串级液位控制系统的改进粒子群神经网络PID控制研究
|
摘要
串级液位控制系统是工业过程控制中最典型的系统,针对该系统具有不确定性和时变性的特点,首先在分析串级液位控制系统的结构和控制原理的基础上,采用实验建模的方法建立了系统的数学模型;然后充分利用PID结构简单、抗干扰能力强的特点以及神经网络具有自学习和自适应的特长,引入粒子群算法对网络权值进行优化,提出了一种基于改进粒子群算法的神经网络PID控制器,既克服了BP神经网络收敛速度慢,容易陷入局部极小值的缺点,也克服了粒子群算法容易陷入局部最优的缺点;最后在MATLAB环境下进行串级液位系统的仿真试验对系统的性能进行分析,实验结果表明,该控制算法具有良好的实时性、鲁棒性,抗干扰能力强,显著提高了系统的性能指标。
String loop liquid control system is the typical system in industry process control.The system has the characteristics of uncertainty and time-variable.Based on analysis of the structure and control principle of string loop liquid control system,this paper has built systemic mathematical model by modeling.Then,introduce particle swarm algorithm to optimize the network′s weight combined with the characteristics of PID simple structure,strong anti-interference ability and the advantage of the neural network′s self-learning and self-adaptive.Advance a kind neural network PID controller based on improved particle swarm algorithm.The algorithm conquered slow convergence speed and easily getting into local dinky value of BP neural network,and easily getting into local optimization of particle swarm.Analyze the system′s performance by string loop liquid control system′s experiment in MATLAB.The result indicated that the control system had good realtime,robustness and strong anti-interference ability.The capability of the system is improved notably.
String loop liquid control system is the typical system in industry process control.The system has the characteristics of uncertainty and time-variable.Based on analysis of the structure and control principle of string loop liquid control system,this paper has built systemic mathematical model by modeling.Then,introduce particle swarm algorithm to optimize the network′s weight combined with the characteristics of PID simple structure,strong anti-interference ability and the advantage of the neural network′s self-learning and self-adaptive.Advance a kind neural network PID controller based on improved particle swarm algorithm.The algorithm conquered slow convergence speed and easily getting into local dinky value of BP neural network,and easily getting into local optimization of particle swarm.Analyze the system′s performance by string loop liquid control system′s experiment in MATLAB.The result indicated that the control system had good realtime,robustness and strong anti-interference ability.The capability of the system is improved notably.
引文
[1]王彦,邓勇,王超.基于改进粒子群算法的模糊神经网络PID控制器设计[J].控制工程,2012,19(5):761-764.
[2]方达意,罗伟原,曾斌.基于改进粒子群算法的柴油发电机自适应PID控制[J].计算机测量与控制,2009,17(7):1308-1310.
[3]冯东青,邢广成,费敏锐,等.基于改进PSO算法的多变量PID型神经网络控制[J].系统仿真学报,2011,23(2):363-366.
[4]徐胜男,周祖德,艾青松,等.基于粒子群优化的神经网络自适应控制算法[J].中国机械工程,2012,23(22):2732-2738.
[5]Lazzus,Juan A.Autoignition Temperature Prediction Using An Artificial Neural Network With Particle Swarm Optimization[J].International Journal of Thermophysics,2011,32(5):1-17.
[6]Xu G Z,Song A G,Li H J.Adaptive Impedance Control for Upper-Limb Rehabilitation Robot Using Evolutionary Dynamic Recurrent Fuzzy Neural Network[J].Journal of Intelligent and Robotic Systems:Theory and Applications,2011,62(3/4):501-525.
[7]Radulovic J.Feedforward Neural Net work and Adaptive Network-based Fuzzy Inference System in Study of Power Lines[J].Expert Systems with Applications,2010,37(1):165-170.
[2]方达意,罗伟原,曾斌.基于改进粒子群算法的柴油发电机自适应PID控制[J].计算机测量与控制,2009,17(7):1308-1310.
[3]冯东青,邢广成,费敏锐,等.基于改进PSO算法的多变量PID型神经网络控制[J].系统仿真学报,2011,23(2):363-366.
[4]徐胜男,周祖德,艾青松,等.基于粒子群优化的神经网络自适应控制算法[J].中国机械工程,2012,23(22):2732-2738.
[5]Lazzus,Juan A.Autoignition Temperature Prediction Using An Artificial Neural Network With Particle Swarm Optimization[J].International Journal of Thermophysics,2011,32(5):1-17.
[6]Xu G Z,Song A G,Li H J.Adaptive Impedance Control for Upper-Limb Rehabilitation Robot Using Evolutionary Dynamic Recurrent Fuzzy Neural Network[J].Journal of Intelligent and Robotic Systems:Theory and Applications,2011,62(3/4):501-525.
[7]Radulovic J.Feedforward Neural Net work and Adaptive Network-based Fuzzy Inference System in Study of Power Lines[J].Expert Systems with Applications,2010,37(1):165-170.