Fault detection and isolation in power converters using deep learning approaches
Recent advances in Artificial Intelligence (AI) have made it possible to enhance the performance and reliability of power conversion systems as well as to reduce their cost. In addition, AI can help the more efficient use of electricity resulting in environmental protection. A promising further application is AI’s use in the diagnosis of complex power electronics systems for detecting fault and pre-fault situations.
1. Study the literature of deep neural networks and their applications for fault detection
2. Get acquainted with a deep learning framework and set it up
3. Study the normal and faulty operation of a selected power electronics converter (e.g. a dc-dc converter)
4. Design and implement a deep learning network capable to identify various fault types
5. Demonstrate a good ability of fault detection by simulations and possibly by laboratory tests
6. Sum up the experiences