基于神经网络学习算法和粒子群算法的改进PID控制在高压静止无功补偿器中的应用  

Application of an Improved PID Control Based on Neural Network Learning Algorithm and Particle Swarm Optimization in High Voltage Static VAR Compensators

作　　者：杨晓峰[1] 罗安[1] 彭楚武[1] 吴敬兵[1] 杨翠翠[1] 马伏军[1] 常亮亮[1] 

机构地区：[1]湖南大学电气与信息工程学院,湖南省长沙市410082

出　　处：《电网技术》2011年第6期65-70,共6页Power System Technology

基　　金：国家自然科学基金项目(60774043)~~

摘　　要：以高压静止无功补偿器(static var compensator,SVC)为研究对象,针对传统比例-积分-微分(proportional integral differential,PID)控制器难以对设定值进行有变化的跟踪和对扰动进行抑制的缺陷,提出在传统PID控制器的基础上加入一个2阶微分控制环节以实现公共连接点的电压稳定控制,并采用改进的神经网络粒子群优化算法对控制器的参数进行优化,使得系统瞬态响应性能和控制性能达到最佳。仿真和实验结果验证了所提出的控制方法能够保证快速、无超调的跟踪电压设定值,具有较强的鲁棒性、适应性,提高了SVC系统的补偿精度。To remedy the defect of traditional proportional integral differential（PID） controller that it is difficult for the controller to trace the set value changeably and suppress the disturbances,it is proposed to add a second-order differential controlling unit in traditional PID controller to implement stable control of voltage at point of common coupling（PCC） in high voltage static var compensator？（SVC）.Parameters of the improved controller are optimized by improved neural network learning algorithm and particle swarm optimization to make the transient response and control performance of the controller optimal.Results of simulation and experiments show that the improved controller possesses strong robustness and adaptability,and can ensure rapid and non-overshoot tracking of set value,so the compensation accuracy of SVC can be improved.

关 键 词：高压静止无功补偿器 神经网络 粒子群优化算法 

分 类 号：TM76[电气工程—电力系统及自动化]