Residential power demand side management optimization based on fine-grained mixed frequency data

Demand-side response (DSR) measures, which facilitate the management of system reliability and maintain system resource adequacy of power grids, are gaining prominence. Traditional optimization methods for residential electricity usage in peak-demand hours often lack flexibility and effectiveness in solving practical problems due to a large amount of data and a high degree of data uncertainty. Our study aims to obtain a high degree of responsiveness for numerous and highly dispersed residential customers by integrating data-driven and model-driven methods. In this study, we conducted large-scale DSR controlled trials, collected 15-min intervals of electricity consumption data, and matched the survey data. A dynamic time-warping clustering-based difference-in-difference method was proposed for empirical analysis. The results indicate that monetary incentives induced a statistically significant reduction in electricity usage during peak periods in summer by 18.6%, while the response effect was not significant in winter. The heterogeneity analysis suggests that the reduction was mainly contributed by the middle-income group and the elderly group. Our study provides a basis for policymakers to formulate tailored policies to optimize residents' electricity consumption during peak hours.