无刷直流电机调速系统神经网络自适应滑模变结构控制

Neural Adaptive Sliding Model Variable Structure Control for Brushless DC motor Speed Systems

为了提高无刷直流电机调速驱动系统的性能,提出神经网络自适应滑模变结构控制策略。推导无刷直流电机端电压与转速之间的微分方程,运用滑模变结构控制理论,通过调节端电压来实现转速控制;为了有效抑制系统在滑模切换面上的抖振采用自适应算法调整滑模增益的大小;从实际应用的角度出发,利用神经网络对非线性函数的任意精度拟合性,设计径向基函数神经网络估计器对控制量中广义扰动进行动态估计。仿真和实验结果表明采用本文提出的方法控制无刷直流电机,超调量小,速度响应快,控制精度高,且系统对各种干扰和参数摄振具有较强的鲁棒性,动、静态性能均优于PID控制。

In this paper a novel neural adaptive sliding model variable structure control strategy is proposed to improve the perfor- mances of brushless DC speed systems. Firstly, a sliding model variable structure speed controller of BLDCM is designed according to its mathematical model, in which an adaptive algorithm for regulating the switching gain is adopted to restrain the chattering around the sliding plane. And then radial basis function neural network （RBFNN） is devised to estimate the generalized disturbance item of the control variable dynamically. Finally, some simulation and experimental results are provide to indicate that the speed sys- tem of BLDCM by using the proposed control method has less overshoot, quick velocity response, higher control precision and good robustness, which is insensitive to the parameter chattering and many disturbances.