Effective control of optimum dressing according to gap increase by using multilayered neural networks

Abstract

Dressing of superabrasive wheels capable of producing a good mirror finish on brittle materials is a current requirement. A neural identifier and a neural controller for optimum control of electro-discharge dressing systems are proposed for this purpose. The modelling of the system and an actual plant control system for mirror-like grinding is obtained from a neural identifier and a neural control structure giving satisfactory stability is proposed. The results of this study using multilayered neural networks show that the proposed neural identifier not only gives accurate results but can also find the relationship parameters for the electro-discharge dressing system. Additionally, the proposed neural controller gives very effective control by gap increase using a learning process in spite of the nonlinear characteristics of the electro-discharge conditions.