Research On Date Compression Technology For Network Packets Of The Smart Substation

With the continuous development of smart substation and constant improvement of the high-speed network communication technology, network packets of the smart substation become the main interaction and sharing information of the intelligent devices, in which SV packets and GOOSE packets are two of the higher real-time ones, these packets contain the normal operation of the power system,accidents, protection action command which is issued, the removal of power system fault, etc, which are related to many important aspects such as the sampling, handling, controlling of the protection and monitoring device. Capturing, storing them and reproducing the procedure after an accident can paly an important role in the processing and analysis of power system faults.The density and flow of the network packets of the smart substation especially the SV packets is very large, a merging unit (MU) network bandwidth is about6Mbit/s, a smart substation often has dozens of MU, real-time recording and storing the packets can generate massive data, thus higher requirements are proposed to the data storage devices. In order to improve the capacity and efficiency of data storage, this paper studies the compression algorithm which is adapted to the characteristics of the network packets, relieves stress of the storage of the network packets.Firstly, this paper analyzes the principle of several classic compression methods and uses them to compress the network packet files,with the results of the compression,it selectes deflate algorithm as the basic compression algorithm; secondly,elaborateds the implement of every module and main function of the deflate algorithm, analyzes the factors of the LZ77compression module such as window size, characters matching scheme and code identification, and analyzes the structure and characteristics of network packets through packet analysis tool-wireshark; finally, this paper makes the analysis of the impacts on the deflate algorithm LZ77module through the characteristic of the network packets, and proposes a context-based prediction deflate algorithm according to its data structure, character handling of the context-based prediction is added to the character matching of the LZ77algorithm. Compression test results show that the context-based prediction deflate algorithm improves the compression of the network packets. And the window size is set to16KB,32KB and64KB to compress the network packets, the test results show that the use of64KB size window gets better results, meanwhile dosen’t increase the running time of the algorithm.So the window size of the improved algorithm is set to64KB. Then the compression test results show that with the improved algorithm,the statistics compression ratio of the SV packets is11%, the statistics compression ratio of the GOOSE packets is about5%. Compared to the original algorithm, the statistics compression ratio reduces about3%, the algorithm improves the deflate compression of the network packets, while the SV packets compression comes to9times and the GOOSE gets20times, greatly reduces the pressure of the network packets storage of the smart substation.