18777 KB323 KB0 KB2012Balič, JožeBrezočnik, MiranKlančnik, SimonXIII, 188 f., 30 cmCOBISSID:15941398URN:URN:NBN:SI:doc-9T11ISZ6slS. Klančnikvisokošolska delacomputer simulationDisertacijegenetic algorithmsgenetski algoritmiinteligentni obdelovalni sistemintelligent machining systemneural networksnevronske mrežeObdelovalni strojioptimizacija z rojem delcevparticle swarm optimizationProgramiranjeskupna inteligencaswarm intelligenceModel inteligentnega CAD/CAM sistema za programiranje CNC obdelovalnih strojev| doktorska disertacija|Modern machine systems are highly automated, requiring a lot of flexibility. They strive toward complete autonomy. Since the programming of machine tools is a very complex process that consists out of many, among each other, dependant problems, despite tremendous efforts, it has not been able to automate it yet. A review of researches has shown that current developed systems are very narrowly limited and can be used just as an accessory, when machining procedure is being prepared. In this dissertation automated programming of CNC-machine tools with the help of artificial intelligence was suggested. The developed intelligence is not only partially, but entirely ableto solve a complex problem of automated programming of machine tools. Based on a CAD-model of the product, the system, without any help of an expert, automatically prepares a NC-program, so that the machining is safe, accurate and time efficient. In the same time certain technological requirements are also fulfilled. In the developed CAD/CAM- system swarm intelligence, NSGA-II multi-objective optimization and feed-forward neural network are used. In the same time the system exploits benefits and power of informatization, so that with the distributed computing architecture, it achieves greater efficiency in solving such a complex problem completely. The system consists out of prediction and evaluation module. In a prediction module the artificial intelligence suggests solutions which include information about tool path, selected tools and suggested cutting parameters. The evaluation module estimates the suggested solutions considering geometrical, technological and time criterion and the criterion of efficiency of machining. A discrete and continuous simulation model was developed for searching optimal solutions. The proposed system is in general suitable for different types of machining. Because of the scope of dissertation, our testing of the system is focused only on milling. The test results have confirmed that with the help of the method of artificial intelligence, machine tools can be automatically programmedSodobni obdelovalni sistemi so visoko avtomatizirani, zahtevajo veliko fleksibilnost in težijo k popolni avtonomnosti. Ker je programiranje obdelovalnih strojev zelo kompleksen proces, ki ga sestavlja več med seboj odvisnih problemov, ga kljub velikim naporom do danes še ni uspelo avtomatizirati. Pregled raziskav je pokazal, da so do sedaj razviti sistemi zelo ozko omejeni in lahko strokovnjaku služijo le kot pripomoček, pri pripravi postopka obdelave. V disertaciji predlagamo samodejno programiranje CNC-obdelovalnih strojev s pomočjo umetne inteligence. Razvita inteligenca je sposobna, ne le delno, ampak v celoti, reševati kompleksen problem samodejnegaprogramiranja obdelovalnih strojev. Sistem na podlagi CAD-modela izdelka samodejno, brez pomoči strokovnjaka, pripravi NC-program obdelave, in sicer tako, da je obdelava varna, pravilna, časovno učinkovita in hkrati zadosti določenim tehnološkim zahtevam obdelave. Inteligentni CAD/CAM-sistem za svoje delovanje uporablja skupinsko inteligenco, NSGA-II večkriterijsko optimizacijo in usmerjeno nevronsko mrežo, hkrati pa koristi prednosti ter močinformatizacije in tako s porazdeljeno arhitekturo dosega večjo učinkovitost pri celovitem reševanju tako kompleksnega problema. Sistem je sestavljen iz napovedovalnega in evalvacijskega modula. V napovedovalnem modulu umetna inteligenca predlaga rešitve, ki vsebujejo informacije o trajektorijah rezov, izbranih orodjih in predlaganih rezalnih parametrih. Evalvacijski modul, na podlagi razvitih simulacijskih modelov, oceni predlagane rešitve glede na geometrijski, tehnološki in časovni kriterij ter kriterij učinkovitosti obdelave. V okviru raziskave smo razvili diskreten in tudi zvezen simulacijski model, ki ga razviti sistem uporablja pri iskanju optimalne rešitve. Predlagani sistem je v splošnem primeren za različne vrste obdelav, v disertaciji pa se zaradi obsega dela pri testiranjih omejimo zgolj na rezkanje. Rezultati testiranj so potrdili, da je z uporabo metod umetne inteligence mogoče samodejno programirati obdelovalne strojeTEXTvisokošolska delatheses and dissertationsSlovenian National E-content AggregatorNational and University Library of SloveniaUniverza v Mariboru, Fakulteta za strojništvo