基于AFSA-BP算法的模糊PID控制器在DMF回收系统中的应用  

Application of Fuzzy PID Controller Based on AFSA-BP Algorithm in DMF Recovery System

作　　者：王华强[1] 张博扬 WANG Hua-qiang;ZHANG Bo-yang(School of Electrical Engineering and Automation,Hefei University of Technology,Hefei 230009,China)

机构地区：[1]合肥工业大学电气与自动化工程学院

出　　处：《测控技术》2019年第10期80-84,共5页Measurement & Control Technology

基　　金：合肥工业大学产学研校企合作项目(12-069,13-214)

摘　　要：针对DMF(N,N-二甲基甲酰胺)回收系统存在时滞性、难以确定温度控制的数学模型等问题,提出了用人工鱼群算法(AFSA)优化BP神经网络,设计了基于AFSA-BP算法的模糊PID控制器。对模糊BP神经网络各层之间的连接权值以及隶属函数相关参数进行训练,克服了反向传播过程中最速下降法易陷入局部最优解的缺点,能够得到PID参数的最优解。研究结果表明,采用AFSA-BP算法的迭代次数少并且能接近给定值,在DMF回收系统中应用相关的PID参数,精馏塔的温度保持稳定,液位高度保持动态稳定且误差更小。Because of the time delays of DMF(N,N-Dimethylformamide)recovery system,it is hard to determine the mathematical model of temperature centrol.Inspired by this,the artificial fish-swarm algorithm(AFSA)is proposed to optimize BP neural network.A fuzzy PID controller based on AFSA-BP algorithm is designed.The connection weights and membership function parameters of the neural network are trained to overcome the shortcoming of falling into the local optimal solution caused by the gradient descent method in BP neural network back propagation.The result shows that the iteration of AFSA-BP algorithm is reduced and it can approach the given value.The temperature and liquid level of distillation column remain dynamically stable.

关 键 词：DMF回收系统 温度控制 先进PID控制器 AFSA算法 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]