Short-term Prediction of Photovoltaic Power Generation for Distribution Management System based on Spatial Correlation Analysis

Abstract

This paper presents a prediction method of photovoltaic power generation amount as a fundamental technology for distribution system. This method is based on correlation analysis of insolation between two locations. The analysis was made using data measured by actinometers which were installed in several places around a distribution area. Based on the analysis, parameters for prediction which are moving direction and speed of insolation fluctuation were identified. And photovoltaic power generation amount was predicted based on the parameters. From the evaluation using measured insolation data, it found the method considerably improved the prediction accuracy in the case of high correlation through comparison with the conventional method.