不同时间尺度下风电功率波动信息的损失及其影响的研究

A Research for the Information Loss and Its Influence of Wind Power Fluctuations in Different Time Intervals

摘要: 短期预测的时间间隔为15分钟，但从秒级到分钟级的风电功率波动信息对电力系统的控制尤为重要，本文从幅值、小波、信息熵三方面，研究了不同时间间隔下风电波动信息损失的规律。首先，基于某风电场的实测数据，对比了不同时间间隔下风电功率波动的变化情况，发现较大时间间隔下损失的多为高频信息。然后对风电功率的幅值进行分析，指出了随时间间隔的增大，信息损失会越来越大。再基于小波变换和相关性分析，发现了每台机组损失频段的一致性，证明了高频分量的损失及其在波动分析中的重要性。接着运用信息熵理论，说明了随时间间隔的增加，波动信息的混乱程度会降低，同时价值也会降低。接下来，通过对不同时间间隔功率预测误差的对比，发现时间间隔越大，预测误差越大。最后，基于以上的分析，总结了风电波动信息损失的影响。为风电功率波动的进一步研究，提供了新思想。

Abstract: The time interval of short-term wind power forecasting is 15 minutes, but the information of wind power fluctuation from minute to second is important for the controlling of power system. Considering three aspects of the amplitude, wavelet and entropy of information, the paper studies the law of the information loss of wind power fluctuation in different time interval. It firstly compares change of wind power fluctuation in different time scale based on measured data of certain wind farm, finding most of the loss is the information of high frequency in large time scale. Then, for the analysis of the amplitude, the paper proposes that the information loses more as time interval is increasing. Next, the consistency of the loss band of each unit is found and the loss and importance of the part of high frequency are proved according to the wavelet analysis and the interdependency. The degree of confusion about the volatility of information is reduced. And the value of information decreases as time interval increases based on the entropy of information. Furthermore, through comparing with the errors of predicting power in different time scale, the error of predicting power is more as the time scale increases is found. Finally, based on the above analysis, the influence on the information loss of wind power fluctuation is summarized, providing a new idea for the further research of wind power fluctuation.