PSO混合优化FNN的磨煤机控制研究  

Research on Coal Mill Control Based on Particle Swarm Optimization Neural Network

作　　者：穆海芳[1,2] 韩君 李明[1,2] 何康[1,2] MU Haifang;HAN Jun;LI Ming;HE Kang(School of Mechanical and Electronic Engineering,Suzhou University,Suzhou 234000,China;Suzhou Machinery and Equipment Collaborative Innovation Engineering Technology Research Center,Suzhou 234000,China)

机构地区：[1]宿州学院机械与电子工程学院,安徽宿州234000 [2]宿州市机械装备协同创新工程技术研究中心,安徽宿州234000

出　　处：《新乡学院学报》2019年第12期69-73,共5页Journal of Xinxiang University

基　　金：宿州学院重点科研项目(2016yzd09);安徽省教育厅人文社科研究项目(2016jyxm1031)

摘　　要：磨煤机制粉系统的输入量和输出量之间相互耦合,且系统具有非线性、延迟时间长等特点,采用常规的控制方法难以达到良好的效果,故提出一种采用误差反传与混沌粒子群算法混合优化前向神经网络权值的方法。首先采用混合优化算法调整神经网络的权值,然后采用神经网络自适应调整PID控制器的参数。仿真实验表明:该方法解决了系统在耦合性和时滞性方面的问题,进一步减小了超调量,跟踪效果好,具有良好的稳态性。The input and output of the pulverizing system are coupled,and the system has the characteristics of nonlinear,latency,and so on.Using conventional control method is difficult to achieve good results.Therefore,this paper proposes a method,which adopts back propagation and chaotic particle swarm optimization to optimize the weights of forward neural networks.The hybrid optimization algorithm was used to adjust the weights of the neural circuits.The PID controller parameters were turned by neural network adaptive adjustment then.The simulation experiment showed that the method solves the coupling and time delay problems of the system,reduces the overshoot amount further,and has good tracking effect and steady-state performance.

关 键 词：磨煤机 神经网络PID 粒子群 

分 类 号：TD453[矿业工程—矿山机电]