Inductance-based Inter-Turn Fault Detection in Permanent Magnet Synchronous Machine Using Magnetic Equivalent Circuit Model

This paper presents an inductance-based method for detection of inter-turn fault in permanent magnet synchronous machines (PMSM). The reason for using inductance for fault detection and its severity estimation is the obvious impact of the fault on the inductances. This fault reduces the effective number of turns and therefore inductances. The PMSM is modeled using magnetic equivalent circuit (MEC). This method takes into account the impacts of stator teeth, rotor yoke, and stator winding distribution precisely. So, it is shown that the MEC method is superior to other analytical methods such as winding function. The governing equations in different parts of the machine are proposed, and the MEC equations of the PMSM are developed. First, healthy and faulty PMSM is simulated using MEC for different severities of the fault. In the 2nd stage, the required analytical equations for self-inductance and mutual inductance estimation are extracted, and the impact of the fault severity upon these inductances is determined. Finally, inter-turn short-circuit fault is detected, and its severity is determined using the amplitude of the inductance of the faulty phase. Simulation results and the proposed fault diagnosis method are validated using finite elements method.