car paint thickness control using artificial neural network and regression method

struggling in world's competitive markets, industries are attempting to upgrade their technologies aiming at improving the quality and minimizing the waste and cutting the price. industry tries to develop their technology in order to improve quality via proactive quality control. this paper studies the possible paint quality in order to reduce the defects through neural network techniques in auto industry production lines. the inputs as effective factor in paint spray process identified for each thin layer on a plate. in the paint shop, defects generate that correlate with film thickness in paint process. in this work, a sheet of metal in demonstrated 50*20 using as a sample when saipa company, iranian auto market leader, is considered as a case study. in the present paper two models of nn is presented. the first model shows prediction of film thickness by10 input for bell, air layers and 12 inputs variables for dry film thickness or final paint thickness and 6 output points for three layers and second model is predicting of paint appearing uniformity by average and standard deviation of film thickness. finally the application of neural network and statistical method (regression) in predicting paint thickness and the comparison of the results are presented.