摘要
在高性能、高精度的控制要求条件下,针对无刷直流电机(BLDCM)利用经典比例-积分-微分(PID)无法满足相关要求的问题,提出了一种神经网络在线整定的PID控制算法,并针对神经网络的缺陷,利用粒子群算法(PSO)进行了优化,力图使BLDCM在复杂多变的情况下响应性能更好、速度波动更小。设计了BLDCM控制板,搭建了实验平台,分析了BLDCM的控制模型,建立了神经网络控制模型,结合PSO对神经网络进行改进优化。实验结果表明优化后的算法收敛速度更快,BLDCM具有更好的动态响应性能,速度波动更小。
Aiming at the control situation that the high performance and high precision classic proportion-integral-differential(PID)of brushless DC motor(BLDCM)can not meet the performance requirements,a neural network online tuning PID control algorithm is proposed,and the defect based on neural network is optimized by particle swarm optimization(PSO).It is trying to make the BLDCM have better response performance and less fluctuation in speed under complicated and variable conditions.The BLDCM control board is designed,the experimental platform is built,the control model of the BLDCM is analyzed,and the neural network control model is established.The neural network is optimized by the PSO.The experimental results show that the optimized algorithm converges faster,and the BLDCM has better dynamic response performance and less speed fluctuation.
作者
田海林
宋珂炜
董铂龙
方辉
TIAN Hai-lin;SONG Ke-wei;DONG Bo-long;FANG Hui(Sichuan University,Chengdu 610065,China)
出处
《电力电子技术》
CSCD
北大核心
2019年第12期106-110,共5页
Power Electronics
基金
国家自然科学基金(51175356)
四川省科技支撑计划(2016GZ0187)~~
关键词
无刷直流电机
神经网络
粒子群算法
brushless direct current motor
neural network
particle swarm optimization