Rotor Speed Identification of PMSM on DTC System Based on PSO and CMAC Neural Network Algorithms

Both the PSO (particle swarm optimization) for global search algorithm and the CMAC (cerebella model articulation controller) algorithm are used in the speed loop of direct torque control system in the permanent magnet synchronous motor. Firstly, the PSO algorithm is applied to search the optimal PID parameters in the domain space, and then the CMAC neural network is adopted to learn and train of the results which derive from the output of the PSO algorithm to furthermore optimize the PID parameters which can improve stability of the system. The DTC control system based on PSO with CMAC algorithmic and the traditional DTC (direct torque control) control system are established and simulated in the MATLAB circumstance whose results were all compared. It revels that the DTC control system based on PSO with CMAC algorithmic is fast response, ideal flux, and good anti-jamming.