基于支持向量机回归的光伏发电出力预测

Photovoltaic power generation output forecasting based on support vector machine regression technique

建立总峰瓦值为30MW的光伏电站数学模型,并且基于保定地区气象资料以及美国国家航空和宇航局(NASA)提供的保定地区太阳辐射数据,模拟得到该光伏发电系统出力数据。分析光伏系统出力特性以及影响光伏出力因素。根据影响光伏出力的诸多因子的复杂性和非线性,决定预报因子与预报对象间的非线性关系,建立光伏系统出力的支持向量机(SVM)回归模型,并进行相应的预测。预测结果表明,支持向量机回归(SVR)方法为解决光伏系统出力的预测提供了一种可行路径。

This paper established a large-scale photovoltaic power generation system output model that has the total power of 30MW, and analyzed the PV system output characteristics and the factors that impact the photovoltaic output, and obtained the PV system output based on the meteorological data of Baoding region and the radiation data provided by the National Aeronautics and Space Agency（NASA） for Baoding region. The complexity and non-linearity of the impact of factors contributed to the PV system decide the non-linear relationship between the forecast factors and the forecast object. This paper established a photovoltaic system output forecast model based on the SVM regression method, the practical examples show that the SVM can provide an effective and feasible way to forecast PV system output.