Research On Key Technology Of Marine Seismic Streamer Shipboard Recording System Based On In-memory Computing

Marine seismic exploration streamer shipboard recording system is an important part in marine seismic exploration facility.When there is a small amount of marine streamers and acquisition channels,the issue with data recording is relatively easy and the recording system does not need to be scalable.However,with the expansion of the scale of marine seismic exploration facility,the amount of marine streamers and acquisition channels will be growing exponentially,making it increasingly demond for the software and hardware system to operate in a scalable manner.Traditionally,the shipboard recording system merely undertakes the recording tasks for the streamer data,uses fixed hardware and software organization structure,and little consideration is given to the overall ease of scalable of the hardware and soft-ware in the system.In addition,the interfaces of various components in the system are different,and the connection between the components is complex,which makes it difficult to achieve flexible system expansion and tailoring.During the marine seismic exploration,the above defects of the traditional shipboard recording system will not only increase the system maintenance cost,but also reduce the efficiency of the exploration surveyor.Therefore,with in-memory computing and the laboratory's past development ex-perience in the field of marine seismic exploration systems as the basis,and the scalable shipboard recording system as design goal,the thesis discusses the specific meaning of the software and hardware scalable of the recording system,and puts forward one data interface and processing separated system architecture.By analyzing the technical difficulties in the new system architecture,and breaking through the key technical is-sues,this thesis targets four key technologies,which are general data processing node design technology,high-speed data transmission technology between nodes,memory-based digital logic hardware processing technology,memory-based distributed stream processing software technology.In terms of designing the general data processing nodes,this thesis draws on the design ideas of virtual instruments,and discusses the design theory of general data pro-cessing nodes from two structured data processing and data processing imaging.A general data perspective of Channel Time Spectrum is proposed to analyze the marine seismic exploration system.For the actual board design,this thesis uses the existing chip-level and board-level common interface solutions that are widely used in the in-dustry,to carry out specific chip selection and circuit design for the node.In terms of high-speed data transmission between nodes,this thesis uses SerDes transmission technology and GTX high-speed serial transceiver,with Aurora 64B/66B IP core,FMC and SFP modules,as well as PCIe data transmission links to achieve the high-speed serial transmission link on physical nodes in the system.In terms of memory-based digital logic hardware processing,this thesis studies virtual FIFO,streamer data merge,memory sharing technology between nodes based on the small system of reading or writing to the DDR memory,and the combination with memory interface module,AXI bus interconnect,DMA data transmission engine and MicroBlaze soft core and other components.In terms of memory-based distributed stream processing software,this thesis pro-poses a variety of marine streamer data processing schemes using the existing in-memory stream processing software framework and distributed database system tech-nology to construct an scalable software system in shipboard recording system based on the Hadoop software ecosystem,and the combination with specific streamer data processing tasks,discusses various data processing schemes for marine streamers.Through the above key technology research,the shipboard recording system de-scribed in this thesis can not only achieve the general deployment goal of the main physical nodes in the system,but also provide a set of hardware and software solutions for processing streamer data based on in-memory computing.The proposed system con-cept of scalable software and hardware and the system design scheme with the separa-tion between the interface and processing components jointly contributed to a simplified system structure.Furthermore,by introducing the popular technology in the field of big data,this thesis broadens the range of technology selection in the field of marine seismic exploration equipment.