基于COOT-SVM的短期光伏发电功率预测

Short-term Photovoltaic Power Forecasting Based on COOT-SVM

摘要:

为了提高短期光伏发电功率预测的精度，提出了一种基于白冠鸡优化算法(COOT)优化支持向量机(SVM)的短期光伏发电功率预测模型。首先，分别选取某光伏电站在2017年4月和7月的前21天数据进行仿真分析，计算光伏输出功率和每一个气象因素之间的皮尔逊相关系数；然后，依据皮尔逊相关系数选择太阳总辐射强度、太阳散射辐射强度、太阳直射辐射强度、组件温度和环境温度5个气象因素作为预测模型的输入数据，光伏电站的发电功率作为输出数据。通过与BP和SVM预测模型进行仿真对比可知，对于4月和7月的数据来说，COOT-SVM预测模型的均方根误差、均方误差和平均绝对误差均比BP和SVM预测模型小。因此，所提COOT-SVM预测模型可有效提高短期光伏发电功率的预测精度，具有较高的工程应用价值。

Abstract:

In order to improve the accuracy of short-term photovoltaic power forecasting, a short-term photovoltaic power forecasting model based on support vector machine (SVM) optimized by coot optimization algorithm (COOT) is proposed. Firstly, the data of a photovoltaic power station in the first 21 days of April and July in 2017 are selected for simulation analysis, and the Pearson correlation coefficient between photovoltaic output power and each meteorological factor is calculated. And then the total solar radiation intensity, solar scattered radiation intensity, solar direct radiation intensity, component temperature and ambient temperature are selected as the input data of the prediction model, and the generated output of photovoltaic power station is used as the output data. Through simulation comparison with BP and SVM prediction models, it is obtained that for the data in April and July, the root mean square error, mean square error and mean absolute error of COOT-SVM prediction model are smaller than those of BP and SVM prediction models. Therefore, the proposed COOT-SVM prediction model can effectively improve the prediction accuracy of shortterm photovoltaic power generation, which has high engineering application value.