Design And Key Algorithm Research Of High Performance Power Big Data Server Software System

The existing domestic big data analysis platform can only process large-scale disk historical data and cannot process real-time power data.However,the application research of foreign power big data real-time processing technology in the power industry is relatively in-depth,and there is already the ability to perform massive data on the power system.The actual system of storage and simple processing is put into operation.Moreover,with the gradual maturity of the power system wide-area measurement system(WAMS)based on Phasor Measurement Unit(PMU)and the universal application of power dispatch centers at or above the provincial level,PMU/WAMS has become necessary for power dispatch automation systems.The components of the electric power industry serve as basic energy facilities for the country.Traditional power big data processing methods can no longer meet the development trend of power big data.Designing and implementing high-performance power big data servers enables real-time power big data acquisition,analysis,and processing to become the development of domestic power big data.This article proposes a Reactor server design patternThe software kernel of this system is constructed,and the data parsing layer of power data transmission message format based on the 104 protocol is designed and implemented.It realizes high concurrency in the kernel part of the server software.The asynchronous log module is implemented,so that the log print level can be changed without restarting the system,which reduces the cost of operation and maintenance and debugging of the server.The work done in this article mainly includes the following aspects:1.Analyze the development trend of China's power big data,the current research status of power big data processing at home and abroad,and the future research trends at home and abroad.This paper analyzes the OpenPDC,a real-time analysis software for power big data in foreign countries,combining with the main problems existing in OpenPDC,and proposes the server system software design requirements that suit China's national conditions.2.Design a highly concurrent server kernel architecture based on the Reactor in threads model,detail the structure of the Reactor pattern,and then study the event scheduling strategy of the libev event library I/O.The I/O scheduling strategy of the entire server kernel is Based on the package construction,the data analysis layer based on the plug-in mode is studied,and the message codec layer and message dispatch layer are designed for the 104 protocol power big data packet data format.3.According to the designed server software system kernel layer and the data parsing layer for the 104 protocol,the encoding is implemented,and the encoding implements the asynchronous log module.Two well-known lock-free synchronization algorithms have been studied,laying a theoretical foundation for improving the concurrent performance of server system software.4.Finally compare the system PMU data throughput with OpenPDC,verifying the advantage of the software in concurrent performance.