基于完全集成经验模态分解和模糊熵分频的短期风电功率预测

随着风电接入电力系统的比例日益增大,准确的风电功率预测显得愈发重要。为此,提出了一种基于模糊熵和完全集成经验模态分解(complete ensemble empirical mode decomposition with adaptive noise,CEEMDAN)的短期风电功率预测模型。采用完全集成经验模态分解将原始风电功率序列进行分解,得到一系列不同频率的子序列。再使用模糊熵(Fuzzy Entropy,FE)算法识别各频率分量特征,将子序列分量分为高、中频分量类和趋势项。趋势项为低频分量,具有较为平稳,波动性小的特点,采用麻雀算法(sparrowSsearch algorithm,SSA)优化支持向量回归(support vector regression,SVR)进行预测；高、中频分量的波动性大且特点较为复杂,则采用SSA优化长短期记忆神经网络(Long Short-Term Memory,LSTM),同时引入注意力机制(Attention Mechanism,AM)对重要信息进行更好的权值分配。最后,经过实验结果分析表明,该模型具有更高的风电功率预测精度。

Short-term wind power prediction based on CEEMDAN and fuzzy entropy frequency division

With the increasing proportion of wind power connected to the power system, accurate wind power prediction becomes more and more important. To this end, a short-term wind power prediction model based on fuzzy entropy and complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) is proposed. The complete ensemble empirical mode decomposition is used to decompose the original wind power series to obtain a series of sub-series with different frequencies. The Fuzzy Entropy (FE) algorithm is then used to identify the characteristics of each frequency component, and the subsequence components are classified into high and medium frequency component classes and trend terms. The trend term is the low frequency component, which has the characteristics of smoothness and low volatility, and the sparrow search algorithm (SSA) is used to optimize the support vector regression (SVR) for prediction; the high and medium frequency components have high volatility and complex characteristics, and SSA is used to optimize Long Short-Term Memory (LSTM) and Attention Mechanism (AM) are introduced to better assign weights to important information. Finally, the analysis of experimental results shows that the proposed method has higher accuracy in wind power prediction.