The Development on the Predictive Model for the Metal Machining Parameters Based on the Artificial Neural Network and Genetic Algorithm

• Main Authors: Yao, Cheng-Fu, 姚成福
• Other Authors: Chien Wen-Tung
• Format: Others
• Language: zh-TW
• Published: 1997
• Online Access: http://ndltd.ncl.edu.tw/handle/05901737705546250483

碩士 === 國立屏東技術學院 === 機械工程技術研究所 === 85 === In this study the artificial neural network was used to build a predictive model for metal machining. The controlling variables are cutting speed, feed anddepth of cut, and the predictive results are the cutting forces and the surface roughness of the workpiece. There are 50 experimental tests were conducted. Forty of them were used to train the network, and other ten of the experimental data were used to validate the accuracy of the predictive model. The error of roughness is 9.4%. The error of the forces were 3.6% on feed force, 4.21% on thrust force and 2.67% on cutting force. In addition, the genetic algorithm was used to find the optimum cutting conditions for the maximum metal removal rate under the constraint of the expected surface roughness. It was found that the optimum cutting conditions are that the cutting speed and depth of cut are close to the upper boundary of the controlling variables, however, the feed varies as the expected surface roughness of the workpiece. Furthermore, the predictive model can be expanded to be a comprehensive data of metal machining by adding more controlling variables and training experimental data. Thus, a convenient tool on deciding the optimum cutting conditions for metal machining is achieved.