کنفرانس مهندسی برق ایران

صفحه اصلی / سی و چهارمین کنفرانس بین المللی مهندسی برق

High Resistance Connection Fault Diagnosis of BLDC motors Using slow feature analysis with NARX neural network

نویسندگان :

Amirhossein Ashgzaran¹ Mohsen Montazeri² Mojtaba Nourimanzar³

1- دانشگاه شهید بهشتی 2- دانشگاه شهید بهشتی 3- دانشگاه شهید بهشتی

کلمات کلیدی :

Brushless direct current (BLDC)،Fault detection،high resistance connection(HRC)،slow feature analysis،recurrent neural networks،NARX model

چکیده :

High resistance connection (HRC) faults, which arise from degraded or loose electrical connections in stator phases, create current/voltage asymmetries that degrade Brushless direct current (BLDC) performance. HRC Faults often Result in unbalanced currents, increased losses, and torque pulsations that may evolve into open-phase faults and thermal hazards. Fault diagnosis can be incorporated into motor drive systems to detect such faults. This article proposes an online neural network based method to detect HRC faults economically and reliably using input current. this method of fault diagnosis is performed by incorporating Slow Feature analysis (SFA) and Nonlinear autoregressive exogenous (NARX) Recurrent neural network to provide meticulous fault detection in all operating conditions of BLDC motor.