Development Of Submarine Cable Integrated Monitoring System And Implementation Of Distributed Computing Platform

The submarine cable is always facing a variety of threats from the sea floor,if it can not find a variety of hidden dangers and repair in time,it will affect the safe and stable operation of the cross-sea power grid and communication network,and even lead to serious accidents,resulting in huge economic losses.In view of the possible threats encountered by submarine cable,this thesis designs and implements a set of integrated submarine cable monitoring system based on optical fiber sensing technology,and develops a distributed computing platform solution for the big data stream processing problem brought by optical fiber sensing equipment and applies it to the integrated submarine cable monitoring system.This thesis mainly completes the following aspects of work:(1)The requirements and technical status of submarine cable monitoring are analyzed,as well as the application status of optical fiber sensing big data and distributed stream processing framework,and the principles of two optical fiber sensing technologies(Φ-OTDR and BOTDA),the principles of ship Automatic Identification System(AIS),Hadoop ecosystem,Spark framework and Kafka stream transmission framework are briefly introduced.(2)In order to solve the weakness of traditional centralized processing scheme for big data stream processing,a distributed stream processing solution for optical fiber sensing data stream processing of submarine cable monitoring system is proposed.Due to the huge data flow generated by the two sets of optical fiber sensing devices(Φ-OTDR,BOTDA)used in the monitoring system,the traditional centralized processing mode of single central processing unit(CPU)or graphics processing unit(GPU)machine has posed a great challenge,making it relatively weak in real-time processing with high reliability.Can’t keep up with their big data generation.Therefore,this thesis proposes a distributed stream processing solution based on Hadoop ecosystem +Spark Structured Streaming distributed stream processing framework +Kafka distributed stream transmission framework,and constructs an overall framework of streaming big data computing,storage,management and application for the submarine cable integrated monitoring system at different levels.The performance of the distributed processing platform is evaluated according to the performance indexes of stream processing,which are common in delay and throughput.(3)The practical application requirements of the submarine cable monitoring system are comprehensively analyzed,and the submarine cable comprehensive monitoring system based on the big data distributed stream processing platform is proposed,and its design and implementation are carried out.The integrated monitoring system is divided into three modules: distributed database,background and foreground.The background is responsible for communication interaction and data processing with the sensor equipment in the monitoring system by means of the distributed stream processing framework,and carries out control interaction with the foreground through the control interface to a certain extent.The front desk is mainly responsible for interacting with users,providing user login and management,data waveform display,alarm display,electronic chart display,video surveillance and other functions,and users can access the web content by accessing the Apache server.In addition,in order to solve the problems such as inaccurate alarm of a single device and difficult matching of ship identification and anchor damage events,a fusion classification algorithm of Φ-OTDR and BOTDA device data is designed.Combined with the operation principle of the entire monitoring system,a linkage classification alarm process from the occurrence of an incident to the positioning and tracking of the accident vessel is designed.In the field test of submarine cable monitoring,the system can run stably for a long time and play the alarm function.