کنفرانس مهندسی برق ایران

صفحه اصلی / سی و چهارمین کنفرانس بین المللی مهندسی برق

Open-Circuit Fault Detection in Induction Motor Drive Using a Multi-Scale Convolutional Neural Network with Residual Temporal-Channel Sparse Attention

نویسندگان :

Ali Khansary¹ Javad Poshtan² Vahid Khansary³ Elia Khansary⁴

1- دانشگاه علم و صنعت ایران 2- دانشگاه علم و صنعت ایران 3- دانشگاه خوارزمی 4- دانشگاه علم و صنعت ایران

کلمات کلیدی :

Open-circuit fault detection

چکیده :

Open-circuit fault detection and localization in power semiconductor switches has been one of the key topics in the field of fault diagnosis over the past decades. In this paper, a deep learning-based method called Residual Temporal-Channel Sparse-Attention Multi-Scale Convolutional Neural Network (RTCSA-MSCNN) is proposed for open-circuit fault detection in a two-level inverter-fed induction motor. The study addresses both single-switch and simultaneous double-switch open-circuit faults, resulting in a total of 22 classification categories. The dataset used for training and validation is obtained from simulations carried out in the MATLAB/Simulink environment. The results demonstrate that, after appropriate preprocessing, the proposed method effectively extracts multi-scale features and adaptively assigns attention weights, outperforming conventional Multi-Scale CNN (MS-CNN) and CNN approaches. The proposed RTCSA-MSCNN achieves high accuracy fault classification within a 25-millisecond observation window, without requiring any additional sensors and relying solely on the three-phase current measurements. This performance is consistently maintained across four different operating scenarios, demonstrating the model’s robustness and generalization capability.