基于多位置NWP和门控循环单元的风电功率超短期预测

Ultra-short-term Prediction of Wind Power Based on Multi-location Numerical Weather Prediction and Gated Recurrent Unit

数值天气预报(NWP)对风电功率超短期预测模型精度有着重要影响。为充分利用NWP信息,考虑多个风电场的空间相关性,提出一种基于多位置NWP和门控循环单元的风电功率超短期预测模型。首先,通过随机森林分析多位置NWP信息对风电场发电功率的重要程度,利用累积贡献率提取NWP中的有效信息,将加权的NWP信息与历史功率数据作为预测模型的输入变量。然后,选取改进的灰狼寻优算法对门控循环单元的参数进行优化,建立多变量时间序列预测模型,进行风电场发电功率的超短期预测。最后,选取中国某风电场的实测数据进行算例分析,验证了所提方法的有效性和可行性。

Numerical weather prediction(NWP) has an important impact on the accuracy of ultra-short-term prediction models for wind power. In order to make full use of the NWP information, considering the spatial correlation of multiple wind farms, an ultrashort-term prediction model of wind power based on multi-location NWP and gated recurrent unit(GRU) is proposed. First, the importance of multi-location NWP information on power generation in wind farms is analyzed through random forest method. The cumulative contribution rate is used to extract the effective information in NWP, and the weighted NWP information and historical power data are used as the input variables in the prediction model. Then, the improved gray wolf optimization algorithm is selected to optimize the parameters of the GRU, and a multi-variate time series prediction model is established for ultra-short-term prediction of power generation in wind farms. Finally, the measured data of a wind farm in China is selected for example analysis,and the effectiveness and feasibility of the proposed method are verified.