Akademik Veri Yönetim Sistemi
4th International Conference on Advanced Engineering, Technology and Applications on Power Systems, Shenyang, Çin, 31 Temmuz - 01 Ağustos 2025, ss.1-10, (Tam Metin Bildiri)
This article presents an effective approach to predict short circuit faults in the stator windings in an induction motor using motor current signature analysis. Traditional sensor-based diagnostic methods require large manpower and a larger number of sensors. In this paper, to come through these drawbacks, fault detection based on current signature is proposed. The main fault detection parameter in this article will be the three-phase current of the stator in healthy and faulty conditions under different loads. Fault detection is, firstly, applying a discrete wavelet transform (DWT) to the output current from the stator under healthy conditions and also under different fault conditions. DWT yields many coefficients in the high-level decomposition required for high resolution. Then, using wavelet coefficients, the energy values in each scale are extracted and used as input parameters for training the artificial neural network (ANN) for fault detection. The proposed method has given 96% accuracy of short circuit fault of stator windings in induction motor of rating 3Phase, 3hp, 460 V and 60 Hz.