An innovative MRAS-based technique for online detection of short-circuit faults in three-phase induction motor windings

Fault diagnosis of induction motors (IMs) through stator stream analysis is a key area in maintenance systems due to its non-invasive nature and low hardware and software requirements. However, the industrial application of this method faces challenges, particularly in estimating stator winding resistance accurately, especially in large IMs operating at shallow slips. The fundamental leakage can mask fault components until the damage is severe, often leading to short circuits. This article suggests a new approach for fault diagnosis based on analyzing the difference in errors between actual stator streams and estimated currents obtained from a virtual model of the IM. These differences are indicators of estimation errors. To enhance fault detection, these error values are multiplied by the estimated stator currents, forming the basis of an adaptive mechanism. This mechanism utilizes proportional–integral regulators to adjust the estimation process, improving the accuracy of detecting asymmetric faults across all stator phases. The estimation of stator winding resistance, which changes due to intern–turn short circuits, is achieved online using a classical model reference adaptive system. However, estimating stator resistance in a reference frame is challenging, particularly when an incipient short circuit fault is present. The stability of the designed method is ensured through the application of Popov's criterion of inequality in its design. Popov's criterion guarantees system stability under various operating conditions. By incorporating Popov's criterion into the design, the designed technique can maintain stability while accurately detecting and estimating different levels of short-circuit faults in various IM operating modes. Empirical results demonstrate the effectiveness and significance of the designed strategy in fault diagnosis of IMs, highlighting its potential for practical industrial application.