An Adaptive Robust Square Root Unscented Kalman Filter for Forecasting Aided State Estimation in Power Systems

Le résumé fourni par la source

In this article, an adaptive robust square-root unscented Kalman filter (ARSQUKF) has been suggested for performing the forecasting aided state estimation (FASE) in power systems. Initially, square root unscented Kalman filter (SQUKF) has been implemented to perform power systems FASE. However, due to its inability to provide robust estimates against outliers, the proposed approach further combines the adaptive Huber’s M-estimation approach with square root unscented Kalman filter (SQUKF). The M-estimation technique is made adaptive by introducing the concepts of normalized innovation (NI) and asymmetry index (AI) in it. With adaptive M-estimation being employed, the proposed ARSQUKF provides robust state estimation discriminating bad data and sudden changes in load. Further, the proposed approach can also deal with process noise of bimodal Gaussian mixture (BGM) nature. Simulations results conducted on IEEE 118 bus test system and one Indian real test network i.e. the Northern Regional Power Grid (NRPG) 246 bus system are discussed to portray the robustness of the proposed ARSQUKF throughout the entire time samples. Further, the proposed methodology has been assessed by comparing with various non-robust and robust Kalman filter-based FASE. The results presented explicitly exhibit the supremacy of the proposed technique.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.