High Dynamic Response Full Order Stator Flux Linkage Observer for IPMSM Drives

This paper presents an improved full-order stator flux-linkage observer for the Permanent Magnet Synchronous Machine (PMSM) drives employed for electromagnetic power conversion in the Electric Vehicle (EV) powertrain. The parameters of a typical PMSM are influenced by constantly changing operating conditions leading to significant errors when torque estimation is performed using an a-priori parametric model, also known as a current model. This issue is usually addressed using a voltage model-based flux-linkage estimation. However, this approach suffers from inaccuracy due to the inverter-generated disturbances. The significance of this disturbance also grows as the operating speed reduces. A conventional full-order flux-linkage observer relies upon a current model for low operating speed and gradually shifts to the voltage model as the machine accelerates. Thus, the flux-linkage, and hence the torque estimation suffers from the errors in the parameter knowledge for low-speed conditions. The presented observer consists of an improved closed-loop voltage model which benefits from a hybrid nonlinear corrective action based on the sliding-mode observation principle, which yields finite time convergence while maintaining a low chattering effect. Moreover, the estimation process is restricted within a feasible estimation period attributed to the differential nature of the controller voltages and measured currents. By virtue of these improvements, the proposed flux-linkage observer boasts high robustness towards the errors in parameter information across a wide operating speed range. The performance of the proposed flux-linkage observer is studied for fast dynamic change in torque at low-speed conditions in a comparative analysis with a conventional open-loop type Gopinath style estimator and a reduced-parameter sensitivity closed-loop observer.