Research On Thermal Error Detection And Modeling Of CNC Machine Tools

During the CNC machine tools' machining process, the machine tools' components may produce thermal deformation under the influence of heat sources such as motor, friction, coolant and ambient temperature, etc., which leads to the change of the relative position between tool and workpiece, therefore causes an error which is defined as thermal error. Many studies proved that thermal error accounts for 40% to 70% of the machine tools'total error, and is the main error source during precision and ultra precision machining process, so it is of great importance to improve the machine tools' accuracy through reducing thermal error. Aiming at thermal error detection, thermal error sensitivity points selecting, thermal error modeling and other issues, this dissertation did some studies from the following aspects:1) The background and significance of the subject was described. The current research and problems about thermal error detection, modeling and compensation technology at home and abroad was introduced at detail, and the main focus of this project was illuminated.2) A multi-channel temperature measuring system for thermal error detection of CNC machine tools was built based on virtual instrument platform LabVIEW. The system can access PCI8310 data acquisition card by calling the dynamic link library file, realize the communication between the LabVIEW and the Excel through using ActiveX module, and improve its anti-jading performance by integrated hardware filtering and software filtering, therefore the measurement and storage of temperature data of CNC machine tools was achieved.3) The selecting of thermal error sensitivity points is a grey problem, which was solved by grey relational analysis method based on grey system theory, and was validated by experiments conducted on a typical machining center. The grey relational analysis method can effectively reduce the thermal error measuring points. Consequently, the complexity of thermal error model is reduced and its real-time performance and precision are improved.4) Based on the analysis of characteristics of thermal error of CNC machine tools, a thermal error model was built using the dynamic fuzzy neural networks, which the network structure was improved. To avoid redundancy of fuzzy rules and enhance the model's prediction accuracy and robustness, a dynamic generation method for fuzzy rule was adopted in the learning algorithm of the dynamic fuzzy neural networks. 5) Taking the temperature data measured by thermal error sensitivity as input, aiming at predicting the positioning accuracy change of machine tools, a thermal error prediction model was built based on the dynamic fuzzy neural networks, and its performance was verified by experiments and comparative analysis. Taking the temperature measuring problem as a grey issue, it was analyzed by the grey system theory. The grey relational analysis theory and its implementation flow were introduced. Experiment showed that the grey relational analysis theory can effectively select the thermal error sensitivity points, therefore solve the problem of existing excessive number of temperature sensors during thermal error measuring process.