Modeling And Compensation Of Geometric And Thermal Errors For A Three-axis Ultra-precision Machine Tool

As the rapid development of modern industries, such as modern machine,semiconductor, electronic, measuring techniques and space technology, ultra-precisionprogress technology plays more and more important role in manufacturing field. As thebasic of ultra-precision progressing technology, ultra-precision equipments are highlyvalued in many industrialized countries, which becomes the criteria to evaluate the levelof manufacturing industries. At present, the research about the ultra-precision machinetool is mainly focused on the development of grind machine and lathe machine, while themilling machine, which has significant application in aspheric freeform surface andmicrostructural optical components, isn’t fully studied. The thesis mainly studies thegeometric and thermal errors of three-axis ultra-precision milling machine tool.In the second chapter, the Finite Element Method(FEM) is adopted to analyze thestatic and dynamic performance of two different machine tool configurations, includingstatic analysis, modal analysis and harmonic analysis. The better configuration is used todevelop the ultra-precision machine tool. In addition, the experiment measuring thenatural frequencies of the machine tool is carried out, whose results are applied tooptimize the NC system.Based on the multi-body theory, the model of the volumetric geometric errors isstudied. The errors in XYZ directions of the tool tip are divided into twelve errors. Thetwelve errors are linear function of position coordinate so that they can be computed in theworking space. The Linear coefficients of the error model are inputted into the PLCsystem to compensate the geometric errors.Due to the precision cooler system of the linear motor drive system, the heatgenerated by linear motors is blocked and not able to be conducted to the machine tool. Inthe fourth chapter, the thermal errors in axial direction are measured by laser displacementsensor. The model of the thermal error is developed by regression model and neuralnetwork respectively. The residual errors of these two models are compared to obtain thefinial model of the thermal error, which is embedded into PLC system to achieve real-timecompensation of the spindle thermal error.