基于动态时间规整的负荷主要生产时段特征提取方法

A method of characteristic extraction for the main production period of load based on dynamic time warping

针对经济活跃程度和复工复产重要评估依据中负荷主要生产时段提取偏差大的问题,文章提出了基于动态时间规整(DTW)的负荷主要生产时段特征提取方法,通过构建参考负荷曲线,将测试负荷曲线与参考负荷曲线进行时间维度的伸缩变换,达到测试与参考负荷曲线间的最佳映射,进而实现了测试负荷主要生产时段的测算。文章以西南某区域典型商业负荷功率数据为案例进行负荷主要生产时间段提取,分析了异常测量数据对动态时间规整算法辨识主要生产时段的影响,实验结果证明了文章所提方法在辨识主要生产时段方面的有效性和准确性。

Aiming at the problem of large deviation in the extraction of the main production period of load in the evaluation of economic activity level and work resumption degree,this paper proposes a method of characteristic extraction for the main production period of load based on dynamic time warping(DTW).After constructing a reference load curve,the DTW method is adopted to realize the optimal mapping of the test curve and the reference curve by the stretch out or draw back the two curves in time dimension.Then,the main production period of test load could be calculated by searching in the optimal warping path table.This paper takes the typical commercial load power data in southwest China as a case to extract the main production period of the load,and analyzes the influence of abnormal measurement data on the recognition of main production period.Finally,the experimental results show that the proposed method is effective and accurate in identifying the main production period.