基于粒子群算法的精轧AGC系统智能参数优化及实验设计  

The intelligent parameter optimization and experimental design for finish rolling AGC system based on PSO

作　　者：万春秋 李擎[1] 崔家瑞[1] 李希胜[1] 徐银梅[1] WAN Chunqiu;LI Qing;CUI Jiarui;LI Xisheng;XU Yinmei(School of Automation and Electrical Engineering,University of Science and Technology Beijing,Beijing 100083,China)

机构地区：[1]北京科技大学自动化学院,北京100083

出　　处：《实验技术与管理》2023年第5期31-37,99,共8页Experimental Technology and Management

基　　金：国家自然科学基金项目(62273033);教育部第二批新工科研究与实践项目(E-ZDH20201602);教育部高等学校自动化类专业教学指导委员会专业教育教学改革研究课题(202149);教育部产学合作协同育人新工科建设项目(202101320001,202002309017);北京科技大学教育教学改革与研究重点项目(JG2021ZD02,JG2019Z04);中央高校基本科研业务费项目(FRF-TP-20-014A1);佛山市人民政府科技创新专项资金项目(BK20AF006)。

摘　　要：将经典粒子群(particleswarmoptimization,PSO)算法和4种不同改进形式的PSO算法应用于精轧自动厚度控制(automatic gauge control,简称AGC)系统PI控制参数的优化整定,设计了基于PSO-PI控制策略的精轧AGC智能优化控制系统。在控制参数的整定过程中,通过加权系数法将控制精度、动态响应特性等多项性能指标的优化问题转化为单目标优化问题,并通过仿真实验研究了各项控制指标权重系数对控制效果的影响。最后,基于PSO算法研究成果构建了精轧AGC智能优化控制虚拟仿真实验系统,为学生解决冶金自动化领域复杂工程问题能力培养提供了有力支撑。The classical particle swarm optimization(PSO)algorithm and four PSO algorithms with different improved forms are employed to optimize the PI control parameters of the automatic gauge control(AGC)of the finish rolling system.In this way,an intelligent optimal control system for finish rolling AGC based on PSO-PI control strategy is constructed.In the process of control parameters tuning,the optimization problem of multiple performance indicators,such as control accuracy and dynamic response characteristics,is transformed into a single objective optimization problem through the weighted coefficient method,and the influence of the weight coefficient of each control indicator on the control effect is studied through simulation experiments.Finally,based on the research results of PSO algorithm,an intelligent optimal control virtual simulation experimental system for finish rolling AGC is constructed.And it provides strong support for students to cultivate their ability to solve complex engineering problems in the field of metallurgical automation.

关 键 词：粒子群算法 精轧AGC系统 PSO-PI智能优化控制 加权系数法 虚拟仿真实验系统 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]