Abstract: In this talk, I will present my research work on fast and accurate methods for fault detection and isolation as a pre-emptive measure in multi-terminal dc systems (MTdc). This is essential for subsequent removal of dc faults to make the system reliable and stable. Present detection methods are sensitive to changes in network parameters and load variations and lead to false positives or have slow detection rates. My research pertains to the fast and accurate detection of faults to trigger specially designed dc breakers, followed by a fault location approach. Three specific research contributions will be emphasized in my presentation: (a) application of wavelet transform that looks at specific fault frequencies using multi-resolution analysis, (b) methods which use local measurements to isolate faults within the breaker protection zones without the need for a communication channel, and (c) a non-intrusive approach for fault location in MTdc networks. Various sensitivity analysis experiments were performed on the MTdc systems to quantify the efficacy of the developed fault detection and location algorithms.
Bio: Bhaskar Mitra is currently a PhD student in the Electrical and Computer Engineering Department of University of North Carolina at Charlotte. He received his B.Tech from West Bengal University of Technology, West Bengal, India in 2013 and MS from University of North Carolina at Charlotte in 2018. He has worked at PNNL as a summer intern in 2017, 2018, and 2019 at PNNL, with work spanning data-driven and machine learning methods, addressing data deficiencies, and data engineering in power systems, in addition to his work on system protection. His research interests include protection of multiterminal systems and developing new protection coordination and fault localization schemes for MTdc systems.