Primjena umjetne inteligencije u optimizaciji rada obradnih sustava

diplomski rad

• Naslov (engleski): Application of artificial intelligence in the optimization of machining system
• Autor: Nikola Pavušek
• Mentor: Toma Udiljak (mentor)
• Ustanova koja je dodijelila akademski / stručni stupanj: Sveučilište u Zagrebu; Fakultet strojarstva i brodogradnje; Zagreb
• Datum i država obrane: 2010-10-14, Hrvatska
• Znanstveno / umjetničko područje, polje i grana: TEHNIČKE ZNANOSTI; Strojarstvo

Sažetak

U radu su prikazane mogućnosti korištenja nekonvencionalnih tehnika optimizacije obradnih sustava. Najkorištenije tehnike uključuju primjenu umjetnih neuronskih mreža, genetskih algoritama, te tehnika optimizacije baziranih na matematičkom opisu prirodnih procesa (optimizacija kolonijom mrava, optimizacija rojem čestica (inteligencija rojeva)). Matematički modeli procesa koji se javljaju u procesu obrade prikazani su sa ciljem pobližeg upoznavanja sa problem optimizacije. Dan je prikaz optimizacije pojedinih procesa obrade, te je dana usporedba dobivenih rezultata ovisno o korištenim algoritmima optimizacije. Primjena genetskih algoritama prikazana je u rješavanju problema optimizacije putanje alata. Također je prikazano jedno od mogućih rješenja nadzora stanja alata korištenjem umjetnih neuronskih mreža i tehnike spajanja signala senzora.

Sažetak (engleski)

The theme of this work is application of artificial intelligence in optimization of manufacturing systems work. Modern manufacturing is based on application of techniques whose objective is improvement of current capability of manufacturing systems. Many researchers conducted in the World are based on possibilities of implementation artificial neural networks in manufacturing supervision systems. Application of genetic algorithms is possible in all areas, where is system optimization needed; in which conventional optimization methods doesn´t provide satisfactory result. Genetic algorithms have property of function global extreme searching, which in some conventional optimization methods isn´t possible. Base of this work is comparison of conventional optimization techniques and optimization techniques based on artificial intelligence, and their implementation in manufacturing systems. In second part of this work, mathematical models of manufacturing processes are given as fundamental models on which are based conventional optimization methods. In part three are presented and described conventional approaches and optimization results of some manufacturing processes. Part three is meant to provide possibility to compare conventional optimization methods with optimization methods based on artificial intelligence. In part four is presented tool path optimization model based on genetic algorithms. Given that, the tool positioning time in some manufacturing processes is great, optimization enabled reducing of needed tool positioning time, and with that faster production. Part five represents concept of intelligent manufacturing system, with representation of all vital functioning parts of manufacturing system, and are given bases of manufacturing system automatization control. In part six are given research results obtained by application of artificial neural networks in tool wear monitoring systems, and in chapter seven is conclusion of this work.

• Ključne riječi: nekonvencionalne tehnike optimizacije; inteligentni obradni sustavi; umjetne neuronske mreže; genetski algoritmi
• Ključne riječi (engleski): artificial intelligence; optimization techniques; artificial neural networks; genetic algorithms; ant colony optimization; particle swarm optimization
• Jezik: hrvatski
• URN:NBN: urn:nbn:hr:235:966674
• Studijski program: Naziv: Strojarstvo; Vrsta studija: sveučilišni; Stupanj studija: diplomski; Akademski / stručni naziv: magistar/magistra inženjer/inženjerka strojarstva (mag. ing. mech.)
• Vrsta resursa: Tekst
• Način izrade datoteke: Izvorno digitalna
• Prava pristupa: Pristup korisnicima matične ustanove
• Datum i vrijeme pohrane: 2019-05-21 20:39:52