A NEW APPROACH OF MINIMIZING COMMUTATION TORQUE RIPPLE FOR BRUSHLESS DC MOTOR BASED ON ADAPTIVE ANN

In this paper, the principle of commutation torque ripple (CTR) for brushless DC motor is analyzed, and a new approach of minimizing CTR is presented, which is based on adaptive artificial neural network (ANN) control.Two three layer forward artificial neural networks are trained both in off line and in on line ways, and the weights in networks are updated using the error back propagation (BP) algorithm. One of the networks is used to estimate the commutation parameters of the motor online. The other is used to regulate the terminal voltages using the parameters estimated by the former network during commutation. In this way, the whole model can be considered as a voltage self tuning regulator (STR). Through regulating the terminal voltages of motor, the STR makes the rising ratio and dropping ratio of the phase currents be approximate in order to keep the amplitude of the total current in the circuit constant, so as to minimize CTR. This approach is unnecessary to know the accurate system parameters, and can modify the model adaptively while parameters are changed. The experimental results show that the proposed method is highly precise and robust.