Composite error prediction of multistage machining processes based on error transfer mechanism

Abstract

Nowadays, modern manufacturing enterprises are paying more and more attention to machining process quality control in order to ensure the product quality. Since machining quality fluctuation is performed as composite error, then error control is becoming the key point in product quality assurance. However, as a prevention method, error prediction is not used effectively. In this paper, a new method of composite error prediction based on error transfer mechanism is proposed to help control the error sources (man, machine, material, method, measurement and environment, namely 5M1E) in multistage machining processes in order to improve the product quality. Firstly, the formation process of quality fluctuation is introduced, and the quality fluctuation network is established. Secondly, after two kinds of error are defined, the single process independent error formation process and the multistage error transfer mechanism are then analyzed deeply. Thirdly, the single process independent error prediction model is established by using the LS-SVM method and error separation principle, according to which, the composite error of multistage processes is predicted based on error transfer function. Finally, an example of the real specific machining process is given to illustrate the effectiveness and correctness of this methodology.