Research on the Prediction of Solar Energy Generation based on Measured Environmental Data

Abstract

As a kind of renewable energy, solar power becomes more and more widely used as the power supply for large-scale datacenters to save the brown energy consumption and to reduce the overall cost. The prediction accuracy of solar energy generation becomes a fundamental issue in the research of how to efficiently manage the renewable energy resources. This paper explores the possible ways to predict the solar radiation intensity based on the assumption that it impacts the solar power generation proportionally. Through the analysis and research of photovoltaic power generation system, we explore the influence factors for solar radiation intensity, establish a relationship of solar radiation intensity and ambient temperature, time, humidity, wind speed in the forecasting model, and finally established the multivariate linear regression model and artificial neural network model. According to the two models, the environmental monitoring data measured at the Qinghai University are employed as the basis of the prediction of solar radiation intensity, and compared with the actual measurement data monitoring system. Experimental results show that, by using the BP neural network prediction model, the achieved accuracy is higher than other empirical model. The prediction method and good results provide a necessary foundation for future related research based on solar radiation values forecasts.