Selective Spectral Coding for TPS selection in Thermal Error Analysis

Abstract

Machine tools are crucial in the manufacturing industry for producing products with high quality and accuracy while reducing expenses and production time. One significant problem affecting precision machining accuracy is thermal inaccuracy. It is mostly produced due to thermal expansion and contraction in machine while it is operating, particularly thermal deformation, which considerably influences the end product's accurateness. Currently, thermal deformation is predicted using mathematical models, and real-time compensation is achieved by Computer Numerical Control (CNC) systems. Temperature Sensitive Points (TSP) must be identified in order to create precise thermal fault forecast models. Due to the nonlinear allocation of high temperature sources in machine equipment, the present approach employs simulations by means of the limited element means and cluster analysis to identify temperature locations that are closely associated with thermal errors. However, the dynamic adjustment of thermal properties leads to a numerical control error that damages the tool. To improve the controlling system's accuracy, a selective learning approach for temperature analysis is suggested.