改进快速非支配排序遗传算法热连轧过程负荷分配的智能优化  

Intelligent optimization of load distribution in hot strip rolling process based on improved non-dominated sorting genetic algorithm

作　　者：丁敬国[1] 宋梦雪 张志鸿 郝学斌 任连波 魏子超 Ding Jingguo;Song Mengxue;Zhang Zhihong;Hao Xuebin;Ren Lianbo;Wei Zichao(State Key Laboratory of Rolling and Automation,Northeast University,Shenyang 110819,China;Handan Iron and Steel Group Company,Handan 056003,China)

机构地区：[1]东北大学轧制技术及连轧自动化国家重点实验室,沈阳110819 [2]河钢邯钢邯宝热轧厂,河北邯郸056003

出　　处：《材料与冶金学报》2022年第3期228-234,共7页Journal of Materials and Metallurgy

基　　金：中央引导地方科技发展资金(2022JH6/100100019);国家自然科学基金-区域创新发展联合基金(U21A20475)。

摘　　要：在热连轧的非稳态轧制过程中,负荷分配不均会导致板形控制精度的下降.针对负荷分配不合理的情况,以传统机理模型为基础,建立了板形和轧制均衡的目标函数,开发了改进快速非支配排序遗传算法(NSGA-Ⅱ)的负荷分配智能优化方法,利用该优化算法对热连轧精轧机组进行了负荷分配优化.通过工业试验将经验法、GA与NSGA-Ⅱ优化得出的负荷分配数据进行对比,发现采用新方法的优化结果明显比经验法更合理,GA和NSGA-Ⅱ的优化结果将机架F5和F6相对凸度误差从4.63%分别降低至1.06%和0.79%,且NSGA-Ⅱ在末机架的相对凸度差为-0.000 009 88,小于GA的-0.000 050 47.应用结果表明,该负荷分配优化算法明显优于GA,对非稳态轧制过程中的负荷分配优化及板形控制具有理论指导意义和实际应用价值.In the unsteady rolling process of hot continuous rolling, the uneven load distribution would lead to the decrease of the accuracy of shape control. Based on the traditional mechanism model, the objective function of strip shape and rolling load balance was established for the unreasonable load distribution, an intelligent optimization method of load distribution based on improved fast Non-dominated sorting genetic algorithm(NSGA-Ⅱ) was developed, which was used to optimize the load distribution of hot strip finishing mill. Comparing the load distribution data obtained by an empirical method, GA and NSGA-Ⅱ optimization with the load distribution data obtained by an empirical method through industrial tests, it is indicated be seen that the optimization results of the new method are more reasonable than that of the empirical method. The optimization results of GA and NSGA-Ⅱ reduce the relative convexity errors of F5 and F6 from 4.63% to 1.06% and 0.79%. Furthermore, the relative convexity difference of NSGA-Ⅱ in the final frame is-0.000 009 88, which is less than-0.000 050 47 of GA. The optimization result of load distribution is better than that of GA in terms of shape. The research result has theoretical guiding significance and practical application value for load distribution optimization and shape control in unsteady rolling process.

关 键 词：热连轧 负荷分配 NSGA-Ⅱ 多目标优化 

分 类 号：TG335.5[金属学及工艺—金属压力加工]