基于近似熵的风电功率可预测性研究

PREDICTABILITY OF WIND POWER BASED ON APPROXIMATE ENTROPY

首先提出风电功率时间序列的可预测性概念,并利用近似熵和可预测系数对可预测性进行定量分析,在此基础上分析不同空间尺度下近似熵和可预测系数的变化规律,最后利用吉林省西部某风电场的实测数据进行验证,几种典型预测方法下的计算结果表明,该文提出的指标能有效表征风电功率的可预测程度。

The predictability concept of wind power on the time series was presented and the approximate entropy algorithm and predictable coefficient were used to quantificationally analyze the predictability of wind power on time series, then the approximate entropy and predictable coefficient variation at different spatial scale were analyzed. Finally, the measured data of a wind farm in western Jilin were used for testing and confirming the result. The results of several typical prediction methods show that the proposed method can effectively characterize predictability of wind power on time series.