Skip to Main Content
This paper presents a general method based on a permeance network (PN) in order to study stator turn faults in permanent-magnet synchronous machines (PMSMs). This approach includes two main parts: first, static computations are performed to identify the various parameters of a faulty machine. Leakage inductances as well as self and cross saturations are taken into account to improve the results without increasing the CPU time compared to finite element analysis (FEA). Then, a faulty model of the PMSM is used rather than time-stepping finite element method (TSFEM) which requires large CPU time. The model gives dynamical results and it is suitable to studying PMSM's behavior under faults, testing, or investigating fault detection methods under parameter variations. The method can be applied for any type of winding arrangements and different fault configurations. Various comparisons validate the proposed approach to determine precisely the parameters of the PMSM model under turn faults including the saturation effect and leakage inductances.