基于集成机器学习模型的短期光伏出力区间预测

Short Term Photovoltaic Output Interval Prediction Based on Integrated Machine Learning Model

为全面深挖影响光伏出力因素之间的关联信息,进一步提高机器学习模型在短期光伏出力区间预测的精度,提出一种基于集成机器学习模型的短期光伏出力区间预测方法。首先,利用快速相关性过滤(FCBF)的特征选择算法对多维的历史光伏数据及气象数据进行最优特征的提取;然后,在集成多个机器学习模型的基础上,收集训练过程中的预测误差,通过最大似然估计获取预测误差的概率分布,得到预测区间的上下限;最后,结合集成学习模型预测得到光伏出力曲线,进而得到最终的日前光伏出力预测区间。最后通过算例验证了所提模型的可靠性与优越性。

To comprehensively explore the correlation information between different factors affecting photovoltaic output,the paper proposes a short term photovoltaic output interval prediction method based on integrated machine learning model to further improve the accuracy of the short-term photovoltaic output interval prediction with the machine learning model.Firstly,the fast correlation based filter(FCBF)is used to extract the optimal features from multidimensional historical photovoltaic data and meteorological data.Then on the basis of integrating multiple machine learning models,the prediction errors during the training process are collected,and the probability distribution of the prediction errors is obtained through maximum likelihood estimation,thereby obtaining the upper and lower limits of the prediction interval.Finally,the photovoltaic output curve is got by combining the integrated learning model prediction,and the final day-ahead photovoltaic output prediction interval is obtained.The reliability and superiority of the proposed model are verified through the examples.