A novel wavelet neural network based robust control of interior permanent magnet motor drives

The paper presents a novel wavelet neural network (WNN) based speed control of interior permanent magnet (IPM) motor drives. The proposed speed controller uses adaptive control algorithm of the WNN. The speed error and change of speed error are used as inputs to the WNN speed controller. The WNN controller parameters are updated continuously on-line for drive operating conditions using the back propagation training algorithm. The Lyapunov-based stability criterion is used for robust operation of the WNN speed controller. In order to operate the IPM motor above the rated speed, the flux-weakening control technique is used. The maximum torque per ampere control technique is used below the rated speed. The proposed WNN speed controller is implemented for the IPM motor drive system in simulation and experiment. The WNN controller-based drive system is implemented in real-time using a DSP controller board. The performances of the WNN controller are compared to those of conventional fixed-gain and adaptive speed controllers. The WNN controller is found to be more robust and quicker than conventional fixed-gain and adaptive speed controllers.