High-resolution determinant analysis of Japanese residential electricity consumption using home energy management system data

This paper examines seasonal panel data analysis at a higher resolution using commercial home energy management system data to identify the parameters that determine Japanese household electricity consumption. The electricity consumption data from 532 detached houses and 208 apartment-style houses is aggregated by use and time period and regressed with explanatory variables to indicate house and occupant attributes. Predictable significant impact factors such as outdoor temperature, floor area, household size, presence of a central air conditioning system, and a variety of appliances are estimated quantitatively. The differences due to appliance possession are estimated as 844 kWh/year for a water server, 885 kWh/year for an additional refrigerator, 491 kWh/year for a portable humidifier, and 443 kWh/year for an air purifier. We found previously unknown correlation between variables such as central air conditioning and hot water demand through this model. We also identified several important parameters that explain electricity demand by data collecting at different sites and time periods. Obtained knowledge will contribute to promoting further energy efficiency program in the Japanese residential sector. Households in our study were newer houses and additional data are required for a more stable and reliable model. In addition, a reverse estimation of household attributes from electricity load dynamics is an issue for future study.