A Substation Live Detection Method on Communication Protocol Data Parsing and Data-level Fusion of Neural Networks

Abstract

In this paper, a data parsing method based on database communication protocol technology and a data fusion method based on the neural network are proposed, which solves the shortcomings in the data communication process of the substation, and provides an intelligent method for data analysis and fusion. The data parsing, transforming method, process, and implementation scheme based on the communication protocol is described in this paper. Firstly, the digital expression method of the data message is proposed. Secondly, the message recognition method based on the feature field is introduced. Through the analysis and heuristic of the feature fields, the protocols of the messages are determined. then the data interface relationships in the system are stored in a separate database. When the parsing process is completed, we use data pre-processing, data cleaning, data integration, and other methods to fuse the data of each source. Finally, the unstructured data fusion method based on the VGG-19 model is used to fuse the data. At the same time, this paper also describes the scheme of some principles step by step and introduces the relevant technical details.