基于BP神经网络和遗传算法的TWIP钢热处理工艺参数优化  

Optimization of Heat Treatment Process Parameters of TWIP Steel Based on BP Neural Network and Genetic Algorithm

作　　者：童希 王荣吉[1] 王亚祥 俞杰 邹伟 TONG Xi;WANG Rongji;WANG Yaxiang;YU Jie;ZOU Wei(School of Mechanical and Electrical Engineering,Central South University of Forestry and Technology,China)

机构地区：[1]中南林业科技大学机电工程学院

出　　处：《热加工工艺》2018年第16期176-179,共4页Hot Working Technology

基　　金：湖南省教育厅科学研究重点项目(14A157)

摘　　要：为了提高Fe-Mn-C-Al系TWIP钢的力学性能,采用BP神经网络与遗传算法对热处理工艺参数优化。以3个热处理工艺参数为优化对象,以抗拉强度与伸长率之积的强塑积作为优化目标,建立3-4-1的BP神经网络的非线性映射模型,再通过遗传算法的全局寻优,得到具有最优强塑积的热处理工艺参数的最优配置组合。预测结果表明,其最优强塑积热处理工艺为：退火温度为863℃、保温时间为26 min、冷却方式为炉冷,并通过试验验证了预测结果的准确性。In order to improve the mechanical properties of Fe-Mn-C-Al TWIP, the optimization of heat treatment process steel based on BP neural network and genetic algorithm was carried out. Taking three heat treatment process parameters as the optimized objective and the product of tensile strength and elongation as optimized target, the 3-4-1 nonlinear mapping model of BP neural network model was established. The optimal composition of heat treatment process parameters was obtained by the genetic algorithm. The prediction results show that the optimal composition of heat treatment process parameters with best product of tensile strength and elongation was annealing temperature of 863 ℃, holding time of 26 min and furnace cooling method. And the correctness of prediction results was verified by experiments.

关 键 词：TWIP钢 BP神经网络 遗传算法 热处理工艺 参数优化 

分 类 号：TG156.2[金属学及工艺—热处理] TG161[金属学及工艺—金属学]