- 學位論文
應用多元迴歸分析建立溫升熱補償技術－以車床的工件主軸為例
Application of Multiple Regression to Thermal Errors Compensation Technology－Experiment on Workpiece Spindle of Lathe
摘要
本研究為智能化工具機之ㄧ的溫升熱補償技術，此技術在日德等先進工具機大廠的機台上已經普遍應用，而國內工具機廠真正將此技術商品化的依舊稀少，大部分尚在實驗階段。 車床的加工特性為機頭的工件主軸夾持工件，藉著圓柱型工件的旋轉產生動力，以固定的刀具車削加工。由於變動的速度及切削負載發生的頻率較高，當面對連續長時間的加工而產生熱變形，會影響刀具與工件之間相對位置的準確度，此等微小的變化會損失工件的加工精度。故本研究於車床上量測溫度及工件主軸的熱變形情況，將實驗所得到的數據作統計分析，以多元迴歸分析的方法，分析熱源與變形的關係，依此建立車床三個正交軸方向的熱變形模型。並藉由數據的分析與再利用，於模型中不斷地加入自變數的多次項、交乘項及落後項等檢定模型，最後再依據模型預測能力的優劣，選擇最佳的熱變形模型。本研究成功地建立熱變形模型之預測誤差的標準差約在5 μm內。
並列摘要
This study is one of the intelligent machine tools technology–Thermal Errors Compensation Technology. This technology has been used widely in advanced machine tools manufacturers such as Japan, Germany etc. However, there’re only few domestic factories that commercialize this technology, most of them are still at the experimental stage. The lathe turns cylindrical workpiece with fixed tools. Through the rotating workpiece, and manufactures with fixed cutting tools. Due to the high changing speed and the frequency of the cutting load, long working hours results in heat deforming and affects the accuracy of the relative position between the tool and the workpiece. These small changes can damage the machining precision. Therefore, this study measures the temperature in the lathe, and the thermal errors of the workpiece spindle. Then we gathered the experimental statistics and analysis, by the multiple regression to analyze the relationship between the heat source and deformation, and made the three axis thermal errors model. Also, through analyzing the data, we keep adding items from polynomial, interaction and cyclical series in the model. Finally, based on the merits of the predictive ability of the model, we choose the best thermal errors model. This study successfully made thermal errors model standard deviation of the forecast error within 5 μm.