Energy consumption prediction for households in a society with an ageing population

Abstract

Social aging significantly impacts household energy consumption patterns and demand, particularly in megacities like Shanghai. This study addresses the gap in understanding high-frequency impacts of aging on energy use by employing advanced machine learning techniques. Using Gaussian Mixture Models (GMM) and Finite Mixture Models (FMM), we analyze high-frequency hourly energy consumption data from 14,000 households in Shanghai (2016–2023) to identify distinct consumption patterns and their relationship with household characteristics. The study also simulates future scenarios incorporating demographic aging and income growth. The results reveal that an aging society not only increases overall energy demand but also significantly alters hourly consumption patterns, amplifying disparities between peak and non-peak hours. These shifts, compounded by income growth, highlight the need for tailored energy policies addressing demographic transitions. This research contributes to sustainable energy planning by providing actionable insights into the intersection of aging demographics, economic development, and urban energy consumption. The findings align with the United Nations Sustainable Development Goals (SDGs) by promoting efficient and inclusive energy strategies.