Development of an inteligent system for modeling of laser deposition of material

Lasersko nanašanje materialov predstavlja sodobno dodajalno tehnologijo, ki ima v primerjavi s preostalimi postopki nanašanja kovinskih materialov številne prednosti. Poleg minimalnega vnosa energije, kvalitetnega spoja in majhnega toplotno vplivnega območja to tehnologijo odlikuje še dobra mehanska trdnost nanesenega materiala, ki je posledica hitrega ohlajanja. Kljub perspektivnosti tehnologije pa je ta še vedno v razvojni fazi. Vpeljujejo se novi materiali ter tehnike za določitev optimalnih procesnih parametrov. V disertaciji je prikazan postopek izdelave empiričnih modelov, s katerimi lahko na podlagi uporabljenih procesnih parametrov stroja napovemo lastnosti nanesenega materiala. Za izdelavo modelov je bilo uporabljeno genetsko programiranje ter regresijska analiza, ki za iskanje regresijskih koeficientov uporablja genetske algoritme. Na podlagi eksperimentalnih podatkov sta bila izdelana modela za napovedovanje volumna in modela za napovedovanje hrapavosti. Verifikacija dobljenih modelov je bila izvedena z uporabo testne množice eksperimentalnih podatkov. Na osnovi dobljenih modelov sistem določi optimalne parametre nanašanja materiala s stališča hitrosti nanašanja, izkoristka materiala in hrapavosti površine. Določitev optimalnih procesnih parametrov stroja je bila izvedena z uporabo nedominiranega sortiranja. Rezultati ponujajo operaterju niz optimalnih nastavitev procesnih parametrov, kar omogoča izdelavo kakovostnih izdelkov.Laser deposition of materials represents a modern additive technology, which has a number of advantages over remaining technologies for depositing metallic materials. Besides a small energy input, a quality bond and minimal heat affected zone, this technology is also characterized by good mechanical properties of the deposited material, which is a result of rapid cooling. Despite the prospects, this technology is still in the developing phase. New materials are being introduced and techniques for determining optimal process parameters. In the dissertation presented empirical modelling procedure can be used for predicting the properties of the deposited material. Genetic programming and regression analysis, where regression coefficients are determined with a genetic algorithm, were used for modelling. Based on experimental data, models for predicting the volume and roughness of deposited material were made. Verification of the models was made with experimental data, which was not included in the modelling process. Optimal process parameters are chosen in terms of deposition speed, powder efficiency and surface roughness. The optimal process parameters were determined with non-dominated sorting. The results offer the operator of the machine a set of optimal process parameters, which enables the production of high quality products