An Intelligent Optimizing Approach For Spinning Production Based On Immune Neural Network Expert System

With the advantage of little investment and cost, large capacity polyester has been widely used around the word. The process optimization of spinning plays a great role for improving production quality and developing new productions.In this paper, it makes the production parameters as an input, the indicators of the key quality as the neural network's output. First of all, it establishes the neural network model of spinning process optimization, where it uses the radial basis function (Radial Basis Function, RBF). Secondly, it uses the immune algorithm to optimize the neural network mode. According to reproduction, cross and mutation operations, it makes the immune optimization algorithm as the objective function antigen, and the neural network weight as antibodies to obtain the optimal solution.After the establishment of the RBF neural network for spinning process optimization, it sent an output to the host computer with the spinning expert system. According to the knowledge base and inference engine of product parameters, the expert system will obtain a group of key quality indicators. It makes compare between the target and the capture value. According to the error, the expert system will make an adjustment of production parameters and explanation.The method is application to not only the short fibers but also to the long fiber and carbon fiber. Also it makes a guide for the new product development and process optimization, and presumes a new idea for the spinning process optimization.