Abstract: Power system fault recording data is the basis of power grid fault analysis; and the fault data compression is beneficial to reduce the data storage capacity and improve the data transmission efficiency. In the view of the electric power fault recording data format and its structure characteristics, this paper proposes a novel data compression algorithm based on Fourier transform and wavelet packet transform. Fourier transform is used to compress and reconstruct data of section B of analog signal channel recording. According to the reconstruction error, the fault channels are picked up. The section data with transient disturbance of the fault channel is compressed by wavelet packet transform, and the normal channels data and fault channel other sections are compressed using Fourier transform. A large number of actual recording files are compressed, and the results show that the algorithm can obtain both high compression rate and good compression precision. The algorithm has been applied in the power grid fault analysis system.

Key words: fault recording, data compression, Fourier transform, wavelet packet transform

摘要： 电力系统故障录波数据是分析电网故障的主要依据，录波数据压缩有益于减小数据存储容量和提高数据传输效率。针对电力故障录波数据的格式及构成特点，提出了一种基于傅里叶变换和小波包变换的数据压缩新算法。采用离散傅里叶变换对录波模拟量通道的B时段数据进行压缩和重构，根据重构误差判断该通道是否为故障通道；对故障通道的暂态扰动时段采用小波包变换进行压缩，对正常通道及故障通道的其他时段采用傅里叶变换进行压缩。大量录波文件的压缩结果和工程实际应用表明，所提算法可以同时获得很高的压缩率和压缩精度，具有广阔的应用前景。

关键词: 故障录波, 数据压缩, 傅里叶变换, 小波包变换