基于改进相似日和ABC-SVM的光伏电站功率预测

POWER FORECASTING OF PHOTOVOLTAIC PLANT BASED ON IMPROVED SIMILAR DAY AND ABC-SVM

为提高光伏电站输出功率预测的精度，以满足电网调度的高精度要求，提出基于改进相似日和人工蜂群算法优化支持向量机的光伏电站功率预测方法。运用熵权法计算得到各气象因素对光伏出力的影响权重，通过计算历史日与待测日气象因数的加权欧氏距离和加权关联度确定相似日，选取相似日光伏输出功率历史数据、温度和湿度以及待测日温度、湿度作为支持向量机的输入变量，采用矗．折交叉验证和人工蜂群算法相结合的方法优化核函数参数和惩罚因子，最终输出光伏电站各时段发电功率的预测值。实验结果表明该方法可有效提高光伏电站功率预测模型的泛化能力和学习能力，具有较高的预测精度。

The method for power prediction of photovohaic plant based on improved similar day and artificial bees colony support vector machine is proposed to enhance the accuracy of photovoltaic plant output power prediction and to satisfy the high accuracy requirement of power grid scheduling. The use of entropy weight method is to calculate the influence of meteorological factors on the photovohaic output weights. Through calculating the weighted Euclidean distance and weighted incidence degree of history day and measured day meteorological factors to determine similar days, selecting historical data of photovoltaic power output, temperature and humidity of similar days and temperature and humidity of test date as input variables of support vector machine, we adopt the method of combining k-fold cross validation and artificial bees colony to optimize kernel function parameters and the penalty factor, and finally get the output in each period of photovoltaic plant power prediction. The experimental results showed that this method can effectively improve the generalization ability and learning ability of photovoltaic plant power prediction model, and it has higher prediction accuracy.