Neural Network Based Experimental Design Method and an Industrial Application

Authors: Ercan ÖZTEMEL, M. Emin AYDIN

Abstract: It requires a great deal of expertise to satisfy customers under rapid technological changes in a highly competitive market. Quality should be taken into account not only in production lines but also during the design phase of the process. Especially, development of sophisticated CNC (Computer Numeric Controlled) machines makes the products produced comply with design specifications with minimum or near zero variation. Therefore the quality of the products is mostly effected by the specifications defined during the design phase. In this paper, a new method which can be used to achieve the best quality in defining the design specifications is described. The method is based on neural network computing technology.

Keywords: Taguchi Method, Artificial Neural Network, Quality Control, Multilayer Perceptron, Experimental Design, Fractional Factorial Design