Interval prediction of photovoltaic power generation based on cloud theory

Abstract

With the increasing integration of photovoltaic power into the power system, the reliable photovoltaic power generation prediction is of significant to the security and economics of the power system operation. However, the prediction deviation is unavoidable, therefore this paper presents an interval prediction model for the photovoltaic power generation based on cloud theory. Based on the error analysis of the photovoltaic power prediction, the training samples can be selected to establish the cloud model for each small power generation bin. In this way, the predictive cloud distributions for different predictive power generation can be obtained to generate the prediction intervals at each time slot. To take the photovoltaic power plant in German as example, the proposed model is validated. The results show that the proposed model outperforms the other benchmark method. And the calculation process of the proposed model is simple and short in computation time. The analysis results are expected to be used in the field of power grid dispatching and decision making.