基于GA-BP神经网络板材辊式矫直工艺预测模型  

Prediction model of plate roller straightening process based on GA-BP neural network

作　　者：王敬龙 朱晓宇 王效岗[1] WANG Jinglong;ZHU Xiaoyu;WANG Xiaogang(Engineering Research Center Heavy Machinery Ministry of Education,Taiyuan University of Science and Technology,Taiyuan 030024,China)

机构地区：[1]太原科技大学重型机械教育部工程研究中心,太原030024

出　　处：《现代制造工程》2023年第8期115-120,共6页Modern Manufacturing Engineering

基　　金：太原科技大学研究生教育创新项目(BY2022007)。

摘　　要：辊式矫直工艺是轧制生产线上必要精整工艺。为提升生产线的整体智能化生产需求,采用神经网络代替曲率积分矫直模型进行计算,解决其求解难、耗时长和不收敛的缺点。针对反向传播(Back Propagation,BP)神经网络易出现泛化能力弱、陷入局部最优等问题,引入遗传算法(Genetic Algorithm,GA),建立一种基于GA-BP神经网络算法的板材辊式矫直工艺神经网络多输入多输出计算模型。对比结果显示,选用trainscg函数可实现较好的预测结果,并通过贪婪策略对模型结构进行优化,实现了矫直工艺模型的快捷、高精度计算,首尾辊压下误差在0.2 mm以内,残余曲率比误差在5%以内,矫直力误差在7%以内。该神经网络模型对轧制生产线有较高的工程应用价值。The roll leveler is used as a necessary finishing process on the rolling production line.In order to improve the overall intelligent production requirements of the production line,the neural network was used to replace the calculation of the curvature integral straightening model to solve its shortcomings of difficult solution,long time consumption and non-convergence.In view of the weak generalization ability of Back Propagation(BP)neural network and the problem of falling into local optimum,a Genetic Algorithm(GA)was introduced to establish a multi-input and multi-output calculation model of plate straightening process neural network based on GA-BP neural network algorithm.The comparison results show that the selection of the trainscg function can achieve better prediction results,and through the optimization of the model structure through the greedy strategy,the fast and highprecision calculation of the straightening process model was realized,and the error of the first and last roll reduction is within 0.2 mm,and residual curvature ratio is within 5%,the straightening force error is within 7%.The neural network model has high engineering application value to the straightening production line.

关 键 词：矫直机 曲率积分模型 遗传算法 反向传播神经网络 矫直力 

分 类 号：TG333.23[金属学及工艺—金属压力加工] TP183[自动化与计算机技术—控制理论与控制工程]