【摘要】 为减少训练样本中的冗余数据和特征信息,提高风电功率预测的精度,提出了一种基于核主成分分析(KPCA)、动态时间规整法(DTW)及核极限学习机(KELM)相结合的预测方法(KPCA_DTW_KELM)。首先对影响风机出力的各个因素进行核主成分分析,筛选出贡献率较高的特征因素。为优化样本数据,引入动态时间规整法,找出与预测日相似的样本数据,采用核极限学习机(KELM)进行风电功率预测。根据提出的预测方法,对上海某风电场数据进行对比实验。实验证明,KPCA_DTW_KELM预测模型提高了短期风电功率预测的精度,具有一定普适性。更多还原

【Abstract】 In order to reduce the redundant feature information of training data and improve the accuracy of wind power prediction, this paper proposes a prediction based on the combination of Kernel Principal Component Analysis(KPCA), Dynamic Time Warping(DTW) and Kernel Extreme Learning Machine(KELM) Method(KPCA_DTW_KELM). Firstly, the nuclear principal component analysis was carried out on the various factors that affect the output of the wind turbine, and the characteristic factors with higher contribution rate were screened out. In order to optimize the sample data, the dynamic time warping(DTW) was introduced to find the sample data similar to the forecast day, and the Kernel Extreme Learning Machine(KELM) was used to predict the wind power. According to the proposed prediction method, a comparative experiment was conducted on the data of a wind farm in Shanghai. Experiments have proved that the KPCA_DTW_KELM prediction model improves the accuracy of short-term wind power prediction and has a certain universality.更多还原