天气类型聚类的支持向量机在光伏系统输出功率预测中的应用

Application of Support Vector Machine Based on Weather Type Clustering in Power Output Forecasting of Photovoltaic Generation System

光伏发电具有较强的波动性和随机性的特点,大容量光伏发电接入,会对电力系统的安全稳定运行带来严峻挑战。本文分析了温度、湿度等气象因素对光伏发电系统输出功率的影响,结合光伏系统的历史发电数据与气象信息,提出一种基于天气类型聚类的支持向量机预测模型。通过计算合适的权值,确定各气象因素的加权欧氏距离,选择输入样本,使样本能更好地反映预测日的天气属性;在此基础上运用支持向量机进行短期输出功率预测,并利用某地实测数据对训练好的模型进行了测试与评估。结果证明,该方法建立的模型具有较高的精度。

Due to the growing demand of renewable energy, photovoltaic （PV） generation system has developed rapidly in recent years. However, with the strong randomness of the solar radiation, the introduction of large- capacity PV power could pose serious challenges to the operation security and stability of power system. In this paper, the effect of such weather factors as temperature and humidity on the power output of PV generation is investigated. Then, sup- port vector machine forecasting model based on weather type clustering is proposed by combing the history generation data of PV system and weather information. Through the calculating of proper weights, the weighted Euclid distance of each weather factor is determined, and the input samples are selected to better reflect weather characteristics of fore- casting day. Furthermore, short-term power output is pre- dicted by SVM, and the trained model is tested and evalua- ted by measured data. In the end, the results show that pro- posed method has better forecasting accuracy.