سال انتشار: ۱۳۸۶

محل انتشار: دومین کنفرانس بین المللی و هشتمین کنفرانس ملی مهندسی ساخت و تولید

تعداد صفحات: ۸

نویسنده(ها):

Mojtaba Ghaderi – M.Sc. graduate of mechanical engineering, khajeh Nasir Toosi university of technology
Mehrdad Kazerooni – Assistant Professor mechanical engineering, khajeh Nasir Toosi university of technology
Mehdi Zohali – M.Sc. graduate of mechanical engineering, Iran university of science & technology

چکیده:

In this paper, a method of plastic injection parameters optimization to reduce warpage has been presented. A preform part of mineral water bottle which is made of PET is studied. Mold temperature, melt temperature, packing pressure, packing pressure time and filling time are effective process parameters.The base of this method is Computer Aided Engineering (CAE) and a predictive model for warpage is developed by using multi layer perseptron (MLP) artificial neural network by exploiting finite element analysis results. The training data are generated by using Moldflow simulation software. A total of 247 data were collected out of which 237 were used to train the neural network and the remaining to validate neural network. Then artificial neural network is interfaced with a genetic algorithm to achieve the optimum process parameter values. In fact, this method is Interaction of finite element software, artificial neural network program and genetic algorithm to optimize and improve desired properties. In this optimization, due to mass production, in addition of quality factor, commercial benefits are considered and time is added in Cost Function.