Full-order Sliding Mode Observer Based Position Sensorless IPMSM Control Strategy

Interior permanent magnet synchronous motor (IPMSM) has been widely usedin electrical drive fields such as industry, aerospace, transportation and domesticappliances due to the benefits of high efficiency, good speed control performance,high power density, and fitness for flux-weakening control, etc. The positionsensorless control techniques can reduce the cost and increase the reliability of thedrive system, therefore, it’s critical to implement position sensorless IPMSM drivesof high control accuracy, wide speed range, and strong robustness. In this paper, theemphases were put on the sliding mode observer (SMO) applied for the medium-and high-speed range and the rotor initial position detection at standstill.In order to improve the robustness and the dynamic estimation performance, onthe basis of the extended electromotive force (EMF) model, a full-order SMOtaking the stator currents and the extended EMFs as the state variables waspresented. The extended EMF model was introduced, which guaranteed that therotor position information could be decoupled from the inductance matrix. Then, thefull-order state equation in the synchronous rotating reference frame taking thestator currents and the extended EMFs as the state variables was built. Whereby, afull-order SMO was constructed, and a simplified feedback gain matrix designmethod was proposed. The quadrature phase locked loop (PLL) was used to extractthe rotor position from the extended EMF estimates, which can effectively reducethe influence of noises.The rotor initial position detection at standstill based on the hybrid signalinjection was presented. The high-frequency (HF) mathematical model of IPMSMwas built, and the HF fluctuating voltage signal injection method was adopted toestimate the rotor magnetic pole position. Namely that a pulsating carrier-voltagevector was injected on the direct-axis in the estimated rotating reference frame, then,the excited HF current vector in the measurement axis which lags45degrees behindthe estimated rotating reference frame was used to extract the magnetic poleposition information. On the basis of saturation and saliency principle, the pulsevoltage vector injection method was adopted to identify the magnet polarity. Twopulse voltage vectors in the estimated magnetic pole position direction and theopposite direction were applied in the stator windings, and the magnet polaritycould be identified by comparing the amplitude of the two excited currents.Based on the theoretical analysis, the position sensorless IPMSM vectorcontrol system was implemented, and experimental verifications were carried out ona platform of the2.2kW IPMSM drive. The experimental results show that, the extended EMF model based full-order SMO can effectively improve the dynamicestimation performance, and the hybrid signal injection method can accuratelyidentify the rotor pole position and the magnet polarity.