反覆學習控制於工具機空間誤差補償之應用

Application of an Iterative Learning Control Algorithm to Volumetric Error Compensation for CNC Machines

盧盈辰(Ying-Chen Lu)

指導教授 ： 葉賜旭

國立臺北科技大學/機電學院/機電整合研究所/碩士(2014年)

https://doi.org/10.6841/NTUT.2014.00819

摘要

工具機在加工時，會產生的空間誤差共有21項，在現今工具機的補償辦法中，空間誤差的補償僅補了 、 、 ，因此成效有限，又若使用了市面上的空間誤差補償工具，一套設備則要價不斐。因此本研究透過選用反覆學習控制補償軟體的方式來降低成本並使這套軟體的適應性增大，達到可在各式機台中使用並價格便宜的作用。在補償過程中，先是透過了初步的軟體補償將節距誤差由誤差表格的內插法計算得到並進行加工命令補償，而後再利用反覆學習控制在有限的次數內找出最小空間誤差之補償指令。透過模擬一與模擬二來證實模擬軟體之效用；其中模擬一的空間誤差收斂度最高可達88.8%，而模擬二模擬螺旋路徑運動情形，空間誤差收斂度達82.6%，皆是相當不錯的數值，接著透過實際切削的三種實驗進行空間誤差軟體補償驗證，實驗三證實補償成效至少可補正12.99%的誤差，至多可補正98.9%的誤差，且誤差變化情形與量測所得的空間誤差表一致，因此可知本研究之反覆學習控制補償軟體應用在空間誤差上有顯著的效果。

關鍵字

工具機 ； 空間誤差 ； 軟體補償 ； 反覆學習控制

並列摘要

There are many factors, such as straightness of guild way, flatness, pitch error, backlash of three axes, to affect the machining accuracy of CNC machines. Nowadays, CNC machines are only compensated three errors : 、 、 by the manufacturers ignoring other errors causing inefficient compensation. An equipment using for volumetric error compensation on CNC machines usually cost a lot. For improving those defects, the study use iterative learning control software compensation to increase the adaptability on CNC machines and low cost. The process of compensation would get the pitch errors from error tables by using interpolation rule first, and compensate the processing commands. After first step’s compensation, use the iterative learning control to find the compensating commands in the limited running times that would cause the minimum volumetric errors. Setting two simulations test the effect of the compensation software. The simulation 1 is testing about the influence of points, and the convergence of the volumetric errors is 88.8% compared with the original. Simulation 2 is imitating the spiral moving process by using the compensation software, and the convergence of the volumetric errors is 82.6%. After knowing the result of simulation, try three kinds of cutting tests and analyze the variation of the data. The cutting test 3 shows that the volumetric error compensation software is working and at least makes the error compensate 12.99%, at most makes the error compensate 98.9%. Results of those simulations and cutting tests indicate that the volumetric errors can be significantly reduced by applying the compensation software.

並列關鍵字

CNC machine tools ； volumetric error ； software compensation ； iterative learning control

參考文獻

[5] 許富銓、吳金舫、呂育廷、莊殷、林英傑，工具機靜態空間誤差與動態循圓量測之應用研究，金屬工業發展中心，高雄，台灣，2009。

[1] “Methods for Performance Evaluation of CNS Machine Center,” ANSI/ASME B5.54, 2005.

[2] ISO230-2, “Test code for machine tools Part 2: Determination of accuracy and repeatability of positioning numerically controlled axes”, 1997.

[3] “Test Code for Machine Tools Part 4: Circular Tests for Numerically Controlled Machined Tools,” ISO 230-4, 1996.

[6] K. G. Ahn, D. W. Cho, “Analysis of the volumetric error uncertainty of a three-axis machine tool by beta distribution,” International Journal of Machine Tools and Manufacture, vol. 40, pp. 2235-2248, Dec. 2000.