Sensorless Predictive Current Control of PMSM EV Drive Using DSOGI-FLL Based Sliding Mode Observer

Conventional back electromotive force (EMF) estimation based active flux concept for sensorless control has various limitations due to pure integrator problems and performance deterioration at low speeds. The fundamental back EMF component extraction with a conventional low pass filter, is difficult due to the presence of lower order harmonics, dc offset, saturation, etc., and causes significant phase delay in the estimated position and speed. This article deals with a sensorless speed and position estimation technique for a surface mounted permanent magnet synchronous motor (SPMSM) based electric vehicle (EV) drive using a sliding mode observer (SMO) with a dual second order generalized integrator frequency locked loop (DSOGI-FLL). A novel method of utilizing the DSOGI-FLL, along with feedforward term is shown to achieve better harmonic elimination of estimated back EMF components and faster dynamic speed estimation. A delay compensated predictive current controller-based PMSM EV drive with the presented estimation technique is used for improved dynamic response. This system is simulated using Simulink and its efficacy is experimentally validated using dSPACE 1202. Test results show that the DSOGI-FLL-based SMO has better performance compared to the conventional technique.