

Spectral recognition of graphs and networks

Spektralna teorija grafova je matematiˇcka teorija koja grafove prouˇcava pomo´cu sopstvenih vrednosti i sopstvenih vektora matrica koje su im pridruˇzene. Posebno interesantni problemi ovog istraˇzivaˇckog domena jesu problemi spektralnog prepoznavanja grafova. Tu ubrajamo: karakterizaciju grafa sa zadatim spektrom, taˇcno ili pribliˇzno konstruisanje grafa sa zadatim spektrom, sliˇcnost grafova i perturbacije grafova. U disertaciji se u prvom redu razmatraju problemi sliˇcnosti grafova, gde se razlika pravi u zavisnosti od toga da li su ili ne poredbeni grafovi istog reda. Sliˇcnost grafova istog reda ustanovljava se izraˇcunavanjem spektralnih rastojanja, dok se, kada je reˇc o grafovima razliˇcitog reda, izraˇcunavaju i upored¯uju mere sliˇcnosti definisane za njihove ˇcvorove. Mere sliˇcnosti mogu i ne moraju da budu spektralno zasnovane, a poredbeni grafovi su u tom sluˇcaju obiˇcno grafovi sa velikim brojem ˇcvorova, pa se nazivaju mreˇzama. Disertacija sadrˇzi izvesne rezultate koji se odnose na Menhetn spektralno rastojanje grafova bazirano na matrici susedstva, Laplasovoj matrici i nenegativnoj Laplasovoj matrici. Predloˇzena je nova mera sliˇcnosti za ˇcvorove poredbenih mreˇza koja se zasniva na razlici funkcija generatrisa za brojeve zatvorenih ˇsetnji u ovim ˇcvorovima. Brojevi zatvorenih ˇsetnji izraˇcunavaju se, shodno novoj spektralnoj formuli, brojanjem razapinju´cih zatvorenih ˇsetnji u grafletima odgovaraju´cih grafova. Zapoˇceta je analiza spektralno-strukturalnih karakteristika digrafova u odnosu na spektar matrica AAT , odnosno ATA, gde je A matrica susedstva razmatranog digrafa. Generalizovan je pojam kospektralnosti, pa su u tom smislu dati neki rezultati vezani za kospektralnost digrafova i multigrafova u odnosu na matricu AAT i nenegativnu Laplasovu matricu, respektivno.

Spectral graph theory is a mathematical theory where graphs are considered by means of the eigenvalues and the corresponding eigenvectors of the matrices that are assigned to them. The spectral recognition problems are of particular interest. Between them we can distinguish: characterizations of graphs with a given spectrum, exact or approximate constructions of graphs with a given spectrum, similarity of graphs and perturbations of graphs. In this dissertation we are primarily interested for the similarity of graphs, where graphs with the same number of vertices and graphs of different orders are considered. Similarity of graphs of equal orders can be established by computation of the spectral distances between them, while for graphs with different number of vertices the measures of similarity are introduced. In that case, graphs under study are usually very large and they are denoted as networks, while the measures of similarity can be spectraly based. Some mathematical results on the Manhattan spectral distance of graphs based on the adjacency matrix, Laplacian and signless Laplacian matrix are given in this dissertation. A new measure of similarity for the vertices of the networks under study is proposed. It is based on the difference of the generating functions for the numbers of closed walks in the vertices of networks. These closed walks are calculated according to the new spectral formula which enumerates the numbers of spanning closed walks in the graphlets of the corresponding graphs. The problem of the characterization of a digraph with respect to the spectrum of AAT , apropos ATA, where A is the adjacency matrix of a digraph, is introduced. The notion of cospectrality is generalized, and so the cospectrality between some particular digraphs with respect to the matrix AAT and multigraphs with respect to the signless Laplacian matrix is considered.

Istraživanje i razvoj procedure i modela za robustni dizajn menjačkih prenosnika

Research and development of procedures and models for robust design of automotive gearboxes

Razvijene procedure u ovom radu zasnivaju se na korišćenju pouzdanosti mašinskog sistema i komponenata kao granični uslov za određivanje konstrukcijskih i drugih parametara, kao svojstvo komponenata i kao pokazatelj kvaliteta sistema...

Procedures developed in this paper are based on the use of reliability of mechanical systems and components as a boundary condition to determine the construction and other parameters, such as feature of components and indicator of system quality...

Međunarodnopravni subjektivitet Evropske unije

Doktorska disertacija pod naslovom Međunarodnopravni subjektivitet Evropske unije ima za predmet da ispita i analizira međunarodnopravni subjektivitet Evropske unije i da identifikuje konkretan institucionalni i normativni okvir za njegovo ostvarivanje, kao i pojedina prava koja Evropska unija poseduje u međunarodnom javnom pravu zahvaljujući svom subjektivitetu. U tom smislu, ovaj rad teži da razmotri praktične i teoretske aspekte ostvarivanja subjektiviteta Evropske unije u međunarodnom javnom pravu, bazirajući se na opštim saznanjima o međunarodnopravnom subjektivitetu međunarodnih organizacija, kao i posebnim saznanjima u vezi sa Evropskom unijom. Da bi se odgovorilo ovom cilju, ovaj rad posebnu pažnju posvećuje fenomenu regionalne integracije kao formi međunarodnog organizovanja država članica uključenih u proces evropskih integracija. Ovaj pristup je značajan jer je Evropska unija regionalna međunarodna organizacija sui generis koja je u značajnom delu svojih ovlašćenja nadnacionalnog (supranacionalnog) karaktera i koja ima razvijen institucionalni i politički sistem te poseban pravni sistem i širi skup pravnih i političkih tekovina – acquis Unije i značajne nadležnosti za delovanje na međunarodnopravnom nivou. S obzirom na njenu ulogu u međunarodnim odnosima, međunarodnom pravu i na konstituisanje posebnog pravnog sistema ove organizacije, naučno ispitivanje i praktično sagledavanje njenog međunarodnopravnog subjektiviteta ima srazmeran značaj u pravnoj nauci. Stoga, kao predmet koji zahteva sistematsku naučnu obradu i istraživanje, odredili smo analizu instrumenata, procedura i uslova u kojima se ostvaruje međunarodnopravni subjektivitet Evropske unije u kontekstu postojećih pravila međunarodnog javnog prava i važećih konstitutivnih ugovora ove organizacije, a u vezi sa ostvarivanjem vrednosti i ciljeva na kojima počiva celokupan proces evropske izgradnje i nadležnosti dodeljenih Uniji. Istraživanje pitanja međunarodnopravnog subjektiviteta Evropske unije nastoji da potpomogne utvrđivanje korelacije koja pokazuje način na koji Evropska unija 4 deluje u međunarodnopravnom sistemu, sa jedne strane, kao i da identifikuje način na koji međunarodnopravni sistem utiče na razvoj Evropske unije kao oblika organizovanja država. Da bismo proveli ovo istraživanje, pošli smo od osnovne hipoteze prema kojoj je Evropska unija, kao regionalna međunarodna organizacija sui generis, subjekt međunarodnog javnog prava, čiji međunarodnopravni subjektivitet predstavlja bitnu osnovu određenja pravne prirode ove međunarodne organizacije i osnov je njenog nezavisnog delovanja na međunarodnom planu u odnosu na države članice te, kao takav, međunarodnopravni subjektivitet Evropske unije čini bitan uslov obavljanja poverenih zadataka i ciljeva postojanja Evropske unije i daljnjeg razvoja procesa evropskih integracija. Radi provere ove hipoteze, doktorska disertacija je podeljena u sledeća poglavlja. Prvo, analiziran je međunarodnopravni subjektivitet uopšte i međunarodnopravni subjektivitet međunarodnih organizacija. U ovom poglavlju smo ukazali na teoretsko poimanje ove kategorije, sumarno predstavljajući različite teoretske koncepcije, i pojasnili smo razvoj međunarodnopravnog subjektiviteta u međunarodnoj praksi. Drugo, razmotrena je pravna priroda Evropske unije i specifičnosti njenog međunarodnopravnog subjektiviteta. Zaključili smo da, uprkos podeljenosti teoretskih razmišljanja, za potrebe našeg istraživanja, Evropsku uniju treba posmatrati kao regionalnu organizaciju sui generis, i shodno tome je njen međunarodnopravni subjektivitet analiziran u svetlu međunarodnopravnog subjektiviteta međunarodnih organizacija. Treće, sagledan je normativni aspekt međunarodnopravnog subjektiviteta kroz analizu nadležnosti Evropske unije, kao pravnog okvira za ostvarivanje njenog međunarodnopravnog subjektiviteta. S obzirom na područje delovanja, nadležnosti Evropske unije se mogu podeliti na one koje joj omogućavaju delovanje na unutrašnjem planu, i one koje joj dozvoljavaju delovanje na spoljnom planu. Četvrto, kroz ispitivanje uloge institucija Evropske unije u ostvarivanju njenog međunarodnopravnog subjektiviteta predstavljen je njegov institucionalni aspekt. Delovanjem institucija, Evropskoj uniji je omogućeno 5 da u međunarodnopravnom sistemu poseduje, pribavlja i ostvaruje odgovarajuća prava i obaveze koja joj pripadaju kao međunarodnopravnom subjektu. Peto, analizirana su pojedina prava koja Evropska unija ostvaruje u međunarodnopravnom poretku zahvaljujući svom međunarodnopravnom subjektivitetu. Takođe, predstavili smo kako se pojedini elementi međunarodnopravnog subjektiviteta ostvaruju u okviru prethodno analiziranog institucionalnog okvira. Šesto, ukazali smo na značaj međunarodnopravnog subjektiviteta Evropske unije koji je dvojake prirode. Sa jedne strane, on je značajan za ostvarivanje ciljeva evropskog integracionog procesa i utvrđivanje odnosa između Evropske unije i njenih država članica u kontekstu međunarodnopravnog poretka. Sa druge strane, međunarodnopravni subjektivitet Evropske unije je jedan od faktora koji utiču na afirmisanje procesa regionalnih integracija u okviru procesa globalizacije, kao i to da pojedina prava koja potiču iz međunarodnopravnog subjektiviteta Evropske unije omogućavaju zaštitu dejstva unutrašnjeg prava međunarodne organizacije. Na kraju, izvedeni su određeni zaključci bazirani na temeljima provedenog istraživanja. U radu smo pokazali da Evropska unija kao regionalna međunarodna organizacija poseduje međunarodnopravni subjektivitet koji je bitan za ostvarivanje ciljeva evropskog integracionog procesa. Međunarodnopravni subjektivitet je eksplicitnog karaktera jer je predviđen članom 47. Ugovora o Evropskoj uniji i on obuhvata sve tekovine međunarodnopravnog subjektiviteta Evropske zajednice koje su nastale u višedecenijskom razvoju ove organizacije čiji je Evropska unija sukcesor na osnovu Lisabonskog ugovora. U pogledu međunarodnopravnog subjektiviteta Evropske unije smo zaključili da je derivativni, aktivni i konkretni međunarodnopravni subjektivitet; on je funkcionalne prirode i zavisi od poverenih zadataka i ciljeva postojanja Evropske unije i određen je u okvirima eksplicitnih ili implicitnih nadležnosti koje su konstitutivnim ugovorima dodeljene institucijama Evropske unije i čine njegov normativni aspekt. Institucionalni sistem Evropske unije koji je podoban za formiranje samostalne pravno relevantne volje ove organizacije i sa tim povezanog 6 nezavisnog manifestovanja njenog međunarodnopravnog subjektiviteta. Postojanje međunarodnopravnog subjektiviteta Evropske unije, iako je značajno za daljnji razvoj procesa evropskih integracija, ne vodi direktno do uspostavljanja evropske naddržave niti može dovesti do promene podele nadležnosti između Evropske unije i država članica definisane u konstitutivnim ugovorima, bez saglasnosti država članica. S druge strane, u pogledu korelacije pravne prirode Evropske unije i njenog međunarodnopravnog subjektiviteta, snatramo da bi pogrešno bilo shvatanje da postojanje međunarodnopravnog subjektiviteta ukazuje na razvoj Evropske unije prema tvorevini državnog karaktera. Ipak, iako su malo verovatne kvalitativne promene pravne prirode Evropske unije, nije nelogično očekivati da dođe do kvantitativne promene odnosa Unije i država članica koja se može ogledati u povećanom obimu nadležnosti Unije, ali samo i jedino uz pristanak država članica koje su ugovornice konstitutivnih ugovora Unije i konsekventno tome gospodari integracionog procesa. U tom kontekstu, Unija će nastojati da unapredi svoje pravne mogućnosti da direktnije učestvuje u međunarodnim odnosima, posebno u domenu međunarodnog organizovanja, nastojeći da za sebe obezbedi što je moguće veći stepen ovlašćenja unutar međunarodnih organizacija. U provođenju ovog istraživanja i izradi rada, kao izvor saznanja, korišćena su raspoloživa teoretska i stručna razmatranja koja mogu pomoći rasvetljavanju predmeta istraživanja, kao i sagledavanje relevantnih pravnih izvora i prakse institucija Evropske unije, drugih međunarodnih organizacija i međunarodnih sudova koji se odnosi na pravni subjektivitet.

The PhD thesis written under the title The International Personality of the European Union aims at exploring and analysing the international personality of the EU and at identifying a concrete institutional and normative framework established for the exercise thereof, as well as particular rights that the EU is entitled to by virtue of its international personality. In this regard, this study tends to consider both practical and theoretical aspects related to the exercise of the EU legal personality under the international law, based on the general knowledge on the international personality of international organisations, as well as concrete observations on the EU. In order to address this task, the present study is focused on the phenomenon of regional integration which is an international form of organizing Member States involved in the European integration process. This approach is appropriate for the present study because the EU is a regional organization sui generis having considerable supranational powers and being based on the developed institutional and political system with a special legal order including a wider set of legal and political commitments forming the Union acquis. In addition, the EU has been conferred significant responsibilities enabling it to act in the sphere of international law. Taking into account its role in the area of international relations and that of international law, as well as its impact in the establishment of its particular legal system, a scientific approach to examine and to practically observe its international personality has certain importance for the legal doctrine. Therefore, this study is to offer a systematic research and analysis of instruments, procedures and conditions which are needed for manifestation and exercise of the international personality of the EU in the light of relevant international law rules and constituent treaties of the organization in question. These procedures and conditions are taken into account as related to values and objectives of the European integration process, on the one hand, and competences conferred to the European Union, on the other hand. The research related to the international personality of the EU aims at identifying the correlation between the way in which the EU is operating within the 8 international legal order, on the one hand, and the way in which the international legal order exercises its influence on the development of the European Union as a form to organize States at the international level, on the other hand. In order to carry out this research, we have started from the hypothesis that the European Union, as a regional international organization sui generis, has the legal personality in the international law, and this international personality represents an important element in defining the EU’s legal nature and enables the EU to act independently at the international level, and, as such, the international personality of the European Union constitutes an essential factor for completion of tasks and objectives assigned to the European Union and for further development of the European integration process. In order to verify this hypothesis, the present study consists of the following chapters. Firstly, the general notion of the international personality was analyzed with a particular attention to the international personality of international organizations. The theoretical aspects of the international personality were elaborated and different theoretical concepts of personality were briefly explained in this chapter. Also, a particular attention was paid to development of the international personality in the international judicial practice. Secondly, the legal nature of the European Union and specificities of its international personality were considered. A conclusion was made that, despite different theoretical conception on the legal nature of the Union, for the purposes of this study, the EU would be considered as a regional organization sui generis, and consequently, its international personality would be observed in the light of the international personality of international organizations. Thirdly, the normative aspect of the EU’s international personality was observed through the analysis of its competences and responsibilities, which offer a legal framework for the exercise of international personality. Regarding its field of activities, the EU’s competences could be divided as those enabling it to act internally and those empowering it to act externally. Fourthly, the institutional aspect of the international personality of the EU was examined through the analysis of the role of EU institutions in exercise of the international personality. The EU is in a position to act, to possess and to acquire certain 9 rights and duties within the international legal order thanks to the activities of its institutions. Fifthly, certain rights which the EU possesses as a person of the international law were analysed in this chapter. The institutional procedures and legal instruments for the exercise of these rights were elaborated, as well. Sixthly, the importance of the international personality of the EU, which is deemed to have a dual character, was pointed out. The personality is important for the attainment of the EU integration goals and positioning of the EU and its Member States within the international legal order, on the one hand. It is equally one of the factors that influence the affirmation of the regional integration as a model of organizing States in the era of globalisation, on the other hand. Furthermore, certain rights stemming from the international personality provide for the protection of the effects of the internal law of an international organisation. Finally, this research led to a number of conclusions. This study demonstrated that the European Union was a regional international organisation possessing an independent personality under the international law. This personality plays an important role in accomplishment of objectives of the EU integration process. The international personality of the EU has been explicitly defined in the Article 47 of the Treaty on European Union and the personality includes also the set of rights acquired during the existence of the previous European Communities in the period prior to the entry into force of the Lisbon Treaty. Regarding the international personality of the European Union, this study made a conclusion that it is a derived personality, derived from the personality of its Member States, of an active and concrete nature; this personality is functionally defined and relies heavily on tasks and objectives assigned to the European Union in its constituting treaties. The EU competences, either explicit or implicit, provide for a normative framework within which the institutions of the EU exercise certain rights stemming form its international personality. The institutional system of the European Union is deemed to represent an adequate scheme for establishment and manifestation of legally relevant will of the EU which is necessary for exercise of its international personality. The existence of the EU’s international personality, although quite important for the future development of the integration process, is not expected to lead to the transformation of the EU’s legal nature in terms of establishing a superstate 10 nor it would alter the defined division of competences between the EU and its Member States as defined by the founding treaties without the consent of the latter. In other words, regarding the correlation between the European Union and its international personality, in our view it would not be correct to consider that the existence and exercise of this international personality would lead to transformation of the EU into an entity having the statehood nature. Although some qualitative changes in the EU legal nature are hardly expected to occur, the shift in terms of quantitative aspect of competences divided between the EU and its Member States may seem possible which would lead to further empowerment of the EU institutions. However, this shift would only be possible if the Member States give their consent in the manner and following the procedure stipulated in the founding treaties since the Member States are masters of the treaties. In this context, the Union is expected to tend to improve its legal capacity to participate directly in international relations to the highest possible extent, especially through direct participation in other international organizations aiming at acquiring as higher as possible set of powers within organisations in question. Relevant sources for this research were found in different theoretical and professional works and analysis which enabled us to clarify certain aspects of the explored topic. Significant information related to the subject of the study were found in legal and administrative practice of the EU institutions, other international organisations and in the case law of international courts and tribunals related to the international personality.

Dijagnostika dinamičkog ponašanja struktura primenom raspodele kinetičkih i potencijalnih energija

Diagnosis of dynamic behavior of structures using the distribution of kinetic and potential energy

U većini objekata vibracije su nepoželjne. To je zato što vibracije stvaraju dinamičke sile i udarce koji mogu izazvati zamor i otkaz strukture. Odgovor strukture na njenu pobudu zavisi od načina primene i lokacije pobudne sile, kao i dinamičke karakteristike strukture kao što su prirodne frekvencije i nivo prigušenja. Strukturalni odgovor se može poboljšati promenom raspodele masa ili krutosti strukture, pomeranjem izvora pobude na drugu lokaciju , ili povećanjem prigušenja u strukturi . Strukturalna dinamička modifikacija ( SDM ) je veoma efikasna i pouzdna tehnika koja se intenzivno koristi za poboljšanje dinamičkih karakteristika strukture kao što su prirodne frekvencije, glavnih oblika i funkcija frekventnih odziva ( FRFs ) . Dinamičko ponašanje konstrukcije može se poboljšati čineći modifikacije delova kao što su kruta mesta, masa , prigušenja itd. Mnogo puta se desi da struktura ne ispunjava potrebne ograničenja dizajna i da dizajn mora da bude modifikovan nekoliko puta pre nego što ona ispuni sve uslove projektovanja. Suština poboljšanja dinamičko ponašanje objekta jeste povećanje prirodnih frekvencija i povećanje intervala između susednih prirodnih frekvencija. Ovaj zahtev se može postići promenom dizajna parametara strukture . Procedura koje se koriste u ovom radu jesu analize distribucije potencijalne i kinetičke energije i razlike između njih u elementima strukture. Studija distribucije potencijalne i kinetičke energije na glavnim oblicima oscilacija strukture daje očigledno predviđanje koje elemente i kako treba izmeniti da se postigne najbolje dinamičko ponašanje . Cilj predloženog razvijenog metoda reanalise i dijagnostike ponašanja struktura je da se utvrdi stvarno ponašanje konstrukcije u eksploataciji. Tehnika reanalise strukture se izvodi primenom metode konačnih elemenata (MKE ) . Informacije o strukturi kao materijal, geometrija i granični uslovi treba da budu spremni pre nego što generiše model. U ovom radu , procedure strukturne dinamičke modifikacija, u procesu reanalise, brojčano IV su primenjene na poznate jednostavne strukture, kao što su rešetke, grede, i ploče, kao i na složene realne strukture sa ciljem da se poboljša dinamičko ponašanje. Proračun je izveden sa programima ABAKUS i Komips. Proučavajući raspodelu kinetičke i potencijalne energija i razlika između njih kroz sve elemente strukture može se predvideti odgovarajuće strukturne dinamičke modifikacije potrebne da bi izbegli problemi izazvani rezonancom . Osim toga, da bi se proverili rezultati koji su dobijeni numerički pomoću konačnih elemenata, laboratorijski eksperimenti su sprovedeni na prototipu modela koji simuliraju pravu strukturu. Dobijeni rezultati proračuna modela su u prilično dobroj saglasnosti sa rezultatima merenja prototipa modela.

In most structures vibration is undesirable. This is because vibration creates dynamic stresses and strains which can cause fatigue and failure of the structure. The response of the structure to excitation depends upon the method of application and the location of the exciting force or motion, and the dynamic characteristics of the structure such as its natural frequencies and inherent damping level. The structural response can be improved by changing the mass or stiffness of the structure, by moving the source of excitation to another location, or by increasing the damping in the structure. Structural Dynamics Modification (SDM) is a very effective and reliable technique which is extensively used to improve structure's dynamic characteristics such as natural frequency, mode shape and frequency response functions (FRFs). The dynamic behavior of the structure can be improved by predicting the modified behavior making some modifications parts like rigid links, beams, lumped masses, dampers etc. Many times it happens that the structure does not meet the required design constraints and the design has to be modified numerous times before it meets all the design constraints. This repeated analysis for each such modification becomes very expensive and time consuming, especially if there are lots of degrees of freedoms. The main point of improving dynamic behavior of a structure is increasing its natural frequencies and maximizing the interval between adjacent natural frequencies. This request can be achieved by changing the design parameters of the structure. The procedures used in this thesis are concerned with the analysis of the distribution of potential and kinetic energy and the differences between them in elements of the structure. Study of distribution of potential and kinetic energy in main oscillation modes of structure gives obvious prediction which elements need some modifications to achieve the best dynamic characteristics. The aim of developed the proposed method of reanalysis and II diagnostic of structure behavior is to determine real behavior of the construction in exploitation. Reanalysis technique can be done for the structure using finite element methods (FEM). Information about the structure like material, geometry and boundary conditions should be prepared before making FE model. In this thesis, Structural Dynamics Modification procedures, in process of reanalysis, have been numerically applied on well known simple structures, such as trusses, beams, and plates, as well as on complex real structures to improve its dynamic response. FE models, using ABQUS and KOMPIS softwares, were created for each case. Then, by studying the distribution of kinetic and potential energy and the difference between them through whole elements of the structure one can predict the appropriate structural dynamic modification needed to avoid problems caused by resonance. Furthermore, in order to verify the results which were obtained numerically using finite element analysis, laboratory experiments were conducted to prototype model which simulate a case study of a real structure. The obtained results of the FE model were in reasonably good agreement with the measured results of the prototype model. Accordingly, successful investigation was conducted on the real structure model, which emphasis that the proposed method provides effective results.

Geometrija četvorodimenzionih nilpotentnih Lijevih grupa

Geometry of four-dimensional nilpotent Lie groups

U ovom radu izlažemo klasifikaciju levo-invarijantnih metrika proizvoljne signature na četvorodimenzionim nilpotentnim Lijevim grupama. Detaljno ispitujemo njihovu geometriju, sa posebnim naglaskom na grupe holonomija i dekompozabilnost metrika. Takođe, potpuno opisujemo grupe izometrija i nalazimo primere metrika za koje su zadovoljene stroge nejednakosti Isplit < Iaut < I: U sluqaju metrika neutralne signature na nilpotentnim Lijevim grupama sa degenerisanim centrom dobijamo Vokerove metrike. Formulišemo i dokazujemo potreban i dovoljan uslov da one dopuštaju nilpotentnu grupu izometrija. Na kraju, dajemo odgovor na pitanje egzistencije projektivno ekvivalentnih metrika. Pokazujemo da su na četvorodimenzionim nilpotentnim Lijevim grupama sve levo-invarijantne metrike ili geometrijski rigidne ili postoje njima projektivno ekvivalentne metrike koje su istovremeno i afino ekvivalentne. Iako su sve afino ekvivalentne metrike levo-invarijantne, Njihova signatura može biti različita.

In the present work we classify left invariant metrics of arbitrary signature on four-dimensional nilpotent Lie groups. Their geometry is extensively studied with special emphasis on holonomy groups and decomposability of metrics. Also, isometry groups are completely described and we give examples of metrics where strict inequalities Isplit < Iaut < I hold. It is interesting that Walker metrics appear as the underlying structure of neutral signature metrics on the nilpotent Lie groups with degenerate center. We fnd necessary and suffient condition for them to locally admit nilpotent group of isometries. Finally, we solve the problem of projectively equivalent metric on four-dimensional nilpotent Lie groups by showing that left invariant metric is either geometrically rigid or have projectively equivalent metrics that are also affinely equivalent. All affinely equivalent metrics are left invariant, while their signature may change.

Razvoj metoda za povećanje radnog veka i pouzdanosti mašinskih sistema u uslovima zamora

Development of methods for increase of working life and reliability of mechanical systems under fatigue loading

U ovom radu prikazan je razvoj metoda za procenu nastanka i širenja zamornih prslina u cilju povećanja radnog veka i pouzdanosti mašinskih elemenata i mašinskih sistema u uslovima zamora. Detaljno je izložena i analizirana dostupna naučna literatura iz teorijskih postavki fenomena zamora. Posebno je analizirana primena postojećih metoda i teorija na rešavanje problema zamornih oštećenja i otkaza kod mašinskih elemenata. Razvijanje novih metoda za procenu nastanka zamornih prslina izloženo je sa različitih aspekata, s obzirom da se radi o rezultatima multidisciplinarnih istraživanja ovog fenomena, uz primenu i interkorporiranje dostupnih savremenih analitičkih , numeričkih i eksperimentalnih metoda. Posebna pažnja posvećena je verifikaciji razvijenih metoda, kroz nekoliko pristupa. Za verifikaciju su korišćeni sopstveni eksperimentalni rezultati, kao i rezultati drugih autora. U cilju boljeg razumevanja razvijenih metoda, a pre svega metodologije širokih mogućnosti primene, detaljno je opisana primena na konkretnom i veoma složenom problemu otkaza mašinskog elementa u uslovima zamornog opterećenja.

This dissertation presents development of methods for assesment of crack initiation and propagation as part of the goal: increase of working life and reliability of machine elements and systems under fatigue loading. Детаљно je изложена и анализирана доступна научна литература из теоријских поставки феномена замора. Посебно је анализирана примена постојећих метода и теорија на решавање проблема заморних оштећења и отказа код машинских елемената On the basis of available scientific literature, the fenomenon of fatigue is analysed. The emphasize is on this research was application of existing methods and theories on failure prevention. Development of new methods for assessment of crack initiation is presented in detail, due to fact that multidisclipinary research is basis for application, analytical, numerical an experimental methods. The important part of presentation of developed methods, with significant potentials, is application on real and complex problem of failures of mechanical elements due to combination of fatigue and corrosion.

Preduzetništvo kao faktor razvoja agroprivrede Republike Srpske

Entrepreneurship as a factor of development TI agrobusiness of the Republic of Srpska

Predmet istraţivanja ove doktorske disertacije je odreĊivanje znaĉaja preduzetništva kao faktora razvoja agro(privrede) Republike Srpske. Preduzetništvo je vaţan element ekonomskog razvoja i konkurentnosti privrede, gde je Republika Srpska tipiĉan primer zemlje tranziciji. Stoga je aspekt preduzetništva posmatran na makro nivou, da bi se uspešno moglo sagledati i na jednom segmentu privrede – agroprivreda, a zatim i mikro nivou. U konkretnom primeru Republike Srpske, kao tranzicione zemlje, postoje još uvek odreĊena ograniĉenja da bi se uspešno sprovela reforma i agrarnog i prehrambenog sektora. Glavna ograniĉenja u daljem razvoju ova dva sektora su: nezadovoljavajuća agrarna struktura, nerešeno agrarno zakonodavstvo, nekonkurentnost proizvoda, slaba iskorišćenost proizvodnih kapaciteta i sl. U prelaznoj fazi kod transformisanja agrarnog i prehrambenog sektora u savremeniji trţišni model odluĉujuću ulogu ima agro(preduzetništvo). Rezultat ovog istraţivanja ukazuje da Republika Srpska ima znaĉajne potencijale za razvoj oba sektora, ali su nedovoljno razvijeni. Naime, sadašnji agrarni i prehrambeni sektor zadovoljava potrebe domaćeg stanovništva. Sa druge strane postoje znaĉajni potencijali za takav razvoj agroprivrede koji bi obezbedio plasman ovih proizvoda na inostranom trţištu. Rezultati istraţivanja ukazuju da je potrebno dinamiĉnije razviti sektor malih i srednjih preduzeća u agroprivredi Republike Srpske. U prilog tome govore rezultati sprovedene ankete za potrebu istraţivanja u disertaciji. Naime, rezultati ukazuju da znaĉajniji udeo u preduzetništvu agroprivrede Republike Srpske imaju mikro preduzeća u odnosu na mala i srednja preduzeća. U tom kontekstu je opravdanije formiranje malih i srednjih preduzeća u primarnoj proizvodnji, dok u prehrambenoj proizvodnji je opravdanije formiranje srednjih preduzeća. TakoĊe, za uspešno poslovanje MSP sektora u agroprivredi Republike Srpske potrebno je povoljno makroekonomsko okruţenje (bruto društveni proizvod, inflacija, trţište rada, spoljnotrgovinsko poslovanje i priliv stranih direktnih investicija). Odluĉujuću ulogu daljem razvoju agropreduzetništva Republike Srpske imaju institucionalni i pravni okviri razvoja preduzetništva, kroz razvoj ĉitavog niza zakona i strategija, kao i institucija koji su u funkciji razvoja agroprivrede. TakoĊe, razvoj preduzetniĉke infrastrukture je znaĉajan faktor unapreĊenja preduzetništva u agroprivredi Republike Srpske, ali i za ekonomski razvoj lokalne zajednice. Konaĉno, razvoj preduzetništva u Republici Srpskoj višestruko doprinosi razvoju agroprivrede preko: većeg stepena iskorišćavanja poljo(privrednih) potencijala, većeg stepena finalizacije poljoprivrednih proizvoda, revitalizaciji sela, smanjenja depopulacije seoskog stanovništva, rasta konkurentnosti, veće zaposlenosti radno sposobnog stanovništva, pa konaĉno i celokupnog ekonomskog razvoja Republike Srpske.

Subject of this dissertation is to determine the importance of entrepreneurship as a factor in the development of agro(business) of Republic of Srpska. Entrepreneurship is an important element of economic development and competitiveness of the economy, and Republic of Srpska is typical example of a country in transition. Therefore, the aspect of entrepreneurship monitored on a macro level, in order to successfully be seen in a segment of the economy - agrobusiness, and then the micro level. In the specific case of the Republic of Serbian, as well as a country in transition, there are still some limitations in order to successfully implement the reform of agricultural and food sector. The main limitations to the further development of these two sectors are unsatisfactory agrarian structure, unresolved agrarian legislation, deficiency of product competitiveness, low utilization of production capacity, etc. In the transitional phase of transformation the agricultural and food sector in the modern market model the crucial role is played by agro(entrepreneurship). The results of this study indicate that in Republic of Srpska has significant potential in both sectors, but they are under-developed. The current agricultural and food sector satisfies the nutritional needs of the local population. On the other hand, there is significant potential for development of agrobusiness which would be provided distribution of these products on the international market. The results indicate the need for more dynamic develop of sector of small and medium-sized enterprises in agrobusiness in Republic of Srpska This assertion confirmed by the results of the survey for the need of research in this dissertation. The results indicate that more signigicant share in agroentrepreneurship in Republic of Srspka have micro- enterprises than small and medium-sized enterprises. In this context, it is justifiable to formation of small and medium enterprises in primary production, while food production has more justified the formation medium-sized companies. Also, for successful management of SME sector in agrobusiness in Republic of Srpska is necessary to provide favorable macro-economic environment (GDP, inflation, labor market, foreign trade and foreign direct investment). Crucial role in further development of agroentrepreneurship in Republic of Srpska have institutional and legal framework for the development of entrepreneurship, through the development of a series of laws and policies, and institutions which are in in the function of development of agrobusiness. Also, the development of entrepreneurial infrastructure is an important factor in promoting entrepreneurship in agrobusiness of Republic of Srpska, but also for the economic development of local communities. Finally, the development of entrepreneurship in the Republic of Srpska have multiple contributions for the development agrobusiness through: a higher level of exploitation agricultural(economic) potential, the greater the level of finished agricultural products, the revitalization of the village, to reduce rural depopulation, increase competitiveness, increasing employment rate of working-age population, and finally and the overall economic development of the Republic of Srpska.

Grebnerove baze za mnogostrukosti zastava i primene

Gröbner bases for flag manifolds and applications.

o Borelovom opisu, celobrojna i mod 2 kohomologija mnogostru- kosti zastava data je kao polinomijalna algebra posečena po određe- nom idealu. U ovom radu, Grebnerove baze za ove ideale dobijene su u slučaju kompleksnih i realnih Grasmanovih mnogostrukosti, kao i u slučaju realnih mnogostrukosti zastava F(1,...,1; 2,...,2,k,n)...

By Borel's description, integral and mod 2 cohomology of ag manifolds is a polynomial algebra modulo a well-known ideal. In this doctoral dissertation, Gr obner bases for these ideals are obtained in the case of complex and real Grassmann manifolds, and real ag manifolds F(1; : : : ; 1; 2; : : : ; 2; k; n). In the case of Grassmann manifolds, Gr obner bases are applied in the study of Z- cohomology of complex Grassmann manifolds. It is well-known that, besides Borel's description, this cohomology can be characterized in terms of Schubert classes. By establishing a connection between this description and Gr obner bases that we obtained, a new recurrence formula that can be used for calculating (all) Kostka numbers is derived. Using the same method for the small quantum cohomology of Grassmann manifolds (instead of the classical), these formulas are improved. In the case of real ag manifoldsF(1,...,1; 2,...,2,k,n), Gr obner bases are used to obtain new results on the immersions and embeddings of these manifolds, and for the calculation of the cup-length of some manifolds of this type.

Individualni obrasci reagovanja na simptome bolesti kod obolelih od hroničnih zapaljenskih bolesti creva

Individual patterns of reaction to illness symptoms in patients with inflammatory bowel disease

Hronične zapaljenske bolesti creva (engl. Inflammatory Bowel Disease - IBD), koje uključuju ulcerozni kolitis (engl. Ulcerative colitis - UC) i Kronovu bolest (engl. Crohn’s disease - CD), su hronična recidivirajuća oboljenja nepoznate etiologije i nedovoljno razjašnjene patogeneze. Tok bolesti je nepredvidiv i na njega utiču brojni psihosocijalni faktori, uključujući i ličnost obolelog. Cilj: Ciljevi ovog istraživanja su bili ispitivanje karakteristika ličnosti obolelih od hroničnih zapaljenskih bolesti creva, ispitivanje individualnih obrazaca reagovanja na simptome bolesti u fazi remisije i procena moderirajućeg efekta pojedinih crta ličnosti na individualne obrasce reagovanja. Metodologija: Studija je obavljena u Univerzitetskom Kliničko bolničkom Centru “Zvezdara” u Beogradu. Prvi deo studije je dizajniran kao longitudinalna opservaciona studija preseka a drugi deo je dizajniran po principima studije vremenskih serija tipa dnevnika sa dnevnim intervalima. U studiju je bilo uključeno 150 ispitanika, oba pola, sa dijagnozama UC (50.7%) i CD (49.3%) u remisiji, starosti od 18-65 godina, bez drugih hroničnih somatskih oboljenja i bez psihijatrijskih oboljenja. Sociodemografski podaci i podaci o bolesti su prikupljeni iz bolničke medicinske dokumentacije. Procenjivanje crta ličnosti obavljeno je primenom upitnika za samoprocenu The Revised NEO personality inventory (NEO PI - R) i DELTA 10. Tokom 14 uzastopnih dana ispitanici su popunjavali Inventare dnevnih događaja, misli i emocija koji su konstruisani za potrebe ove studije i tako su za svakog ispitanika dobijene multiple dnevne samoprocene u vidu vremenskih serija. U analizi dobijenih podataka korišćeni su odgovarajući parametarski testovi i seri- ja hijerarhijskih modelovanja slučajnim koeficijentima (engl. Multilevel random coeffi- cient modeling – MRCM). Podaci su analizirani primenom statističkog programa IBM SPSS 21 i softvera HLM 6.06. Rezultati: U poređenju sa normativnim uzorkom, oboleli od IBD su imali značaj- no više skorove Neuroticizma i značajno niže skorove Dezintegracije. Na nivou faceta, oboleli od IBD su imali značajno više skorove na facetima Anksioznost, Asertivnost, Blaga narav i Dužnost i značajno niže skorove na facetima Toplina, Potraga za uzbuđen- jem, Pozitivne Emocije, Akcija, kao i na svim facetima Dezintegracije osim Depresije, Somatoformne Disregulacije i Socijalne Anhedonije. Razlike između UC i CD nađene su samo na nivou faceta. Faceti koji su najviše doprinosili prediktivnoj snazi diskrimi- nativne funkcije su bili Opšta Egzekutivna Disfunkcija, za kojom slede Toplina, Samo- disciplina, Depresija i Manija. Nađene su značajne interindividualne i intraindividualne razlike u učestalosti posmatranih dnevnih varijabli. Značajni prediktori veće učestalo- sti dnevnih stomačnih simptoma su bili: odsustvo dnevnih pozitivnih događaja, dnevne negativne emocije, dnevne nefokusirane misli i visoki skorovi na crtama Neuroticizam, Savesnost i Dezintegracija. Prediktori manje učestalosti dnevnih stomačnih simptoma su bili dnevni pozitivni događaji. Dnevni negativni događaji, dnevne pozitivne emocije, dnevne fokusirane misli i skorovi na crtama Ekstraverzije, Otvorenosti i Saradljivosti nisu bili značajni prediktori učestalosti stomačnih simptoma. Moderirajući efekat odnosa između posmatranih dnevnih varijabli i stomačnih simptoma imali su samo visoki skor- ovi na crtama Savesnost i Otvorenost. Zaključak: Između obolelih od IBD i opšte populacije, kao i između UC i CD postoje razlike u karakteristikama ličnosti. Pojedine crte ličnosti imaju značajne direktne i indirektne efekte na dnevno funkcionisanje obolelih. Zbog toga bi rutinsko ispitivanje crta ličnosti kod obolelih omogućilo prepoznavanje individualnih faktora rizika za loše ishode bolesti. To bi omogućilo osmišljavanje adekvatnih psihosocijalnih intervencija us- merenih na pojedinca sa ciljem poboljšanja psihosocijalnog funkcionisanja i toka bolesti.

Inflammatory Bowel Disease (IBD), including Ulcerative Colitis (UC) and Crohn’s disease (CD), are chronic relapsing diseases of unknown aetiology and insufficiently explained pathogenesis. The course of disease is unpredictable and it is influenced by many psychosocial factors, including personality of the patient. Purpose: The purpose of this study was to observe the personality traits of the pa- tients suffering from IBD, to examine individual patterns of reaction to the symptoms of diseases in remission stage and to evaluate the moderating effects of personality traits on individual patterns of reaction. Methodology: The study has been conducted at the University Clinical Hospital Centre “Zvezdara” in Belgrade. The first part of the study was designed as a longitudinal observational cross-section study, while the second part was designed according to the principles of time-based diary study as daily-diary study. The study included 150 par- ticipants of both genders, suffering from UC (50.7%) and CD (49.3%) in the remission, from 18 to 65 years of age. None of the participants suffered from other chronic somatic diseases or psychiatric disorders. Socio-demographic data and the disease related data were collected from the clinical medical documentation. The personality traits related data were collected using the self-assessment survey - The Revised NEO personality inventory (NEO PI - R) and DELTA 10. During the period of 14 consecutive days, the subjects were filling in the Inventories of daily events, thoughts and emotions, specially designed for the needs of this study. This provided us with the multiple daily self-as- sessments formed as time series for each subject. The analysis of the obtained data used appropriate tests of parameters and the serial of multilevel random coefficient modelling (MRCM). The data were analysed by means of statistical software IBM SPSS 21 and software HLM 6.06. Results: In comparison to the normative sample, the subjects suffering from IBD had significantly higher scores of Neuroticism and significantly lower scores of Disin- tegration. On the facet level, the subjects suffering from IBD had significantly higher scores in the following facets: Anxiety, Assertiveness, Tender-mindedness and Dutiful- ness, and significantly lower scores in facets Warmth, Excitement Seeking, Positive Emo- tions, Activity, and in all other Disintegration facets, with the exception of Depression, Somatoform Dysregulation and Social Anhedonia. The difference between UC and CD was found only at the level of facets. The facets that mainly contributed to the predictive strength of discrimination function were General Executive Dysfunction, followed by Warmth, Self-discipline, Depression and Mania. There are also significant inter-individ- ual and intra-individual differences in the frequency of the observed daily variables. The significant predictors of higher frequency of abdominal symptoms were: lack of daily positive events, daily negative emotions, daily unfocused thoughts and high scores on the following traits: Neuroticism, Conscientiousness and Disintegration. The predictors of less frequent daily abdominal symptoms were daily positive events. Daily negative events, daily positive emotions, daily focused thoughts and scores in traits of Extraver- sion, Openness and Agreeableness were not significant predictors of abdominal symptom frequency. The moderating effect of the relationship between the observed daily variables and the abdominal symptoms was only reported in high scores of traits of Conscientious- ness and Openness. Conclusion: Differences have been reported in personality characteristics among the subjects suffering from IBD and the general population, and also among UC and CD. Certain personality traits have mutual direct and indirect effects on daily functioning of the subjects. This is why a routine examination of personality traits of patients would enable recognizing the individual risk factors responsible for negative outcomes of the disease. It would result in conceiving the adequate psychosocial interventions focused on individuals, with the purpose of improving their psychosocial functioning and the course of the disease accordingly.

Istraživanje interakcije dejonizovane vode sa hidrofilnim i hidrofobnim materijalima, biomolekulima i hidrogenizovanim ugljeničnim nanomaterijalima

Investigation of interaction of deionized water with hydrophilic and hydrophobic materials, biomolecules and hydrogenized carbon nanomaterials

Predmet i cilj istraživanja ove doktorske disertcije je utvrđivanje postojanja interakcije između dejonizovane vode (18,2 MΩ) i različitih tipova materijala: hidrofobni i hidrofilni materijali, biomolekuli, hidrogenizovani ugljenični nanomaterijali. Da bi se ovaj cilj realzovao izvršene su četiri različite vrste eksperimenta u kojima su korišćene različite koncentracije rastvorenih ispitivanih supstanci, pri čemu su sve promene posmatrane putem mikroskopije sa velikim uvećanjima i rezolucijom, a podaci obrađivani kompjuterskim softverom. Istražujući interakcije dejonizovane vode objašnjen je fenomen eksluzivne zone sa aspekta vodoničnih veza, potvrđen je fenomen postojanja eksluzivne zone koji se može snimiti mikroskopom na osnovu migracije mikrosfera prečnika 1 µm u vodenim rastvorima, utvrđeno je da veličina eksluzivne zone zavisi od koncentracije mikrosfera u rastvoru, da se u kontaktu hidrogenizovanih ugljeničnih nanomaterijala i hidrofobnog materijala stvara zona vode sa drugačijim karakteristikama i da se pored sočiva kao hidrofilne membrane u dodiru sa hidrofilnim mikrosferama formiraju eksluzivne zone. Prvi put je posmatran fenomen struktuiranja i stvaranje uređenih slojeva pri interagovanju hidrogenizovanog ugljeničnog nanomaterijala fulerola, hidrofilnih čestica mikrosfera i hidrofobnog materijala nafiona. Fenomen stvaranja eksluzivne zone ima višestruke primene i značaj u ispitivanju biološkog sveta, pa se tako koristi u procesima predfiltriranja vode, biomedicini pri stvaranju suznog filma i nanofluidici pri stvaranju transportnih i procesnih elemenata na nano-nivou. Izvršena istraživanja će imati uticaj na razvoj „water based nanomedicine“ oblasti biomedicinskog inženjerstva uz primenu biokompatibilnih nanomaterijala.

Subject and the goal of this doctoral disertation is determination of appereance and existance of the interaction between deionized water (18,2 MΩ) and different types of materials: hydrphobic and hydrofilc materials, biomolecules, hydrogenized carbon nanomaterials. To realize this goal there were conducted four exeriments in which are used different concentrations of diluted examined substances, all changes are observed with high resolution microscopy, and datas are elaborated with computer softvare. By examining the interaction of deionized water it is explained phenomena of exclusion zone from the aspect of the hydrogen bonds, it is cofirmed the phenomena of exclusion zone which can be observed and recorded with mcroscope based on migration of microspheres whose diameter is 1 µm in water solutions, it is determined that size of the exclusion zone depends of the concentration of microspheres in the solution, that in contact of hydrogenized carbon nanomaterials and hydrphobic material forms area of water with different properties, and that next to the contact lens as hydrphlici membrane in contact with hydrophilic microspheres exclusion zone is forming. For the first time is examined and observed the appereance of structuring layers in interaction of hydrogenized carbon nanomaterial fullerol, hydrophobhilic particles microspheres, and hydrophobic material nafion. Phenomena of exclusion zone have multiple implementations and significance in examination of biological systems, so it can be used in processes of prefiltration of water, biomedicine of making tear film, and nanofluidics in making the transport and process elements on nano-level. Experminets that have been done will have the impat on development of „water based nanomedicine“ part of biomedical egineering with application of nanomaterials.

Izvodljivost velikih resekcija jetre kod bolesnika sa hepatocelularnim karcinomom i cirozom jetre primenom radiofrekventne energije

Feasibility of major liver resections in patients with hepatocellular carcinoma and liver cirrhosis using radiofrequent energy

kurativno lečenje primeni na što je moguće većem broju bolesnika. Budući da je transplantacija jetre ograničena definisanim kriterijumima i brojem donora, resekcija jetre predstavlja glavnu terapijsku opciju potencijalno kurativnog lečenja ali je ograničena funkcionalnim kapacitetom preostalog parenhima jetre. Radiofrekventna resekcija jetre uvažava sve principa savremene hirurgije jetre. Kako je hepatična cirkulacija kod ove tehnike prezervirana tokom transekcije parenhima, funkcija preostalog često cirotičnog parenhima jetre nije narušena ishemijsko-reperfuzionom lezijom. Time se resekcija jetre kao potencijalno kurativni modalitet lečenja može primeniti i kod bolesnika koji imaju ograničenu funkciju jetre. Cilj rada: Utvrditi izvodljivost velikih resekcija jetre kod bolesnika sa hepatocelularnim karcinomom na terenu cirotično imenjene jetre određivanjem postoperativnog morbiditeta i mortaliteta; odrediti preoperativne, intraoperativne i patohistološke prognostičke parametre povezane sa najvećim preživljavanjem i ispitati postoperativni oporavak funkcije cirotične jetre primenom statičkih i dinamičkih testova i merenjem nivoa TNFά i IL6 u prvih 48h posle resekcije jetre Metod: Kohortna studija delom retrospektivna delom prospektivna realizovana na Klinici za digestivnu hirurgiju Kliničkog centra Srbije u periodu od novembra 2001. godine do novembra 2012. godine. U navedenom periodu kod 41 bolesnika (20 sa cirozom, 21 bez ciroze jetre) sa histološki potvrđenim hepatocelularnim karcinomom urađena je radiofrekventna resekcije jetre i to velika resekcija (resekcija 3 i više segmenata). Isključeni su bolesnici sa mešovitim hepatocelularnim/holangiocelularnim karcinomom, bolesnici kod kojih je infiltrisana donja šuplja vena, glavno stablo portne vene i bolesnici sa ekstrahepatičnim širenjem bolesti. U studiji su korišćeni demografski podaci, anamnestički podaci, perioperativni parametri, patohistološke karakteristike tumora, laboratorijske analize, vrednosti lidokainskog testa, vrednosti nivoa TNFά i IL6 u serumu i tkivu jetre, praćen je postoperativni morbiditet i mortalitet. Za statističku obradu korišćen je statistički programski paket SPSS for Windows (18.0). Izbor testova za analizu numeričkih obeležja posmatranja vršen je zavisno od prirode njihove raspodele koja je ispitivana korišćenjem Koglomorov–Smirnov-og testa.U slučaju normalne raspodele podataka za testiranje razlike između dve grupe ispitanika korišćen je t-test dok je kod neparametarskih podataka korišćen Mann Whitney U test. Kaplan Majer-ovim krivama preživljavanja praćen je ishod kod ispitanika sa HCCom. Cox-ovom regresionom analizom ispitivani su prediktori preživljavanja. Granična vrednost za prihvatanje radne hipoteze postavljena je na p < 0.05 . Rezultati: Kod 41 bolesnika (20 sa cirozom, 21 bez ciroze) urađena je velika resekcija jetre primenom radiofrekventne energije. U odnosu na pol, starost, segmentnu lokalizaciju tumora, veličinu i broj tumora, status resekcione margine, težinu preparata, dodatnu operaciju, ASA skor, trajanje operacije i resekcije jetre, dužinu postoperativnog lečenja nije nađena statistički značajna razlika između ispitivanih grupa bolesnika. Ukupno prosečno preživljavanje iznosi 60,00 meseci, u grupi bolesnika sa cirozom je 57,40 a u grupi bolesnika bez ciroze jetre je 62,00 bez statistički značajne razlike. U odnosu na učestalost postoperativnih komplikacija nema statistički značajne razlike između dve posmatrane grupe ispitanika. Perioperativna transfuzija data je kod 4 bolesnika bez statistički značajne razlike između grupa. Od ispitivanih prognostičkih parametara jedino je dužina trajanja operacije (više od 300 min) imala nepovoljan uticaj na ishod (veća učestalost komplikacija u grupi više od 300 minuta) dok su parametri uzrast preko 70 godina, resekcija IV segmenta, Child B status ciroze, R1 resekcija, perioperativna transfuzija bez uticaja na učestalost komplikacija ili ukupno preživljavanje. Statički testovi funkcije jetre pokazuju da se 7. postoperativnog dana vrednosti bilirubina, transaminaza, PT vraćaju na preoperativni nivo i da nema statistički značajne razlike između ispitivanih grupa. Lidokainski test u testiranim slučajevima pokazuje dobru povezanost sa ishodom lečenja a klirens lidokaina ne pokazuje statistički značajnu razliku između ispitvanih grupa. Nivo IL6 I TNFα određivan je u serumu i na proteinskom i genskom nivou. U svima ispitivanim kategorijama nije nađena statistički značajna razlika između bolesnika sa cirozom i bez ciroze jetre. Nivo IL-6 u postoperativnom periodu u svim merenjima pokazuje značajno više vrednosti u odnosu na preoperativne što ukazuje na odgovarajuću reprezentativnost ovog citokina kao markera akutne faze regeneracije jetre. Ovaj rezultat nije nađen za TNFα. Zaključak: Radiofrekventna resekcija jetre je bezbedna hirurška procedura koja se može izvesti i kod bolesnika sa cirozom jetre u uslovima kada je potrebno obaviti veliku resekciju jetre. Ova hirurška tehnika obezbeđuje brz oporavak funkcije jetre uz očuvan regenerativni potencijal u akutnoj fazi regeneracije čak i u uslovima ciroze. Interleukin-6 je važan medijator regeneracije jetre u ovoj fazi.

Background: Potentially curative treatment options should be addressed to as many as possible patients with hepatocellular carcinoma (HCC). Since liver transplantation is limited by defined criteria and by scanty donor pool liver resection is the main potentially curative treatment option. The extent of liver resection is limited by functional capacity of liver remnant. Radiofrequent (RF) liver resection respects all the principles of modern liver surgery. Since liver blood flow is maintained during transection of liver parenhyme when RF resection is used, functional capacity of liver remnant even being cirrhotic is not impaired by ishemic-reperfusion injury. Having this advantage RF resection as potentially curative treatment can be offered to patients with compromised liver function. Objective: To assess feasibility of major liver resection in patients with hepatocellular carcinoma in cirrhotic liver by analyzing postoperative morbidity and mortality; to determine preoperative, intraoperative and histology parameters associated to best clinical outcome; to assess postoperative recovery of liver function in cirrhotic livers using static and dynamic liver function tests and measuring TNFά and IL6 level during the first 48h after liver resection Method: The cohort study (partly retrospective, partly prospective) was done at University clinic for digestive surgery from November 2001 until November 2012. In that period 41 patients (20 pts with liver cirrhosis and 21 pt wtihout cirrhosis) with histology proven HCC had major liver resection using RF-assisted technique. The excluding criteria were patients with mixed hepatocellular and cholangiocellular carcinoma, patients with infiltration of major blood vessels and patients with extrahepatic spread. The following parameters were used – demographic data, anamnestic data, perioperative parameters, histopathology data, laboratory analysis, lidocain test, TNFά and IL6 level in serum and in liver tissue, postoperative morbidity and mortality. The statistical software package SPSS for Windows (18.0) was used for statistical analysis. Selection of tests for analysis of numerical observations was carried out depending on the nature of their distribution, which is examined using Koglomorov-Smirnov test. In case of normal distribution of data difference between the two groups was assessed by t-test; for nonparametric data Mann Whitney U test was used. Kaplan-Meier survival curves were used to assess the outcome of patients with HCC. Cox's regression analysis examined predictors of survival. Threshold for accepting working hypothesis was set at p <0.05. Results: Radiofrequent, major liver resection was performed in 41 patients (in 20 pts with liver cirrhosis and in 21 pt without cirrhosis). Statistically significant difference between the two patient groups was not found in regard to their gender, age, segmental tumor localisation, size and number of tumor nodes, resection margin, weight of specimen, additional surgical procedures, ASA score, duration of operation and liver resection and lenght of postoperative hospital stay. Overall survival for all patients was 60,00 months, for patients with liver cirrhosis 57,40 and for patients without cirrhosis 62,00 months. Statistically significant difference was not found. In regard to postoperative complications no statistical difference was found between the two groups of patients. Perioperative transfusion was given to 4 patients again without statistical difference between the two groups. Among prognostic parameters duration of operation (longer than 300 min group) was the only parameter associated to impaired outcome while other parameters like age more than 70 years, resection of segment IV, Child B status of liver function, R1 resection, perioperative transfusion have demonstrated no influence on overall survival and postoperative complications rate. Statical liver function tests have shown return to preoperative values for bilirubin, liver transaminase, and prothrombin time on day 7. No difference was found between the two groups. The lidocain test have shown good correlation with the clinical course; results of lidocain clearance have found no difference between the two groups of patients. Interleukin-6 and tumor necrosis factor alpha was measured in serum and was assessed on protein and genetic level. In all assessed categories no difference was found between the patients with liver cirrhosis and without cirrhosis. Postoperative level of IL-6 was significantly higher in all tested categories compared to preoperative level suggesting that IL-6 has important role in acute phase of liver regeneration. This result was not found for TNFα. Conclusion: Radiofrequent liver resection is a safe procedure feasible in patiens with liver cirrhosis even in cases when major liver resection is required. This procedure provides quick recovery of liver function maintaining regenerative capacity of cirrhotic liver during acute phase of liver regeneration. Interleukin-6 is an important mediator of acute-phase liver regeneration.

Proračun podzemnog toka metodom konačnih zapremina

Computation of subsurface flow using finite volume method

U disertaciji su razmatrane numerićke metode za rešavanje problema podzemnog strujanja, transporta mase i energije u anizotropnoj i deo po deo neprekidnoj sredini. Ovakvi problemi sreću se u hidrologiji, naftnoj industriji, ekologiji i drugim oblastima. Podzemno strujanje u zasićenoj sredini opisano je linearnom parcijalnom diferencijalnom jednačinom, dok je u nezasićenoj sredini opisano Ričardsovom nelinearnom parcijalnom diferencijalnom jednačinom. Transport mase i energije opisan je advektivno-difuznim jednačinama...

The thesis considers numerical methods for the computation of subsurface flow and transport of mass and energy in an anisotropic piecewise continuous medium. This kind of problems arises in hidrology, petroleum engineering, ecology and other fields. Subsurface flow in a saturated medium is described by a linear partial differential equation, while in an unsaturated medium it is described by the Richards nonlinear partial differential equation. Transport of mass and energy is described by advectiondiffusion equations. The thesis considers several finite volume methods for the discretization of diffusive and advective terms. An interpolation method for discretization of diffusion through discontinuous media is presented. This method is applicable to several nonlinear finite volume schemes. The presence of a well in the reservoir determines the subsurface flow to a large extent. Standard numerical methods produce a completely wrong flux and an inaccurate hydraulic head distribution in the well viscinity. Two methods for the well flux correction are introduced in this thesis. One of these methods gives second-order accuracy for the hydraulic head and first-order accuracy for the flux. Explicit and implicit time discretizations are presented. Preservation of the maximum and minimum principles in all considered schemes is analyzed. All considered schemes are tested using numerical examples that confirm teoretical results.

Fetalni biohemijski markeri u predviđanju nepovoljnog maternalnog i fetalnog ishoda trudnoće

Fetal biochemical markers in prediction of adverse maternal and fetal pregnancy outcome.

Biohemijski markeri prvog i drugog trimestra, osim u skriningu hromozomopatija, pokazuju vaţnu i nesumnjivu povezanost sa maternalnim i fetalnim ishodom trudnoće kod pacijentkinja gde se radi o fetusu sa euplodnim brojem hromozoma. Smatra se da u osnovi najozbiljnijih komplikacija trudnoće, kao što su hipertenzija, abrupcija placente, intrauterusna restrikcija rasta ploda, mala telesna masa za gestacijsku dob i intrauterusna smrt ploda, osim sliĉnih faktora rizika, postoji i ishemijska bolest placente. Morfološke promene u graĊi placente direktno utiĉu na promenu u nivoima sekrecije biohemijskih markera:βHCG-a(beta humanog horionskog gonadotropina), PAPP-a (plasma protein A u vezi sa trudnoćom) u periodima od 11.do 14. nedelje, kao i HCG-a, AFP (alfa feto proteina), uE3 (nekonjugovanog estriola) i inhibina A u periodu od 16.do 19. nedelje gestacije. Raznolikost biohemijskih markera je velika, ali se prednost daje onima koji se rutinski rade u kliniĉkoj praksi, ranije detektuju nego što je pojava prvih kliniĉkih manifestacija bolesti, kao i to da su u dobroj korelaciji sa ozbiljnošću komplikacija, uz što veću senzitivnost i specifiĉnost i manje troškove analize. Cilj:Cilj rada je bio utvrditi korelaciju vrednosti biohemijskih markera skrininga prvog i drugog trimestra sa razliĉitim komplikacijama trudnoće, analizirati period trudnoće kada se komplikacije javljaju, ispitati maternalni i fetalni morbiditet u vezi sa markerima, kao i neonatalni ishod trudnoća gde su pacijentkinje imale razliĉite kategorije vrednosti navedenih markera. Materijal i metod: Studija je bila kliniĉka i prospektivna, sprovedena u periodu od 1.1.2013. do 01.07.2014.godine na Klinici za ginekologiju i akušerstvo (KGA) Kliniĉkog centra Srbije. U studiju su bile ukljuĉene 104 trudnice, sa monofetalnom trudnoćom nastalom prirodnim putem. U prvoj studijskoj grupi su bile 54 ispitanice kod kojih su jedan ili više fetalnih biohemijskih markera prvog ili drugog trimestra imali vrednosti ispod 0,5 i preko 2 MoM-a, a u drugoj studijskoj grupi je bilo 50 ispitanica kod kojih su vrednosti ovih markera bile u opsegu od 0,5 do 2 MoM-a...

Biochemical markers of the first and second trimester of pregnancy, show an important and indisputable correlation with the maternal and fetal outcome of pregnancy in patients with euploid chromosomic fetuses. It is believed that, in the basis of the most severe complications of pregnancy, such as hypertension, placental abruption, intrauterine restriction of fetal growth, low birth weight for gestational age and intrauterine fetal death, besides similar risk factors, there is also ischemic placental disease. Morphological changes of placental structure directly affect the changes in the secretion levels of the biochemical markers: βHCG (beta human chorionic gonadotrophin), PAPP-A(pregnancy associated plasma protein A) in the period of 11.th-14.th gestational weeks, and HCG, AFP (alpha-feto protein), uE3 (unconjugated estriol) and inhibin A in the period from 16.th to 19.th weeks of gestation. The diversity of biochemical markers is high, but the advantage is given to those who are routinely being done in clinical practice, those who are previously detected than the onset of the first clinical manifestations of the disease, and which correlate with the seriousness of complications, with greater sensitivity and specificity and less analysis costs. Objective: The aim of this study was to determine the correlation of the biochemical markers values of the first and second trimester screening with different complications of pregnancy, to analyze the period of pregnancy when complications occurred, to examine maternal and fetal morbidity related to markers, as well as the neonatal outcome of pregnancy, where the patients had different categories of markers values. Material and method: The study was clinical and prospective, conducted in the period from 01.01.2013.th until 01.07.2014.th at the Clinic for Gynecology and Obstetrics of the Clinical Center of Serbia...

Primena mehanike loma na procenu integriteta zavarenih konstrukcija od legura aluminijuma

Fracture mechanics aplication on integrity assessment of welded construction made from aluminium alloys

U radu je prikazana praktična primena mehanike loma na nastanak i rast zamornih prslina u konstrukcijama zavarenim primenom frikcionog zavarivanja mešanjem (FSW). Poseban naglasak je dat na tankozidne strukture od legura aluminijuma. Razmatran je rast prsline u neojačanamim i ojačanim konstrukcijama. Pri tome je korišćena numerička metoda bazirana na konačnim elementima - tzv. Proširena Metoda Konačnih Elemenata (PMKE). Analiziran je rast prsline u FSW spoju, kroz zone spoja. Pri tome je korišćen software Abaqus, kao i Morfeo za prikaz dobijenih podataka. Primenjeno je zatezno zamorno opterećenje, pri koeficijentu asimetrije cikusa opterećenja R  0 . Svi modeli su sačinjeni od legure aluminijuma 2024-T351. Karakteristike materijala u zonama spoja, kao i geometrijske mere samog FSW spoja su usvojene iz već postojećih eksperimenata. Razmatrano je više modela: Primer br.:1. najjednostavniji primer ploče, linearno-elastična i elasto- plastična analiza. Primer br.:2. dvokomponentna ploča, linearno-elastična i elasto-plastična analiza. Primer br.:3. FSW spoj-2D izložen dejstvu većeg, odnosno manjeg zateznog opterećenja. Primer br.:4. ploča sastavljena od dva FSW spoja izložena dejstvu većeg, odnosno manjeg zateznog opterećenja. Primer br.:5. ploča ojačana sa dva FSW T-spoja izložena dejstvu većeg, odnosno manjeg zateznog opterećenja. Kao rezultati proračuna dobijeni su sledeći podaci:  naponsko-deformaciono stanje u konstrukciji, pri čemu je naročita pažnja posvećena zoni u okolini vrha prsline,  koordinate tačaka fronta prsline za svaki korak rasta  x,y z ,  raspodela faktora intenziteta napona- I K , II K , KIII i Kekv duž fronta prsline za svaki korak rasta,  vek konstrukcije u formi promene broja primenjenog ciklusa opterećenja-N za svaki korak rasta prsline. Na bazi dobijenih podataka analizirana je stabilnost širenja prsline kroz strukturu. Takođe, posmatrano je ponašanje linearno-elastičnog i elasto- plastičnog materijala. Razmatran je rast prsline iz jedne zone u drugu. Data je uporedna analiza rasta prsline kod neojačanih i ojačanih ploča.

In this work aplication of fracture mechanics principels on initiation and growth of fatigue crack in friction stir welded (FSW) strucutres is presented. Special attention is dedicated to thin-walled structures made from aluminum alloys. Crack growth in unstiffened and stiffened construction is considered by using the numerical method based on finite elements - eXtended Finite Element Method (XFEM). Crack growth in FSW joint is analized through all its zones. Thereby is used software Abaqus, as well as Morfeo for results display. Tension fatigue load is applied, with stress intesity ratio R  0 . All analyzed models are made from aluminum alloy 2024-T351. Material properties in joints zones, and geometry mesures of FSW joint are adopded from available experiments. Following examples are analized: Example No.:1. simple plate, analisis in the area of linear-elastics and elasto-plastics. Example No.:2. plate with two regions, analisis in the area of linear- elastics and elasto-plastics. Example No.:3. FSW joint-2D, exposed to effect of higher, apropos lower tension load. Example No.:4. plate made of two FSW joints, exposed to effect of higher, apropos lower tension load. Example No.:5. plate stiffened with two FSW T-joints, exposed to effect of higher, apropos lower tension load. Following results are obtain in numerical computations:  stress-displacement state in construction, where special attention is dedicated to crack tip zone,  points coordinates of crack front  x,y z for every progression step,  distributions of stress intensity factors- I K , II K , KIII i Kekv along crack front for each progressions step,  construction life in form of chage of applied load number-N for each step of propagation. On the basis of obtained data, analysis of crack growth stability is made. Also, linear-elastics and elasto-plastics behavior of material is observed. Crack growth from one to another zone is considered. Finally, comparative analisys of crack growth in unstiffened and stiffened plates is made.

Multivariate model for vehicles' and machines' interior space anthropometric design

Multivarijantni model za antropometrijsko projektovanje unutrašnjeg prostora vozila i mašina

Iako je poznato da istraživanje interakcije čoveka i mašine u sistemu, sa aspekta njegovog unapređenja, odnosno prilagođavanja u cilju poboljšanja efikasnosti funkcionisanja, smanjenja zamora i očuvanja zdravlja čoveka i obezbeđivanja optimalnih uslova radne sredine, predstavlja izazov brojnim inženjerima konstruktorima mašina i vozila, kao i mnogim drugim stručnjacima koji se bave ovom problematikom, pregledom postojeće literature dolazi se do zaključka da postoji veoma mali broj istraživanja koja se bave problemom koji je predmet ove disertacije, kako u svetu, tako i kod nas. Tako, usklađenost antropometrijskih karakteristika vozača/rukovaoca vozila i mašina sa dimenzijama kabine, kao i sa dimenzijama i položajem opreme u kabini, direktno utiče na na samog korisnika sa aspekata komfora, zdravlja i radne sposobnosti, a posledično na radni učinak, produktivnost i finansijske gubitke kompanije kao i na bezbednost radnog okruženja, šire posmatrano. Imajući u vidu i druge brojne razvojne probleme regiona Srbije i Libije, očekuje se da će uspostavljanje i provera originalnog modela za antropometrijsko projektovanje unutrašnjeg prostora vozila i mašina na uzorcima srpskih i libijskih vozača i rukovaoca transportnim mašinama predstavljati i koristan alat koji će donosiocima odluka u predmetnim industrijama omogućiti mnogo efikasnije funkcionalno upravljanje na globalnom nivou. U skladu sa tim definisane su inicijalne hipoteze, koje su u disertaciji obrađene i potvrđene, a zatim je uspostavljena baza antropomera predmetnih populacija na osnovama načela statičke antropometrije, koja uključuje 1514 vozača i 133 rukovaoca dizalicom, a sa ciljem potvrde prisutnih demografskih razlika. Najpre je primenom korelacione i regresione analize, kao i testiranjem hipoteza dokazano da postoje značajne razlike u antropometrijskim merama razmatranih viii srpskih i libijskih populacija, a zavisno od pola, nacionalnosti i zanimanja (vozača i rukovaoca), što ukazuje na potrebu projektovanja za specifičnu populaciju korisnika, odnosno nalaže uključivanje svih specifičnih populacija korisnika za razliku od dosadašnje prakse projektovanja za opštu populaciju. Zatim je predložen i proveren originalni model za antropometrijsko projektovanje unutrašnjeg prostora vozila i mašina. Pokazano je da pri rešavanju problema projektovanja koji uključuju više dimenzija treba tragati za novim modelom koji ne treba da koristi univarijantni percentilni metod, već multivarijantno modeliranje. Putem predloženog integralnog modela za antropometrijsku adaptaciju zasnovanom na metodama multivarijantne statistike moguće je višedimenzionalni problem svesti na trodimenzionaln, prostorni model. Tako je ispunjen cilj da se ograničeni, što manji, a čoveku što bolje prilagođen prostor ograničen visinom, dužinom i širinom unutrašnjeg prostora odredi novom originalnom metodologijom, tako što kao antropometrijsko ograničenje uzimamo kombinacije ekstremnih veličina parova i nizova antropomera kojima treba prilagoditi kabinu, uz primenu teorije mehanizama i biomehanike. Proverom modela pokazano je da je model dovoljno tačan i precizan, sa obuhvatom 95% populacije od interesa, te su na taj način potvrđene postavljene hipoteze ovog istraživanja. Na osnovama multivarijantnog modela za antropometrijsku adaptaciju određene su dimenzije minimalnog potrebnog prostora za komforan i bezbedan smeštaj rukovaoca i vozača i one iznose 1327×1123×1926 mm za srpske i 1203×1090×1838 mm za libijske rukovaoce dizalicom, odnosno 1500×561×1230mm za srpske i 1400×591×1155mm za libijske vozače putničkog automobila. Dati rezultati su u skladu sa prethodno potvrđenim demografskim razlikama između razmatranih populacija. Generalizacijom modela definisanog u ovoj disertaciji uspostavlja se platforma za širu primenu predloženog i potvrđenog modela istraživanja u drugim ix kontekstima, kao i mogućnost daljeg razvoja i unapređenja modela, što je i predlog daljih istraživanja u predmetnoj oblasti.

Although it is known that the study of human-machine interaction in a system, in terms of its improvement and adjustments is a way to improve the efficiency of functioning, reduce fatigue, preserve human health and ensure optimum working environment conditions, it is still a challenge for many engineers - machine and vehicle constructors and experts who deal with this problem. Thus, the compatibility of the anthropometric characteristics of the driver/operator of the vehicle and machinery with cab dimensions, as well as the dimensions and position of the equipment in the cabin, directly affects the user from the aspect of comfort, health and working ability, and consequently influence the performance, productivity and financial losses as well as safety of the work environment, in a very broad scope. By reviewing the existing literature, it can be concluded that there is very little research dealing with the problem that is the subject of this dissertation. Bearing in mind other numerous development problems of the regions of Serbia and Libya, it is expected that the establishment and verification of the original model for the anthropometric design of the interior space of vehicles and machines on samples of Serbian and Libyan drivers and operators for transport machines will be a useful tool for decision- makers in subjected industries that will enable better functional management on a global scale. In accordance with this, the initial hypotheses were then defined, processed and confirmed in the dissertation using collected anthropometric measurements by static anthropometry, on the specific populations, involving samples of 1,514 drivers and 133 crane operators to confirm the present demographic differences. By applying correlation and regression analysis, as well as by testing the hypothesis, the first was confirmed. There are significant differences in the anthropometric measurements of the Serbian and Libyan populations, according to gender, nationality and occupation (drivers and operators), which indicates the need for the design for a specific population of users or requires the inclusion of all specific user populations as opposed to the previous design practices for the general population. vi The the original model for the anthropometric design of the interior space of vehicles and machines was next proposed and verified. It has been shown that when dealing with design problems involving more dimensions, a new model based on multivariate statistical modeling should be used instead of the commonly used univariate percentile method. Through the proposed integral multivariate model for anthropometric adaptation, it is possible to reduce the multi-dimensional problem to a three-dimensional, spatial model. Thus, the goal is to determine a limited, as small as possible and the most adapted three- dimensional space for a person, with the new original methodology that takes, as an anthropometric constraint, combinations of extreme pairs of dimensions and uses the theory of mechanisms and biomechanics for user accommodation. By checking the model, it was shown that the model is accurate and precise, since it covers 95% of the population of interest and, in that manner, all posted hypotheses have been confirmed. On the basis of the multivariate model for anthropometric adaptation, the dimensions of the minimum space required for the comfortable and safe accommodation are set to 1327 × 1123 × 1926 mm for Serbian and 1203 × 1090 × 1838 mm for Libyan crane operators and 1500 × 561 × 1230mm for Serbian and 1400 × 591 × 1155mm for Libyan passenger car drivers. Those results are in line with previously shown demographic differences between these populations. A generalization of the model defined in this dissertation establishes a platform for wider application of the proposed and confirmed model in other contexts, as well as the possibility of its further development and improvement, which is a proposal for further research in the subject area.

Kontrola kvaliteta praćenjem centralne tendencije negausovih slučajnih veličina

Quality control by monitoring central tendency of nongaussian random variables

Razmatran je dizajn kontrolnih grafika za praćenje centralne tenden- cije negausovih slučajnih veličina sa simetričnim ili pozitivno asimetričnim raspode- lama. U slučaju negausovih simetričnih raspodela, u ovoj disertaciji je predložen modifikovani kontrolni grafik uzoračke sredine. Za odabranu Studentovu, Laplasovu, logističku i uniformnu raspodelu izračunata je teorijska raspodela standardizovane uzoračke sredine i aproksimirana je Pirsonovom raspodelom II tipa ili Pirsonovom raspodelom VII tipa. Izračunate su širina kontrolnih granica i moć kontrolnog grafika uzoračke sredine, za datu verovatnoću greške prve vrste. Dobijeni rezul- tati ukazuju da odgovarajuća Pirsonova raspodela može poslužiti kao veoma dobra aproksimacija raspodele standardizovane uzoračke sredine. Za primenu kontrolnog grafika uzoračke sredine u praksi, poređene su mere koeficijenta spljoštenosti i upotre- ba predloženog grafika ilustrovana je na podacima. U slučaju pozitivno asimetričnih raspodela, u ovoj disertaciji je predložen jed- nostrani kontrolni grafik medijane za praćenje centralne tendencije karakteristike kvaliteta. Za odabranu eksponencijalnu, gama i Vejbulovu raspodelu izvedena je teorijska raspodela uzoračke medijane i aproksimirana je Pirsonovom raspodelom I tipa ili Pirsonovom raspodelom VI tipa. Dobijene vrednosti za gornju kontrolnu granicu i moć kontrolnog grafika medijane za teorijsku raspodelu medijane i odgo- varajuću Pirsonovu raspodelu su vrlo bliske. Za primenu predloženog kontrolnog grafika medijane, upoređene su mere koeficijenata asimetrije i spljoštenosti i zatim je konstruisan kontrolni grafik medijane za date podatke. Pored statističkog dizajna kontrolnih grafika za praćenje centralne tendencije ne- gausovih slučajnih veličina, razmatran je i njihov optimalni ekonomsko-statistički dizajn. U ovoj disertaciji je predložena upotreba genetskog algoritma za mini- mizaciju funkcije očekivanog gubitka pod ograničenjima. Odabrane su iste simetrične raspodele kao u slučaju statističkog dizajna kontrolnog grafika uzoračke sredine, odnosno iste pozitivno asimetrične raspodele kao u slučaju statističkog dizajna kontrolnog grafika medijane. Za sve odabrane raspodele karakteristike kvaliteta, odgovarajuća Pirsonova raspodela daje rezultate vrlo bliske rezultatima za teorijsku raspodelu standardizovane uzoračke sredine (uzoračke medijane).

Design of control chart for monitoring central tendency of nongaus- sian random variables with symmetric or positively skewed distributions is consid- ered. In the case of nongaussian symmetric distributions, modified X bar control chart is proposed in this dissertation. For chosen Student, Laplace, logistic and uniform distributions, theoretical distribution of the standardized sample mean is calculated and approximated with Pearson type II or Pearson type VII distributions. Width of control limits and power of X bar control chart are established, for a given probability of type I error. The results imply that the corresponding Pearson distri- bution represents very good approximation of the distribution of the standardized sample mean. For implementation of X bar control chart in practice, measures of sample kurtosis are compared and the usage of proposed chart is illustrated on given data. In the case of positively skewed distributions, one sided median control chart for monitoring central tendency of quality characteristics is proposed in this dissertation. For chosen exponential, gamma and Weibull distributions, theoretical distribution of sample median is calculated and approximated with Pearson type I or Pearson type VI distributions. Calculated values of upper control limits and power of me- dian control chart for theoretical distribution of sample median and corresponding Pearson distribution are very close. For implementation of median control chart in practice, measures of sample skewness and sample kurtosis are compared and then proposed median chart is constructed for given data. Besides the statistical design of control charts for monitoring central tendency of nongaussian random variables, their optimal economic statistical design is also considered. Use of genetic algorithms for constrained minimization of expected loss function is proposed in this dissertation. Same symmetric distributions as in the case of statistical design of the X bar control chart and positively skewed distributions as in the case of statistical design of median control chart are chosen. For all chosen distributions of quality characteristic, a corresponding Pearson distribution gives results very close to results based on the theoretical distribution of the standardized sample mean (sample median).

Korelacija fizioloških, biohemijskih i morfoloških parametara i ishoda nativnih arteriovenskih fistula za hemodijalizu

Correlation of physiological, biochemical and morphological parameters and outcomes of native arteriovenous fistulas for hemodialysis

Preduslov za uspešnu hemodijalizu (HD) je postojanje adekvatnog vaskularog pristupa sistemskoj cirkulaciji koji omogućava dovoljan dotok krvi u vantelesnom sistemu. Nativna podlaktična radiocefalična arteriovenska fistula (RCAVF) je preferirani i željeni vaskularni pristup za hemodijalizu od strane vaskularnih hirurga i nefrologa. Međutim, visoki procenat neuspeha maturacije je barijera koju treba prevazići. Mnoge studije su ispitivale faktore povezane sa neuspehom maturacije AVF. Starost, ženski pol, dijabetes, rasa, prisustvo cerebrovaskularne ili periferne vaskularne bolesti su samo neki od prediktivnih faktora za neuspeh AVF. Međutim, optimalni dijameter krvnih sudova za kreiranje AVF ostaje i dalje ključno pitanje. Pored dijametra, neki funkcioni-fiziološki parametri izgleda da imaju dodatnu ulogu u maturaciji AVF. Ultrazvuk kao neinvazivna metoda nam može dati značajnu informaciju o morfološkim i funkcionim parametrima krvih sudova kao i o spremnosti za korišćenje AVF. Vreme od momenta kreiranja AVF do prve punkcije se zove period maturacije. Postoje razlike u mišljenju vezano za vreme prve kanulacije. Optimalno vreme je kada je process maturacije završen, a to znači da je pristup spreman za korišćenje, da može da obezbedi adekvatnu dozu dijalize (Kt/V) sa minimalnim rizikom od oštećenja. Zbog svega navedenog, cilj ovog istraživanja je bio da se ispita korelacija različitih faktora (morfoloskih, funkcionih i biohemijskih) i ishoda nativnih AVF za hemodijalizu; da se definišu faktori od najvećeg uticaja kao i da se na osnovu dobijenih rezultata predloži klinički algoritam predikcije ishoda nativnih podlaktičnih AVF. Ispitivanje je obuhvatilo 122 uzastopna bolesnika u terminalnoj fazi hronične bubrežne slabosti koji su započeli HD u periodu od Oktobra 2008 do Aprila 2010. godine. Svi bolesnici su bili hospitalizovani u Klinici za nefrologiju, Kliničkog centra Srbije radi kreiranja primarne AVF i započinjanja hroničnog programa hemodijalize. Svi bolesnici su bili informisani o procedure planiranih ispitivanja za koja su dali svoju saglasnost. Studija je odobrena od strane Etičkog komiteta Kliničkog centra Srbije. Svim bolesnicima je preoperativno obavljen fizikalni i kolor Doppler ultrazvučni (CDU) pregled krvnih sudova nedominantne ruke sa merenjem dijametra cefalične vene (CVd), radijalne arterije (ARd), brahijalne arterija (ABd), indeksa rezistencije (RI), venskog distenzibiliteta (VD), endotelom posredovane dilatacije brahijalne arterije (FMD). Sve AVF su bile radicefalične po tipu lateroterminalne anastomoze. Maturacija AVF je bila praćena merenjem protoka krvi (Qa) kroz AVF i CVd 0, 14 I 28 dana. U zavisnosti od vremena kada su dostignuti kriterijumi maturacije AVF (Qa ≥500 mL/min, CVd ≥5 mm), bolesnici su podeljene u tri grupe: (U) uspešna maturacija AVF nakon 4 nedelje, (Pr-U) prolongirano uspešna maturacija (unutar 8 nedelja) i (N) neuspešna maturacija. Krv za određivanje biohemijskih parametara i proinflamatornih i prokoagulantnih biomarkera uzimana je našte pre kreiranja AVF. Za obradu podataka korišćeni su softverski paketi STATISTIKA i MedCalc. Kumulativni uspeh maturacije AVF je bio 89% (109/122 bolesnika), što uključuje uspešno maturisane AVF unutar četiri nedelje (65 bolesnika[53%]) i produženo maturisane unutar osam nedelja (44 bolesnika [36%]). Neuspešno maturisanih AVF je bilo 11% (13 bolesnika). Značajna korelacija je dobijena između ishoda AVF i pola, ARd, AB, CVd, VD, triglicerida (trg) i alkalne fosfataze (ALP). ROC analiza je izdvojila granične dijametre varijabla od značaja za uspeh AVF (CVd >1.8 mm, ARd >1.6 mm, AB>3.43 mm, VD >0.4 mm, trg>1.39 mmol/l, ALP>67u IU/l, C-reaktivni protein ≥12 g/l). Arterijski dijametri su bili značajno manji kod žena (1.83 vs. 2.01 mm, P=0.01) dok se venski nisu razlikovali. Među polovima nije bilo razlike u godinama starosti (P=0.37) i VD (P=0.57). Ženski pol je povezan sa prolongiranim ishodom maturacije (OR 0.35, 95% CI=0.17-0.72; P=0.005), ima značajno manje dijametre arterijskih krvnih sudova i bolju funkciju endotela u poređenju sa muškarcima(2981.5 vs. 2689.5, P=0.02). Muškarci imaju više uspešno maturisanih AVF nakon četiri nedelje (65% vs. 39%), dok žene imaju više prolongirano uspešnih ishoda (46% vs.27%). Nakon produženog perioda od osam nedelja nema razlike u kumulativnom uspehu među polovima (92% vs. 85%). Primenom višestruke regresione analize dobijeni su prediktivni faktori ishoda nativnih AVF, formiran je matematički model a potom i tabela sa skorom koja može imati kliničku primenu u smislu predviđanja ishoda nativnih radiocefaličnih AVF. Primenjene analize nisu potvrdile postojanje značajnog uticaja prokoagulantnih i proinflamatornih biomarkera na rani ishod AVF. Prisustvo promena na krvnim sudovima u smislu proliferacije vasa vasorum, vaskulitisa vasa vasorum i upale vena su povezane sa većim rizikom od neuspešnog ishoda AVF. Ukratko, distenzibilitet vene je značajan faktor za predviđanje ishoda maturacije AVF. Granični dijametar cefalične vene od >1.8 mm i dobijena granica od >1.6 mm za radijalnu arteriju razdvajaju neuspešan ishod AVF od produženog i uspešnog ishoda AVF. Većina RCAVF u muškaraca dostiže kriterijume maturacije nakon četiri nedelje ali neke mada neke zahtevaju produženi period od osam nedelja. Žene imaju manje dijametre arterija nego muškarci ali bolju funkciju endotela; ishod RCAVF je bio lošiji nakon četiri nedelja, ali nakon osam nedelja one se izjednačavaju sa muškarcima u pogledu dostigle maturacije. Zato bi trebalo žene razmatrati za kreiranje RCAVF ranije od muškaraca. Od svih ispitivanih parametara od uticaja na ishod maturacije AVF, morfološki i funkcioni su se izdvojili kao najznačajniji. Dalja istraživanja treba usmeriti ka boljem razumevanju odnosa između navedenih parametara i praćenju njihovog uticaja na dugoročno preživljavanje i ishod nativnih radiocefaličnih AVF. U tom smislu prospektivne, randomizirane, multicentrične studije na većem broju bolesnika treba da daju odgovore na mnoga pitanja.

A prerequisite for successful hemodialysis (HD) is the existence of adequate access to the systemic circulation, which allows sufficient blood flow in the extracorporeal system. A native, radiocephalic arteriovenous fistula (RCAVF) is the preferred option for hemodialysis and the desired vascular access for vascular surgeons and nephrologists. Nevertheless, a significant failure-to-mature rate is a barrier to overcome. Several studies have assessed factors associated with AVF failure. Older age, female sex, diabetes, race, preexisting cerebrovascular or peripheral vascular diseases are some of the predictive factors. However, optimal vessel size for successful AVF creation is still a key question. In addition, data from the literature also indicate that some functional blood vessel parameters (arterial response to reactive hyperemia, vein dilatation under tourniquet and endothelial function) appear to play additional roles in maturation. Ultrasound can give us valuable information about vessel diameters, their functional characteristics and AVF readiness for use. The time from AVF creation to the moment of the first cannulation is called the period of maturation. There are different opinions regarding the timing of the first AVF cannulation. The optimal time is when the period of maturation is completed, which means that access is ready for use, and can provide adequate dialysis dose (Kt/V) with minimum risk of damage. For the above reason, the aim of this study was to investigate the influence of different (morphologic, functional and biochemical) parameters on AVF, to define the most important factors and suggest (according to obtained results) a clinical algorithm for prediction of the radiocephalic AVF maturation. The study included 122 consecutive end-stage kidney disease patients, who commenced hemodialysis between October 2008 and April 2010. All patients were hospitalized at the Clinic of Nephrology, Clinical Center of Serbia to create a primary native AVF, in order to be prepared for the chronic hemodialysis program. All patients were informed about the planned procedure and gave informed consent. The study was approved by the Clinical Center of Serbia Ethics Committee. All patients underwent pre-operative physical and Color Doppler Ultrasound (CDU) assessment of blood vessels of the non-dominant arm. Measurements of internal diameters of cephalic vein (CVd), brachial (ABd) and radial artery (ARd), venous distensibility (VD), resistance index (RI) and endothelial function by flow mediated dilatation (FMD) were determined by ultrasound examination before AVF placement. All AVF created were radiocephalic between the side of the radial artery and the end of the vein. AVF maturation was observed by measuring blood flow (Qa) and CVd 0, 14 and 28 days after creation. Depending on the time when AVFs attained maturity (Qa ≥500 mL/min, CVd ≥5 mm), patients were divided into three groups: (S) successful maturation (after four weeks), (PS) prolonged maturation (within eight weeks) and (FTM) failure to mature. Blood samples for analysis of biochemical, proinflammation and procoagulant biomarkers were drawn from fasting patients before the AVF creation. MedCalc and STATISTICA software packages were used for statistical analysis of data. The overall cumulative success of AVF maturation was 89% (109 of 122 patients). This included successful AVF maturation in four weeks (65 patients [53%]) and prolonged maturation during eight weeks (44 patients [36%]). FTM rate was 11%. Statistically significant correlations were found between AVF outcome and sex, ARd, AB, CVd, VD, triglycerides and alkaline phosphatase. ROC analysis defined the limits of variables relevant for AVF success (CVd >1.8 mm, ARd >1.6 mm, AB>3.43mm, VD >0.4 mm, trg>1.39mmol/l, ALP>67u IU/l, C-reactive protein ≥12g/l). The arterial diameters were significantly lower in women than in men, although venous diameters did not differ. There were no statistically significant differences between the sexes in age (P=0.37), vein distensibility (P=0.57). Female sex was associated with prolonged maturation (OR 0.35, 95% CI=0.17-0.72; P=0.005) having a significantly smaller ARd (1.83 vs. 2.01 mm, P=0.01) but better FMD (2981.5 vs. 2689.5, P=0.02) compared to men. Men had more successful matured AVFs (65% vs. 39%) and women had a prolonged AVF outcome more often than men (46% vs. 27%). After an extended period of eight weeks there were no differences in cumulative AVF outcome between the sexes (92% vs.83%). Therefore, on multiple variable logistic regression modeling, radial artery diameter, cephalic vein diameter and vein distensibility together were the predictive factors of early success AVF vs. relative failure. The performed analysis didn’t show significant impact of pro-coagulant and pro-inflammatory biomarkers on early AVF outcome. The presence of changes in blood vessels in terms of proliferation of vasa vasorum, vasa vasorum vasculitis and inflammation of the vein are associated with a higher risk of an unsuccessful outcome of AVF. In short, vein distensibility is a factor of significant importance for the prediction of a successful AVF maturation. The marginal value of CVd >1.8 mm and the limit of >1.6 mm obtained for ARd were the thresholds for failed AVF outcomes in relation to prolonged and successful outcomes. Most RCAVFs in men attain maturation criteria after four weeks but some require an extended period of eight weeks. Women have smaller arterial diameters than men but better endothelial function; the RCAVF outcome was worse after four weeks but after eight weeks they equaled men in achieving maturation. Therefore, they should be considered for RCAVF placement earlier than men. Out of all examined parameters influencing the outcome of AVF, the morphological and functional parameters were singled out as the most significant ones. Further research should be directed toward better understanding of the relationship between these parameters and to the monitoring of the outcome and long-term survival of the AVF. Therefore, the prospective, randomized, multi-center study on a larger number of patients should provide answers to many questions.

Kvantifikacija rizika na projektima osvajanja industrijskog proizvoda

Risk quantification in industrial product development projects

Predmet doktorske disertacije je postavka originalnog modela za kvantifikaciju rizika na projektima osvajanja industrijskih proizvoda, koja je usledila nakon sistematizacije postojećih znanja iz oblasti upravljanja projektnim rizicima uz adaptaciju istih na koncept razvoja industrijskog proizvoda, kroz praćenje projektnih rezultata - tehničkih karakteristika, vremena (plana) i troškova, a sa novouspostavljenom metodologijom kvantifikacije rizika. Razvoj modela kao svoj ultimativni cilj ima potrebu da unapredi procese osvajanja proizvoda u domaćim industrijskim preduzećima. U cilju verifikacije postavljenog modela sprovedeno je eksperimentalno istraživanje na uzorku 52 projekta osvajanja industrijskog proizvoda istog tipa. Eksperimentalno su određene veličine neophodne za izvedeni matematički model, statistički obrađene primenom metode glavnih komponenata, a zatim je izvedena Monte Karlo simulacija dobijenih rezultata i analiza osetljivosti u cilju provere modela. U modelu je procena rizika predstavljena iz perspektive projektnog tima, tako da su odabrane tehnike za procenu faktora rizika prilagođene kompetencijama projektnog tima i predstavljaju nadogradnju redovnim projektnim aktivnostima. Na ovaj način osigurana je efikasnost upravljanja projektnim rizicima upotrebom dostupnih resursa, bez ugrožavanja ostalih projektnih aktivnosti. Primena modela omogućava sistemski pristup kvantitativnoj analizi projektnih rizika u preduzećima, koja kao svoj poslovni poduhvat realizuju osvajanje industrijskog proizvoda. Rezultati istraživanja, kao i njihova generalizacija, potvrđena dokazom osnovnih istraživačkih hipoteza, koje se odnose na ukupan iznos rizika prema svakom od projektnih rezultata u vidu plana, budžeta i tehničkih karakteristika projekta, pružaju jasan uvid u rizike prisutne prilikom procesa osvajanja II industrijskog proizvoda u malom i srednjem preduzeću kao i njihovom uticaju na definisane projektne ciljeve. Testiranjem osnovnih istraživačkih hipoteza potvrđeni su sledeći rezultati:  Vrednosti iznosa rizika pripadaju približno normalno raspodeljenoj populaciji: o Približno 68% vrednosti iznosa rizika nalazi se u intervalu od jedne standardne devijacije sa obe strane matematičkog očekivanja. o Približno 95% vrednosti iznosa rizika nalazi se u intervalu od dve standardne devijacije sa obe strane matematičkog očekivanja. o Približno 99, 7% vrednosti iznosa rizika nalazi se u intervalu od tri standardne devijacije sa obe strane matematičkog očekivanja.  Promenljive veličine koje određuju verovatnoću pojave greške ne pružaju isti doprinos varijansi raspodele iznosa rizika kao i promenljive veličine koje određuju težinu greške. Ovi rezultati mogu biti od velike koristi prilikom donošenja odluka, kako tokom samog projekta u cilju proaktivnog reagovanja na rizike, tako i odluke o pokretanju ili odustajanju od projekta osvajanja određenog proizvoda.

The doctoral dissertation aims to propose an original model for risk quantification in industrial product development projects. The model is built through the systematization of the existing body of knowledge in the field of project risk management and its application to the concept of industrial product development along with monitoring project results – technical performances, schedule and costs – using the newly-established methodology of risk quantification. The primary motivation behind the development of this model is to contribute to the process of product development in domestic industrial enterprises. In order to verify the proposed model, the author has conducted experimental research on the sample of fifty-two projects of development of industrial product of the same kind. The sizes required for the derived mathematical model were experimentally determined and statistically processed by means of principal component analysis. Afterwards, Monte Carlo Simulation of the obtained results, as well as the sensitivity analysis, was conducted in order to verify the model. In the model, the risk assessment is presented from the perspective of the project team and is the upgrade of the standard project activities. Thus, the efficiency of managing project risks is secured by using the available resources without jeopardizing the remaining project activities. The implementation of this model enables a systematic approach to the quantitative analysis of project risks in enterprises whose business endeavor is industrial product development. The results of the research, as well as their generalization verified by proving the main research hypotheses, pertain to the total risk related to each of the project results, i.e. plan, budget and technical performances of the project. They, accordingly, provide a clear insight into the risks present in the process of industrial product development in small and medium-sized enterprises, as well as into their impact on the defined project goals. IV By testing the main research hypotheses, the following conclusions were reached:  The values of total risk are approximately normally distributed: o Approximately, 68% of total risk is distributed within the interval of one standard deviation away from its mathematical expectation. o Approximately, 95% of total risk is distributed within the interval of two standard deviations away from its mathematical expectation. o Approximately, 99.7% of total risk is distributed in the interval of three standard deviations away from its mathematical expectation.  Variable sizes which determine the probability of error, as well as the variable sizes which determine the error weight, do not contribute to the same extent to the variance of the distribution of total risk. These results can be very useful in the process of decision making, both during the project in order to proactively respond to risk, and when it comes to deciding on whether to initiate or to abandon the project of industrial product development.

Konvejeva notacija u teoriji čvorova i njena primena u metodima za određivanje rastojanja čvorova

Conway notation and its appliance in knot distance determination methods, in knot theory

Glavni sadržaj ovog rada je konstrukcija novih metoda za određivanje različitih tipova rastojanja čvorova - rastojanja čvorova nastalih promenama preseka (Gordijeva rastojanja) i rastojanja čvorova nastalih zaravnjivanjem (s-rastojanja). U radu su predstavljeni različiti načini prikazivanja čvorova, a posebno model ogledalskih krivih. Prikazana je primena ovog modela, kodiranje čvorova u njemu, uveden metod za određivanje čvorova predstavljenih ovim modelom i izvedeni svi čvorovi koji mogu biti smešteni u mreže dimenzija p × q (p ≤ 4, q ≤ 4). Detaljnije su opisane i različite notacije čvorova, a poseban akcenat je postavljen na Konvejevu notaciju i njena topološka svojstva. Konvejeva notacija ima glavnu ulogu u dobijanju novih rezultata u ovom radu...

A main focus of the paper is construction of new methods for defining diverse knot distance types - the distance of knots made by crossing changes (Gordian distance) and the distance among knots made by crossing smoothing (smoothing distance). Different ways of knots presentation are introduced, with objective to a mirror curve model. It is presented a purpose of the model, coding of knots, by using the model preferences, as well as introduction of a method to determinate a knots presented by the model and derived all the knots that could be placed to a nets dimensions p×q (p ≤ 4, q ≤ 4). Diverse knot notations are described into details, with a focus to Conway’s notation and its topological characteristics. As it is known, a present algorithms are based on an algebra of chain fractions, that are in close relation with a presentation of rational knots, which results in an absence of a huge number of non-rational knots, in an existing Gordian’s distance tables. The subject of the paper is an implementation of methods with bases on determination of new distances equal 1. The methods are based on a non-minimal presentation of rational and non-rational knots, generation of algorithms established on geometrical characteristics of Conway’s notation and a weighted graph search. The results are organized into Gordian’s distance knots tables up to 9 crossings, and have been enclosed with the paper. In order to append the table with knots having a bigger number of crossings, it has been suggested a method for extension of results for knot families. Using facts of relation among Gordian’s numbers and smoothing numbers, a new method for smoothing number determination is presented, and results in a form of lists for knots not having more then 11 crossings. In conjunction with Conway’s notation concept and the method, algorithms for a smoothing distance are generated. New results are organized in knot tables, up to 9 crossings, combined with previous results, and enclosed with the paper. A changes and smoothing to a knot crossing could be applied for modeling topoisomerase and recombinase actions of DNA chains. It is presented the method for studying changes introduced by the enzymes. A main contribution to the paper is the concept of Conways notation, used for all relevant results and methods, which led to introduction of a method for derivation a new knots in Conways notation by extending C-links. In a lack of an adequat pattern for an existing knot tables in DT-notation, there is usage of a structure based on topological knot concepts. It is proposed a method for knot classification based on Conways notation, tables of all knots with 13 crossings and alternated knots with 14 crossings has been generated and enclosed. The subject of the paper takes into consideration Bernhard-Jablan’s hypothesis for a determination of unknotting number using minimal knot diagrams. The determination is crucial in computation of diverse knot distances. The paper covers one of main problems in knot theory and contains a new method of knot minimization. The method is based on relevance of local and global minimization...

Metodologija održavanja pogonskih agregata specijalnih vojnih vozila

Methodology of maintenance of special military vehicles engines

U disertaciji je definisana originalna metodologija za izbor koncepcija održavanja sklopova i podsistema tehničkih sistema. Kod novo definisane metodologije bazirane na metodologiji „Održavanje prema pouzdanosti (RCM)“, koraci pri izboru koncepcije održavanja su jasno definisani odgovarajućim originalnim algoritmom, što na kraju rezultuje usaglašavanjem intervala izvođenja izabranih koncepcija održavanja. Rezultati „Analize logističke podrške (LSA)“ se koriste kao osnova za izbor odgovarajućih koncepcija održavanja za pojedine elemente, sklopove i podsisteme tehničkog sistema ili njihove modernizacije, ukoliko nije moguće definisati odgovarajuću koncepciju održavanja u skladu sa definisanim zahtevima za raspoloživost, pouzdanost i bezbednost datog tehničkog sistema. Kod „Analize načina i efekata otkaza (FMEA)“ koja se radi u okviru LSA, po prvi put uvodi se formiranje matrice rizika otkaza sklopova i podsistema, na osnovu posledica otkaza i intenziteta otkaza, radi izbora koncepcija održavanja. Metodologija predviđa da se koncepcija održavanja razmatra još tokom razvoja tehničkog sistema tako da se projektant tehničkog sistema suočava sa znatno većim brojem zahtevnih parametara nego što je to ranije bio slučaj a u cilju stvaranja određenih preduslova za efektivnije i efikasnije održavanja predmetnih tehničkih sistema. Adekvatna primena podrazumeva da se tokom faze upotrebe tehničkog sistema, vrši ažuriranje LSA i po potrebi revizija postojećih koncepcija održavanja pojedinih sklopova i podsistema tehničkog sistema a nakon toga i redefinisanje programske šeme održavanja. Definisanom originalnom metodologijom za izbor koncepcija održavanja sklopova i podsistema tehničkih sistema mogu se adekvatno izabrati koncepcije održavanja i bez podataka iz eksploatacije, na način da se koriste ekspertska znanja od strane korisnika i održavalaca tehničkih sistema što ovu metodologiju čini kvalitativnom. U drugom delu disertacije predstavljena je metodologija za reviziju postojeće koncepcije održavanja pogonskih agregata specijalnih vojnih vozila, koja u sebi xiii sadrži metodologiju za izbor koncepcija održavanja sklopova i podsistema tehničkih sistema. Primena ove metodologije je pokazana na podsistemu pogonskog agregata za dovod goriva, koji se nakon sprovedene Pareto analize pokazao kao najkritičniji sa stanovišta broja evidentiranih otkaza u posmatranom osmogodišnjem periodu eksploatacije, kao i na klipnjači pogonskog agregata, kao pokretnom elementu pogonskog agregata čiji je otkaz prouzrokovao najveću štetu na pogonskom agregatu. Nakon izvršenih određenih analiza u okviru LSA i prolaska kroz originalni algoritam metodologije, određene su koncepcije održavanja za elemente i sklopove podsistema za dovod goriva pogonskog agregata. Intervali za izvođenje aktivnosti održavanja u okviru izabranih koncepcija održavanja su određeni primenom metode višekriterijumskog odlučivanja, uz prethodno definisani uslov da kriterijum zahtevane raspoloživosti ima koeficijent značajnosti 0,9 a kriterijum ekonomičnosti koeficijent značajnosti 0,1. U poređenju sa postojećom koncepcijom preventivnog održavanja podsistema za dovod goriva pogonskog agregata, novom koncepcijom preventivnog održavanja se predlažu dodatne aktivnosti u vidu održavanja prema stanju kritičnih elemenata i sklopova podsistema pogonskog agregata koje nisu obuhvaćene postojećom koncepcijom preventivnog održavanja. U cilju optimizacije intervala održavanja, predlog je da se predložene dodatne aktivnosti održavanja obuhvate već postojećim I i II tehničkim pregledom pogonskog agregata uz neznatno produženje vremenskog intervala za njihovo izvršenje. Pored navedenog, definisan je i originalni postupak za otkrivanje uzroka otkaza klipnjače pogonskih agregata specijalnih vojnih vozila, kao specijalnog događaja, čija analiza se predviđa u okviru LSA kod Analize zadatka održavanja (MTA). Analizom uzroka otkaza je zaključeno da su neadekvatna mašinska obrada, izostanak poliranja i najviše naponsko stanje u delu klipnjače gde se desio lom klipnjače glavni uzroci nastalog loma. Predložene su određene aktivnosti na usavršavanju izrade klipnjače kako se ovakav događaj ne bi više ponovio u buduće. Na osnovu navedenih rezultata analize uzroka otkaza klipnjače može se izvesti opšti zaključak da je definisani postupak primenjiv ne samo kod klipnjače, već i kod ostalih mašinskih elemenata izloženih promenljivim dinamičkim xiv opterećenjima. Definisani postupak omogućuje forenzičkim inženjerima da utvrde tačan uzrok otkaza i na osnovu toga predlože adekvatne preventivne i korektivne aktivnosti. Sprovedenom analizom loma klipnjače pogonskog agregata je utvrđeno da nije opravdano menjati određene elemente i sklopove pogonskog agregata po isteku definisanog eksploatacionog resursa pogonskog agregata od 500 mč, a da se prethodno ne izvrši analiza karakteristika otkaza takvog elementa i sklopa i njihovog trenutnog tehničkog stanja. Obzirom da intenzitet krive otkaza klipnjače ne odgovara obliku za koje odgovara preventivna zamena, prevremena zamena postojeće klipnjače sa novom klipnjačom lošijeg kvaliteta je imala za posledicu havariju pogonskog agregata. Ovo je jedan od razloga koji je uticao na definisanje predloga da se generalni remont pogonskog agregata ne vrši striktno nakon 500 mč rada, već da se tokom II Tehničkog pregleda (nakon 400 mč) izvrši snimanje spoljne brzinske karakteristike pogonskog agregata na probnom stolu u laboratoriji. Ukoliko bi odstupanje spoljno-brzinske karakteristike bilo veće od dozvoljenog predlog je da se pristupi demontaži pogonskog agregata do sklopova ili elemenata, radi njihove defektacije i eventualne opravke ili zamene u skladu sa propisanim aktivnostima generalnog remonta. Na ovaj način, pogonski agregat bi se rastavljao tek kad bi se za to ukazala stvarna potreba. Zaključak je da bi se primenom novo definisane originalne metodologije za izbor odgovarajućih koncepcija održavanja sklopova i podsistema pogonskog agregata i optimizacijom intervala njihovog izvođenja stvorili uslovi za znatno poboljšanje efektivnosti sistema održavanja a samim tim i pouzdanosti i raspoloživosti predmetnih pogonskih agregata specijalnih vojnih vozila.

The dissertation describes the original methodology developed for the selection of the assemblies and subsystems maintenance concepts. Newly defined methodology, based on "Reliability Centered Maintenance (RCM)", has the steps in selection of maintenance concept clearly defined by the appropriate original algorithm, which ultimately results in harmonization of the execution intervals of chosen maintenance concepts. The results of the "Logistics Support Analysis (LSA)" are used as a basis for appropriate maintenance concepts selection for certain elements, assemblies and subsystems of the technical system or their modernization, if there is no possibility to define the appropriate maintenance concept in accordance with the defined technical system’s availability, reliability and safety requirements. In the "Failure mode and effects analysis (FMEA)", which is done within the framework of the LSA, the formation of a risk matrix for failures of assemblies and subsystems is introduced for the first time, based on the consequences of the failures and their intensity, in order to select the maintenance concepts. The methodology envisages that the maintenance concept is considered during the technical system development so that the system designer faces a significantly larger number of requested parameters than it was practice before, in order to create certain preconditions for more effective and more efficient maintenance of the subjected technical systems. Adequate application implies that during the phase of the technical system usage, the LSA is updated and, if necessary, the existing maintenance concepts of individual assemblies and subsystems of the technical system are revised, and finally, the program maintenance schemes are to be redefined. Defined original methodology for the assemblies and subsystems of technical systems maintenance concepts selection enables adequate maintenance concept selection without exploitation data by using of users and maintainers of technical systems experts’ knowledge, which is a characteristic of the qualitative methodology. In the second part of the dissertation, it is presented a methodology for the revision of the existing maintenance concept of special military vehicles engines, which includes a methodology for the selection maintenance concepts of assemblies and subsystems of technical systems. The application of this methodology has been represented on the fuel xvi supply subsystem, which, according to conducted Pareto analysis, proved to be the most critical in terms of the number of registered failures in the observed eight-year exploitation period, as well as on the engine connecting rod, as a moving element whose failure caused the largest damage to the engine. After carrying out certain analyses within the LSA and passing through the original algorithm of the methodology, maintenance concepts for the elements and assemblies of the engine fuel supply subsystem have been defined. Intervals for carrying out of maintenance activities within selected maintenance concepts are determined by using the multi-criteria decision method, with a predefined condition that the required availability criterion has a significance coefficient of 0.9 and the criterion of economy has a significance coefficient of 0.1. Compared to the existing preventive maintenance concept of the engine fuel supply subsystem, the new preventive maintenance concept proposes additional maintenance activities according to the condition of the critical elements and the assemblies which are not covered by the current preventive maintenance concept. In order to optimize the maintenance intervals, it is proposed that the additional maintenance activities include existing I and II Technical Inspection of the engine with a slight extension of the time interval for their execution. In addition, the original procedure for detecting the cause of the special military vehicle engine’s connecting rod failure is defined as a special event, whose analysis is foreseen within the LSA in the Maintenance Task Analysis (MTA). The analysis of the failure cause concluded that inadequate machining, the absence of polishing and the highest stress in the part of the connecting rod where fracture occurred, were the main causes of the resulting fracture. Certain activities have been proposed in order to improve the construction of the connecting rod so that this event does not repeat in the future. Based on the results of the connecting rod failure cause analysis, a general conclusion is that the defined procedure is applicable not only to the connecting rod, but also to other mechanical elements exposed to variable dynamic loads. The defined procedure allows forensic engineers to determine the exact cause of failure and accordingly to propose adequate preventive and corrective actions. The analysis of the engine connecting rod failure found that it is not justified to change certain elements and assemblies of the engine after the defined exploitation resource of 500 Hours of Engine Operations (HEO) expiration, without previous analyzing of the failure characteristics and the current technical state of that element and its assembly. xvii Since the intensity of the connecting rod failure curve does not correspond to the shape for which the preventive replacement corresponds, the premature replacement of an existing connecting rod with a new poor quality connecting rod caused the failure of engine with catastrophic consequences. This is one of the reasons that led to definition of proposal that the overhaul of the engine should not be carried out strictly after 500 HEO, but the recording of the engine’s external speed characteristics on the test bench during the II Technical Inspection (after 400 HEO) is rather to be performed. If the deviation of the external-speed characteristics of engine is greater than allowed, it is necessary to start with disassembling of the engine to the assemblies or elements, for the purpose of their defectation and eventual repair or replacement in accordance with the prescribed activities of the overhaul. In this way, the engine would be disassembled only if the actual need for that has been shown. The conclusion is that by using the newly defined original methodology for selection of the appropriate maintenance concepts of assemblies and subsystems of engines and their execution interval optimizing, the conditions to significantly improve the effectiveness of the maintenance system would be created and, therefore, the reliability and availability of the relevant engines of special military vehicles would also improve.

Mogućnosti i optimizacija sletanja na južnu hemisferu Marsa

Possibilities and optimization of landing on the southern hemisphere of Mars

Skoro 40 godina nakon prvog uspesnog sletanja na Mars, istrazivanje njegove povrsine pomocu kosmickih letelica i dalje predstavlja ogroman naucni i tehnoloski izazov. Ovo se pre svega odnosi na Marsove juzne visoravni, koje se protezu na velikom delu njegove juzne hemisfere, na koje do sada nije izvrseno nijedno uspesno sletanje. Sa druge strane, na severne nizije, koje se prostiru na velikom delu severne hemisfere, do sada je izvrseno 7 uspesnih sletanja. Ova razlika je posledica specicne topograje Marsove povrsine usled takozvane globalne ili hemisfericne dihotomije Marsa, koja predstavlja jedinstvenu karakteristiku u Suncevom sistemu. Usled razlike u elevaci- jama povrsine, letelice koje slecu na juzne visoravni imaju na raspolaganju tanji sloj atmosfere koji mogu da iskoriste za disipaciju kineticke energije. Buduci da letelice sa interplanetarnih putanja ulaze u atmosferu Marsa hiperbolickim brzinama, nji- hova kineticka energija je veoma velika i mogucnost da se ona smanji kroz interakciju sa atmosferom je od odlucujuceg znacaja za sletanje na lokacije sa velikom elevaci- jom. Osim topografske razlike izmed-u severnih nizija i juznih visoravni, postoji i velika razlika u broju udarnih kratera izmed-u ovih oblasti, sto ukazuje na veliku razliku u starosti. Procenjuje se da su juzne visoravni nekoliko milijardi godina starije nego severne nizije, koje su u relativno bliskoj proslosti bile izlozene snaznoj vulkanskoj aktivnosti. Ovako velika razlika u starosti predstavlja motivaciju za istrazivanje celokupne povrsine Marsa u cilju razumevanja njegove slozene geoloske evolucije. Osim toga, veliki broj otvorenih basena, koji su otkriveni na juznim visoravnima, ukazuju na to da su ove regije u proslosti mogle biti prekrivene tecnom vodom, sto predstavlja jedan od najznacajnijih razloga za istrazivanje ovih lokacija...

Almost 40 years after the rst successful landing on Mars, exploration of its surface still presents enormous scientic and technological challenge. This particularly holds for the Mars southern highlands, which cover large part of the southern hemisphere where no successful landings have been achieved up to date. On the other hand, in the northern lowlands, which spans most of the northern hemisphere, there has been 7 successful landings. This difference is consequence of the specic topography of the Martian surface due to so-called global or hemispheric dichotomy, which is a unique characteristic in the Solar system. Due to the difference in the surface elevations, the vehicle trying to land on the southern highlands has thinner layer of the atmosphere that can use for the dissipation of its kinetic energy. Since the vehicles from the interplanetary trajectories enter the Martian atmosphere with hyperbolic velocities, their kinetic energy is extremely large and possibility for its dissipation through the interaction with the atmosphere has crucial importance for the landing on the locations with high elevations. Beside the differences in topography between northern lowlands and southern highlands, there is also a big difference in the number of impact craters between these two regions. This fact indicates that the southern highlands are several billion years older than the northern lowlands, which were exposed to strong volcanic activity in the relatively recent past. Such huge difference in age is strong motivation to explore the entire surface of Mars in order to understand its complex geological evolution. In addition, large number of open basins discovered in the southern highlands suggests that these regions could contain liquid water in...

Jedan prilaz informatičkom modeliranju teksta i algoritmi njegove transformacije

One approach to text modelling and algorithms of text transformations

U tezi se predstavljena dva pristupa modeliranju teksta: prvi se tiče njegove logičke strukture koja se informatički može opisati jezicima za obeležavanje teksta. Detaljno se prikazuje standardni jezik za obeležavanje teksta, SGML, daje se njegova formalna struktura kao i struktura parsera koji ga obrađuju. Dalje su u vidu primera date primene SGML jezika: prva, jednostavna, koja predstavlja osnovu Web-a, HTML, i druga znatno složenija, TEI, koja postaje de facto standard za obeležavanje najraznovrsnijih tekstova (a pre svega literarnih) za različite potrebe . Drugi pristup se tiče modeliranja sadržaja e-tekstova a detaljno se opisuje model koji se zasniva na e-rečnicima. E-rečnici su rečnici namenjeni automatskoj obradi teksta u kojima su „reči“ jezika detaljno opisane: njihova morfološka, sintaksična, semantička dijalekatska i druga svojstva. U tezi se detaljno razlaže model e-rečnika za srpski jezik, a posebno njegov segement koji pokriva glagole, zamenice i brojeve, njihove paradigme. Takođe se opisuje kako bi se varijacije rečnika koje potiču pre svega od različitog izgovora mogle sistematski ugraditi u e-rečnik srpskog kroz korišćenje „super-leme“. Konačno se daju primeri primene ovako obogaćenog e-teksta, na primer kroz prizvodnju bogatijih konkordanci u kojima se pojam ključne reči redefiniše kroz upotrebu (super)lema e-rečnika.

The thesis presents two approaches to the text modeling: the first concerns its logical structure that can be described using mark-up languages. The standard mark-up language, SGML, is described in details, its formal structure and the structure of SGML parsers are thoroughly discussed. The examples of application of SGML language are given as well: first, the simple one, which is the basis of the Web, HTML, and the other is more complex, TEI, which is becoming the de facto standard for annotating diverse texts (primarily literary) for different purposes. The second approach is aimed to modeling the content of e-texts and a model based on e-dictionaries is described in detail. E-Dictionaries are designed for automatic processing of text in which "words" of a language are described in detail: their morphological, syntactic, semantic, dialectal and other properties. The model of an e-dictionary for the Serbian language is presented in full details, particularly its segment covering verbs, pronouns and numbers, and their paradigms. It is also described how the vocabulary variations which come primarily from a variety of pronunciations could be systematically incorporated into an e- Dictionary of Serbian through the use of so-called "super-lemma". Finally, the examples are given that illustrate the application of this enriched e-text, for example through the production of richer concordances in which the concept of keywords is redefined through the use of e-dictionary (super-)lemmas.

Fraktalna i teksturalna analiza strukture hromatina limfocita krvi kod pacijenata sa različitim stepenom depresije

Fractal and textural analysis of blood lymphocyte chromatin structure in patients with different grades of depression

Imajući u vidu dokazanu povezanost između promena u strukturnoj kompleksnosti tkiva i pojedinih fizioloških i patoloških procesa, pretpostavili smo da u toku depresije dolazi do smanjenja strukturne kompleksnosti nukleusnog hromatina limfocita, odnosno da sa povećanjem depresivnih simptoma dolazi do smanjenja hromatinske fraktalne dimenzije. Takođe smo pretpostavili da su depresivni simptomi u pozitivnoj korelaciji sa nivoom hromatinske neuređenosti, odnosno da teksturalna entropija hromatina limfocita raste sa povećanjem depresije kod pacijenata. Ciljevi istraživanja su bili: (1) Utvrditi postojanje razlike u fraktalnim i teksturalnim parametrima hromatina limfocita između pacijenata sa depresijom, i ispitanika koji ne zadovoljavaju dijagnostičke kriterijume za depresiju i (2) Utvrditi postojanje i jačinu korelacije između stepena depresije i fraktalnih, odnosno teksturalnih parametara hromatina limfocita Metod. Istraživanje je dizajnirano po tipu studije preseka i obavljeno na uzorku od 100 ispitanika, od kojih 50 pacijenata obolelih od rekurentnog depresivnog poremećaja lečenih u Klinici „Dr Laza Lazarević“ u Beogradu. Kontrolnu grupu je činilo 50 ispitanika koji su pripadali drugim dijagnostičkim kategorijama, a nisu ispunjavali dijagnostičke kriterijume za depresiju. Eksperimentalni protokol je sproveden na sledeći način: uzimanje detaljne istorije bolesti, primena Bekove skale za procenu depresije i Skale 5 depresivnosti, anksioznosti i stresa DASS–21, uzimanje krvi za standardne hormonske (tireoidni hormoni) i biohemijske analize. Uporedo sa uzimanjem krvi za analize, pristupilo se pravljenju krvnih razmaza na predmetnim staklima i njihovom fiksiranju u metanolu. Krvni razmazi su bili obojeni modifikacijom metode Gimza bojenja po uzoru na prethodno publikovanu studiju u kojoj su autori radili fraktalnu analizu hromatina krvnih ćelija. Nakon bojenja razmaza, kreirani su digitalni mikrografi limfocita i svaki ispitivani limfocit je izračunata prosečna vrednost hromatinske fraktalne dimenzije, lakunarnosti, entropije, angularnog drugog momenta, teksturalnog kontrasta, varijanse, teksturalne korelacije i inverznog momenta razlike, po ugledu na prethodno publikovane studije. Nakon toga za svakog pacijenta izračunali smo prosečne vrednosti ovih parametara. Kao dodatak, u in vitro delu studije, ispitivana je mogućnost da dopamin, odnosno njegov derivat oksidopamin ima određenu ulogu u promenama fraktalnih i teksturalnih (GLCM) parametara hromatina limfocita. Rezultati. Postojala je značajna razlika u pogledu fraktalnih parametara jedra limfocita periferne krvi između pacijenata sa depresijom i kontrolne grupe. Sa druge strane, nije bilo značajne razlike u pogledu teksturalnih parametara hromatina limfocita između pacijenata sa depresijom i kontrolne grupe. Kod pacijenata obolelih of rekurentnog depresivnog poremećaja, fraktalni parametri hromatina limfocita su bili u korelaciji sa rezultatima skale za depresiju. Pokazano je da oksidopamin značajno menja teksturalne parametre hromatina limfocita periferne krvi. Zaključak. Fraktalna i teksturalna analiza su potencijalno značajni dodaci budućim tehnikama i algoritmima za evaluaciju pacijenata sa rekurentnim depresivnim poremećajem. 6

Introduction. Having in mind already proven connection between changes in tissue complexity and certain physiological and pathological processes, we assumed that during the depression the structural complexity of the nucleic lymphocytes chromate is reduced, that is, with the increase in depressive symptoms, the chromatin fractal dimension decreases. We also assumed that depressive symptoms were in a positive correlation with the level of chromatic disorder, that textural entropy of lymphocyte chromatin was increasing with increasing depression in patients. The objectives of the research were: (1) to determine the existence of a difference in the fractal and textural parameters of lymphocyte chromatin between patients with depression and subjects who did not meet the diagnostic criteria for depression; and (2) determine the existence and strength of the correlation between the degree of depression and fractal or textural parameters of chromatin in lymphocytes. Methods. The study was designed as a cross section study, and performed on a sample of 100 subjects, of which 50 patients with recurrent depressive disorder treated at the Clinic "Dr Laza Lazarević" in Belgrade. The control group consisted of 50 subjects who belonged to other diagnostic categories and did not fulfill the diagnostic criteria for depression. The experimental protocol was carried out as follows: taking a detailed history of the disease, applying the Beck scale to assess depression, and Depression, anxiety 8 and stress scales DASS-21, taking blood for standard hormone (thyroid hormones) and biochemical analysis. Along with taking blood for analysis, blood smears were created on glass slides and fixated in methanol. The blood smears were colored using the modification of the Gimza staining method based on a previously published study in which the authors performed a fractal analysis of blood cell chromatin. After dyeing the smears, digital micrographs of the lymphocytes were created, and for each tested lymphocyte we calculated the average value of the chromatin fractal dimension, laccunarity, entropy, angular second moment, textural contrast, variance, textural correlation, and inverse moment of difference, according to the previously published studies. After that, for ea ch patient, we calculated the average values of these parameters. In addition, in the in vitro part of the study, we investigated the possibility the dopamine derivative oxidopamine, has a particular role in the changes in fractal and textural (GLCM) parameters of lymphocyte chromatin. Results. There was a significant difference in nuclear fractal parameters of the peripheral blood lymphocytes between patients with depression and control group. On the other hand, there was no significant difference in the textural parameters of lymphocyte chromatin between patients with depression and control group. In patients suffering from recurrent depressive disorder, the fractal parameters of lymphocyte chromatin were in correlation with the results of the depression scale. It was shown that oxidopamine significantly changes the textural parameters of peripheral blood lymphocyte chromatin. Conclusion. Fractal and textural analyses are potentially significant additions to future techniques and algorithms for evaluating patients with recurrent depressive disorder. 9

Primena robusne redukcije reda sistema u modelovanju i upravljanju realnim objektima u mašinstvu

Application of robust order reduction in modeling and control of real systems and objects in mechanical engineering

Disertacija PRIMENA ROBUSNE REDUKCIJE REDA SISTEMA U MODELOVANjU I UPRAVLjANjU REALNIM OBJEKTIMA U MAŠINSTVU je posvećena tehnikama i metodama za redukciju reda linearnih modela, predstavljenih u prostoru stanja, kao i uslovima i ograničenjima za njihovu primenu. Prikazan je koncept singularnih perturbacija i njegova primena u redukciji složenosti modela sistema. Razmotrene su različite karakterizacije linearnih modela velikog reda, čije reprezentacije modelima u prostoru stanja imaju različite vremenske skale ili i mali parametar (imaju matrice velikih dimenzija sa puno nula-elemenata ili sa elementima, čiji se redovi veličine veoma razlikuju). Predstavljeni su: singularno perturbovani slabo povezani sistemi, slabo povezani sistemi, singularno perturbovani sistemi, kvazi singularno perturbovani sistemi i kvazi slabo povezani sistemi. Svaka od ovih klasa predstavljena je odgovarajućim modelom u prostoru stanja nekog realnog sistema. Detaljnije je izložen proračun regulatora za linearan singularno perturbovan sistem. Opisana je metoda balansiranja, kao i tehnike za redukciju reda linearnih modela, poznate iz literature, koje zahtevaju primenu transformacije balansiranja: balansirano odsecanje, balansirana rezidualizacija, generalisana balansirana rezidualizacija, korigovano balansirano odsecanje, metoda zasnovana na brzom podsistemu uz odbacivanje sporog, modifikovana generalisana balansirana rezidualizacija, kao i reverzna tehnika rezidualizacije. Svaka od ovih tehnika daje po jedan model redukovanog reda, polazeći od balansiranog modela punog reda. Tehnike izložene u ovoj tezi, primenjene su na redukovanje reda četiri modela realnih sistema. Prvi je model binarne destilacione kolone sa pregrevačem i devet podova. Drugi je model u prostoru stanja borbene letelice L-1011. Treći je matematički model katalitički kontrolisane reakcije, iz procesne tehnike. Četvrti je matematički model dela elektroenergetskog sistema Srbije, sačinjen od dve mašine. Izložena je i redukcija reda modela nestabilnih linearnih sistema, na osnovu izabrane literature. Multimodelovanjem uz primenu balansiranja se linearni model punog reda aproksimira pomoću dva podsistema, nazvana prvi i drugi, bez gubitka promenljivih stanja. Redovi prvog i drugog podsistema se redukuju nekom od ranije izloženih tehnika. Dva načina su primenjena na modelu sistema iz prakse (model u prostoru stanja za elektroenergetski sistem, sastavljen od dve identične oblasti, gde je svaka oblast jedna elektrana) – jednom su prvi i drugi podsistem redukovani primenom balansiranog odsecanja, a drugi put primenom balansirane reziduali- zacije. Kvalitet aproksimacije balansiranog modela punog reda, kako u otvorenom kolu dejstva tako i u zatvorenom kolu dejstva, sa optimalnim LKG regulatorom u povratnoj sprezi, pokazao se kao dobar. Izložen je novi pristup multimodelovanju na bazi balansiranja, koji se zasniva na primeni principa superpozicije. I ova tehnika primenjena je na model realnog sistema – binarne destilacione kolone sa pregrevačem i devet podova. -

The dissertation Application of Robust Order Reduction for Modeling and Control of Real Systems in Mechanical Engineering is dealing with techniques and methods of order reduction for linear models in the state space representation, as well as with conditions and limitations of their applicability. Well known concept of singular perturbations is described with its application in reduction of complexity of the system’s model. Different characterizations of large scale linear models are overviewed, having state space representations who exhibit different time scales or small parameter, too. These models have matrices of large dimensions with many zero-elements or elements with different size order, some very large and others very small. The singularly perturbed and weakly coupled systems, the weakly coupled systems, the singularly perturbed systems, the quasy singularly perturbed systems and the quasy weakly coupled systems are listed and represented here. Each class of these is represented with corresponding state space model of one real system. The design of the regulator for the linear singularly perturbed system is described in detail. The balancing method is analyzed, as well as techniques for the order reduction of linear model’s, known from the literature. These require application of the balancing transformation: the balancing truncation, the balancing residualization, the generalized balancing residualization, the corrected balancing truncation, the method based on the fast subsystem with rejection of the slow subsystem, the modified generalized balancing residualization, as well as the reversed residualization technique. Each of these techniques mentioned gives one model of reduced order, starting from the balanced full order model. Techniques represented in this thesis are applied on order reduction of four real system’s models. The first is the state space model of binary distillation column, with condenser, reboiler and nine plates. The second is the state space model of L-1011 fighter aircraft. The third is a mathematical model of the controlled catalytic reaction, from process engineering. The fourth part is a mathematical model of the part of electric power system of Serbia, consisting of two machines. Described here is the order reduction for models of unstable linear systems as well, based on the chosen references. Multimodeling based on balancing application results in approximation of the linear full order model with two subsystems, namely the first and the second, without losing any of the state space variables. Order of the first and of the second subsystem can be further reduced, using some of the techniques mentioned earlier. Two ways are applied on the state space model known from engineering practice (state space representation of the model of the power system composed of two interconnected identical areas, where each area is one power plant) – once the first and the second subsystems have their orders reduced through application of the balanced truncation, then applying the balanced residualization. The precision obtained in approximation of the balanced full order model, both in open loop and in closed loop (closed using the optimal LQG regulator) was acceptable in both cases. The new approach is described - the multimodeling based on balancing, with application of the superposition principle. This technique too is applied on the real system’s model – the one of binary distillation column, with condenser, reboiler and nine plates. -

Istraživanje obrazaca u određivanju karakteristika proteina

Mining sequential patterns for determination of protein characteristics

Belančevine ili proteini su važni biološki makromolekuli polimerne prirode (polipeptidi), koji se sastoje od amino kiselina i predstavljaju osnovnu gradivnu jedinicu svake ćelije...

Proteins are signicant biological macromolecules of polymeric nature (polypeptides), which contain amino acids and are basic structural units of each cell. Their contents include 20+3 amino acids and, as a consequence, they are presented in biological databases as sequences formed from 23 dierent characters. Proteins can be classied based on their primary structure, secondary structure, function etc. One of possible classications of proteins by their function is related to their contents in a certain cluster of ortholologous groups (COGs). This classication is based on the previous comparison of proteins by their similarities in their primary structures, which is most often a result of homology, i.e. their mutual (evolutionary) origin. COG database is obtained by comparison of the known and predicted proteins encoded in the completely sequenced prokaryotic (archaea and bacteria) genomes and their classication by orthology. The proteins are classied in 25 categories which can be ordered in three basic functional groups (the proteins responsible for: (1) information storage and processing; (2) cellular processes and signaling; and (3) metabolism), or in a group of poorly characterized proteins. Classication of proteins by their contents in certain COG category (euKaryote Orthologous Groups- KOG for eukaryotic organisms) is signicant for better understanding of biological processes and various pathological conditions in people and other organisms. The dissertation proposed the model for classication of proteins in COG categories based on amino acid n-grams (sequences of n- length). The set of data contains protein sequences of genomes from 8 dierent taxonomic classes [TKL97] of bacteria (Aquicales, Bacteroidia, Chlamydiales, Chlorobia, Chloroexia, Cytophagia, Deinococci, Prochlorales), which are known to have been classied by COG categories. The new method is presented, based on the generalized systems of Boolean equations, used for separation of n-grams characteristic for proteins of corresponding COG categories. The presented method signicantly reduces the number of processed n-grams in comparison to previously used methods of n-gram analysis, thus more memory space is provided and less time for protein procession is necessary. The previously known methods for classication of proteins by functional categories compared each new protein (whose function had to be determined) to the set of all proteins which had already been classied by functions in order to determine the group which contained most similar proteins to the one which was to be classied. In relation to the previous, the advantage of the new method is in its avoidance of sequence-sequence comparison and in search for those patterns (n-grams, up to 10 long) in a protein which are characteristic of the corresponding COG category. The selected patterns are added to a corresponding COG category and describe sequences of certain length, which have previously appeared in that COG category only, not in the proteins of other COG categories. On the basis of the proposed method, the predictor for determination of the corresponding COG category for a new protein is implemented. Minimal precision of the prediction is one of the predictors arguments. During the test phase the constructed predictor shown excellent results, with maximal precision of 99% reached for some proteins. According to its properties and relatively simple construction, the proposed method can be applied in similar domains where the solution of problem is based on n-gram sequence analysis.

Određivanje nivoa koncentracije glukoze u vodenim rastvorima na osnovu promeneoptičkih osobina kontaktnih sočiva dopiranih nanomaterijalima

Determining the level of glucose in aqueous solutions based on changes optical properties of nanomaterial-based contact lenses

Predmet ove disertacije je istraživanje kako nanostrukturni materijali ugrađeni u meka kontaktna sočiva bogata vodom utiču na promenu vidljivog spektra pod uticajem različitih koncentracija glukoze u vodi, polazeći od činjenice da je merenje koncentracije glukoze u vodenim rastvorima na bazi materijala kontaktnih sočiva osnova za merenje glukoze u suznom filmu i krvnom sudu oka. Konvencionalne metode određivanja glukoze u krvi su invazivne i skupe te istraživanje primena neinvazivnih metoda postaje veoma aktuelna oblast proučavanja što se može uočiti i kroz porast patentne dokumentacije iz ove oblasti. U osnovni materijal mekih kontaktnih sočiva SL38 dodati su fulerenski nanomaterijali fuleren C60 (fullerene) , fuleren hidroksilat C60(OH)24 (fullerenol) i fuleren metamorfen hidroksilat C60(OH)12(OC4N5H10)12 (metformin hydroxylated fullerene). Posmatrana su tri različita rastvora koncentracije glukoze 5mmol/l,10mmol/l i 15mmol/l. Fulereni su korišćeni zbog svojih dobrih transmisionih karakteristika u oblasti ultraljubičaste, vidljive i bliske infracrvene svetlosti. Ovako dopirana nanofotonska kontaktna sočiva predstavljaju biosenzore za kontinuirano, neinvazivno praćenje nivoa glukoze koja se nalazi u suzama. Nanofotonična kontaktna sočiva i kontaktna sočiva izrađena od osnovnih materijala su potopljena u rastvore tokom određenog vremenskog perioda, sa ciljem da se utvrdi njihov uticaj na rastvore glukoze. Rezultati ispitivanja novo dobijenih sočiva OMS metodom pokazuju da intenzitet pomeraja zavisi od koncentracije glukoze, ali i vremena uticaja glukoze na ispitivano sočivo. Pomoću UV-VIS spektroskopije detektovane su absorpciono/transmisione karakteristike za svaki ispitivani slučaj i na osnovu dobijenih spektara data je analiza kontaktnih sočiva koja bolje optički reaguju na datu koncentraciju glukoze. Dobijeni rezultati istraživanja omogući će dalja in vivo istraživanja identifikacije glukoze u suznom filmu.

The aim of the dissertation is to study nanostructured materials embedded in soft contact lenses rich in water which resulting in changes in the visible spectrum under the influence of different concentrations of glucose in water, proceeding from the fact that the measurement of the concentration of glucose in the aqueous-based contact lens materials basis for the measurement of glucose in the tear film of the eye and the blood vessel. Conventional methods for determining glucose levels are invasive and expensive, therefore the application of non-invasive method becomes very actual area of study which can be seen in the increase of patent documents in this field of research. In the basic material of soft contact lenses SL38 were added the fullerene C60, fullerenol C60(OH)24 and metformin hydroxlate fullerene C60(OH)12(OC4N5H10)12. The study focused on three different glucose concentrations 5mmol / l, 10mmol / l and 15 mmol / l. Fullerenes were used because of their good transitive characteristics in ultraviolet, visible and near infrared light spectrums. Such doped nanophotonic contact lenses are biosensors for continuous, non-invasive glucose monitoring, which can be found in tears. The nanophotonic contact lens and contact lens made from base material were dipped in solutions over a specific period of time, in order to determine their influence on the glucose solutions. Test results obtained by OMS were observed that the intensity of the shift depends on the concentration of glucose, and time the influence of glucose to the tested lens. By using UV-VIS spectroscopy have been detected the absorption / transmission characteristics of each case and related to the obtained spectra, contact lenses have been analyzed according to the optically better respond to a given glucose concentration. The results of research will enable further in vivo studies of identification of glucose in the tear film.

Vihorno strujanje u pravim difuzorima kružnog poprečnog preseka

Swirl flow in a conical diffusers

Ovim radom obuhvaćena su kompleksna teorijska i eksperimentalna istraživanja vihornih turbulentnih strujanja, u pravim difuzorima kružnog poprečnog preseka. Generator vihornog strujanja, u okviru ovog rada, je obrtno kolo aksijalnog ventilatora, koje je postavljeno u početnom delu kratke cevne deonice kružnog poprečnog preseka na koju se nastavlja ispitivani difuzor. Prvi cilj istraživanja je izučavanje karakteristika vihornog strujanja u difuzoru, odnosno određivanje odnosa koeficijenta gubitaka pri vihornom strujanju sa koeficijentom gubitaka pri čisto aksijalnom strujanju u pravim difuzorima kružnog poprečnog preseka. Drugi cilj, je formiranje algebarskog modela turbulentnog vihornog strujanja, baziranog na Rejnoldsovim jednačinama i saznanjima o ukupnim naponima turbulentnog strujanja do kojih se došlo u okviru istraživanja. Ukupni naponi turbulentnog strujanja računaju se na osnovu prosečnih vrednosti veličina stanja fluida. Prosečne vrednosti određene su merenjem profila totalnog (ukupnog) pritiska, statičkog (strujnog) pritiska i strujnog ugla, odnosno profila obimske i aksijalne komponente prosečne brzine, u različitim presecima duž tri difuzora (različitih uglova širenja), za više različitih režima. Analizom sračunatih ukupnih napona turbulentnog vihornog strujanja, dobijeni su opšti izrazi i zakonitosti njihovih promena po radijusu duž difuzora. Na osnovu algebarskog modela određuju se polje brzine i pritiska u nizstrujnim presecima difuzora, kada su dati granični i početni uslovi nestišljivog strujanja fluida.

This doctoral dissertation comprises complex theoretical and experimental investigations of the turbulent swirl flow in conical diffusers. The incompressible swirl flow field is induced by the axial fan runner, which is set in the initial part of the straight pipe section followed by a conical diffuser. The first goal of this research is studying of swirl flow caracteristics in conical diffusers, and determination of the ratio of swirl flow loss coefficients and loss coefficients in pure axial flow in conical diffusers. The second goal is to form an algebraic model for the calculation of turbulent swirl flow, which is based on the Reynolds equations and turbulent stresses determined in the experimental investigations. Turbulent stresses are calculated on the basis of the average values of fluid state parameters. Average values were determined by measuring the profiles of total and static pressure, as well as flow angle, i.e. profiles of the average circumferential and axial velocity components at different sections along the three conical diffuser geometries (various diffuser angles), for different regimes. Аnalysis of calculated turbulent stresses provides general expressions and their distributions along the radius and diffuser axis. Velocity and pressure fields in the downstream sections of the diffuser are determined on the basis of algebraic model, when boundary and initial conditions of incompressible fluid flow are defined.

Istraživanje i razvoj metode za analizu radnog procesa motora na osnovu merenja trenutne ugaone brzine kolenastog vratila

Proces sagorevanja kod motora SUS najčešće se simulira Vibeovim parametarskim modelom. Parametri modela zavise od geometrijskih karakteristika komore i termo- i gaso-dinamičkih pojava u cilindru, zbog čega se korišćenjem empirijskih relacija njihove vrednosti ne mogu odrediti lako i sa dovoljnom tačnošću. Osim determinističkih metoda koje ne omogućuju sistemski pristup, za odreĎivanje parametara modela oslobaĎanja toplote koriste se i indirektni postupci zasnovani na minimizaciji greške izmeĎu vrednosti modeliranog i izmerenog pritiska u cilindru. U ovom radu prkazan je postupak identifikacije parametara modela oslobaĎanja toplote na primeru višecilindarskog benzinskog motora primenom optimizacione metode Levenberg-Markvart sa fiksnim ograničenjima za minimizaciju sume kvadrata razlika izmerene i modelirane trenutne ugaone brzine kolensatog vratila. Pritisak u cilindru i moment gasne sile u ugaonom domenu simuliran je pomoću nelinearnog dvozonskog, nultodimenzionalnog modela radnog procesa. Moment mehaničkih gubitaka u ugaonom domenu simuliran je pomoću detaljnog, višekomponentnog analitičkog modela trenja u tribološkim sistemima motora. Razvijeni modeli iskorišćeni su za analizu uticaja parametara dinamičkog modela motora na tok trenutne ugaone brzine kolenastog vratila i istraživanje odstupanja u merenjima radi definisanja ograničenja u procesu identifikacije parametara modela.

Most frequently used approach to describe combustion process in IC engines is based on Wiebe function. Its parameters are strongly influenced by geometric and in-cylinder gas- and thermodynamic processes and can’t be predicted easily and with sufficient accuracy using available empirical relations. Along with deterministic methods lacking in systematic approach, indirect analysis methods based on minimization of error between simulated and measured in-cylinder pressure have been successfully demonstrated. In this paper the same concept is introduced to identify combustion model parameters of a multi-cylinder SI engine based on box constrained Levenberg-Marquardt minimization of nonlinear Least Squares (LSQ) given for measured and simulated instantaneous crankshaft angular speed determined from the solution of the engine dynamics torque balance equation. For predictions of gas pressure and engine torque, a nonlinear, detailed, angle resolved two-zone, zero-dimensional multi-cylinder SI engine combustion model has been developed as well as a detailed analytical component model of engine friction and mechanical losses. The models are used to analyze the influences of engine combustion and dynamics model parameters on instantaneous crankshaft speed and to investigate biases in measurement in order to recognize and establish the constraints in process of combustion model parameters identification.

Apstrakcija i primena matematičke analize u nastavi na tehničkim fakuletima

Abstraction and application of mathematical analysis in teaching on technical faculties

Osnovni koncept rada jeste povezanost apstraktne teorije i primenjene matematičke analize u univerzalnom matematičkom sistemu. Matematika je doživela veliku praktičnu primenu, na primer, primenu matematičke statistike, numeričke analize, primene u elektrotehnici, doprinos razvoju računarstva i dr. Istovremeno, u naučnom pogledu, uzdigla se do neslućenih apstrakcija (topološki prostori, vektorski prostori i drugo). Zbog ovih činjenica neophodno je unapredivati nastavu matematičke analize na tehničkim fakultetima ali i same metode nastave.

The main concept of this thesis is the connection of abstract theory and applied mathematical analysis in the universal mathematical system. The Mathematics is being applied practically in great number of cases, for example, in the application of mathematical statistics, numerical analysis, application in electrical engineering, contribution to computer developing etc. At the same time, in scientific perspective, it reached the level of incredible abstractions (topological spaces, vector spaces et al.). Due to these facts, it has become necessary to advance the teaching of mathematical analysis in technological universities, but also the very methods of teaching as well. This thesis contains both a theoretical and an empirical research. The theoretical research sheds light onto the notions of abstraction and application and gives examples from next teaching subjects: Lagrange’s Theorem, Convexity and Consequences. Taylor’s Formula. Hardy’s Approach to Calculating Surfaces of Flat Shapes. Fourier’s Series and their Application. Banach Fixed-Point Theorem and its’ Application. The empirical research consists of two parts: a questionnaire and a test. This research shows how the relation between theory and application is seen by students, how they perceive the teaching of mathematics and which learning instruments they use. The research also shows how students solve simple problems and which type of problems they solve more sucessfully. The sample consists of 429 students of electrical engineering, construction and mechanical engineering for the Questionnaire and 450 students of the same universities for the Test. Students that participated in the research are studying at the University of Belgrade, University of Novi Sad and University of Niˇs. The results of the research confirmed that students from technological universities have a favourable attitude towards mathematics and that they see its significance in its application, i.e. in its use value. Students have clearly defined attitude on the idea that a good lecture by a professor is one which can be understood, which is well articulated, and which motivates students to take part in it. They point out the significance of examples that have elements of application. Visual presentations also enhance the success in solving problems. The research shows that students did not acquire the skill to apply their knowledge of mathematical analysis in solving tasks and problems. Theoretical clarification of notions of abstraction and application, followed by a displayfive topics of mathematical analysis confirm that abstract theory and applied mathematical analysis are interconnected and conjoint in the universal mathematical system. On the basis of results recommendations are defined concerning innovational approaches to teaching, such as planning the lectures and meliorating the contents, asking questions, and also an intelligent prospect, lecture improvement and learning instruments application. This way, it is confirmed that a methodically well organized lecture helps a better understanding of the relation between abstraction and application of mathematical analysis.

Ocenjivanje parametara pouzdanosti dvokomponentnog sistema

Estimation of the reliability parameter of a two-component system.

Prvi radovi iz tzv. stress-strength problematike pojavili su se polovinom dvadesetog veka, a ova oblast koja pripada teoriji pouzda- nosti je veoma živa i danas, čemu u prilog govori i to da se svake godine objavi po desetak radova na ovu temu. U disertaciji su predstavljeni metodi za ocenjivanje parametra pouzdanosti kod sistema sa nezavisnim stresom i snagom. Takođe su uvedena dva nova modela i izvedene su razne ocene parametra pouzdanosti svakog od modela.

Early papers dealing with so-called stress-strength problems were published in the middle of the 20th century. This topic, which belongs to the reliability theory, is still very active nowadays, which can be seen through the number of published papers dealing with it - around ten each year. In this dissertation, some methods for estimation of the reliability parameter for a system with independent stress and strength are presented. Also, two new models are introduced and some estimators of the reliability parameter for each of them are derived. The dissertation is divided into four chapters. In the rst chapter, some basic terms are introduced and some examples from real life, illustrating big possibilities for application of the results from this scientic eld, are described. Sorted based on the stress and strength distributions, a chronological overview of all research activities dealing with these topics, to the author's best knowledge, is presented. Some special func- tions, which are later used for calculations, along with their main properties are shown. The expressions for the reliability parameter for some stress and strength distributions are either derived or listed. The second chapter is devoted to different methods used for point esti- mation, as well as for interval estimation of the reliability parameter of a system. For each methods estimators of the reliability parameter for some stress and strength distributions are either derived or listed. In the third chapter, a new model is introduced. In this model, the stress has geometric, while the strength has Poisson distribution. This is one of the rst, if not the rst, appearances in the literature, where the stress and strength distributions do not belong to the same family of distributions. For this model, the reliability parameter is estimated using different methods and decision on optimal estimators for usage in practice is based on the simulations. In the fourth chapter, another model is introduced, with the stress and strength distributions which are not only from different families of distribu- tions, but also do not belong to the same type of distributions. The stress has geometric, while the strength has exponential distribution...

Nelinearna dinamika satnih mehanizama

Nonlinear dynamic of clock mechanisms edoctoral dissertation

Satni mehanizmi pripadaju oblasti precizne mehanike, koji visokim stepenom ravnomernosti svoga hoda mere proticanje vremena. Ključni podsklop satnog mehanizma predstavlja zaprečno-impulsni mehanizam koji svojim funkcijama unosi poremećaje u oscilatorni proces tako da oscilacije više nisu sopstvene, već prinudne sa frekvencom koja je podložna promeni. Zato se korektan matematički i fizikalni opis funkcionisanja satnih mehanizama ne može ispravno ostvariti zanemarivanjem malih poremećaja višeg reda, linearizacijama diferencijalnih jednačina i aproksimacijama uobičajenim za klasični pristup u izučavanju mehanizama. Predmet istraživanja ove disertacije jesu upravo ti mali poremećaji višeg reda koji se javljaju u zaprečno-impulsnom mehanizmu i narušavaju ravnomernost hoda satnog mehanizma. Korišćenjem teorije perturbacija biće izvedeni matematički izrazi za izračunavanje grešaka koje generišu dve vrste zaprečno-impulsnih mehanizama. Korektnost kako ovih matematičkih izraza, tako i onih opštih, integralnih formula, proverena je kompjuterskom simulacijom i analizom kretanja 3D modela odgovarajućih zaprečno-impulsnih mehanizama. Prikazana je kvalitativna i kvantitativna analiza grešaka zaprečno-impulsnih mehanizama odnosno, poremećaja frekvence oscilatora časovnika koga izazivaju njihovi zaprečno-impulsni mehanizmi. Pored istraživanja poremećaja rada zaprečno- impulsnog mehanizma, disertacija daje prikaz svih nelinearnih dinamičkih svojstava oscilatora satnog mehanizma kao i opis delova sata koji su korišćeni za simulaciju sata. Numerički rezultati ovih simulacija pokazali su visok stepen slaganja sa rezultatima teorijskih numeričkih proračuna, čime je potvrđena korektnost matematičkih izraza izvedenih primenom perturbacionog računa.

Clock mechanisms belongs in the field of precision mechanics, and they measures the flow of time with a precision uniformity of its work. A key subassembly of the clock mechanism is a escapement mechanism that, by its functions, creates disruptions in the oscillatory process so that the oscillations are no longer their own, but are forced with a frequency that is subjective to change. Therefore, the correct mathematical and physical description of the functioning of clock mechanisms cannot be correctly achieved by neglecting small disorders of higher order, linearization of differential equations and approximations common to the classical approach in the study of mechanisms. The subject of the study of this dissertation are precisely these small disorders of higher order that occur in the escapement mechanism and undermine the uniformity of the work of the clock mechanism. Using the theory of perturbation, mathematical expressions for the calculation of errors are derived for two types of impulse mechanisms. The correctness of both these mathematical expressions and those general integral formulas was verified by computer simulation and the analysis of the motion of the 3D model of the corresponding escapement mechanisms. A qualitative and quantitative analysis of errors of the escapement mechanisms, i.e. frequency disturbances of the clock oscillator caused by their escapement mechanisms are presented. In addition to investigating the disturbance of the work of the escapement mechanism, the dissertation provides an overview of all the non-linear dynamic properties of the clock mechanism oscillator as well as a description of the parts of the clock that were used to simulate the clock. The numerical results of these simulations showed a high degree of agreement with the results of the theoretical numerical calculations, which confirmed the correctness of mathematical expressions performed using the perturbation calculus.

Naponska analiza ugaonih zavarenih spojeva nosača konstrukcija savremenim numeričko-eksperimentalnim pristupom

Stress analysis of corner welded joints structure by modern numerical-experimental approach

Otvoreni i zatvoreni tankozidi preseci imaju široku primenu u industrijskim aplikacijama za dizajn mnogih mašina i strukturnih komponenti. Ove komponente su često fabrikovane zavarivanjem, a ne livenjem ili kovanjem. Tankozidi profili su obično povezani pomoću ugaonih zavarenih spojeva. Takvi spojevi su takođe korišćeni u drugim inženjerskim aplikacijama kao što su građevinske mašine, mostovi, ramovi, šasije vozila i dr.. U ovoj disertaciji ponašanje zatvorenih profila (kutija, okrugla cev) i otvoreniih profila (L, Z, C i K profili) šavnih profila, proučavani su tako što su izlagani statičkim opterećenjima i numeričko-eksperimentalnim pristupom. Sa tačke gledišta strukturne analize, uprkos široko rasprostranjenim ugaono zavarenim spojevima koji efikasno nose opterećenja na elementima, ne postoji praktičan, jednostavan i precizan pristup za njihov dizajn i analizu. U tu svrhu , inženjeri često moraju pripremiti relativno komplikovan model konačnih elemenata ploče ili zapremine. To je zato što navedeni eleementi izloženi opštim opterećenjima proizvode koncentracije napona u blizini spojeva. Jedan od doprinosa ovog istraživanja je bolje razumevanje ponašanja ugaono zavarenih spojeva izloženi naprezanju savijanja i naprezanja savjanja sa uvijanjem. Primena metode konačnih elemenata predstavlja osnovni numerički pristup. Posmatrani profili su medelilrani različitim konačnim elementima (greda, ploča i zapremina) i sa različitim slučajevima opterećenja. Primećuje se da na kratkom rastojanju od spoja putem ugaono zavarenih spojeva struktura nosećeg elementa pod izloženim opterećenjem ponaša se kao greda. Zavareni spojevi su generisani različitim tehnikama modelrianja sa primenom različitih tipova i veličina konačnih elemenata elemenata. Kompletni rezultati deformacija su, takođe, eksperimentalno dobijeni pomoću tehnike primene korelacije digitalne slike (DIC). Eksperimentalni testovi se izvode za proveru numeričkih simulacija u cilju ispitivanja mehaničkih performansi zavarenh spojeva izloženi kombinovanim opterećenjima. Rezultati numeričkog prroačuna pokazuju dobro slaganje sa eksperimentalnim rezultatom . vi U numeričkom modeliranju primenom metode konačnih elelemenata zavareni spojevi se modeluju pomoću greda, ploča i zapremina korišćenjem softvera - Abakus. Performanse zavarenih spojeva dalje su proučavani korišćenjem jednostavnih analitičkih metoda sa ciljem dobijanja brzih informacija o uticaju zavarenog spoja na predviđanju strukturalnih performansi. Eksperimenti na mnogim spojevima na gredi koji su podvrgnuti savijanju se koriste za dalju procenu FE modela razvijenih u ovoj studiji. Studija se obavlja sa fokusom na predviđanje nosivosti strukture zavarenih spojeva. Rezultati proračun na grednom modelu pokazuju prostiranje linearnog napona posmatranog preko debljine grede. Zato je, upotrebna ploča za analizu spojeva odgovarajuća. Takođe, na relativno maloj udaljenosti od spoja, efekti lokalnih napona i deformacija nestaju. Rezultati pokazuju da normalan napon u jednom ili dva pravca su najviši na ivici zavarenog spoja na gornjoj površini i na horizontalnoj liniji zavarenog spoja poprečnog preseka profila. Primetno je da komponenta normalnog napona u jednom ili dva pravca ima najveću vrednost. Normalni naponi u jednom ili dva pravca mnogo veći od ostalih komponenti napona. To znači da su naponi istezanja u zavarenom spoju uglavnom odgovorni za pojavu plastičnosti i naprslina, kao i njihovo širenje.

Hollow and non hollow section members are widely used in industrial applications for the design of many machine and structural components. These components are often fabricated by welding rather than by casting or forging. For example, in agricultural machinery, the hollow tubes are typically connected together through welding to form a corner welded joints. Such joint connections are also employed in other engineering applications such as construction machinery, offshore structures, bridges, and vehicle frames. In this dissertation, the behavior of tubular (box and circle profile) and non tubular (L, Z, C and X profiles) joint connections profiles, subjected to static loads were studied both experimentally and numerically. From a structural analysis point of view, despite of the wide use of corner welded joints as efficient load carrying members, there is no available practical, simple and accurate approach for their design and analysis. For this purpose, engineers must often prepare relatively complicated and time consuming Finite Element models made up of shell or solid elements. This is because unlike solid-section members, when hollow section members are subjected to general loadings, they may experience severe deformations of their cross-sections that results in stress concentrations in the connection’s vicinity. One of the objectives/contributions of this research work is the better understanding of the behavior of the corner welded joint connections under out-of- plane bending and torsion loading conditions. Through a detailed Finite Element Analysis (FEA) using shell and solid elements, the stress distribution at the connection of the tubular and non tubular corner welded joints were obtained for different loading conditions. It is observed that at a short distance away from the connection of the corner welded joints, the structure behaves similar to beams when subjected to loadings. Finite element models with different modeling techniques and meshing with various size and types of elements were created and analyzed. The full displacement field results were obtained experimentally by using the digital image correlation (DIC) technique. Experimental tests were performed to validate numerical simulations in order to investigate the mechanical performance of a series of fillet-welded connections under combined loading. The full displacement field results show good agreement comparing with the experimental results. iv In the numerical study, the connections were modeled using beam, shell and solid elements using software for Finite Element Analysis - Abaqus. The performance of the connections under combined loading was further studied by a simple analytical method aiming for rapid information on the influence of the weld modeling on prediction of the structural performance. Experiments of many beam joints subjected to bending were used for further assessment of the FE models developed in the present study. The study was performed with focus on the prediction of the structural strength of fillet welded joints. The finite element results of the solid element model show that a generally linear stress distribution across the thickness of the beams is observed. Therefore, the use of shell elements for the analysis of the beam-joint connection is appropriate. Also, at a relatively short distance from the connection, the effects of the local stresses and deformations disappear. The results show that the normal stresses in the 1 or 2-direction are the highest at the weld path (top of the path) on the top surface, and at horizontal weld path of the beam. It can be observed that the 1 or 2-direction component of normal stress causes the most damage. The values of the Von-Mises equivalent stresses are very close to the values of the stresses in the 1 or 2-direction (absolute value). This is expected as the stresses in the 1 or 2-direction are much larger than the other components of stress. This means that the strains and stresses normal to the weld (i.e., 1 or 2 -direction in this study) are mainly responsible for plasticity/crack initiation and propagation.

Identifikacija parametara ekvivalentnog dinamičko - torzionog sistema kolenastog vratila dizel motora na osnovu promenljivog toka ugaone brzine

Identification of parameters of diesel engine crankshaft equivalent dynamic-torsion system based on crankshaft variable angular velocity

Trenutna ugaona brzina motora SUS se najčešće koristi za analizu radnih procesa u cilindrima motora, pošto takve informacije u njoj dominiraju. Kod većih motora elastične deformacije kolenastog vratila usled torzionih oscilacija ostavljaju mnogo dublji trag na trenutnu ugaonu brzinu nego što je to slučaj kod manjih, ne mnogo opterećenih motora. Sa stanovištva torzionih oscilacije, dve najvažnije informacije procenjene sa velikom tačnošću, mogu se dobiti jednostavnom harmonijskom analizom trenutne ugaone brzine kolenastog vratila. To su informacije o vrednostima sopstvenih frekvencija oscilovanja i potencijalno kritičnim brojevima obrtaja. Dalji proračuni vezani za probleme torzionog oscilovanja dinamičkog sistema sa motorom kao njegovim delom, zasnivaju se na idealizaciji i uvođenju ekvivalentnog dinamičko- torzionog sistema (EDTS). Parametri EDTS koje je potrebno definisati su inercijalne mase, torzione krutosti i koeficijenti unutrašnjeg i spoljašnjeg prigušenja. Pored toga, potrebno je poznavati momente koji pobuđuju kretanja (uključujući i torzione oscilacije) pojedinih inercijalnih masa dinamičkog sistemu. Za ovakav sistem se mogu postaviti diferencijalne jednačine sa ciljem da opišu kretanja svake inercijalne mase. Ni za jedan od nabrojanih parametara EDTS se ne može tvrditi da je prethodno relativno tačno procenjen, pošto postoje metode koje obezbeđuju samo njihovu orijentacionu procenu. U ovom radu je prikazan postupak identifikacije parametara EDTS na primeru dizel- motora sa dvanaest cilindara u V gradnji, primenom optimizacione metode Markvarta (Marquardt) baziranoj na minimizaciji sume kvadrata razlika izmerene i modelirane trenutne ugaone brzine kolenasatog vratila. Koeficijenti unutrašnjeg prigušenja su poslužili da se proceni ukupni srednji moment trenja u motoru. Momenti gasnih sila su procenjeni na osnovu izmerenih pritisaka u dva cilindra na suprotnim stranama ispitivanog motora. Na ovom primeru se jasno pokazalo da je pretpostavka o ujednačenosti radnih procesa po cilindrima ključni faktor koji utiče na odstupanja izmerene i modelirane trenutne ugaone brzine kolenastog vratila. Procenom pobudnih momenata inercijalnih sila na osnovu razvijene procedure određivanja promenljivog momenta inercije složenog krivajnog mehanizma ispitivanog motora, isključena je mogućnost da se sa te strane unese greška u proračun. Dodatno, momenti koji se od kolenastog vratila oduzimaju za pogon značajnijih podsistema motora su kombinacijom matematičkog modeliranja i neposrednog merenja procenjeni u ugaonom domenu kolenastog vratila i direktno ugrađeni u matematički model. Razvijeni modeli i postupci su iskorišćeni za procenu nivoa torzionih oscilacija u ispitivanom dinamičkom sistemu i mehničkog stepena korisnosti motora.

The current angular velocity of IC engines is most commonly used to analyze the work processes in engine cylinders, since such information dominate in it. In larger engines, the crankshaft elastic deformations, due to torsional vibrations, leave a much deeper mark on the current angular velocity than in the case of smaller engines which are not heavily stressed. From the viewpoint of torsional oscillations, the two most important information, estimated with high accuracy, can be obtained by simple harmonic analysis of the crankshaft current angular velocity. Those are information about the values of natural oscillation frequencies and potentially critical rpm. Further calculations related to the problems of torsional oscillation of a dynamical system with an engine as its part, are based on idealization and introduction of an equivalent dynamic-torsional system (EDTS). The EDTS parameters which must be defined are the inertial masses, the torsion stiffnesses and the coefficients of internal and external damping. In addition, it is necessary to know the moments that cause motions (including torsional oscillations) of individual inertial masses of the dynamic system. Differential equations can be set up for such a system in order to describe the motion of each inertial mass. For none of the EDTS parameters mentioned above can be said that it was previously estimated with relative accuracy, since there are only methods that provide their approximate estimation. This paper presents a procedure to identify the parameters of the EDTS, on the example of diesel engine with twelve cylinders in V layout, by applying the Marquardt optimization method based on the minimization of the squares sum of the differences between the measured and modeled current angular velocity of the crankshaft. Internal damping coefficients were used to estimate the total mean friction moment in the engine. Moments of gas forces are estimated based on the measured pressures in two cylinders on opposite sides of the tested engine. This example clearly shows that the assumption of work processes uniformity per cylinders is the key factor that affects the deviations of the measured and modeled current angular velocity of the crankshaft. By evaluating the excitation moments of inertial forces on the basis of developed procedure for determining the variable moment of inertia of the complex crank mechanism of the tested engine, any possibility of introducing an error in the calculus from this side was excluded. In addition, by combining the mathematical modeling and direct measurements, the moments taken from the crankshaft to drive the engine major subsystems are estimated in the crankshaft angular domain and incorporated directly into the mathematical model. The developed models and procedures were used to estimate the level of torsional oscillations in the tested dynamic system and engine mechanical efficiency.

Unapređenje hibridizacijom metaheuristika inteligencije rojeva za rešavanje problema globalne optimizacije

Improvement by hybridization of swarm intelligence metaheuristics for solving global optimization problems.

Te²ki optimizacioni problemi, nere²ivi u prihvatljivom vremenu izvr²avanja deterministi £kim matemati£kim metodama, uspe²no se poslednjih godina re²avaju populacionim stohasti£kim metaheuristikama, me u kojima istaknutu klasu predstavljaju algoritmi inteligencije rojeva. U ovom radu razmatra se unapre enje metaheuristika inteligencije rojeva pomo¢u hibridizacije. Analizom postoje¢ih metaheuristika u odre enim slu£ajevima uo£eni su nedostaci i slabosti u mehanizmima pretrage prostora re²enja koji pre svega proisti£u iz samog matemati£kog modela kojim se simulira proces iz prirode kao i iz nedovoljno uskla enog balansa izme u intenzikacije i diversikacije. U radu je ispitivano da li se postoje¢i algoritmi inteligencije rojeva za globalnu optimizaciju mogu unaprediti (u smislu dobijanja boljih rezultata, brºe konvergencije, ve¢e robustnosti) hibridizacijom sa drugim algoritmima. Razvijeno je i implementirano vi²e hibridizovanih metaheuristika inteligencije rojeva. S obzirom da dobri hibridi ne nastaju slu£ajnom kombinacijom pojedinih funkcionalnih elemenata i procedura razli£itih algoritama, ve¢ su oni utemeljeni na sveobuhvatnom izu£avanju na£ina na koji algoritmi koji se hibridizuju funkcioni²u, kreiranju hibridnih pristupa prethodila je detaljna analiza prednosti i nedostataka posmatranih algoritma kako bi se napravila najbolja kombinacija koja nedostatke jednih neutrali²e prednostima drugih pristupa. Razvijeni hibridni algoritmi verikovani su testiranjima na standardnim skupovima test funkcija za globalnu optimizaciju sa ograni£enjima i bez ograni£enja, kao i na poznatim prakti£nim problemima. Upore ivanjem sa najboljim poznatim algoritmima iz literature pokazan je kvalitet razvijenih hibrida, £ime je potvr ena i osnovna hipoteza ovog rada da se algoritmi inteligencije rojeva mogu uspe²no unaprediti hibridizacijom...

Hard optimization problems that cannot be solved within acceptable computational time by deterministic mathematical methods have been successfully solved in recent years by population-based stochastic metaheuristics, among which swarm intelligence algorithms represent a prominent class. This thesis investigates improvements of the swarm intelligence metaheuristics by hybridization. During analysis of the existing swarm intelligence metaheuristics in some cases deciencies and weaknesses in the solution space search mechanisms were observed, primarily as a consequence of the mathematical model that simulates natural process as well as inappropriate balance between intensication and diversication. The thesis examines whether existing swarm intelligence algorithms for global optimization could be improved (in the sense of obtaining better results, faster convergence, better robustness) by hybridization with other algorithms. A number of hybridized swarm intelligence metaheuristics were developed and implemented. Considering the fact that good hybrids are not created as a random combination of individual functional elements and procedures from dierent algorithms, but rather established on comprehensive analysis of the functional principles of the algorithms that are used in the process of hybridization, development of the hybrid approaches was preceded by thorough research of advantages and disadvantages of each involved algorithm in order to determine the best combination that neutralizes disadvantages of one approach by incorporating the strengths of the other. Developed hybrid approaches were veried by testing on standard benchmark sets for global optimization, with and without constraints, as well as on well-known practical problems. Comparative analysis with the state-of-the-art algorithms from the literature demonstrated quality of the developed hybrids and conrmed the hypothesis that swarm intelligence algorithms can be successfully improved by hybridization...

O nekim transmisionim problemima u disjunktnim oblastima

About some transmission problems in disjoint domains

U primenama, naročito u inženjerstvu, često se sreću kompozitne ili slojevite strukture, pri čemu se osobine pojedinih slojeva mogu značajno razlikovati od osobina okolnog materijala. Slojevi mogu imati strukturnu, termičku, elektromagnetsku ili optičku ulogu itd. Matematičkim modelovanjem prenosa energije i mase u oblastima sa slojevima dobijaju se tzv. transmisioni problemi. Na samom početku, u disertaciji se razmatra transmisioni spektralni problem u oblasti koja se sastoji od dva disjunktna intervala. Na svakom intervalu zadat je problem sopstvenih vrednosti, dok se interakcija između njihovih rešenja opisuje nelokalnim uslovima saglasnosti. Dokazana je egzistencija prebrojivog niza generalisanih rešenja, pri čemu uređeni parovi sopstvenih funkcija pripadaju odgovarajućim prostorima Soboljeva. Opisana je struktura spektra i asimptotsko ponašanje sopstvenih vrednosti. Konstruisana je diferencijska shema za njihovo rešavanje. Pored transmisionog spektralnog problema, u disertaciji se razmatraju klase nestandardnih eliptičkih i paraboličkih transmisionih problema u disjunktnim oblastima. Kao modelni primer uzeta je oblast koja se sastoji iz dva nesusedna pravougaonika. U svakoj podoblasti zadat je granični problem eliptičkog tipa, odnosno početno-granični problem paraboličkog tipa. Interakcija između rešenja opisuje se pomoću nelokalnih uslova saglasnosti na granicama posmatranih podoblasti. Razmotreno je više primera žičkih i inženjerskih zadataka koji se svode na transmisione probleme sličnog tipa. Za modelne probleme dokazana je egzistencija i jedinstvenost rešenja u odgovarajućim prostorima Soboljeva. Takođe su konstruisane diferencijske sheme za njihovo rešavanje i dokazana njihova konvergencija.

In applications, especially in engineering, often are encountered composite or layered structures, where the properties of individual layers can vary considerably from the properties of the surrounding material. Layers can be structural, thermal, electromagnetic or optical, etc. Mathematical models of energy and mass transfer in domains with layers lead to so called transmission problems. At the beginning, in this dissertation we consider a transmission spectral problem in the area, which consists of two disjoint intervals. At each interval was given a eigenvalue problem, while the interaction between their solutions is described by means of the nonlocal integral conjugation conditions. The existence of countable series of generalized solutions is proved, whereby ordered pairs of eigenfunctions belong to the corresponding Sobolev spaces. The structure of the spectrum and asymptotic behavior of eigenvalues is described. A diference scheme for solving them is constructed. In addition to the spectral transmission problems, in this dissertation we consider a class of non-standard elliptic and parabolic transmission problems in disjoint domains. As a model example it is taken an area consisting of two non-adjacent rectangles. In each subarea was given a boundary problem of elliptic type, as well as initial-boundary problem of the parabolic type. The interaction between their solutions is described by means of the nonlocal integral conjugation conditions. It was considered more examples of physical and engineering tasks which are reduced to transmission problems of similar type. For the model problems the existence and uniqueness of its weak solution in appropriate Sobolev-like space is proved. They are also constructed a finite diference schemes for solving them and proved their convergence.

Mogućnost primene peleta i briketa od biomase za proizvodnju toplotne energije u sistemima daljinskog grejanja Beograda

Possibilities for utilization of wood briquetts and pellets for Belgrade district heating system heat production

Izvršena je detaljna analiza mogućnosti primene peleta i briketa od biomase za proizvodnju toplotne energije u postojećim kotlovima sistema daljinskog grejanja koji su namenski konstruisani i razvijeni za sagorevanje uglja. Na bazi te analize postavljena je hipoteza da je u kotlovima namenski konstruisanim za sagorevanje ugljeva moguće uspešno sagorevati i drvne brikete i pelete uz povećan stepen korisnosti i smanjenu emisiju polutanata. Na bazi pažljivo analiziranog koncepta i odabira kotlovskih postrojenja izvršena je serija eksperimentalnih istraživanja performansi na više kotlova različitih konstrukcija. Izvršena su sistematska ispitivanja koja su obuhvatala praćenje svih ključnih performansi kotlova, uključujući i sastav produkata sagorevanja. Dobijeni rezultati su detaljno analizirani što je u potpunosti potvrdilo opravdanost hipoteze. Takođe, postavljen je empirijski model kojim se predviđa emisiju ugljen monoksida za kotlove sa nepokretnom rešetkom malih i srednjih snaga.

A detailed analysis on possibilities for heat production in coal fuelled boilers for district heating systems making use of wood briquettes and pellets was performed. Based on this analysis a hypothesis was proposed that these boilers could be fuelled with wood briquettes and pellets resulting in improved efficiency and emissions. A careful analysis of this concept resulted in selection of different types of boilers and testing methodology. Detailed systematic experimental research was performed which covered key boiler performance and flue gases pollutants measurements. The obtained results were analyzed which confirmed the hypothesis of improved efficiency and emissions. An empirical model of CO emissions for low and medium thermal power boilers with fixed grates was developed, as well.

Granične raspodele parcijalnih maksimuma ravnomernih AR(1) procesa

The limiting distributions of the partial maxima in the uniform AR(1) processes

Tema doktorske disertacije odnosi se na probleme ekstremnih vrednosti u strogo stacionarnim slučajnim nizovima. Ona pripada veoma aktuelnoj oblasti verovatnoće i statistike sa mnogobrojnim primenama. Oslanja se na veliki broj izvornih naučnih radova i monografija. Predmet doktorske disertacije je određivanje asimptotskog ponašanja maksimuma nekompletnih uzoraka iz autoregresionog procesa prvog reda sa ravnomernim marginalnim raspodelama. Disertacija se sastoji iz tri poglavlja.

The subject of this doctoral dissertation is related to the problems of extreme values in strictly stationary random sequences. It belongs to the topical area of probability and statistics, broadly applicable to real life situations and in many scientic elds. It relies on large number of seminal articles and monographs. The main aim of the dissertation is to determine the asymptotic behavior of maxima of some incomplete samples from the rst-order auto-regressive processes with uniform marginal distributions. The dissertation consists of three chapters.

Ispitivanje promene kvaliteta života i stepena depresivnosti bolesnika kod različitih metoda lečenja kamena u urinarnom traktu

The examination of the quality of life and degree of depression among patients with urinary tract stones treated by different methods

Urolitijaza je jedno od najčešćih oboljenja u urologiji sa stalnim porastom incidence i prevalence u populaciji. Patogeneza ovog oboljenja je multifaktorijalna, a za posledicu ima veoma ozbiljne probleme koje mogu imati značajan uticaj na kvalitet života ovih bolesnika. Hronicitet urolitijaze ima za rezultat višu incidencu depresije kod ovih bolesnika. Tokom prethodnih godina hirurški modaliteti lečenja urolitijaze pretrpili su evolutivne promene u korist minimalno invazivnih tehnika lečenja ove bolesti u cilju postizanja bolje efikasnosti lečenja i poboljšanja kvaliteta života bolesnika. Cilj studije bio je ispitati i proceniti kvalitet života i stepena depresivnosti bolesnika sa urolitijazom u odnosu na pimenjenu metodu lečenja. Metode: Ispitivanje je dizajnirano kao panel studija (kombinacija studije preseka i kohortne studije). U studiju su bili uključeni bolesnici sa urolitijazom lečeni ESWL ili Ureterorenoskopskom litotripsijom ( URSL aparatom Lithoclast) u periodu od 01.02.2017-01.02-2018. godine od kojih je sto ispitanika ispunilo kriterijume za uključivanje u studiju. Oni su podeljeni u dve grupe: Grupa 1 ili Lithoclast grupa (URSL) i Grupa 2 ili ESWL grupa . U svakoj grupi bilo je 50 ispitanika. Kao instrument za merenje kvaliteta života korišćen je SF 36 upitnik i za procenu stepena depresivnosti korišćena je Bekova skala depresivnosti koju su bolesnici popunjavali neposredno preoperativno i jedan mesec postoperativno, prilikom dolaska na kontrolni pregled. Za obradu podataka korišćen je SPSS program. Rezultati: Statističkom obradom i analizom skora SF 36 upitnika preoperativno, utvrdili smo da postoji smanjenje kvaliteta života kod bolesnika sa urolitijazom u gotovo svim dimenzijama života (uloga fizičkog funkcionisanja, telesni bol, vitalnost, opšte zdravlje, uloga emocija i mentalno zdravlje).Nije bilo statistički značajne razlike u preoperativnom SF skoru između dve grupe bolesnika izuzev domene uloge fizičkog zdravlja u kojoj je URSL (Lithoclast) grupa imala statistički značajno viši SF skor od ESWL grupe i mentalnog zdravlja u kojoj je ESWL grupa imala statistički značajno viši skor od Lithoclast grupe. Statističkom anlizom SF 36 upitnika postoperativno i ispitivanjem uticaja modaliteta lečenja na kvalitet života utvrdili smo da Lithoclast metoda postiže mnogo viši, statistički značajan skor SF 36 upitnika i to u više dimenzija života odnosno statistički značajno utiče na poboljšanje kvaliteta života bolesnika postoperativno i to u više dimenzija života (uloga fizičkog funkcionisanja, telesni bol,vitalnost i socijalno funkcionisanje) u odnosu na ESWL metodu koja čak beleži i blagi pad u skoru za neke dimenzije kvaliteta života postoperativnno. Statističkom analizom Bekove skale depresivnosti utvrdili smo veću učestalost depresije u obe grupe bolesnika u odnosu na zdravu populaciju. Nije bilo statistički značajne razlike između dve grupe. Zaključak: Procena kvaliteta života dobar je put za evaluaciju modaliteta lečenja u kliničkoj praksi. Jedan od načina da se to uradi jeste procena kvaliteta života bolesnika preko različitih dimenzija života primenom upitnika SF 36 visoke pouzdanosti. Primenom SF 36 upitnika u našoj studiji utvrdili smo da je ureterorenoskopska litotripsija (Lithoclast metoda) modalitet lečenja urolitijaze koja postoperativno daje mnogo bolje, statistički značajno poboljšanje kvaliteta života u nekoliko domena zdravlja u odnosu na ESWL metodu. Primenjeni metodi lečenja nisu imali uticaja na simptome depresije postoperativno.

Urolithiasis is one of the most common urological illnesses with a continual rise in incidence and prevalence in the population. Its pathogenesis is multi-factor; hence, its consequences are serious problems that can significantly impact the quality of life of the patients. Patients with urolithiasis have higher incidence of depression because its chronicity. In the last years, operational modes of urolithiasis treatment had undergone evolution changes towards minimally invasive treatment techniques aimed at improving its efficacy and patients' life quality. The purpose of the study is to examine and evaluate the quality of life and degree of depression of the patients with urolithiasis depending on the applied treatment method. Methods:This research is designed as a panel study – a combination of a cross- sectional and cohort study. The sample included patients with urolithiasis treated with ESWL or ureteroscopic lithotripsy (URSL apparatus Lithoclast), on the basis of the decision of the Calculosis Consilium. The research was carried out between February 1, 2017 and February 1, 2018, and 100 respondents met the inclusion criteria. They were divided into two equal groups considering the applied method of the stone disintegration: Group 1 or Lithoclast Group (URSL) and Group 2 or ESWL Group, each with 50 respondents. The instrument used for measuring the quality of life was SF 36 questionnaire and the Beck Depression Inventory for depression. It was administrated to the patients immediately before the operation and one month after the operation during the control checkup. The SPSS program was used for the data analysis. The results: The statistical analysis of the scores obtained pre-operationally on the SF 36 questionnaire have revealed the decrease in the quality of life of patients with urolithiasis in almost all dimensions of life (the role physical functioning, bodily pain, overall health, the role emotions and mental health). The statistically relevant difference in preoperative SF scores between the two groups of patients was not established except in the domain of the role of physical health. In this domain, the URSL (Lithoclast) group had a statistically significant higher score than the ESWL group. Furthermore, in the domain of mental health, the ESWL group had a statistically significant higher score than the Lithoclast group. The post-operative statistical analysis of SF questionnaires and the examination of the impact of the treatment mode on the quality of life have shown that the use of the Lithoclast method has resulted in the much higher, statistically significant score at SF36 questionnaires regarding several life dimensions. Hence, it has a statistically more significant impact on the enhancement of the quality of life of patients post- operatively regarding several life dimensions (the role of a physical functioning, bodily pain, vitality and social functioning) than the ESWL method. The application of the ESWL method even resulted in the decrease in the post-operational score for some life dimensions.The statistical anlysis The Beck Depression Inventory scale showed higher incidence of depression in both groups with no difference between groups. Conclusion: The assessment of the quality of life is an adequate tool for the evaluation of treatment modes in the clinical practice. One method of accomplishing it is to assess the patient's quality of life divided into different life dimensions through the highly reliable SF 36 questionnaire. By using the SF 36 questionnaire in this study, we established that the ureteroscopic lithotripsy (the Lithoclast method) is a method that results in much higher and statistically significant improvement of the life quality in several health domains than the ESWL method. Methods of treatment did not have influenze on appearance and simptoms of depression.

Aerodinamička opterećenja i optimizacija vetroturbine za specifične resurse vetra na lokacijama u Srbiji

Aerodynamic loads and optimization of wind turbine for specific wind resources on locations in Serbia

U okviru istraživanja obrađena je i definisana metodologija određivanja aerodinamičkih opterećenja lopatica vetroturbina i optimizacije prema uslovima vetra na određenim karakterističnim lokacijama u Srbiji. Proračun aerodinamičkih opterećenja izveden je primenom teorije elementa lopatice sa savremenim korekcijama koje se odnose na popravke za konačan broj lopatica, radijalno strujanje, radne režime pri skretanju rotora i teško opterećeni rotor. Takođe je formulisana metoda određivanja aerodinamičkih opterećenja vetroturbine primenom Rejnoldsovih jednačina sa izvornim članovima kojima je modeliran uticaj aerodinamičkih sila rotora na strujanje. Izvorni članovi se računaju primenom teorije elementa lopatice. Poređenjem rezultata sa dostupnim eksperimentalnim ispitivanjima vetroturbina zaključeno je da se definisana metodologija može primeniti za potrebe optimizacije. Metoda konačnih zapremina primenjena je za proračun strujanja nestišljivog viskoznog fluida u cilju određivanja aerodinamičkih karakteristika aeroprofila i definisanja ulaznih podataka kako za aerodinamički proračun vetroturbine tako i za proces optimizacije oblika aeroprofila. Validnost primenjene metode testirana je na aeroprofilima za koje su objavljeni rezultati eksperimentalnih ispitivanja visokog kvaliteta. Za potrebe obrade rezultata primenjene su savremene metode obrade i vizuelizacije velikog broja podataka. U okviru metodologije optimizacije koja se u slučaju aeroprofila zasniva na metodi optimizacije rojem čestica, a u slučaju vetoturbine na genetskom algoritmu primenjena je metoda generalizacije na osnovu koje su izvedeni opšti zaključci o mogućnostima primene razvijene metodologije za određivanje aerodinamičkih opterećenja za potrebe optimizacije lopatica vetroturbina. Parametrizacija aeroprofila izvršena je CST metodom koja za relativno mali broj parametara daje dobre rezultate, čime je omogućeno brže postizanje optimalnog rešenja. Konačno, izvršena je aerodinamička optimizacija lopatica dve vetroturbine za tri izabrane lokacije u Srbiji prema kriterijumu maksimalne godišnje proizvodnje energije. Prva lopatica je namenjena za vetroturbinu nominalne snage 2.6 MW koja je regulisana promenljivom brzinom obrtanja i korakom rotora, dok je druga lopatica za vetroturbinu snage 10 kW regulisana promenom brzine obrtanja i slomom uzgona.

Methodology of determination of wind turbine aerodynamic loads was defined as well as optimization of wind turbine blades for wind conditions on certain locations in Serbia. Calculation of wind turbine aerodynamic loads was performed using the results of blade element momentum theory with contemporary corrections in order to account for finite number of blades, radial flow, yaw and highly loaded rotor. In addition, the method for determination of wind turbine blade aerodynamic loads using Reynolds equations with source terms that account for rotor aerodynamic forces influence on flow was developed. Source terms are calculated using blade element theory. Comparison of the results for both methods with available experiment researches of wind turbine aerodynamics has showed that the results are applicable for wind turbine blade optimization. Finite volume method was used for flow analysis of incompressible viscous fluid in order to determine aerodynamic characteristics of airfoils and input data definition both for calculation of blade aerodynamic loads as well as optimization of airfoil shape. Validity of applied methods was tested on airfoils for which there are results of high quality experiments. In order to process obtained data methods of data processing and visualization were used. In the optimization framework which is based on particle swarm optimization for airfoil optimization and genetic algorithm for wind turbine blade optimization method of generalization was applied in order to derive general conclusions on abilities to use developed methodology for wind tubine aerodynamic loads calculation in order to concuct optimization. Parameterization of airfoil was done using Class Shape Transformation which is able to provide good parameterization for relatively small number of parameters. The speed of optimization method was increased by using this method. Finally, aerodynamic optimization was conducted for two wind turbines and three selected locations in Serbia based on maximum annual energy production criteria. The first blade was intended for 2.6 MW wind turbine controlled by variable rotor speed and pitch, and the second one is intended for 10 kW wind turbine regulated by variable rotor speed and stall...

Modifikacija pristupa upravljanju projektima u montaži mašinskih instalacija i opreme

Modification of approach to project management in mechanical equipment installation projects

U disertaciji je razmatrana modifikacija projektnog pristupa u montaži mašinskih instalacija i opreme kod kvalifikovanih preduzeća koja se bave montažom mašinskih instalacija i opreme u izgradnji investicionih objekata, sagledano postojeće stanje i dati kvantitativni pokazatelji ocene projektnog pristupa. Prilagođena je i organizacioni dijagram investicionog projekta u kome se koriste Internet tehnologije. Na osnovu rezultata istraživanja zaključeno je da preovlađuje neformalizovan pristup upravljanju projektima, pa je predložena modifikacija projektnog pristupa, tako da se u većoj meri primenjuju savremene informaciono-komunikacione tehnologije. Pored toga, modifikovana je i metoda za upravljanje rizikom, i predložena formalizacija metoda upravljanja rizikom u projektu.

This dissertation considers the project approach to mechanical equipment installation projects for qualified Serbian companies that perform the installation of mechanical equipment in the construction of investment facilities. It also reviews the current situation and provides quantitative indicators of the project approach. Furthermore, the dissertation presents a adjusted diagram of the investment project that use Internet technologies. Then, on the basis of research results, it was concluded that non- formalized approach of project management prevails, so it is proposed to modify the project approach in order to use modern information-communication technologies and the Internet to a larger extent. Also, the method for risk management is modified and suggested formalization of method for project risk management.

Strukturna predikcija funkcije proteina i odnos funkcionalnih kategorija i neuređenosti

Structured prediction of protein function and relationship between functional categories and disorder

Proteini predstavljaju najvazniju grupu biomolekula u zivom svetu. Ra- zlicite funkcije koje imaju u svakom organizmu jedinstvene su i nezamenljive, pocev od raznovrsnih celijskih procesa, preko njihove strukturalne uloge, uloge katalizatora, velikog broja metabolickih funkcija i slicno. Poznavanje i razumevanje funkcija proteina je stoga esencijalno u istrazivanju bilo kog bioloskog procesa, sa posebnim naglaskom na oboljenja ljudi, s obzirom da se mnoga od njih mogu pojaviti zbog funkcionalnih mutacija. U ovom radu bice predstavljeno istrazivanje ovog domena kroz dva razlicita pristupa. U prvom, funkcija proteina posmatrana je kroz GO ontologije, koje podrazumevaju predstavljanje funkcije proteina kroz tri velika usme- rena aciklicka grafa funkcija: jedan je vezan za bioloske procese, drugi za celijske komponente, a treci za molekulsku funkciju. Svaki od njih sadrzi vise hiljada cvorova, pri cemu svaki cvor odreduje specicniju funkciju od svojih predaka. Zadatak ovog dela istrazivanja je razvoj prediktora funkcije proteina na osnovu njene primarne sekvence primenom metode struktural- nih podrzavajucih vektora koja predstavlja generalizaciju poznate metode podrzavajucih vektora na strukturalni izlaz. Jedno od osnovnih nacela molekularne biologije predstavlja paradigma struk- tura-funkcija po kojoj je 3D struktura proteina blisko povezana sa njegovom ulogom u organizmu. Utvr eno je da su neure eni proteini (kojima ne- dostaje 3D struktura) kao i neure eni delovi proteina u vezi sa teskim savre- menim bolestima i kao takvi su predmet aktuelnih istrazivanja. U drugom pravcu razmatrana je veza funkcionalnih kategorija proteina sa njihovom neuredenoscu, kao i sa drugim zicko-hemijskim karakteristikama proteina. Funkcija proteina ovde je posmatrana kroz 25 osnovnih funkcija proteina koje su grupisane u 4 funkcionalne grupe. U ovoj disertaciji su prikazani rezul- tati detaljne analize nad velikim skupom proteina za koje je neure enost odre ena primenom javno dostupnih alata...

Proteins represent the most important groups of biomoleculs. Dierent functions that they carry out in each organism are unique and irreplaceable, including versatile cellular processes, structural role of proteins, catalytic function, a number of metabolic functions and so on. Knowing and under- standing protein function is therefore essential in investigation of any biolo- gical process, especially of human diseases since a lot of them are caused by functional mutations. In this paper, we represent investigation of protein function domain through two dierent approaches. In the rst one, protein function is represented by GO ontologies with the structure of a directed acyclic graph. There are three GO ontologies: one for functions regarding biological processes, one for functions regarding cellular components and one for molecular functions. Each ontology contains several thousands of nodes, where every node deter- mines more specic function than his ascendants. The task of this part of research was to develop a software for predicting protein function from its primary sequence based on structural support vector machines method which represents generalization of well-known support vector machines method on structural output. Structure-function paradigm is one of basic concepts in molecular biology, stating that 3D proten structure is closely connected to its role in organism. It has been detected that disordered proteins (the ones that lack 3D struc- ture) and disordered regions of proteins are related with severe contemporary illnesses, which contributed to their popularity in modern research. In an- other aspect, we investigated the relationship between proteins' functional categories and their disorder, as well ad with other physico-chemical char- acteristics of proteins. Here, protein function has been observed through 25 elementary functions grouped in 4 functional groups. In this work, we present results of thorough analysis over large protein dataset where dis- order has been determined computationally, using publicly available tools...