Selection of laser bending process parameters for maximal deformation angle through neural network and teaching-learning-based optimization algorithm

作     者：Omidvar, Mahyar Fard, Reza Kashiry Sohrabpoor, Hamed Teimouri, Reza 

作者机构：Golpayegan Shohada Univ Engn & Technol Golpayegan Iran Islamic Azad Univ Kordestan Branch Sci & Res Dept Mech Engn Sanandaj Iran Islamic Azad Univ Dezfull Branch Dept Mech Engn Dezful Iran Babol Univ Technol Dept Mech Engn Babol Sar Iran 

出 版 物：《SOFT COMPUTING》 (Soft Comput.)

年 卷 期：2015年第19卷第3期

页      面：609-620页

核心收录：

学科分类：08[工学] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

主　　题：Laser bending L-25 Taguchi design Radial basis network Teaching-learning-based optimization algorithm 

摘      要：The present study focused on selecting optimal factors combination that causes maximum bending angle in laser bending of AA6061-T6. For this purpose a L-25 Taguchi orthogonal design (four factors-five levels) is used to design experiments. Here, the process main factors are laser power, spot diameter, pulse duration and scanning speed and the main response was bending angle. To correlate relationship between process factors and bending angle, a radial basis function neural network (RBFNN) was utilized. Then the developed RBFNN model was used as an objective function for maximizing deformation through teaching-learning-based optimization algorithm. Results indicated that the laser power of 3.6kW, spot diameter of 2 mm, pulse duration of 0.9 ms and scanning speed of 2 mm/s lead to maximal bending angle about 28.7 degrees. Hereafter the optimal results have been verified by confirmatory experiments.