Model-Based and Data-Driven Contingency Filtering for Real-Time Corrective Security-Constrained Economic Dispatch

Donghai Han,Yaowen Yu,Dingxin He
DOI: https://doi.org/10.1109/ei252483.2021.9713313
2021-01-01
Abstract:Security-constrained economic dispatch (SCED) is one of the most important problems in power system operations. Corrective SCED (CSCED) is a type of SCED that considers corrective capabilities of the power system and adjusts base case decisions according to post-contingency states. Because of the large scale of the CSCED problem, it is difficult for purely model-based approaches to meet the time limits in real-time practical operations. To leverage historical operation data, this paper develops a novel hybrid model-based and data-driven framework to accelerate the solution process of CSCED. In the offline stage, our previous model-based contingency filtering approach is utilized to label active statuses of contingencies against historical net load samples. In the online stage, a multi-label classifier based on the K-Nearest Neighbor (KNN) algorithm quickly generates an active contingency set corresponding to the real-time net load. This active contingency set can be used to calculate the economic dispatch decisions in one shot. In addition, the accuracy of our framework can be further improved by adding an optional contingency filtering procedure at the end. Numerical testing results on the IEEE RTS-96 system demonstrate the accuracy and computational efficiency of the hybrid framework as compared to a purely model-based approach.
What problem does this paper attempt to address?