电动汽车无离合器神经网络自校正PID控制  

Neural Network Self-turning PID Control for Electric Vehicles without Clutch

作　　者：佘致廷[1] 杨婷[1] 秦亚胜 董旺华 

机构地区：[1]湖南大学电气与信息工程学院,长沙410082

出　　处：《控制工程》2017年第6期1124-1130,共7页Control Engineering of China

基　　金：国家自然基金(61203207);863计划(2012AA111004);湖南省科技计划项目(2013FJ4038)

摘　　要：由于传统的PID控制器是建立在被控对象精确已知的情况下,通常难以满足多变量、非线性、强耦合的无离合器电动汽车表贴式永磁同步电机(PMSM)精确的速度跟踪控制和高动态响应的性能要求,所以提出一种基于RBF神经网络辨识的自校正PID控制器,以提高负载转矩波动下的PMSM速度自动跟踪能力。仿真和实验结果表明,该系统减小了速度调节的超调量,加快了系统的速度响应,实现了无离合器电动汽车精确的速度跟踪控制。The traditional PID controller, which is based on the precise control objects, usually fails to satisfy the performance requirements of precise speed tracking control and quick dynamic state responds of a multivariable, nonlinear and strong coupled surface permanent magnet synchronous motor （PMSM） of electric vehicles without clutch. Therefore, a self-tuning PID controller based on RBF neural network identification has been proposed to improve the auto speed tracking ability of PMSM when the load torque ripples. Simulation and experimental results show that the system not only shortens the overshoot of the speed regulation, but also fastens the speed responses, realizing the precise speed tracking control of electric vehicles without clutch.

关 键 词：无离合器电动汽车 速度跟踪 矢量控制 RBF网络辨识 自校正PID控制 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]