Robust adaptive sliding mode observer for self-sensing IPMSM control based on optimized HF injection method

Abstract

Classical tracking algorithms often employed for HF injection techniques used to estimate position of ac machine depend on machine inductances. To overcome this dependency, a new approach to extract the sign of the rotor position estimation error is developed and then, a new adaptive step-by step sliding mode observer is proposed as an alternative solution to estimate the rotor position, speed and acceleration of the machine. The gains of this observer are made variable in order to reduce the chattering effects. Moreover, Low-pass-filters usually used in the pulsating-wave signal injection technique are removed leading to reduce the cost and complexity of implementation and improve the estimation in term of delays. The performance of the proposed algorithm is compared with respect to the existing tracking algorithms and the step by step sliding without gains variation in simulations and experiments.