DSP-based optimal sensor-less vector control drive for axial-flux PM motor

This paper presents a practical consideration to implement a sensor-less system drive on the Axial Flux Permanent Magnet (AFPM) motor with a high performance Digital Signal Processor (DSP) in real time. An on-line Extended Kalman Filter (EKF) is applied to estimate motor position and angular speed by measuring motor voltages and currents. In order to achieve desirable dynamic and robustness motor performance, Genetic Algorithm (GA) is used to tune coefficients of PI controller. This algorithm arranged these coefficients with considering importance of three factors in speed response: overshoot, steady state error and rise time. For investigation, entire system drive including EKF, Vector control, SV-PWM and AFPM motor are simulated in MATLAB/SIMULINK After that, the control system is executed on the TMS320F2812 controller platform.