CN-120934615-B - Optical fiber network intelligent fixed inspection and fault closed loop processing system, method and equipment

CN120934615BCN 120934615 BCN120934615 BCN 120934615BCN-120934615-B

Abstract

The application discloses an intelligent fixed inspection and fault closed-loop processing system, method and equipment for an optical fiber network, and belongs to the technical field of intelligent operation and maintenance of optical fiber communication networks. The method comprises the steps of collecting and fusing multi-source heterogeneous data such as resource management, fault network management, distributed sensing and geographic information, analyzing the fused data by utilizing a multi-mode model based on a convolutional neural network and a graph neural network, realizing the meter-level accurate positioning of faults, adopting a long-period memory network model to dynamically optimize a fixed-check plan, automatically creating, circulating and associating fault lists, mode lists and maintenance lists based on analysis decision results, and realizing the whole-flow closed loop from diagnosis list dispatching to repair verification. By the aid of the scheme, the problems of data dispersion, single analysis dimension and lack of closed loops in the process in the prior art are solved, and the accuracy, efficiency and automation level of operation and maintenance of the optical fiber network are remarkably improved.

Inventors

LIAO YONGQIANG
LI YUJIE
LIU HONGBIN
GAO ZHICHENG

Assignees

广东信通通信有限公司

Dates

Publication Date: 20260512
Application Date: 20250827

Claims (7)

1. An intelligent optical fiber network fixed-inspection and fault closed-loop processing system is characterized by comprising: the multi-source heterogeneous data acquisition and preprocessing unit is configured to be connected to a resource management system, a fault network management system, a distributed optical fiber sensing system and a geographic information system so as to acquire and perform standardized processing on various operation and maintenance data, thereby generating a group of structured data streams with accurate time stamps and geographic position labels; The intelligent analysis and decision unit is configured to receive the structured data stream, and carry out deep analysis and decision on the state of the optical fiber network through at least one preset intelligent analysis model so as to output a decision instruction containing a high-precision diagnosis result and dynamic strategy suggestions; the collaborative work order and closed loop processing unit is configured to automatically create, associate and circulate three types of work orders, namely a mode order, a fault order and an overhaul order according to the decision instruction, and construct and execute a full-flow automatic closed loop from fault diagnosis, dispatching order, restoration to effect rechecking; The comprehensive control and presentation unit is configured to systematically integrate the decision instruction, the flow state of the automatic closed loop and the various operation and maintenance data, and generate a multi-dimensional monitoring cockpit and a visual report for operation and maintenance management personnel; the intelligent analysis and decision unit specifically comprises: the multi-mode fusion fault positioning model is used for receiving the data snapshot when a fault triggering event is monitored, and comprehensively analyzing the optical time domain reflectometer test curve data, the optical fiber network topology structure data, the network management alarm content and the distributed sensing event data contained in the data snapshot to output a fault point geographic coordinate with the highest confidence coefficient; The dynamic fixed inspection decision model based on the long-short-period memory network is used for learning the historical state data of the optical cable as a time sequence so as to predict the health degree evolution trend and the fault probability of the optical cable, and dynamically calculating and outputting the adjusted optical cable fixed inspection period according to the health degree evolution trend and the fault probability; The service-driven optical path automatic design engine is used for inquiring available standby fiber cores or routing resources in the network in real time when faults occur, and automatically designing a standby optical path which meets the service level protocol requirements and conforms to the risk avoidance principle; the internal structure of the multimode fusion fault location model comprises: A one-dimensional convolutional neural network, which is designed for processing one-dimensional time sequence signals, namely the test curve data of the optical time domain reflectometer, so as to automatically extract waveform characteristic vectors representing fault physical forms; A graph neural network designed for modeling the physical topology of the optical fiber network, wherein network facilities are abstracted into nodes of the graph, optical cable segments are abstracted into edges, and network management alarms and distributed sensing events acquired from other systems are attached to the corresponding nodes or edges of the occurrence positions as dynamic attributes so as to learn and extract topology-event characteristics; And the fusion layer is configured to splice and jointly infer the waveform characteristic vector and the topology-event characteristic so as to synthesize multidimensional information and output the geographic coordinates of the fault point.
2. The system of claim 1, wherein the collaborative worksheet and closed-loop processing unit is configured to: After receiving the accurate positioning result output by the intelligent analysis and decision unit, automatically creating a new fault list, and automatically filling information such as fault ID, geographic coordinates, inferred fault types, estimated influence business scope and the like; Automatically inquiring and associating all the mode sheets directly affected by the fault by utilizing the light path ID in the resource library; Based on the generated fault list and the mode list set related to the fault list, a service list with detailed contents is automatically generated and distributed to a preset operation and maintenance team.
3. The system of claim 2, wherein the collaborative work order and closed loop processing unit is further configured to perform a closed loop verification mechanism in the full flow automated closed loop, the closed loop verification mechanism comprising: after the maintenance personnel updates the work order state into the repair completion, automatically issuing an instruction to the test system, and requiring the light path after the repair completion to be immediately retested once; Acquiring a retested optical time domain reflectometer curve, and intelligently comparing the retested optical time domain reflectometer curve with a historical reference curve of the optical path and a fault curve before repair; If the comparison result shows that the optical path performance is restored to the preset acceptance standard, automatically updating the fiber core state in the related mode list to be normal, closing the related fault list and the maintenance list at the same time, and finally generating a closed-loop filing report containing the full record of the processing process.
4. The intelligent optical fiber network fixed inspection and fault closed loop processing method is characterized by comprising the following steps: Establishing a collaborative data view, continuously acquiring and fusing operation and maintenance data from a plurality of source systems such as a resource management system, a fault network management system, a distributed optical fiber sensing system, a geographic information system and the like through a multi-source heterogeneous data acquisition and preprocessing unit so as to form and dynamically maintain a unified multi-dimensional optical fiber network real-time state view; triggering fault diagnosis and accurate positioning, when a preset fault triggering event is monitored, immediately acquiring data snapshots in a time window before and after the occurrence time of the fault event by an intelligent analysis and decision unit, inputting the data snapshots into a multi-mode fusion fault positioning model, and calculating and outputting a high-precision fault geographic coordinate through a joint reasoning mechanism; Automatically generating and associating worksheets, automatically creating a new fault sheet by a collaborative worksheet and a closed-loop processing unit according to the high-precision fault geographic coordinates output in the last step, and automatically inquiring and associating all the mode sheets affected by the fault; the intelligent dispatching maintenance task is realized, a maintenance list is automatically generated by the collaborative work list and the closed loop processing unit based on the generated fault list, and the maintenance list is automatically filled with a standby resource scheme recommended by an optical path automatic design engine driven by a service; performing repair and closed loop verification, automatically triggering retesting of the repaired light path after the field repair operation is finished, and automatically updating and closing a related work order according to the comparison condition of retesting results and preset acceptance criteria so as to finish fault processing closed loop; The multi-source heterogeneous data acquisition and preprocessing unit specifically comprises: The communication network management data acquisition module is used for acquiring static resource data, real-time performance indexes, alarm information and massive historical fault records from the resource management system and the fault network management system through a standard protocol or a special adapter; The distributed sensing data interface module is used for being specially accessed into the distributed optical fiber sensing system, performing preliminary intelligent processing on the received original high-frequency vibration or temperature data stream, and refining the primary high-frequency vibration or temperature data stream into a structured event record; The space geographic information acquisition module is used for actively acquiring the accurate coordinates of the optical cable along the network facilities and the environment information layers along the lines from the geographic information system through the standardized space query interface; The data fusion and alignment engine is used for carrying out uniform format loading and space-time alignment on the static resource data, the real-time performance index, the alarm information, the historical fault record, the event record and the accurate coordinate and environment information layers so as to form a multi-mode space-time synchronous data snapshot; the intelligent analysis and decision unit specifically comprises: The multi-mode fusion fault positioning model is used for receiving the data snapshot when a fault triggering event is monitored, and comprehensively analyzing optical time domain reflectometer test curve data, optical fiber network topology structure data, network management alarm content and distributed sensing event data contained in the data snapshot to output a fault point geographic coordinate with highest confidence coefficient; The dynamic fixed inspection decision model based on the long-short-period memory network is used for learning the historical state data of the optical cable as a time sequence so as to predict the health degree evolution trend and the fault probability of the optical cable, and dynamically calculating and outputting the adjusted optical cable fixed inspection period according to the health degree evolution trend and the fault probability; The service-driven optical path automatic design engine is used for inquiring available standby fiber cores or routing resources in the network in real time when faults occur, and automatically designing a standby optical path which meets the service level protocol requirements and conforms to the risk avoidance principle; the internal structure of the multimode fusion fault location model comprises: A one-dimensional convolutional neural network, which is designed for processing one-dimensional time sequence signals, namely the test curve data of the optical time domain reflectometer, so as to automatically extract waveform characteristic vectors representing fault physical forms; A graph neural network designed for modeling the physical topology of the optical fiber network, wherein network facilities are abstracted into nodes of the graph, optical cable segments are abstracted into edges, and network management alarms and distributed sensing events acquired from other systems are attached to the corresponding nodes or edges of the occurrence positions as dynamic attributes so as to learn and extract topology-event characteristics; And the fusion layer is configured to splice and jointly infer the waveform characteristic vector and the topology-event characteristic so as to synthesize multidimensional information and output the geographic coordinates of the fault point.
5. The system of claim 4, wherein the data fusion and alignment engine is configured to use a memory computing framework, use a high-precision timestamp as a core alignment key, and use a sliding time window algorithm to instantaneously correlate and align test events, network management alarm events, sensing events, and static resource information and geographic environment information related to an event occurrence point within a certain minute time window.
6. The system of claim 4, wherein the long and short term memory network-based dynamic routing decision model is configured to calculate the adjusted cable routing period via a configurable weighted scoring mechanism, the inputs to the weighted scoring mechanism including a future failure probability predicted by the long and short term memory network model, a level of importance of traffic carried by the cable, and a level of environmental risk of the location of the cable obtained from the geographic information system.
7. An intelligent optical fiber network checking and fault closed loop processing device, which is a server or a server cluster deployed in a data center physically, is characterized in that the device comprises a processor, a memory and a network interface in communication connection with the processor on hardware; The memory is fixedly stored with a computer program which, when being executed by the processor, realizes all the steps of the intelligent optical fiber network checking and fault closed-loop processing method as claimed in claim 4.

Description

Optical fiber network intelligent fixed inspection and fault closed loop processing system, method and equipment Technical Field The invention belongs to the technical field of intelligent operation and maintenance of an optical fiber communication network, and particularly relates to an intelligent fixed-inspection and fault closed-loop processing system, method and equipment of the optical fiber network. Background Along with the deep transformation of the power system to the digital and intelligent directions, the power communication network is used as a 'neural network' for supporting the safe and stable operation of the power grid, and the reliability and the operation and maintenance efficiency of the power communication network become core elements for guaranteeing the energy supply safety. The optical fiber communication network forms a physical foundation for bearing key services such as relay protection, dispatching automation, intelligent inspection and the like by the inherent advantages of high bandwidth, strong electromagnetic interference resistance, long transmission distance and the like. Therefore, ensuring the health status of the optical fiber communication network, realizing the rapid response and accurate positioning of the faults of the optical cable is a major topic faced by the modern power operation and maintenance system. Currently, aiming at operation and maintenance guarantee of an electric power communication optical fiber network, the technical scheme represented by periodic detection and single-system fault analysis is mainly formed in the field. On the one hand, the operation and maintenance personnel adopts professional instruments such as Optical Time Domain Reflectometer (OTDR) and the like to carry out periodic off-line test on the optical cable path. The method can judge the on-off state and the loss condition of the optical cable to a certain extent by analyzing the reflection and attenuation events of the OTDR curve, and plays a fundamental role in guaranteeing the physical connectivity of the basic communication link of the optical fiber network. On the other hand, in order to improve the automation level of fault diagnosis, an algorithm such as a neural network is used to analyze OTDR data to identify a fault type. The technical means jointly form a basic stone of the operation and maintenance mode of the existing optical fiber network, and a basic tool is provided for daily maintenance and fault investigation. However, with the continuous improvement of the requirements of the power grid service on the reliability and the real-time performance of the communication, and the increasing size and the complexity of the network, the above-mentioned operation and maintenance manner relying on the fixed period and the single data source is increasingly limited. First, the prior art generally has the problem of "data islands". Resource management, fault network management, and inspection system are often independent service systems, and data standards and storage architectures (such as SQL and NoSQL databases) are not unified, so that key information cannot be effectively fused. The fracture directly leads to single fault analysis dimension, for example, only analysis of OTDR data but neglecting strong association information such as network management alarm, resource topology and the like, so that fault positioning accuracy is poor, and average positioning deviation is generally more than 200 meters. Secondly, the intelligent degree and the automatic closed-loop capability of the prior art are insufficient. The fixed inspection plan depends on a fixed time period (for example, 6-12 months), the inspection strategy is difficult to dynamically adjust according to the external environment threats such as typhoons, icing and the like, and in the fault processing flow, the verification link from diagnosis, dispatch to the effect after repair lacks an effective automatic association and closed loop verification mechanism, and data show that more than 78% of fault processing cases do not automatically recheck the repair effect. Finally, the existing scheme has weak adaptability to the power service scene, the severe requirements of advanced services such as relay protection on the communication recovery time (service switching time in the existing scheme often exceeds 5 minutes) cannot be fully considered, and a mechanism for performing collaborative risk assessment and protection on a plurality of optical cables laid in the same channel is lacking. Therefore, how to break through the bottleneck of data dispersion and single analysis dimension of each operation and maintenance system at present, a collaborative processing method capable of fusing multi-source heterogeneous data and realizing intelligent analysis decision and automatic closed loop of worksheet process is constructed, so that the accuracy of fault location of the power optical fiber network, the dynamic adaptabilit