Newton-Raphson Predictive PLL Based Position Estimation Method for Sensorless SynRM Control

A rotor position observer is essential in the sensorless control of synchronous reluctance motors (SynRMs). However, conventional fixed-parameter position observers require repeated trials for parameter tuning and have poor dynamic performance. This paper proposes a novel Newton-Raphson based predictive phase-locked loop (NR-PPLL) to improve the dynamic performance, which does not require parameter tuning. A cost function is established based on the error between the stator flux estimated by the voltage equation and the one calculated with currents. For the search strategy of the conventional PPLL with multiple iterations and large computation, the NR method is adopted, which requires only three iterations to estimate the accurate position. Moreover, in response to the erroneous convergence of NR caused by inappropriate initial value selection, a simple correction approach is proposed. Finally, the effectiveness of the proposed NR-PPLL is verified on a 3-kW SynRM drive platform.