基于UFELM的DHP在水泥回转窑系统中的应用  

Application of DHP Based on UFELM in Cement Rotary Kiln System

作　　者：李鑫[1] 程勇 陈薇[1] 陈熙鹏 高翔 张宏图 LI Xin;CHENG Yong;CHEN Wei;CHEN Xi-peng;GAO Xiang;ZHANG Hong-tu(School of Electrical and Automation Engineering,Hefei University of Technology,Hefei 230009,China;Hefei Cement Research&Design Institute Corporation Ltd,Hefei 230051,China)

机构地区：[1]合肥工业大学电气与自动化工程学院,安徽合肥230009 [2]合肥水泥研究设计院有限公司,安徽合肥230051

出　　处：《控制工程》2023年第2期193-199,共7页Control Engineering of China

基　　金：安徽省重点研究与开发计划项目(202104a05020054)。

摘　　要：水泥是建筑等行业必不可少的原材料之一,回转窑是水泥烧成的核心设备。水泥回转窑烧成系统是典型的多扰动、强耦合、大时滞的非线性系统,其导致回转窑温度难以控制,为克服这一难题,建立了基于遗传算法改进BP网络水泥烧成系统模型。在此基础上,利用全反馈极限学习机的双启发式动态规划(UFELM-DHP)算法设计回转窑烧成系统优化控制器。仿真结果表明,此算法可以实现回转窑烧成系统的稳定控制,且具有更快的控制速率。Cement is one of the essential raw materials for construction and other industries,and the rotary kiln is the core equipment for cement firing.The cement rotary kiln firing system is a typical non-linear system with multiple disturbances,strong coupling and large time delay.In order to overcome the problem of temperature control of the rotary kiln,a cement firing system model is established based on genetic algorithm and improved BP network.On this basis,the dual heuristic dynamic programming(UFELM-DHP)algorithm of the full feedback extreme learning machine is used to design the optimized controller of the rotary kiln firing system.The simulation results show that this algorithm can realize the stable control of the rotary kiln firing system and have a faster control rate.

关 键 词：水泥回转窑烧成系统 遗传算法 BP网络 极限学习机 双启发式动态规划 

分 类 号：TP355[自动化与计算机技术—计算机系统结构]