Research On The Robustness Of Thermal Error Model On CNC Machine Tools

CNC machine tools as the tool of processing a variety of products, more and more high machining accuracy is required. And thermal error is the main factor that affect the machining accuracy of CNC machine tools, it gets more and more attention of the researchers. Researchers have proposed a variety of thermal error models on CNC machine tools, but the research on the robustness of thermal error model is relatively less. For thermal error model, its robustness is that the thermal error model can still meet the ability of a certain prediction accuracy when the external environment, working conditions, instrument measurement precision and other factors changes. The better robustness, thermal error model can be accepted, and widely used. So, it is necessary to put forward a method to analysis and evaluate the robustness of thermal error model, for obtaining its robustness.This paper research the robustness of the commonly used thermal error models on CNC machine tools. Main research work in this paper is as follows:(1) researches on robustness for thermal error model on CNC machine tools, put forward some robustness evaluation performance of thermal error model including anti-noise performance and ability to adapt emergency.(2) chooses RBF neural network and Least Squares Support Vector Machine(LS-SVM) establish a thermal error model respectively, test experiment was carried out on the anti-noise ability and ability to adapt emergency of thermal error model.(3) studies the basic theory of cloud model, based on analyses of cloud model dealing with uncertainty problem, the cloud model theory is introduced into the robustness evaluation mechanism of thermal error model, proposed robustness evaluation method of the thermal error model based on cloud model. Through experiments to prove that the evaluation method of its robustness for thermal error model is effective. This research has a certain theoretical significance and use value.