摘  要：针对采用常规PID控制器很难取得很好的控制效果，提出了单神经元PI复合控制的开关磁阻电机调速系统的新方法，利用具有自学习和自适应能力的单神经元来构成开关磁阻电机的单神经元自适应控制器，不但结构简单，而且能适应环境变化，具有较强的鲁棒性。以速度误差为系统外环输入，大偏差时采用单神经元控制，小偏差时采用PI控制。外环的输出变量为内环的目标转矩，送入60 kW三相6／4结构的开关磁阻电机直接转矩调速系统内环。仿真结果表明，这种复合控制方法解决了常规控制方法因电机数学模型难以精确确定而无法确定控制参数的问题，并克服了常规PI控制存在静差、无抗干扰能力的缺点，很好的解决了系统上升时间与超调的矛盾。 
关键词：开关磁阻电机；非线性；BP-PID；直接转矩控制

Abstract: It is difficult to obtain good control effect for normal PID controller. The paper proposed a new control for BP_PID.using self-learning and adaptive capacity form the switched reluctance motor neuron adaptive controller, not only simple, but can adapt to environmental changes, has strong robustness. The outer loop input is speed error, while the outer loop output is torque which enters the inner loop of the direct torque control system to control 60KW three phase 6/4 pole SRM.For the outer loop, it adopts single-neuron controller when the speed error is big, and adopts PI controller when the error is small. Simulation results indicate that this compound control strategy can settle the parameter easily, and it overcomes the steady-state error of speed in normal PID controller, and the paradox between the response time and overshoot of speed is managed. The system has a good anti-interference performance.
Key words: switched reluctance motor；nonlinear；BP-PID；direct torque control