Study On Data Compression For Fault Recorder In Power System

In power system, the quantity of the data recorded by digital fault recorder (DFR) over the period of disturbance incident is very large. The time of the data transmission to the control center is very long under the condition of current communication equipment. To solve the problem, this paper has discussed theoretically and made experiments on the recorded data compression.On the base of analyzing the popular data compression algorithm, the paper discusses two approach of the recorded data compression in detail. The first one is a lossless algorithm called liner prediction differential and static Huffman joint encoding. The algorithm first removes the data correlative between the samples by replacing the sampling data with their differential value and then reduces the data redundancy by static Huffman encoding. The second one is a lossy algorithm called wavelet transform and LZW joint encoding. Wavelet transform has good multi-resolution ability both on the time domain and frequency domain, so it is very useful for recorded data compression. In order to get better compression effect, LZW algorithm is introduced to encode the quantized wavelet coefficients.The results show that the first approach is easy to be carried out and its compression ratio is about 20%. Compared with the first approach, the second approach is a lossy algorithm and it has a better compression effect, but it:, procedure is too complicated.