FPGA-realization of an RBF-NN tuning PI controller for sensorless PMSM drives

Abstract

The estimation and control of the rotor position for sensorless permanent magnet synchronous motor (PMSM) drives based on extended Kalman filter (EKF) and artificial neural network (ANN) is presented in this paper. The EKF is a rotor position estimator which is a full-order stochastic observer for the recursive optimum state estimation of a nonlinear dynamic system in real time by using signals that are in the noisy environment. An ANN constructed by radial basis function neural network (RBFNN) and a parameter adjustable mechanism is applied to speed control loop of the PMSM drives to cope with the effect of the system dynamic uncertainty and external load. In this paper, firstly, a mathematical model for PMSM is derived, and a sensorless FOC is built up. Secondly, the rotor position and rotor speed which are estimated by using EKF is described. These estimated values are feed-backed to the current loop for FOC and to the speed loop for RBFNN-based self-tuning PI control. Thirdly, a very high-speed IC hardware description language (VHDL) is presented to describe the behavior of the adopted control and estimation algorithm. Fourthly, to verify the correctness of the designed VHDL code of the control and the estimation algorithm, based on electronic design automation (EDA) simulator link, a co-simulation work is constructed by Simulink and ModelSim. And some simulation results verify the correctness and effectiveness. Finally, an experimental system with a PMSM, a motor driver circuit, and a field programmable gate array (FPGA) board are set up to implement the proposed rotor position estimation and speed control algorithm.