基于时间窗口聚类的电力时序数据压缩研究

Research on power time series data compression based on time window clustering

为节约电力时序数据的存储空间,有效实现对电力网络的扩容处理,提出了基于时间窗口聚类的电力时序数据压缩方法。根据时间窗口聚类条件,分别求解时间子序列与窗口子序列条件,从实时存储库中提取电力时序数据样本。定义压缩性能指标,通过稀疏变换数据矩阵的方式,实现对电力时序数据的压缩处理。对比实验结果表明,时间窗口聚类模型可将电力时序数据存储空间控制在3.0×10~7bit内,对于节约数据存储空间、实现电力网络的扩容起到了一定的促进性作用。

In order to save the storage space of power time series data and effectively expand the capacity of power network,a method of power time series data compression based on time window clustering is proposed.According to the time window clustering conditions,the time subsequence and window subsequence conditions are solved respectively,and the power time series data samples are extracted from the real-time repository.The compression performance index is defined,and the power time series data is compressed by sparse transformation of the data matrix.The comparison experiment results show that the time window clustering model can control the storage space of power time series data within 73.0×10 bit,which plays a certain role in saving data storage space and realizing the expansion of power network.