| Summary: | 碩士 === 國立高雄第一科技大學 === 電機工程研究所碩士班 === 106 === CNC machine tools are indispensable tools for precision machining in today's industry. Machining accuracy is a key indicator of their performance, and machining accuracy is affected by various factors. The thermal deformation error caused by temperature rise is the most serious, and the error may even be as high as 100 microns. Therefore estimation and compensation of thermal deformation errors of machine tools is a problem that cannot be ignored.
For the prediction of thermal deformation error, this study believes that there is a considerable deviation in directly estimating the thermal deformation error based on the local and a few of measured temperatures. In order to reduce this prediction error, the practice of this study is to establish a heating source model, a temperature rise model, and then establish a thermal deformation model. That is, according to the working instructions of the machine and the dynamic model of the system, the heat sources generated by the motors, bearings, slide rails, bushings, machining, etc. are estimated, and then these heat sources are injected into the thermal temperature rise model to grasp the temperature of the complete parts. Then, according to the temperature rise of the complete parts to estimate the thermal deformation error. In order to grasp the temperature rise of the complete parts, this study divides the whole machine into institutional parts according to the mechanism, and then divides the institutional parts into blocks. Each block is regarded as an isothermal body. The thermal temperature rise model regulates the heat transfer relationship between those parts and blocks and the temperature rise dynamic behavior with the heat capacity heat source and heat dissipation of each block source and heat dissipation. A few measured temperatures are used to calibrate the parameters of the temperature rise model and the heating source model. According to this, the temperatures of the completely distributed locations of the machine can be obtained. Based on those completely distributed temperatures artificial neural network was used to establish a thermal deformation model.
|
|---|
