基于改进免疫克隆算法的轧制规程多目标优化研究
发布时间：2018-11-11 14:37
【摘要】：在冷连轧的计算机控制系统中,轧制规程的设定是冷连轧生产控制参数设定的核心内容。合理的轧制规程可以使冷连轧制高效运行,达到更好的控制效果,并能降低生产能耗及提高产品质量。随着轧制技术的不断发展和品质要求的不断提高,基于单目标制定轧制规程已无法满足生产的需要,目前广泛采用的是多目标优化的轧制规程。本文根据生产的实际需要,研究了基于改进免疫克隆算法的轧制规程多目标优化方法。首先,分析了传统免疫克隆多目标进化算法的进化机制,并针对算法收敛速度过慢的问题,改进了算法的变异策略。利用差分变异及云模型方式对种群个体进行自适应的变异,使个体的变异更有针对性,增大了算法的搜索范围,提高了算法的收敛性能。其次,分析研究各种保持多目标个体分布性的保留裁剪机制,根据免疫克隆多目标算法中进化种群和记忆种群的特点及在优化过程中的作用,设计了混合的裁剪选择机制,分别对进化种群与记忆种群进行个体选择,防止种群中的个体过于密集的情况,提高了算法中个体的分布度。最后,针对轧制规程优化过程中多目标函数之间存在耦合,相互制约,难以选择目标侧重的问题,以唐山某钢厂五机架冷连轧生产线为例,构造了由等功率裕度、板形良好和防打滑目标函数组成的多目标负荷函数,并采用本文提出的改进算法对其进行多目标优化计算。试验结果表明,此算法能够很快的收敛到Pareto前沿,并且能够获得分布性良好的解集;优化后的不同偏好轧制规程组合可以满足不同的选择要求。并与原规程相比,各轧机的利用更加合理,提高了轧机的利用效率,改善带钢板形和表面质量,并且减少了划痕的产生概率,对二级设定计算起到了指导的作用。
[Abstract]:In the computer control system of cold continuous rolling, the setting of rolling schedule is the core of the control parameter setting of cold tandem rolling. Reasonable rolling schedule can make cold continuous rolling run efficiently, achieve better control effect, and reduce production energy consumption and improve product quality. With the continuous development of rolling technology and the continuous improvement of quality requirements, making rolling schedule based on single objective can no longer meet the needs of production. At present, multi-objective optimization rolling schedule is widely used. In this paper, the multi-objective optimization method of rolling schedule based on improved immune clone algorithm is studied according to the actual needs of production. Firstly, the evolutionary mechanism of traditional immune clone multi-objective evolutionary algorithm is analyzed, and the mutation strategy of the algorithm is improved to solve the problem of slow convergence. The difference mutation and cloud model are used for adaptive mutation of individual population, which makes individual mutation more targeted, increases the search range of the algorithm and improves the convergence performance of the algorithm. Secondly, based on the characteristics of evolutionary population and memory population in immune clone multiobjective algorithm and its role in the optimization process, a hybrid clipping selection mechanism is designed. The individual selection of evolutionary population and memory population is carried out respectively to prevent the individuals in the population from being too dense and to improve the distribution degree of the individuals in the algorithm. Finally, aiming at the problems of coupling and mutual restriction between multi-objective functions in the process of rolling schedule optimization, it is difficult to select the target focus. Taking a five-stand cold continuous rolling line in Tangshan steel plant as an example, the equal power margin is constructed. The multi-objective load function with good shape and anti-skid objective function is used to calculate the multi-objective load function by using the improved algorithm proposed in this paper. The experimental results show that the proposed algorithm can converge to the Pareto front quickly and obtain a well distributed solution set, and the different preferred rolling schedule combinations after optimization can meet different selection requirements. Compared with the original regulation, the use of each rolling mill is more reasonable, the utilization efficiency of the mill is improved, the strip shape and surface quality are improved, and the probability of the scratch is reduced, which plays a guiding role in the calculation of the two-stage setting.
【学位授予单位】：燕山大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TG334.9;TP18
本文编号：2325142
[Abstract]:In the computer control system of cold continuous rolling, the setting of rolling schedule is the core of the control parameter setting of cold tandem rolling. Reasonable rolling schedule can make cold continuous rolling run efficiently, achieve better control effect, and reduce production energy consumption and improve product quality. With the continuous development of rolling technology and the continuous improvement of quality requirements, making rolling schedule based on single objective can no longer meet the needs of production. At present, multi-objective optimization rolling schedule is widely used. In this paper, the multi-objective optimization method of rolling schedule based on improved immune clone algorithm is studied according to the actual needs of production. Firstly, the evolutionary mechanism of traditional immune clone multi-objective evolutionary algorithm is analyzed, and the mutation strategy of the algorithm is improved to solve the problem of slow convergence. The difference mutation and cloud model are used for adaptive mutation of individual population, which makes individual mutation more targeted, increases the search range of the algorithm and improves the convergence performance of the algorithm. Secondly, based on the characteristics of evolutionary population and memory population in immune clone multiobjective algorithm and its role in the optimization process, a hybrid clipping selection mechanism is designed. The individual selection of evolutionary population and memory population is carried out respectively to prevent the individuals in the population from being too dense and to improve the distribution degree of the individuals in the algorithm. Finally, aiming at the problems of coupling and mutual restriction between multi-objective functions in the process of rolling schedule optimization, it is difficult to select the target focus. Taking a five-stand cold continuous rolling line in Tangshan steel plant as an example, the equal power margin is constructed. The multi-objective load function with good shape and anti-skid objective function is used to calculate the multi-objective load function by using the improved algorithm proposed in this paper. The experimental results show that the proposed algorithm can converge to the Pareto front quickly and obtain a well distributed solution set, and the different preferred rolling schedule combinations after optimization can meet different selection requirements. Compared with the original regulation, the use of each rolling mill is more reasonable, the utilization efficiency of the mill is improved, the strip shape and surface quality are improved, and the probability of the scratch is reduced, which plays a guiding role in the calculation of the two-stage setting.
【学位授予单位】：燕山大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TG334.9;TP18
【参考文献】
中国期刊全文数据库 前7条
1 杨光辉;曹建国;张杰;王聪;贾生晖;;SmartCrown冷连轧机板形控制新技术改进研究与应用[J];钢铁;2006年09期
2 张清东，陈先霖，徐乐江，华建新，，徐耀寰，陈华昶;CVC四辊冷轧机板形预设定控制研究[J];钢铁;1997年10期
3 王战权,赵朝义,云庆夏;进化策略中基于柯西分布的变异算子改进探讨[J];系统工程;1999年04期
4 包仁人;张杰;李洪波;宫贵良;贾生晖;刘海军;;冷轧板形平坦度计算方法[J];钢铁;2014年12期
5 林浒;彭勇;;面向多目标优化的适应度共享免疫克隆算法[J];控制理论与应用;2011年02期
6 陈基漓;;基于高斯变异的生物地理学优化模型[J];计算机仿真;2013年07期
7 戚玉涛;刘芳;刘静乐;任元;焦李成;;基于免疫算法和EDA的混合多目标优化算法[J];软件学报;2013年10期
本文编号：2325142
本文链接：https://www.wllwen.com/kejilunwen/jinshugongy/2325142.html
教材专著
