Digital twin-assisted fault diagnosis of induction motors with thermal analysis of inter-turn short circuit faults

Abstract Inter-turn short circuit faults (ITSCFs) in induction motors (IMs) can lead to significant motor damage and system failure, particularly in electric drive systems. This paper proposes a novel digital twin (DT)-assisted method for ITSCF diagnosis, incorporating an analysis of the temperature rise characteristics induced by the faults. First, a DT model of the IM is developed, with state parameters updated in real-time using measured current data to accurately replicate motor behavior under various operating conditions. The DT model then simulates motor currents and temperature rise characteristics under healthy and faulty states, including inter-turn short circuit conditions. Combined with limited measured data, these synthetic data are used to create a physically-virtually fused dataset for training a deep learning-based fault diagnosis model. Additionally, the temperature rise characteristics are analyzed to reveal the distinctive thermal behavior of the stator windings under ITSCF conditions. Experimental results demonstrate that the proposed method not only achieves high diagnostic accuracy but also provides critical insights into the thermal effects of ITSCF in IMs. This approach offers a promising solution for real-time fault monitoring and thermal analysis in electric drive systems.