基于有限混合Laplace模型的风功率波动特性研究

Wind Power Fluctuation Characteristic Analysis Based on Finite Laplace Mixture Model

风电功率具有波动性,准确描述风电场功率波动规律对于研究风电大规模并网运行具有重要意义。首先基于风功率波动概率分布的重尾特性,提出有限混合Laplace分布(finite Laplace mixture,FLM)的概率分布模型,并求解了模型最优参数集。然后提出评价模型精确度指标体系,并基于大量风电场实测数据,用FLM模型分别与混合高斯模型及单一分布模型作对比,论述并证明了FLM模型能更精确地描述风场功率波动特性。在此基础上,分析在不同时空尺度下的风功率波动情况,总结了风功率波动在多时空尺度下的规律特征。实验结果表明,FLM模型相较于其他分布模型,具有更高的拟合精度和更普遍的适用性。

Wind power is characterized with fluctuation. Accurate description of wind power fluctuation characteristics is of significance for large-scale grid-connected wind power operation. Firstly, based on heavy-tailed characteristics of wind power fluctuation distribution, this paper proposed finite Laplace mixture （FLM） probability model and solved optimal model parameters. Then, a series of evaluation indices were designed to indicate model accuracy. Based on large amount of wind farm measurement data, and compared FLM model with Ganssian Mixture Model and other single distribution function model, it was proved that FLM model could more accurately describe wind power fluctuation characteristics. Therefore, through analysis of wind power fluctuation with different temporal and spatial scales, valuable rules about wind power fluctuation were discovered. Experimental results show that FLM model has higher fitting accuracy and more general applicability compared with other distribution models.