基于VMD-GWO-ELMAN的光伏功率短期预测方法

以进一步提高光伏输出功率短期预测的准确性和可靠性为目标，针对传统Elman神经网络权值和阈值盲目随机的缺点以及光伏输出功率信号波动性和非平稳性的特点，提出一种基于变分模态分解（VMD）和灰狼优化算法（GWO）优化Elman神经网络的光伏输出功率短期预测模型。首先，使用K-means算法对原始数据按天气类型进行聚类；然后，使用VMD对每一类型天气光伏输出功率数据进行分解，分别将各分解子序列输入经GWO优化的Elman神经网络进行光伏输出功率预测；最后，将各预测结果进行叠加。实例证明：该模型的预测精度有所提升。

PV Power Short-Term Forecasting Method Based on VMD-GWO-ELMAN

This paper aims to further improve the accuracy and reliability of short-term photovoltaic (PV) output power forecasting. Considering the blindness and randomness of weights and thresholds of traditional Elman neural networks and the fluctuation and nonstationarity of PV output power signal, the paper proposes a short-term prediction model of PV output power based on variational mode decomposition (VMD) and an Elman neural network optimized by grey wolf optimization (GWO) algorithm. Firstly, the K-means algorithm is used to cluster the original data according to weather types. Then, VMD is employed to decompose the PV output power data of each weather type, and the decomposition subsequences are input into the Elman neural network optimized by GWO for PV output power forecasting. Finally, the forecasting results are superimposed. An example shows that the model has improved forecasting accuracy.