基于神经网络PID的电动汽车轮毂电机调速设计与仿真

Design and Simulation of Electric Vehicle In-hub Motor Speed Control based on Neural Network PID

轮毂电机式电动汽车在启动和运转过程中,电机控制系统经常要接收随机调速控制信号。传统PID控制难以实现快速、精确的速度调节。为解决此不足,提出采用神经网络PID(NNPID)进行控制的方法,首先对无刷直流电机进行建模分析,然后以BP算法训练神经网络并搭建控制系统,最后在Matlab/Simulink仿真环境下对该系统进行多种运转条件下的仿真并与传统控制策略进行比较,结果证明:基于神经网络的控制策略的电机控制系统启动平稳,能有效减少不稳定信号的干扰,对期望输出能实现较好的跟踪,可以满足一般电动汽车运行的需要.

The control system of the in-hub motor often receives random signals during its startup and normal operation of an electric vehicle. It is difficult for a traditional PID control to achieve rapid and precise speed regulation. In order to solve this problem,a neural network PID（ NNPID） is used to control the DC motor. A mathematic model of the DC motor is built,the neural network is trained by using BP algorithm and its control system is established. The system is simulated under various operating conditions in Matlab / Simulink. Compared with the traditional control strategy,the results indicate that the motor with new control strategy based on NNPID has a better performance in steady startup,less unstable signal interference and better tracking to the desired output,and that the designed control system can meet the needs of the electric vehicle.