Partial Least Squares Regression Analysis of Thermal Distortion Behavior for Linear Motor Drive Feed System
In order to explore the effect factors of thermal distortion behavior of the high speed feeding system, a methodology with partial least squares linear regression (PLSR) was brought forward to analyze the thermal process on the basis of experiments. The feeding system test-bed and the thermal behavior measurement system were established for the experiments. The experiments were conducted under both the natural heating and force cooling condition to show the feasibility of the method. The experimental scheme and procedure were also introduced in detail. On the basis of the experiments and the regression calculation, the distribution characteristics of temperature for the high speed feeding drive system and the rule of the thermal distortion behavior were explored with further discussion. The results show that the proposed method can effectively recognize the experimental data and reach high distortion identification accuracy. The resultant factors and their proportions of thermal deformation were presented through regression calculation. The prediction of the distortion at some random axis positions was done with the regression model. The result shows good performance for recognition of the axis deformation. As a result, the study effort provides a theoretical foundation for optimum design and error compensation in machining process.