摘要:针对五相开关磁阻电机,进行了直接转矩控制方法的分析与研究,提出了一种基于BP神经网络的PID与传统PI相结合的新型控制策略。在大误差下采用神经网络控制,小误差下采用传统PI控制,并且在传统PI控制环节上,采用了依据负载转矩误差限构成的变参数法,对系统速度误差进行快速调整,可以更好的适应外界干扰。对五相开关磁阻电机进行了仿真,仿真结果表明,这种控制方法能够有效的减小转矩脉动,并且避免了电机多变量、强耦合、非线性等问题,系统动静态性能良好。
关键词:五相开关磁阻电机;直接转矩控制;BP神经网络;变参数 Abstract: For five-phase switched reluctance motor, the direct torque control strategy was researched and analyzed. A novel control strategy based on back propagation neural network PID combined with traditional PI control was proposed. It adopted neural network controller when the speed error was big, and adopted traditional PI controller when the error was small. In the traditional PI control link, the speed error was adjusted quickly and can better adapt to outside interference by using a variable parameter method based on the load torque error limits. The simulation results of the five-phase switched reluctance motor direct torque control show that this control strategy can effectively suppress torque fluctuation and avoid the motor multivariable, strong coupling, nonlinear problems, static and dynamic characteristics of a good system.
Key words: five-phase switched reluctance motor; direct torque control; BP neural network; variableparameter
|