基于混沌粒子群的神经网络主汽温控制系统  

Main Steam Temperature Control System Based on CPSO and Neural Network

作　　者：王勇[1] 

机构地区：[1]江苏科技大学张家港校区,江苏张家港215600

出　　处：《计算机仿真》2012年第8期322-325,共4页Computer Simulation

基　　金：江苏省道路载运工具新技术应用重点实验室开发基金(BM2008206009)

摘　　要：研究火电厂锅炉主蒸汽温度控制优化问题,针对主汽温对象具有大惯性、大迟延、时变性和非线性系统,由于存在实时性和实时性差,传统的PID控制难以获得很好的控制效果,提出一种混沌粒子群优化神经网络的主汽温控制方法。采用RBF神经网络对PID参数进行在线整定,并通过混沌粒子群算法对RBF神经网络初始参数进行优化,不仅具有RBF神经网络的自适应能力,同时具有常规PID串级控制的特性,增强了系统对不确定因素的适应性。仿真结果表明,控制算法具有较好的鲁棒性和控制品质,抗干扰能力强,可为锅炉主蒸汽温度优化控制提供参考。The main steam temperature object is the system with large inertia, large time delay, time -varying and nonlinear, and the traditional PID control cannot obtain satisfactory control performances. This paper proposed a control method based on CPSO and RBF neural network. PID parameters were tuned by RBF neural network, and the initial parameters of RBF neural network were optimized by chaos particle swarm optimization algorithm. The pro- posed control algorithm not only has adaptive ability and the characteristics of conventional PID control, but strength- ens the adaptability of system. The simulation results show that the proposed control algorithm has better robustness and control quality, and strong ability of resisting disturbances.

关 键 词：主汽温控制 混沌粒子群算法 神经网络 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]