GCA-BULF: A Bottom-Up Framework for Short-Term Load Forecasting Using Grouped Critical Appliances

Researchers have developed GCA-BULF, a bottom-up framework for short-term load forecasting that groups only critical appliances to improve accuracy while reducing monitoring costs. The framework uses three key modules: Critical Appliance Filtering ranks appliances by power consumption, switching frequency, and usage pattern periodicity, then identifies critical ones through iterative load decomposition. Next, Related Appliance Grouping clusters these critical appliances based on spatial and temporal correlations for group-level forecasting. Finally, Collaborative Load Forecasting combines multiple group-level forecasts to refine total load predictions.

In tests on residential and office building datasets, GCA-BULF improved hourly total load forecasting by 20.85%-57.88% compared to existing top-down methods and by 33.03%-92.48% compared to bottom-up methods. This approach supports peak-shifting strategies for time-of-use and tiered electricity pricing, helping consumers lower energy costs while enhancing grid stability. The work is detailed in a paper on arXiv (2604.24766).