Abstract
Researches have been done on systems that use heat pump water heaters and other appliances to compensate for PV imbalances. In such systems, pinpoint solar irradiance forecast is needed for each PV facility in order to carry out the planning of electricity supply and demand. However, there are still few studies that cover pinpoint solar irradiance forecast. Therefore, in this study, using the forecasted and measured solar irradiance data from 21 locations in Miyama City, the accuracy of pinpoint solar irradiance forecast has been verified and analyzed from several different perspectives. We confirmed that smoothing effect exists even within small areas. The relationship between season, lead time, data release time and forecast accuracy has been investigated. We also found out that Laplace distribution is better suited than normal distribution to represent the forecast error distribution.
