摘要
针对开关磁阻电机非线性动态特性不易控制的缺点,提出了一种积分型滑模变结构与神经网络补偿相结合的复合控制策略。应用一个具有积分型式的滑模切换面的变结构控制器,使用积分补偿技巧降低系统的振动与稳态误差。为减小滑模面的抖动,引入神经网络补偿控制环节。建立滑模变结构控制的数学模型,并给出神经网络补偿滑模面抖动的控制律表达式。利用该积分型滑模变结构与神经网络补偿相结合的复合控制策略,可有效地改善开关磁阻电机动态响应问题。为验证控制方案的可行性,对系统进行了仿真与实验研究。实验结果表明:系统具有良好的动静态性能,且在无需知道电机精确模型的情况下可有效克服转矩脉动。
As switched reluctance motor is difficult to be controlled due to its nonlinear,a hybrid control scheme is proposed.The control scheme was implemented by using siding mode variable structure control with integral action and neural network compensation control.Its variable structure controller with a integral sliding mode switching surface can achive good performances of reducing vibration and static errors in switched reluctance motor driving system.Moreover,neural network compensation control was introduced to reduce jitter of sliding-mode surface.The mathematic models of sliding-mode variable structure control were built,and the control rate express for reducing jitter of sliding-mode surface based on neural network compensation control was given.The dynamic performances of switched reluctance motor were improved by the proposed hybrid control scheme.For verifying feasibility of the proposed control scheme,simulation and experimen tests were achieved.Both simulation rusults and experimental results verify the feasibility of the proposed control scheme,better dynamic and static performances of the driving system without accurate mathematical model,and smaller torque ripple of the switched reluctance motor.
出处
《电机与控制学报》
EI
CSCD
北大核心
2011年第1期33-37,43,共6页
Electric Machines and Control
基金
湖南省科技计划项目(2010GK3100)
湖南省自然科学湘潭联合基金(10JJ9009)
关键词
开关磁阻电机
滑模变结构
积分
转矩脉动
神经网络
switched reluctance motors
sliding mode variable structure
integral
torque ripple
neural network
