Speed sensorless control of the permanent-magnet synchronous motor based on wavelet neural networks

Permanent-magnet synchronous-motor is a multi-variable, non-linear, strong coupling system that is highly sensitive to the outer interfere and inter perturbation. To improve the system robustness, the non-linear mapping is built based on the properties of local time frequency characteristics, zooming index, adaptive study, strong robustness and non-linear approaching of wavelet neural network. By the wavelet neural network the position and speed of rotor is estimated. In the control strategy the sliding mode control is introduced. In this paper the simulation of control strategy is compared with the PI control strategy of PMSM under the same condition. The simulation result shows that the strategy can improve the dynamic and static performance of system.