Neural network-based model for estimation of solar power generating capacity

Solar energy is one of the most promising renewable energy sources. The generating capacity of this source however is highly dependent on the available sunlight, its duration and intensity. In order to integrate these types of sources into an existing power distribution system, system planners need an accurate model that predicts its generating capacity with the usage of easily accessible information. In this paper, three methods are used to estimate global irradiation received on a tilted surface; mathematical model, regression models and neural network analysis. From the results obtained, the regression model provides the most superior performance.