基于CEEMD-BiLSTM-RFR的短期光伏功率预测

由于短期光伏预测中气象因素的时间尺度不同，直接分析其对光伏功率的相关性，易忽略时间尺度的影响，进而导致预测模型误差。为提高光伏功率预测精度，构建了预测模型。首先，利用互补集合经验模态分解(complementary empirical mode decomposition, CEEMD)将光伏序列进行分解，得到在不同时间尺度上的光伏分量；然后，通过Pearson相关系数分析各光伏分量与空气温度、太阳辐射度、风速、风向和空气湿度的关系，对于强相关分量建立关于气象因素的随机森林回归(random forest regression, RFR)预测模型，弱相关分量直接通过双向长短期记忆网络(bidirectional long short-term memory neural network, BiLSTM)进行预测；并将预测求和输出。通过安徽省蚌埠市光伏电站7月实测数据进行验证，实验结果表明，所提预测模型CEEMD-BiLSTM-RFR相比传统预测模型有较好的预测精度。

Short Term Photovoltaic Power Prediction Based on CEEMD-BiLSTM-RFR

Since the time scales of meteorological factors in short-term photovoltaic power prediction are different, time scales are usually ignored in the analysis of the correlation between time scales and photovoltaic power, leading to errors in the prediction models. To improve the prediction accuracy of photovoltaic power, the CEEMD-BiLSTM-RFR prediction model is then constructed. Firstly, the photovoltaic power is decomposed by complementary empirical mode decomposition (CEEMD) to get the modalities on different time scales; then, the relationship between each photovoltaic component and meteorological factors is analyzed by Pearson correlation coefficient. Strongly correlated components are predicted by the random forest regression (RFR) prediction model, Weakly correlated components perform prediction through bidirectional long short-term memory neural network (BiLSTM); Finally, the results of each component prediction are combined to obtain the final prediction result. It is verified by using the measured data of a photovoltaic station in Bengbu, Anhui Province, in July. The results show that the proposed prediction model has better prediction accuracy than the traditional prediction model.