Research Projects

The broad objective of the proposed project is “Intelligent Approaches for Incipient Fault Detection and Remaining Life Estimation of PMSM”. With advancements in science, engineering, and technology, the variety of electrical machines continues to expand, making them the most widely used machines in our daily lives. Nowadays, different kinds of electric motor drives are broadly utilized in Electric Vehicles (EVs) and Hybrid Electric Vehicles (HEVs), for example, Induction Motor (IM), brushless Permanent Magnet (PM) synchronous motor such as Permanent Magnet Synchronous Motor (PMSM), Brushless Direct Current (BLDC) motor, and Switched Reluctance Motor (SRM). Regardless of how well-designed, calculated, or manufactured electric motor drives are, they inevitably have a tendency to fail over time. Neglecting to monitor the condition of the motor during use or to perform maintenance can result in failures that pose significant economic and safety risks. Even minor faults that may initially seem insignificant can ultimately lead to catastrophic consequences. In real-world applications, motors are exposed to various forms of stress, including environmental, physical, and thermal stress, which can impact their normal structure and operation and potentially lead to faults. Recently, several Computationally Intelligent approaches have been successfully explored for condition monitoring and fault diagnosis. Hence, this project aims to employ Computationally Intelligent approaches like Kernel Principal Component Analysis (KPCA), hierarchical Convolution Neural Networks, Deep Belief Networks, Deep Neural Networks (DNN) Recurrent Neural Networks (RNN), Support Vector Machine (SVM), Compressive Sampling and Subspace Learning (CS-SL), Compressive Sampling and Deep Neural Network (CS-DNN), etc. to perform suitable experimental studies on the Machine Condition Monitoring (MCM) for incipient fault detection and thereafter remaining life estimation of PMSM based electric drives. The project intends to create a suitable framework for the implementation of especially Deep Learning based approaches for MCM for rotating machines by developing suitable ‘deep’ models and validation of results experimentally.