结合数据增强及组合算法的短期光伏功率预测

Short-term Photovoltaic Power Prediction Using Data Augmentation and Combined Algorithms

针对光伏发电数据完备性低、预测精度低的问题,提出结合数据增强及组合算法的短期光伏功率预测模型。首先,利用K-means++聚类算法对光伏数据进行天气分型;其次,利用条件生成对抗网络对光伏数据的分布规律进行学习,生成高质量样本;然后,优化变分模态分解的分解数和惩罚因子,进一步降低子序列的模糊熵值;最后,通过正余弦算法对深度极限学习机的输入权重和偏置进行优化,分别对各子序列进行建立预测模型。实验结果表明,所提模型具有一定的优越性。

A short-term photovoltaic(PV)power prediction model based on data augmentation and combined algorithms is proposed to address the problems of insufficient data completeness and low power prediction accuracy in PV power generation data.First,the PV data is divided into different weather types using the K-means++ clustering algorithm.Second,conditional generative adversarial networks are used to learn the distribution patterns of PV data and generate high-quality samples.Third,the decomposition number and penalty factor of variational mode decomposition are optimized to further reduce the fuzzy entropy value of subsequences.Finally,the input weights and biases of the deep extreme learning machine are optimized by using the sine cosine algorithm,and predictive models are established for each subsequence.Experimental results show that the proposed model has its superiority。