Application of neural network to 24-hours-ahead generating power forecasting for PV system

Abstract

In recent years, there have been focus on environmental pollution issue resulting from consumption of fuel, e.g., coal and oil. Thus, introduction of an alternative energy source such as solar energy is expected. However, insolation is not constant and output of photovoltaic (PV) system is influenced by meteorological conditions. In order to predict the power output for PV system as accurate as possible, it requires method of insolation estimation. In this paper, the authors take the insolation of each month into consideration, and confirm the validity of using neural network to predict insolation by computer simulations. The proposed method in this paper does not require complicated calculation and mathematical model with only meteorological data. © 2008 The Institute of Electrical Engineers of Japan.