Data-driven Mid-Term Load Forecasting Based on Topological Data Analysis
Yang Han, Ruiyao Jia, Wenpeng Luan, Bochao Zhao, Hanju Cai, Bo Liu · 2024
Load forecasting has always held a pivotal position within the strategic planning and day-to-day management of electric utilities, encompassing both transmission and distribution entities. As technology evolves, economic conditions fluctuate, and myriad other influential factors come into play, the significance of accurate load forecasting continues to escalate. Moreover, as emerged as an indispensable tool for modern power systems, particularly in nations where the power sector operates within a deregulated framework. In our paper, we propose a data-driven method based on topological data analysis and we apply it to predict load level in a span of 15 days quarterly hours. Our result show it is feasible to use topological data analysis as feature extractor in load forecasting