Abstract: In order to improve the performance of permanent magnet synchronous motor sliding mode control system, a fal function was used to replace the traditional symbolic function, which was based on the traditional exponential approach law sliding mode control. And an improved exponential approach law was designed, and its stability was verified by Lyapunov function. Then, the parameters of the improved exponential approach law were dynamically optimized by fuzzy neural network, and the sliding mode speed controller of permanent magnet synchronous motor was designed. Through simulation analysis, the results show that: compared with the traditional exponential approach law, the improved exponential approach law reduces the overshoot by 67.1%; the speed drop amplitude under load disturbance is reduced by 22.2%, and the speed recovery time is shortened by 0.01s. The improved exponential approach law optimized by the fuzzy neural network further reduces the overshoot by 50%; the speed drop amplitude under load disturbance is reduced by 8.9% again, and the speed recovery time is shortened by 0.0032s again.

Key words: vehicle engineering; permanent magnet synchronous motor; sliding mode control; exponential approach law; fal function; fuzzy neural network

摘要： 为进一步提高永磁同步电机滑模控制调速系统性能，在传统指数趋近律滑模控制基础上，采用了一种fal函数来代替传统符号函数，设计了一种改进型指数趋近律并用李雅普诺夫函数验证了其稳定性。借助模糊神经网络对改进指数趋近律参数进行动态优化，设计出永磁同步电机滑模转速控制器。仿真分析结果表明：与传统指数趋近律相比，改进指数趋近律减小了67.1%的超调，在负载扰动时的转速下降幅度减小了22.2%，转速恢复时长缩短了0.01 s；经模糊神经网络优化的改进指数趋近律又进一步减小了50%的超调，负载扰动时的转速下降幅度再次减小了8.9%，转速恢复时长再次缩短了0.003 2 s。

关键词: 车辆工程；永磁同步电机；滑模控制；指数趋近律；fal函数；模糊神经网络