The study of method of electric demand prediction for liberalized household electricity market

In Japan, household electricity market is going to be liberalized in 2016. After the liberalization, the needs of demand and supply adjustment by demand sides will rise. Concretely speaking, electric trading style in which demand sides buy electricity on the basis of a demand prediction will be more common in Japan. In this study, we proposed the method of hourly or half-hourly electric demand prediction of the following day in order to control the balance between supply and demand through the electric power system, and test the accuracy of the method by applying it to 2 demand data. One data is the aggregated demand data of the whole consumers contracting with TEPCO (Tokyo Electric Power Company, contracting with about 29 million consumers), and the other data is the household demand data aggregating about 400 households in Kyushu, Japan. The former data is collected every hours, and the latter one is collected every 30 minutes. In the case study using TEPCO data, we cloud predict the demand curve with 3.08% error. On the other hand, in the household case, the error is bigger than 10%.