Household-Level Temperature Sensitivity Extraction for Residential Consumer Segmentation

As power grids become increasingly decentralized, understanding the temperature sensitivity at household level with high temporal resolution is critical for accurate consumer segmentation and effective demand response implementation. While existing studies have explored temperature-based load modeling for system-level impact analysis, they mostly overlook variations in household-level responses. Additionally, the limited availability of high-resolution real-world data has constrained efforts to quantify the combined effects of temperature and residential demand variations. This paper presents a data-driven approach for extracting household-level temperature sensitivity using 15-minute smart meter data from approximately 9,000 households in the Northern Plains of the U.S. By analyzing consumer-specific heating and cooling consumption patterns, the proposed method identifies temperature sensitivity-based load characteristics for temperature-based consumer segmentation. The insights of this study can further support the development of targeted demand response programs, customized tariff structures, enhancement of load forecasting, and optimized energy management strategies for utility companies.