Application of a Bayesian Network to Thermal Error Modeling and Analysis for Machine Tool

Abstract

CNC machine tool dynamic thermal error compensation has always been a hot issue to improving precision. This dissertation proposes a method of machine tool thermal error modeling during processing, based on Bayesian network theory, by describing the correlation between the various factors of generated the heat error, through the sample data, analyzed and simplified the intrinsic correlation between these various factors, established the basic thermal error compensation model, and used the network’s good characteristic of self-studying, combining the result of update collection data, continually modify the model to reflect the machining process condition changes. Finally, the experimental results show the feasibility of Bayesian network model, it was a stronger application for achieving the thermal error compensation.