A study on estimation of aggregated electricity demand for one-hour-ahead forecast

Abstract

Japanese electricity retail market is fully liberalized in April 2016 and new entry electricity supplier companies make an electricity supply schedule, which target several hundred or more of households. A real time measurement (or now-casting) of the aggregated electricity demand of all households or all distributed micro generation’s outputs are important for them to reduce the electricity supply and demand imbalance. They are able to do it by totaling each household’s demand data or each micro generation power output data that are measured by smart meters or energy management system. However, it seems very costly to now-cast the electricity demand of several hundred or more of households because they need to process a large amount of data.For reducing such processing loads, we have been developing an “upscaling model” which estimates the aggregated electricity demand of several hundred households using a small number of representative households’ electricity demand. The proposed upscaling model consists of a method of selecting representative households and a method of estimating aggregated demand.In this paper, we modify the upscaling model that is proposed in our preceding study. In addition, we combine the upscaling model with one-hour-ahead forecast model to modify the forecasted electricity demand by a day-ahead forecast model.