Dynamic Position Estimator for a PMSM Using Regular Current Sampling and Numerical Optimisation

Easy-to-implement and easy-to-tune self-sensing control of permanent magnet synchronous machines is an ongoing research topic. In this paper, a novel position estimator is proposed which pursues these characteristics by using a numerical optimisation and regular current sampling (at twice the PWM frequency) even at standstill and low speeds. By defining a residual which is minimised using an online optimisation, the rotor position can be accurately estimated in stationary and dynamic states of the drive system. The proposed estimator achieves a high bandwidth self-sensing control with analytical tuning and low computational effort.