Research On Theory And Application Of Thermal Error Robustness Modeling For CNC Machine Tools

Accuracy is an important performance indicator for high-end CNC machines.The thermal error is caused by the extra offset between the tool and the workpiece resulting from the thermal deformation of the parts during the machine tool processing.It accounts for 40%?70% of the total error of the CNC machine tool,which has seriously affected the accuracy of machine tool processing.Currently,the thermal error compensation technology is t he most effective way to reduce the thermal error of the machine tool.It is necessary to first measure the temperature and thermal error of the machine tool at multiple points synchronously,and then select the points with the highest weight on the therma l error according to the measured data,namely,temperature sensitive point,and establish the mathematical model between the temperature sensitive point and the thermal error.The model is used to predict the thermal error through the machine tool temperature,and feed it back to the numerical control system of the machine tool.The servo system controls the tool for reverse displacement to counteract the offset caused by the thermal error,thereby reducing the influence of the thermal error.The key to th ermal error compensation lies in the accuracy of thermal error prediction.This paper improves the accuracy and robustness of thermal error model from the perspective of mathematics and engineering application.From the mathematical point of view,in addit ion to the algorithm of the thermal error modeling,the selection of temperature sensitive points is also crucial.At present,the commonly used algorithm for temperature sensitive point selection is basically classification and optimization.The core idea of this method is to reduce the collinearity between temperature sensitive points.However,based on the observation of long-term thermal error experiments,it is found that reducing the collinearity will inevitably lead to a decrease in the correlation between temperature sensitive points and thermal errors.The model is susceptible to the interference of changes in external factors,which is not conducive to the long-term maintenance of the prediction accuracy of the model and will lead to poor robustness.Therefore,this paper studies the influence mechanism of collinearity on modeling error,and proposes the use of improved algorithm to suppress the influence of collinearity,thereby using strong correlation temperature sensitive points for modeling and improving the robustness of prediction accuracy of thermal error model from the mathematical perspective.For engineering applications,first of all,it is noted that the current thermal error measurement uses the "Five-point Method" proposed in the international standard [1] " Test code for machine tools-Part 3 Determination of thermal effect"(ISO 230-3:2007 IDT).This method requires the removal of the tool during the measurement,so the thermal error can only be measured when the machine tool is in an idle state.When the machine tool is in a cutting state,it will be affected by coolant,cutting force and other additional factors.This paper proposes an online thermal error detection method which successfully measures the thermal error in the real-cut state and shows a big difference with the characteristics of the idling state.In addition,considering that there are many factors influencing the thermal error in the real cutting state,in order to improve the modeling effect under real cutting condit ions,the Taguchi orthogonal test is used to investigate the influence of four factors,namely,spindle speed,feed speed,cutting depth and ambient temperature,on thermal error,and the optimal combination of parameters is selected to further improve the modeling accuracy and robustness under real cutting conditions.Secondly,when the thermal error is measured using the above-mentioned “five-point method”,the measuring fixture is fixed on the workbench,and only the fixed single-point thermal error on the workbench can be modeled and predicted.Therefore,in this study,the online detection system is used to further improve and realize the rapid measurement of the multi-point thermal error on the workbench,and the surface fitting and interpolation algo rithm is combined to establish a comprehensive model that can predict the thermal error of the whole workbench according to the position coordinates of the workbench.In addition,the thermal error compensation device is improved to read the position coordinates of the workbench through the airport numerical control system while acquiring the temperature of the temperature sensitive points,thereby introducing the compensation of the thermal error of the whole worktable by the model.Finally,this paper proposes a thermal error detection method based on actual cutting.The machine tool to be tested is used to process test pieces of specific specifications through a specific process,thereby mapping the thermal error in the final machining dimensional accurac y,and testing and evaluating the test pieces.The advantage of this method is that it can reflect the thermal error in the real processing state.Furthermore,for manufacturers or users without testing equipments,the processed test pieces can be sent to the testing organization for testing,which does not require the procurement of testing equipments and broadens engineering applicability.