Study on Predicting Solar Radiation Using Pattern Matching for Searching Similar Weather

Abstract

A photovoltaic solar energy system has been widely utilized as an alternative energy source to fossil fuel at a residential area. We have so far introduced a hybrid generation system that is a combination of a photovoltaic cell （PV）and a fuel cell （FC）with electrolyzer （EL）as the energy storage device. It is necessary to make a practical use plan of FC and EL in order to minimize operation costs. Since the output of PV changes according to weather conditions，it often deteriorates the reliability of the plan. In this paper，we propose a method for predicting the amount of solar radiation 24 hours ahead for PV generation. This method extracts data similar on a prediction target day from the past statistical one. The simulation result shows that the accuracy was more than equivalent as compared with the prediction method of others.