Optimization of Parameters in GM(1,1) Model by Taguchi Method

• Main Authors: Chen Chien Cheng, 陳建成
• Other Authors: 紀岍宇
• Format: Others
• Language: zh-TW
• Published: 2009
• Online Access: http://ndltd.ncl.edu.tw/handle/88994846994551998104

碩士 === 清雲科技大學 === 機械工程系所 === 97 === This paper uses Taguchi Method and GM (1,1) to model electric spark machining and to find (copper) the electrode waste degree by changing the background weight Alpha value of GM(1,1). In order to get a set of best working condition, the S/N ratio of Taguchi Method L25 (5 6) and S/N ratio of L81 (9 6) combine with GM (1,1) are used to forecast the electrode waste. The important factors in electric spark machining are material conductivity (1.215), electric discharge potency (0.71), gap between work piece (0.635), electrode working area (0.62), working depth (0.58), medium fluid (0.62) and process period length (0.68). The results show the best set of factors combination is A3 (0.5), B3 (0.5), C2 (0.3), D2 (0.3), E5 (0.9) and F3 (0.5) for S/N ratio equal to 35.6058, The difference between the best choice and α=0.5 is very tiny equal to 0.0136. Moreover simplifies into L81 (9 6), subdivides the weight value for (0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9), the results show the best set of factor combination is A3 (0.3), B2 (0.2), C6 (0.6), D6 (0.6), E6 (0.6), F4 (0.4) for S/N ratio equal to 35.6955. The result difference with L81 is not to be large again in 0.0897. Therefore so long as the suggestion with L25 (5 6) can achieve the optimization design.