基于bp神经网络的电力监测多规约识别系统的设计与实现-通信与信息系统专业论文.docxVIP

江 江 苏 大 学 项 士 学 位 论 文 ABSTRACT In electric power syst锄，the transfer of the electric power communication must comply with the communication protoc01．Now that unifying communication protocol is still a long process，it is imperative to connect the devices with different communication protocols． The main task of this paper is to analyze the mapping among different protocols，work out all mutil-protocol recognise system based on BP Artificial Neural Network．Center around the subject．the main work has been done as follows：Features extraction．研Ⅱl the characteristic of the electric power communication protocol，fix on the Features extraction，which is the data in the front side of the data frame．Pattern recognition．The feature vector are the input of artificial neural network and of Support Vector Machine respectively,is used to recognize．In this paper,BP neural network with different arithmetics is proposed．The results fil e compared and analysed in order to make the most appropriate Net-framework．then emluate to get recognition rate，which is 94．5％．Working-flow chart．With the characteristic of terminal and electric power system and real-time，fixation location，design to Working-flow chart to makes a few improve on performance and breakthroughs some limitation of electric power inspection system．Soft-making．with the VC++,make the soft project of the mufti—procotol recognise system．this paper set out the Class diagrams and main function of the recognise system．The result，this paper compare and analyse performance of the mutil—procotol recognize system according as emluate test．The result indicate that the mutil-procotol recognize system has high recognize rate and convergence· The last，this paper gives the deficiency of the mutil-procotol recognize system．And，In future，which topic the research of this will pay attention OIL KEYWORDS：communication protocol，recognize system,neural network,electric inspection． Ⅱ 学位论文版权使用授权书本学位论文作者完全了解学校有关保留、使用学位论文 学位论文版权使用授权书 本学位论文作者完全了解学校有关保留、使用学