Model Based Online Detection of Inter-Turn Short Circuit Faults in PMSM Drives under Non-Stationary Conditions

In this paper, a model based fault detecting algorithm for detection of inter-turn short circuits in permanent magnet synchronous motors (PMSMs) is presented. Based on the PMSM model, the algorithm compares predicted values of the current space vector with the measurements to calculate an error vector, which will be used to detect the fault. A significant advantage of the developed algorithm is the ability to detect an inter-turn short circuit also during transients, i.e. at variable speed and while changes of dq-axis current space vector. Furthermore, the algorithm does not use any frequency transformation, which results in low computational cost and avoids high storage requirement.