基于特征曲线提取的风电数据跨度选取

Selection of Wind Power Data Span Based on Feature Curve Extraction

为减少风力发电技术中海量数据计算容易导致计算灾害的问题,提出一种以有限跨度数据代替海量风电数据的计算方法。通过K-means算法对风电数据进行分类,在分类曲线的基础上对各类风电数据进行特征提取,得到表征特征曲线的最佳类采样天数,最后通过移动"跨度窗"方法求取出能够代替海量风电数据的典型数据跨度。将计算结果应用于某风电场的容量配置计算,通过对比验证了所提出的理论的正确性。

In order to reduce the problem of calculating disaster caused by the large amount of data calculation in wind power generation technology,this paper presented a method to replace the massive wind power data with finite span data.The method to classify the wind data through the K-means algorithm,based on the classification of various types of wind power curve data for feature extraction,it extracted the best characterizations of characteristic curve of sampling days,finally by moving the“span window”method to compute the typical data span instead of massive wind power data.The results are used to calculate the capacity configuration of a wind farm,and the validity of the proposed theory is verified by the comparison of the calculated results.