Natural gas pipeline maintenance optimization method and system based on big data

Abstract

The invention discloses a big data-based natural gas pipeline maintenance optimization method and system, relates to the technical field of natural gas pipeline maintenance, and aims to construct a structured monitoring data set, construct a graph model of pipeline states and risk factors, establish a prediction model, introduce a memory mechanism to process low-frequency anomalies, and generate and optimize a maintenance scheduling scheme. According to the natural gas pipeline maintenance optimization method provided by the invention, the map model and the prediction model are constructed based on the multi-source monitoring data, the causal path between the pipeline operation state and the external risk factor can be identified, and the maintenance efficiency is improved. Maintenance task generation, early warning triggering and terminal linkage operation are completed through risk level evaluation and scheduling optimization, collected monitoring data are used for model training and map updating after being processed, a closed-loop management process is formed, and the method is suitable for natural gas pipeline state recognition and maintenance scheduling in various operation environments.

Description

Technical Field

[0001] The present invention relates to the technical field of natural gas pipeline maintenance, and in particular to a natural gas pipeline maintenance optimization method and system based on big data. Background Art

[0002] As a clean and efficient energy source, natural gas transmission relies on an extensive, high-pressure pipeline system. As gas transmission scale continues to expand and operating environments become increasingly complex, natural gas pipelines face multiple risks, including geological subsidence, temperature stress, and third-party construction. Conventional manual inspections or fixed-periodic maintenance methods are no longer sufficient to meet the requirements of dynamic risk prevention and control and rapid response.

[0003] Currently, a Chinese invention patent application numbered CN202411318520.7 discloses a natural gas pipeline corrosion prediction and early warning method and system. The method includes: installing sensors at different locations on the natural gas pipeline to collect data, storing it on a central server, preprocessing it, and then extracting features; constructing a long short-term memory network model based on the natural gas pipeline, optimizing the hyperparameters of the long short-term memory network model through an adaptive genetic algorithm to obtain an optimized long short-term memory network model, and predicting the corrosion trend of the natural gas pipeline; constructing a graph neural network model based on the pipeline network according to the location of the natural gas pipeline to predict the future corrosion propagation path; designing early warning thresholds and generating early warning signals. At the same time, based on the early warning signals, intelligent maintenance and repair suggestions are provided to optimize resource allocation, and the early warning thresholds and decision plans are adjusted in real time according to actual implementation conditions.

[0004] The above technologies make it difficult to achieve accurate correlation modeling and dynamic maintenance optimization scheduling between the operating status of natural gas pipelines and multi-source external risk factors. Summary of the Invention

[0005] The technical problem solved by the present invention is that it is difficult to achieve accurate correlation modeling and dynamic maintenance optimization scheduling between the operating status of natural gas pipelines and multi-source external risk factors with existing technology.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: The natural gas pipeline maintenance optimization method based on big data includes the following steps: Step S1, constructing a structured monitoring data set; Step S2, constructing a graph model of pipeline status and risk factors; Step S3, establishing a prediction model and introducing a memory mechanism to handle low-frequency anomalies; Step S4, generating and optimizing a maintenance scheduling plan; Step S5, triggering an early warning and executing a terminal maintenance instruction; Step S6: Collect maintenance results for model and map update.

[0007] Preferably, the step S1 includes the following sub-steps: Step S101, collecting internal operating status data of the natural gas pipeline, wherein the internal operating status data includes pressure monitoring data, flow rate monitoring data, temperature monitoring data and vibration monitoring data; Step S102: collecting external environmental monitoring data of the natural gas pipeline, wherein the external environmental monitoring data includes geological subsidence data, surface temperature data, surface humidity data, meteorological monitoring data, and remote sensing image data; Step S103: collecting human activity monitoring data around the natural gas pipeline, wherein the human activity monitoring data includes construction trajectory data, surface disturbance image data, and drone inspection video data; Step S104 , the internal operating status data, external environment monitoring data and surrounding human activity monitoring data are cleaned, time-synchronized, spatially aligned and format-standardized to form a structured monitoring data set and stored in a database.

[0008] Preferably, step S2 includes the following sub-steps: Step S201: perform integrity check and outlier removal on the structured monitoring data set, and uniformly use standard timestamps and spatial coordinates for correction; Step S202: Mapping the structured monitoring data into a spatiotemporal causal graph based on a graph neural network model. The nodes of the spatiotemporal causal graph are the natural gas pipeline operation status monitoring variables and the external impact monitoring variables. The edges of the spatiotemporal causal graph are the causal relationships between the variables, and the edge weights are calculated based on the degree of historical fluctuation combination. In step S203, a fixed time window rolling analysis method is used to identify path subgraphs with stable causal relationships. The path in which geological subsidence monitoring data triggers changes in vibration monitoring data and causes fluctuations in pressure monitoring data is identified as a high-attention subgraph as input to the risk prediction model.

[0009] Preferably, step S3 includes the following sub-steps: Step S301: input the high-attention path subgraph marked in step S2, and use a multi-layer temporal convolutional network model to extract the state feature sequence of the monitoring variable changing over time; Step S302: Introduce a memory mechanism based on historical events to construct an event memory index table. The frequency of historical abnormal events, the fluctuation range of the monitoring variables involved, and the final risk level are normalized into a memory factor. The similarity of the current input state sequence is matched and the high-risk identification process is determined. Step S303: output the risk level assessment results of each section of the natural gas pipeline, wherein the risk level assessment results include high risk, medium risk and low risk, and synchronously transmit the risk level assessment results to the maintenance scheduling module.

[0010] Preferably, step S4 includes the following sub-steps: Step S401, reading the risk level assessment result output in step S3, and outputting the maintenance priority according to the risk level assessment result; Step S402: combining the maintenance personnel's location information, travel path information, equipment status information, and weather condition information to generate a maintenance scheduling candidate plan based on resource constraints; Step S403: Use reinforcement learning to train the scheduling optimization model, set a multi-objective scoring function including risk level reduction, response time, maintenance path length and maintenance cost, perform simulation training and output a maintenance scheduling plan.

[0011] Preferably, step S5 includes the following sub-steps: Step S501: When the risk level assessment result of any pipeline section exceeds the set threshold, an alarm record is generated and accompanied by the corresponding risk classification, key monitoring variables and reference fluctuation record; Step S502: Push the alarm record to the dispatch command center terminal and the inspection personnel terminal through the unified management platform, and publish it in the form of map, text prompt and voice broadcast; In step S503, based on the optimal maintenance scheduling plan outputted in step S4, a patrol task route map is generated. The patrol task route map includes the starting coordinates, necessary inspection points, estimated operation time, required maintenance equipment and inspection content, and is uploaded to the patrol terminal device for task navigation and status feedback.

[0012] Preferably, step S6 includes the following sub-steps: Step S601: collecting maintenance result data of the inspection personnel during the inspection task, the maintenance result data including actual trajectory data, maintenance start and end time, on-site image data, audio recording data and risk treatment result data, and writing it into the database for archiving; Step S602: incrementally train the risk prediction model using the maintenance result data to update the boundary threshold, classification weight parameters, and feature variable mapping relationships used for state identification in the model; In step S603, the spatiotemporal causal graph constructed in step S2 is structurally iterated, and the graph edge weights and node risk identifiers are corrected using the maintenance result data. The path subgraphs with prediction deviations greater than the set range are submitted to the model correction module for processing.

[0013] Preferably, the memory mechanism based on historical events triggers a high-risk processing channel by constructing a memory factor based on historical abnormal events and matching it with the current state sequence.

[0014] Preferably, when executing step S104, global positioning system coordinate identification and time stamp identification are used for the structured monitoring data set.

[0015] A natural gas pipeline maintenance optimization system based on big data, which is applied to the natural gas pipeline maintenance optimization method based on big data, includes a data acquisition module, a graph construction module, a model prediction module, a solution generation module, an early warning maintenance module, and a maintenance update module; The data acquisition module is used to construct a structured monitoring data set; The graph construction module is used to construct a graph model of pipeline status and risk factors; The model prediction module is used to establish a prediction model and introduce a memory mechanism to handle low-frequency anomalies; The scheme generation module is used to generate and optimize the maintenance scheduling scheme; The early warning maintenance module is used to trigger early warning and execute terminal maintenance instructions; The maintenance and update module is used to collect maintenance results for model and map updates.

[0016] Beneficial effects of the present invention: The natural gas pipeline maintenance optimization method provided by the present invention constructs a graph model and a prediction model based on multi-source monitoring data, can identify the causal path between the pipeline operation status and external risk factors, and completes maintenance task generation, early warning triggering and terminal linkage operations through risk level assessment and scheduling optimization. The collected monitoring data is used for model training and graph update after processing, forming a closed-loop management process, which is suitable for natural gas pipeline status identification and maintenance scheduling under various operating environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A flowchart of the steps of a natural gas pipeline maintenance optimization method based on big data provided by one embodiment of the present invention; Figure 2 A schematic diagram of the basic process of a natural gas pipeline maintenance optimization system based on big data provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0018] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, but not all of the embodiments.

[0019] Example 1, reference Figure 1, provides a natural gas pipeline maintenance optimization method based on big data, including the following steps: Step S1: construct a structured monitoring dataset.

[0020] Step S2: construct a graph model of pipeline status and risk factors.

[0021] Step S3: Establish a prediction model and introduce a memory mechanism to handle low-frequency anomalies.

[0022] Step S4: Generate and optimize a maintenance scheduling plan.

[0023] Step S5: trigger an early warning and execute terminal maintenance instructions.

[0024] Step S6: Collect maintenance results for model and map update.

[0025] Step S1 includes the following sub-steps: Step S101 : collecting internal operating status data of a natural gas pipeline, where the internal operating status data includes pressure monitoring data, flow rate monitoring data, temperature monitoring data, and vibration monitoring data.

[0026] Step S101 acquires core monitoring data reflecting the operating status of the natural gas pipeline, which is used to evaluate the dynamic changes of the pipeline itself.

[0027] Step S102 : collecting external environment monitoring data of the natural gas pipeline, where the external environment monitoring data includes geological subsidence data, surface temperature data, surface humidity data, meteorological monitoring data, and remote sensing image data.

[0028] Step S102 acquires information data reflecting changes in the geographical environment in which the pipeline is located, providing a basis for identifying risks caused by geological factors.

[0029] Step S103 : collecting human activity monitoring data around the natural gas pipeline, where the human activity monitoring data includes construction trajectory data, surface disturbance image data, and drone inspection video data.

[0030] Step S103 obtains data on the impact of surrounding human activities on the pipeline to provide support for identifying potential hidden dangers caused by external disturbances.

[0031] Step S104 , the internal operating status data, external environment monitoring data and surrounding human activity monitoring data are cleaned, time-synchronized, spatially aligned and format-standardized to form a structured monitoring data set and stored in a database.

[0032] When step S104 is executed, the structured monitoring data set is identified by global positioning system coordinates and timestamps.

[0033] Step S104 synchronizes internal, external, and human activity monitoring data across time and space, storing them in a unified format as a structured monitoring dataset. By introducing GPS coordinate identifiers and timestamps, the data is ensured to be traceable and consistent, providing a foundation for subsequent modeling and analysis.

[0034] Step S1 collects and processes multiple types of monitoring data from natural gas pipelines to construct a unified and standardized structured monitoring data set, providing an accurate and temporally consistent data foundation for map modeling and risk prediction.

[0035] Step S2 includes the following sub-steps: Step S201 : perform integrity check and outlier removal on the structured monitoring data set, and uniformly use standard timestamps and spatial coordinates for correction.

[0036] Step S201 improves the integrity and consistency of structured monitoring data, ensures that the data used for subsequent map construction has a unified time and space reference, and eliminates interference from abnormal data.

[0037] In step S202, the structured monitoring data is mapped into a spatiotemporal causal graph based on the graph neural network model. The nodes of the spatiotemporal causal graph are the natural gas pipeline operation status monitoring variables and the external impact monitoring variables. The edges of the spatiotemporal causal graph are the causal relationships between the variables, and the edge weights are calculated based on the degree of historical fluctuation combination.

[0038] Step S202 forms a spatiotemporal causal graph model with monitoring variables as nodes and variable causal relationships as edges, providing structural expression and dynamic representation capabilities for multi-variable relationship modeling.

[0039] In step S203, a fixed time window rolling analysis method is used to identify path subgraphs with stable causal relationships. The path in which geological subsidence monitoring data triggers changes in vibration monitoring data and causes fluctuations in pressure monitoring data is identified as a high-attention subgraph as input to the risk prediction model.

[0040] Step S203 extracts a path subgraph with stable causal logic, converts the causal chain between key variables into model input, and supports the identification and predictive modeling of specific risk evolution processes.

[0041] Step S2 constructs a spatiotemporal causal map between the operating status of the natural gas pipeline and external risk factors, provides a causal correlation data basis for subsequent risk identification and prediction models, and realizes mechanism modeling between pipeline status and environmental interference.

[0042] Step S3 includes the following sub-steps: Step S301: input the high-attention path subgraph marked in step S2, and use a multi-layer temporal convolutional network model to extract the state feature sequence of the monitoring variable changing over time.

[0043] Step S301 extracts the temporal variation characteristics of the monitoring variables in the high-attention path subgraph, obtains a state sequence that can reflect the risk evolution process, and provides dynamic input features for subsequent risk identification.

[0044] In step S302, a memory mechanism based on historical events is introduced to construct an event memory index table. The frequency of historical abnormal events, the fluctuation amplitude of the monitoring variables involved, and the final risk level are normalized into a memory factor, matching the similarity of the current input state sequence and determining to enter the high-risk identification process.

[0045] The memory mechanism based on historical events triggers high-risk processing channels by constructing memory factors based on historical abnormal events and matching them with the current state sequence.

[0046] Step S302 uses known risk features from historical abnormal events to establish an event memory index table, triggering a high-risk identification mechanism when the current state sequence is similar to the historical risk pattern to achieve identification and processing of low-frequency abnormalities.

[0047] Step S303: output the risk level assessment results of each section of the natural gas pipeline, the risk level assessment results including high risk, medium risk and low risk, and synchronously transmit the risk level assessment results to the maintenance scheduling module.

[0048] Step S303 outputs the risk level assessment result of each pipe section and passes the assessment result as input to the scheduling module to provide a basis for the priority sorting of maintenance tasks.

[0049] Step S3 constructs a risk prediction model based on the graph subgraph and combines it with the historical abnormal event memory mechanism to evaluate the risk level of each section of the natural gas pipeline, providing an accurate risk judgment basis for maintenance scheduling.

[0050] Step S4 includes the following sub-steps: Step S401 , reading the risk level assessment result outputted in step S3 , and outputting the maintenance priority according to the risk level assessment result.

[0051] Step S401 determines the priority order of maintenance tasks based on the risk level assessment results of each pipe section, which serves as the basis for scheduling generation.

[0052] Step S402 : generating a maintenance scheduling candidate plan based on resource constraints by combining the maintenance personnel's location information, travel path information, equipment status information, and weather condition information.

[0053] Step S402 forms a scheduling candidate plan that meets the actual execution conditions based on the maintenance task sorting and comprehensively considers limiting factors such as personnel distribution, path accessibility, equipment availability, and weather conditions.

[0054] Step S403: Use reinforcement learning to train the scheduling optimization model, set a multi-objective scoring function including risk level reduction, response time, maintenance path length and maintenance cost, perform simulation training and output a maintenance scheduling plan.

[0055] Step S403 uses a reinforcement learning method to simulate and optimize the candidate solutions, scores them based on multiple scheduling objectives, and outputs an optimal maintenance scheduling solution that meets risk control and resource efficiency requirements.

[0056] Step S4 constructs a maintenance task priority ranking based on the risk level assessment results and resource constraints, and generates a maintenance scheduling plan through a scheduling optimization model to achieve reasonable allocation of maintenance resources and dynamic adjustment of task execution paths.

[0057] Step S5 includes the following sub-steps: Step S501: When the risk level assessment result of any pipe section exceeds the set threshold, an alarm record is generated and accompanied by the corresponding risk classification, key monitoring variables and reference fluctuation record.

[0058] After identifying the high-risk pipeline section, step S501 outputs alarm information including risk classification, key monitoring variables and their fluctuation records, providing complete data support for subsequent response.

[0059] Step S502: Push the alarm record to the dispatch command center terminal and the inspection personnel terminal device through the unified management platform, and publish it in the form of map, text prompt and voice broadcast.

[0060] Step S502 distributes the alarm information to the dispatch center and on-site terminals through the management platform, ensuring that the alarm content is delivered to the execution personnel in various forms, thereby improving the timeliness and accuracy of the response.

[0061] In step S503, based on the optimal maintenance scheduling plan output in step S4, a patrol task route map is generated. The patrol task route map includes the starting coordinates, necessary inspection points, estimated operation time, required maintenance equipment and inspection content, and is uploaded to the patrol terminal device for task navigation and status feedback.

[0062] Step S503 generates a clear inspection task roadmap based on the optimized scheduling plan and sends it to the terminal, guiding the inspection personnel to complete the operation tasks as planned and feedback the task execution status, ensuring the visual execution and process recording of the maintenance tasks.

[0063] In step S5, when the risk level assessment result exceeds the set threshold, the system automatically triggers an alarm, and the management platform and the inspection terminal are linked to complete the release of alarm information and the issuance of inspection tasks, realizing immediate response and closed-loop control from risk identification to maintenance execution.

[0064] Step S6 includes the following sub-steps: Step S601: collect maintenance result data of the inspection personnel during the inspection task. The maintenance result data includes actual trajectory data, maintenance start and end time, on-site image data, audio recording data and risk treatment result data, and write them into the database for archiving.

[0065] Step S601 records the actual maintenance status during the inspection process, and writes the trajectory, time, image, audio and risk treatment result data into the database to build a data foundation that can be used for model training and system tracing.

[0066] Step S602: incrementally train the risk prediction model using the maintenance result data, and update the boundary threshold, classification weight parameter, and feature variable mapping relationship used for state identification in the model.

[0067] Step S602 uses the maintenance result data for incremental training of the risk prediction model, dynamically adjusts the model recognition threshold, classification weight and feature variable mapping, and enhances the model's ability to recognize new states.

[0068] In step S603, the spatiotemporal causal graph constructed in step S2 is structurally iterated, and the graph edge weights and node risk identifiers are corrected using the maintenance result data. The path subgraphs with prediction deviations greater than the set range are submitted to the model correction module for processing.

[0069] Step S603 uses the maintenance result data to optimize the constructed spatiotemporal causal graph structure, correct the edge weights and node risk identifications in the graph, and submit the path subgraphs with prediction errors to the model correction module to improve the representation accuracy and prediction support capabilities of the graph.

[0070] Step S6 collects actual result data during the maintenance execution process to continuously update the risk prediction model and spatiotemporal causal graph, improve the system's recognition accuracy and adaptability under different operating conditions, and form a closed-loop self-learning mechanism.

[0071] Example 2, reference Figure 2 , provides a natural gas pipeline maintenance optimization system based on big data, including data acquisition module, graph construction module, model prediction module, solution generation module, early warning maintenance module and maintenance update module.

[0072] The data acquisition module is used to construct structured monitoring data sets.

[0073] The graph construction module is used to build a graph model of pipeline status and risk factors.

[0074] The model prediction module is used to establish a prediction model and introduce a memory mechanism to handle low-frequency anomalies.

[0075] The plan generation module is used to generate and optimize maintenance scheduling plans.

[0076] The early warning maintenance module is used to trigger early warnings and execute terminal maintenance instructions.

[0077] The maintenance and update module is used to collect maintenance results for model and map updates.

[0078] The present invention realizes the correlation analysis between the operating status of natural gas pipelines and external risk factors by constructing a structured monitoring data set, establishing a graph model and a prediction model. Based on specific data items such as pressure monitoring data, flow rate monitoring data, temperature monitoring data, vibration monitoring data, geological settlement data, remote sensing image data and construction trajectory data, it identifies risk paths with stable causal chains, outputs maintenance scheduling priorities according to risk levels, and issues terminal maintenance instructions through a unified management platform. The invention is suitable for plateau areas affected by freeze-thaw settlement, urban fringe areas with intensive construction activities, and long-distance pipeline sections in uninhabited areas under high-dimensional remote sensing monitoring conditions. It completes alarm generation, path map planning and model self-updating without relying on human intervention, and realizes the closed-loop operation of the natural gas pipeline anomaly identification, maintenance scheduling and prediction mechanism.

[0079] It should be understood by those skilled in the art that the embodiments of the present invention may be provided as methods, systems or computer program products. Therefore, the present invention may take the form of a completely hardware embodiment, a completely software embodiment or an embodiment combining software and hardware aspects. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program code. The storage medium may be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk. These computer program instructions may also be stored in a computer-readable memory that can guide a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including an instruction device, which is implemented in the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0080] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A natural gas pipeline maintenance optimization method based on big data, characterized in that: The steps include: Step S1, constructing a structured monitoring data set; Step S2, constructing a graph model of pipeline status and risk factors; Step S3, establishing a prediction model and introducing a memory mechanism to handle low-frequency anomalies; Step S4, generating and optimizing a maintenance scheduling plan; Step S5, triggering an early warning and executing a terminal maintenance instruction; Step S6: Collect maintenance results for model and map update.

2. The natural gas pipeline maintenance optimization method based on big data according to claim 1, characterized in that: The step S1 includes the following sub-steps: Step S101, collecting internal operating status data of the natural gas pipeline, wherein the internal operating status data includes pressure monitoring data, flow rate monitoring data, temperature monitoring data and vibration monitoring data; Step S102: collecting external environmental monitoring data of the natural gas pipeline, wherein the external environmental monitoring data includes geological subsidence data, surface temperature data, surface humidity data, meteorological monitoring data, and remote sensing image data; Step S103: collecting human activity monitoring data around the natural gas pipeline, wherein the human activity monitoring data includes construction trajectory data, surface disturbance image data, and drone inspection video data; Step S104 , the internal operating status data, external environment monitoring data and surrounding human activity monitoring data are cleaned, time-synchronized, spatially aligned and format-standardized to form a structured monitoring data set and stored in a database.

3. The natural gas pipeline maintenance optimization method based on big data according to claim 2, characterized in that: The step S2 includes the following sub-steps: Step S201: perform integrity check and outlier removal on the structured monitoring data set, and uniformly use standard timestamps and spatial coordinates for correction; Step S202: Mapping the structured monitoring data into a spatiotemporal causal graph based on a graph neural network model. The nodes of the spatiotemporal causal graph are the natural gas pipeline operation status monitoring variables and the external impact monitoring variables. The edges of the spatiotemporal causal graph are the causal relationships between the variables, and the edge weights are calculated based on the degree of historical fluctuation combination. In step S203, a fixed time window rolling analysis method is used to identify path subgraphs with stable causal relationships. The path in which geological subsidence monitoring data triggers changes in vibration monitoring data and causes fluctuations in pressure monitoring data is identified as a high-attention subgraph as input to the risk prediction model.

4. The natural gas pipeline maintenance optimization method based on big data according to claim 3, characterized in that: The step S3 includes the following sub-steps: Step S301: input the high-attention path subgraph marked in step S2, and use a multi-layer temporal convolutional network model to extract the state feature sequence of the monitoring variable changing over time; Step S302: Introduce a memory mechanism based on historical events to construct an event memory index table. The frequency of historical abnormal events, the fluctuation range of the monitoring variables involved, and the final risk level are normalized into a memory factor. The similarity of the current input state sequence is matched and the high-risk identification process is determined. Step S303: output the risk level assessment results of each section of the natural gas pipeline, wherein the risk level assessment results include high risk, medium risk and low risk, and synchronously transmit the risk level assessment results to the maintenance scheduling module.

5. The natural gas pipeline maintenance optimization method based on big data according to claim 4, characterized in that: The step S4 includes the following sub-steps: Step S401, reading the risk level assessment result output in step S3, and outputting the maintenance priority according to the risk level assessment result; Step S402: combining the maintenance personnel's location information, travel path information, equipment status information, and weather condition information to generate a maintenance scheduling candidate plan based on resource constraints; Step S403: Use reinforcement learning to train the scheduling optimization model, set a multi-objective scoring function including risk level reduction, response time, maintenance path length and maintenance cost, perform simulation training and output a maintenance scheduling plan.

6. The natural gas pipeline maintenance optimization method based on big data according to claim 5, characterized in that: The step S5 includes the following sub-steps: Step S501: When the risk level assessment result of any pipeline section exceeds the set threshold, an alarm record is generated and accompanied by the corresponding risk classification, key monitoring variables and reference fluctuation record; Step S502: Push the alarm record to the dispatch command center terminal and the inspection personnel terminal through the unified management platform, and publish it in the form of map, text prompt and voice broadcast; In step S503, based on the optimal maintenance scheduling plan outputted in step S4, a patrol task route map is generated. The patrol task route map includes the starting coordinates, necessary inspection points, estimated operation time, required maintenance equipment and inspection content, and is uploaded to the patrol terminal device for task navigation and status feedback.

7. The natural gas pipeline maintenance optimization method based on big data according to claim 6, characterized in that: The step S6 includes the following sub-steps: Step S601: collecting maintenance result data of the inspection personnel during the inspection task, the maintenance result data including actual trajectory data, maintenance start and end time, on-site image data, audio recording data and risk treatment result data, and writing it into the database for archiving; Step S602: incrementally train the risk prediction model using the maintenance result data to update the boundary threshold, classification weight parameters, and feature variable mapping relationships used for state identification in the model; In step S603, the spatiotemporal causal graph constructed in step S2 is structurally iterated, and the graph edge weights and node risk identifiers are corrected using the maintenance result data. The path subgraphs with prediction deviations greater than the set range are submitted to the model correction module for processing.

8. The natural gas pipeline maintenance optimization method based on big data according to claim 7, characterized in that: The memory mechanism based on historical events triggers a high-risk processing channel by constructing a memory factor based on historical abnormal events and matching it with the current state sequence.

9. The natural gas pipeline maintenance optimization method based on big data according to claim 8, characterized in that: When step S104 is executed, the structured monitoring data set is identified by global positioning system coordinates and timestamps.

10. A natural gas pipeline maintenance optimization system based on big data, which is applied to the natural gas pipeline maintenance optimization method based on big data according to any one of claims 1 to 9, characterized in that: It includes data acquisition module, graph construction module, model prediction module, solution generation module, early warning maintenance module and maintenance update module; The data acquisition module is used to construct a structured monitoring data set; The graph construction module is used to construct a graph model of pipeline status and risk factors; The model prediction module is used to establish a prediction model and introduce a memory mechanism to handle low-frequency anomalies; The scheme generation module is used to generate and optimize the maintenance scheduling scheme; The early warning maintenance module is used to trigger early warning and execute terminal maintenance instructions; The maintenance and update module is used to collect maintenance results for model and map updates.

Images