A geological disaster-oriented long-distance oil and gas pipeline risk intelligent evaluation method

Abstract

The application discloses a kind of geological disaster-oriented long oil and gas pipeline risk intelligent evaluation methods, comprising: based on pipeline working condition (steel grade, wall thickness, weld, corrosion, operating pressure etc.) and multi-source heterogeneous data (DEM, geology, InSAR, rainfall, historical disasters, pipeline monitoring etc.) are segmented dynamically;Adopt multi-model integrated prediction landslide, collapse, debris flow, settlement etc. Disaster-prone E;Construction oil and gas special vulnerability index system and failure consequence index system, and through AHP and entropy weight combination empowerment (including difference coefficient λ) calculate vulnerability index V and consequence index C;According to R = E * V * C Or weighted linear model calculates segmented risk value and is divided according to grade;Based on contribution degree analysis identifies leading factor and dynamically updates risk result under monitoring trigger condition, outputs key risk section, risk change trend and inspection / emergency suggestion.This method considers oil and gas working condition characteristics and data driving ability, is convenient for engineering implementation and decision support.

Description

Technical Field

[0001] This invention belongs to the field of oil and gas pipeline engineering safety and geological disaster prevention technology. Specifically, it relates to an intelligent risk assessment method for long-distance oil and gas pipelines oriented towards geological disasters. It is used to identify, quantify, interpret and dynamically update the geological disaster risks along long-distance natural gas pipelines, crude oil pipelines and refined oil pipelines based on oil and gas operating conditions and multi-source monitoring data, so as to support inspection, operation and maintenance and emergency decision-making. Background Technology

[0002] Long-distance oil and gas pipelines, as key infrastructure for national energy transportation, extensively traverse mountainous areas, river valleys, soft strata, and areas prone to geological disasters. Affected by factors such as topography, geological structure, rainfall, surface deformation, and human activities, oil and gas pipelines are susceptible to threats such as landslides, collapses, debris flows, ground subsidence, and karst subsidence.

[0003] Existing risk assessment methods for oil and gas pipelines have the following main shortcomings: (1) Insufficient coupling between segmentation method and oil and gas pipeline operating conditions Many pipelines use fixed lengths or experience-based segmentation, failing to incorporate oil and gas-specific factors such as pipeline pressure, welds, and corrosion into the segmentation rules, resulting in the dilution or omission of critical risk segments.

[0004] (2) The disaster susceptibility model was not optimized for the characteristics of oil and gas pipelines. Existing models are mostly general geological disaster prediction models, which do not take into account the terrain sensitivity and disaster triggering mechanisms along oil and gas pipelines.

[0005] (3) The vulnerability index system lacks characteristics of the oil and gas industry. Key parameters such as steel grade, weld quality, corrosion level, operating pressure, and risk of explosion after leakage have not been systematically quantified.

[0006] (4) Lack of interpretability and dynamic update mechanism The assessment results are difficult to use directly for inspection prioritization and emergency response, and cannot be automatically updated based on real-time monitoring data.

[0007] To address the aforementioned issues, there is an urgent need for an intelligent geological hazard risk assessment method that is interpretable, updatable, and quantifiable, tailored to the characteristics of oil and gas pipeline projects. Summary of the Invention

[0008] The purpose of this invention is to provide an intelligent risk assessment method for long-distance oil and gas pipelines oriented towards geological hazards to solve the problems mentioned in the background art. This invention achieves accurate identification, quantitative classification and engineering output of risks along the pipeline by introducing oil and gas-specific indicators, subjective and objective combined weighting, interpretable contribution analysis and monitoring-triggered dynamic update mechanism.

[0009] To achieve the above objectives, the present invention adopts the following technical solution: A smart risk assessment method for long-distance oil and gas pipelines oriented towards geological hazards includes the following steps: S1. Multi-source data acquisition and preprocessing: Acquire and standardize multi-source data information, including oil and gas pipeline attributes, operating parameters, geological disaster data, pipeline monitoring data, and environmental exposure data; S2. Dynamic segmentation: Based on pipeline pressure change points, weld locations, corrosion anomaly points, and surface deformation data, the target oil and gas pipeline is dynamically segmented to obtain several pipeline segments. S3. Disaster Susceptibility Identification: A multi-model collaborative mechanism is used to predict the probability of occurrence of various geological disasters and obtain the disaster susceptibility index corresponding to each pipeline section. S4. Vulnerability Assessment System Construction: Based on the properties, operating parameters and geological disaster data of oil and gas pipelines, a vulnerability assessment system is constructed, and the vulnerability index of each pipeline section is calculated. S5. Consequence Calculation: Based on the leakage consequence estimation, a consequence index system is constructed to calculate the consequence index of each pipeline section. S6. Comprehensive Risk Calculation: The evaluation indicators are normalized, and the final weight of each indicator is determined by a combination of subjective and objective weights. Then, the disaster susceptibility index, vulnerability index and consequence index are calculated. S7. Risk Classification and Output: Based on the disaster susceptibility index, vulnerability index and consequence index, the comprehensive risk value of each pipeline section is calculated using a preset risk calculation model, and the risk level is classified according to the obtained comprehensive risk value. S8. Dynamic Update: Perform contribution analysis on the comprehensive risk value to identify the dominant risk factors, and dynamically update the comprehensive risk value and its level when the preset monitoring trigger conditions are met. S9. Threshold Determination: Based on historical data and accident records, a combination of statistical analysis, sensitivity analysis, and expert consultation is used to determine the various trigger thresholds and parameters used for dynamic segmentation and dynamic updates.

[0010] Preferably, the oil and gas pipeline attributes mentioned in S1 include steel grade, wall thickness, weld location and type, corrosion level, burial depth, and pipeline protection measures; the operating parameters include pipeline pressure, temperature, medium type, and historical pressure fluctuation records.

[0011] Preferably, the dynamic segmentation in S2 specifically includes the following: S2.1. The first segmentation is carried out using pipeline pressure change points, weld locations, corrosion anomaly points, and surface deformation data as initial segmentation nodes. S2.2. The first segmentation results are further subdivided by combining topographic data and InSAR surface deformation monitoring data. S2.3 Set minimum / maximum segment length constraints during the segmentation process to balance recognition accuracy and engineering feasibility.

[0012] Preferably, the geological hazards mentioned in S3 include landslides, collapses, debris flows, ground subsidence, and karst collapses; the multi-model collaborative mechanism uses at least two models, including statistical models, random forests, XGBoost, convolutional neural networks, or long short-term memory networks, for prediction, and generates the final hazard susceptibility index through weighted fusion or stacking methods; wherein, the weighting coefficients of the weighted fusion are determined through cross-validation or historical case backtracking.

[0013] Preferably, in step S6, the comprehensive indicators are all calculated using a unified weighted calculation framework, specifically as follows: For the normalized index vector With weight vector Calculated using a weighted linear combination:

[0014] In the formula, Y This indicates the comprehensive index to be calculated, which may be a disaster susceptibility index, vulnerability index, or consequence index. Represents the normalized i-th j Item; Indicates the first j The weight of each indicator; the weight vector w is calculated by combining the subjective weight vector and the objective weight vector, and the calculation formula is:

[0015] In the formula, This represents the active weight vector, which is derived through the analytic hierarchy process and has undergone a consistency test. This represents the objective weight vector, which is calculated using the entropy weighting method. Represents the coefficient of difference. It is determined through methods such as historical accident retrospection, cross-validation, or expert consultation.

[0016] Preferably, in calculating the comprehensive index Y When an individual indicator exceeds a preset threshold, a nonlinear penalty mapping function is used to amplify its risk contribution; wherein, the nonlinear penalty mapping function is a quadratic penalty function or an exponential amplification function.

[0017] Preferably, the contribution analysis in S8 uses Shapley value, feature importance ranking, or feature contribution method based on a locally interpretable model to identify the dominant risk factors for each pipeline segment.

[0018] Preferably, the monitoring triggering conditions in S8 include: the InSAR deformation rate exceeding a preset deformation threshold, the cumulative amount of short-term heavy rainfall exceeding a preset rainfall threshold, abnormal pipeline pressure monitoring values, or pipeline strain monitoring values ​​exceeding a preset strain threshold.

[0019] Preferably, the method is applicable to long-distance natural gas pipelines, crude oil pipelines, and refined oil pipelines, and differentiated processing rules are set in the vulnerability assessment system and consequence indicator system for pipelines with different transport media.

[0020] The present invention further protects a computer device, the computer device including a processor and a memory, the memory storing at least one instruction, at least one program, code set or instruction set, the instruction, program, code set or instruction set being loaded and executed by the processor to implement the above-mentioned intelligent risk assessment method for long-distance oil and gas pipelines oriented towards geological disasters.

[0021] The present invention further provides a computer-readable storage medium storing at least one instruction, at least one program, code set, or instruction set, wherein the instruction, program, code set, or instruction set is loaded and executed by a processor to implement the above-mentioned intelligent risk assessment method for long-distance oil and gas pipelines oriented towards geological hazards.

[0022] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) Oil and gas operating conditions coupling: This invention incorporates segmentation and vulnerability assessment directly into oil and gas-specific parameters, effectively improving identification accuracy.

[0023] (2) Balancing data and expert input: This invention combines subjective and objective weighting, taking into account both engineering experience and data characteristics, and the weights can be adjusted according to the amount of data and the stage of the project.

[0024] (3) Explainable and operable: This invention provides a clear basis for inspection and emergency response through contribution analysis, and proposes a dynamic update mechanism to support real-time risk management.

[0025] (4) The project is feasible: The normalization, weighting, aggregation, and threshold determination methods proposed in this invention are easy to implement and verify in engineering systems. Attached Figure Description

[0026] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings involved in the embodiments are now briefly described. Obviously, the drawings in the following description are merely illustrative of some embodiments of the present invention. For those skilled in the art, other forms of drawings can be constructed based on these drawings without creative effort.

[0027] Figure 1 This is a schematic diagram of the overall architecture of the intelligent risk assessment method for long-distance oil and gas pipelines proposed in this invention. Figure 2 This is a flowchart of the intelligent risk assessment method for long-distance oil and gas pipelines mentioned in this invention; Figure 3 This is a schematic diagram of the risk interpretation and dynamic update mechanism proposed in this invention. Detailed Implementation

[0028] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0029] This invention proposes an intelligent risk assessment method for long-distance oil and gas pipelines oriented towards geological hazards, comprising: 1. Multi-source data acquisition and preprocessing: Collect and standardize pipeline properties (steel grade, wall thickness, weld, corrosion level, burial depth), operating parameters (pressure, temperature, medium), geological and environmental data (DEM, geological map, InSAR, rainfall, historical disasters), pipeline monitoring data (strain, leak detection), and exposure data (population, critical facilities).

[0030] 2. Dynamic segmentation based on operating condition constraints: Using weld seams, pressure surges, corrosion anomalies, and surface deformation as initial nodes, secondary subdivisions are performed by combining terrain and monitoring trigger rules, and minimum / maximum segment length constraints are set to balance identification accuracy and engineering feasibility.

[0031] 3. Disaster susceptibility identification: For different types of disasters (landslides, collapses, debris flows, subsidence, karst), multiple models (statistical models, random forests, XGBoost, CNN, LSTM, etc.) are used in synergy, and a susceptibility index is generated through weighted fusion or stacking. .

[0032] 4. Vulnerability and Consequence Indicator System: A vulnerability index set and a failure consequence index set specific to oil and gas were constructed. After normalization of all indicators, the final weights were obtained by combining subjective weights (AHP) and objective weights (entropy weight method). And calculate each comprehensive indicator according to a unified weighted framework. (Applicable to vulnerability) and consequences ).

[0033] 5. Comprehensive Risk Calculation and Classification: Using a multiplication model Or weighted linear model Calculate segmented risk values ​​and map them to Level 5 risk according to preset levels.

[0034] 6. Explainability and Dynamic Updates: Dominant factors are identified based on Shapley values ​​or feature importance methods; when monitoring trigger conditions (deformation rate, short-duration heavy rainfall, pressure anomalies, etc.) are met, the segmentation is automatically updated and recalculated. , , , Output key risk segments, contribution ranking, and inspection / emergency recommendations.

[0035] 7. Threshold Determination Method: Trigger thresholds and parameters were determined by combining historical accident retrospective analysis, statistical quantile analysis, ROC curve analysis, and sensitivity analysis with expert consultation. Examples of thresholds and preferred engineering values ​​are listed in the implementation examples.

[0036] The following description, in conjunction with the accompanying drawings and specific examples, illustrates the intelligent risk assessment method for long-distance oil and gas pipelines oriented towards geological hazards proposed in this invention.

[0037] Example 1: Please see Figure 1-3 This example proposes an intelligent risk assessment method for long-distance oil and gas pipelines oriented towards geological hazards. The specific implementation steps are as follows: Step 1: Data Acquisition and Preprocessing Data items: Pipeline properties (steel grade, wall thickness, weld location / type, corrosion level, burial depth), operating parameters (pressure, temperature, medium), topography (DEM, slope), geology (faults, lithology), monitoring (InSAR time series, GNSS, pipeline strain), meteorology (rainfall), historical accident records, exposure (population, substations, towns).

[0038] Preprocessing: coordinate unification, rasterization, missing value imputation, outlier removal, normalization (range or logarithmic transformation).

[0039] Step 2: Dynamic Segmentation Initial nodes: weld location, pressure change point, corrosion anomaly point.

[0040] Secondary subdivision rule: If the deformation rate of a certain InSAR segment If the slope / terrain sensitivity exceeds the threshold, then further subdivision is required.

[0041] Segment length constraint: setting and The decision will be made by experts, taking into account feasibility.

[0042] Step 3: Disaster Susceptibility Identification Model set: Statistical Regression, Random Forest, XGBoost, CNN.

[0043] Fusion: The output probabilities of each model are weighted and fused according to the weights determined by cross-validation to obtain... .

[0044] Steps 4-5: Vulnerability and Consequence Calculation Indicator normalization: For each indicator Obtained by range or logarithmic normalization .

[0045] Weight calculation: The expert group constructs a judgment matrix to obtain... Entropy weight method calculation ;according to Obtain the final weights.

[0046] Unified calculation: for any comprehensive indicator (represent or )use

[0047] The contribution is amplified by a quadratic penalty mapping for terms exceeding the threshold.

[0048] Step 6: Comprehensive Risk Calculation Multiplication model: .

[0049] Linear model alternatives: And determine through historical retrospection .

[0050] Step 7: Risk Classification and Output Will The risk level is mapped to 5 levels, and a risk level chart, key risk segments, dominant factors, and inspection recommendations are output.

[0051] Step 8: Dynamic Update When the monitoring trigger conditions are met, S2–S6 will be automatically re-executed and the output will be updated.

[0052] Step 9: Threshold Determination Method Thresholds for deformation rate and rainfall were determined by using historical accident retrospective analysis and ROC curves; the final engineering thresholds were determined by combining sensitivity analysis and expert consultation.

[0053] Example 2: Example of a natural gas pipeline crossing a landslide zone in a mountainous area 1. Project Background: A 220 km long-distance natural gas pipeline traverses a landslide zone in a mountainous area, and InSAR monitoring and pipeline strain sensors are installed along the route.

[0054] 2. Data: DEM resolution 5 m, InSAR revisited for 12 days, pipeline pressure time series, weld location, corrosion detection report, historical landslide points.

[0055] 3. Segmentation: Using the weld seam and pressure change as initial nodes, the InSAR deformation rate vth was determined by the 90th quantile of historical landslide pre-slope data.

[0056] 4. Vulnerability Indicator Example: Steel grade is mapped to 0–1, wall thickness is normalized according to relative thickness, weld quality is mapped according to grade, and corrosion grade is normalized according to corrosion depth.

[0057] 5. Example of consequence indicators: The leakage rate was estimated based on pipe diameter and rupture length and then logarithmically normalized; the probability of combustion and explosion was estimated based on the flammability of the medium and the density of surrounding combustibles.

[0058] 6. Weight Determination: The expert group obtained ws through AHP, and wo through entropy weight method. λ was selected by backtracking through historical accidents to minimize the identification error.

[0059] 7. Output: Several key risk sections were identified and inspection and reinforcement recommendations were given according to priority.

[0060] Example 3: Example of a crude oil pipeline crossing a valley and soft soil area 1. Project Background: Crude oil pipelines face risks of ground subsidence and erosion in the soft soil areas of river valleys.

[0061] 2. Susceptibility Model: By combining groundwater level, rainfall and surface subsidence time series, LSTM is used to predict subsidence trend and it is fused with RF to obtain E.

[0062] 3. Vulnerability and Consequences: Vulnerability assessment focuses on burial depth, sheath integrity, and weld quality; consequences assessment focuses on environmental sensitivity (drinking water sources, wetlands) and economic loss estimation.

[0063] 4. Threshold determination: Trigger thresholds were determined by retrospective analysis of historical settlement events and ROC curves, and sensitivity analysis was used to verify the robustness of the thresholds.

[0064] 5. Dynamic updates: When the groundwater level changes abruptly or short-term heavy rainfall exceeds the threshold, the risk is automatically recalculated and emergency response suggestions are generated.

[0065] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A smart risk assessment method for long-distance oil and gas pipelines oriented towards geological hazards, characterized in that, Includes the following steps: S1. Multi-source data acquisition and preprocessing: Acquire and standardize multi-source data information, including oil and gas pipeline attributes, operating parameters, geological disaster data, pipeline monitoring data, and environmental exposure data; S2. Dynamic segmentation: Based on pipeline pressure change points, weld locations, corrosion anomaly points, and surface deformation data, the target oil and gas pipeline is dynamically segmented to obtain several pipeline segments. S3. Disaster Susceptibility Identification: A multi-model collaborative mechanism is used to predict the probability of occurrence of various geological disasters and obtain the disaster susceptibility index corresponding to each pipeline section. S4. Vulnerability Assessment System Construction: Based on the properties, operating parameters and geological disaster data of oil and gas pipelines, a vulnerability assessment system is constructed, and the vulnerability index of each pipeline section is calculated. S5. Consequence Calculation: Based on the leakage consequence estimation, a consequence index system is constructed to calculate the consequence index of each pipeline section. S6. Comprehensive Risk Calculation: The evaluation indicators are normalized, and the final weight of each indicator is determined by a combination of subjective and objective weights. Then, the disaster susceptibility index, vulnerability index and consequence index are calculated. S7. Risk Classification and Output: Based on the disaster susceptibility index, vulnerability index and consequence index, the comprehensive risk value of each pipeline section is calculated using a preset risk calculation model, and the risk level is classified according to the obtained comprehensive risk value. S8. Dynamic Update: Perform contribution analysis on the comprehensive risk value to identify the dominant risk factors, and dynamically update the comprehensive risk value and its level when the preset monitoring trigger conditions are met. S9. Threshold Determination: Based on historical data and accident records, a combination of statistical analysis, sensitivity analysis, and expert consultation is used to determine the various trigger thresholds and parameters used for dynamic segmentation and dynamic updates.

2. The method according to claim 1, characterized in that, The oil and gas pipeline attributes mentioned in S1 include steel grade, wall thickness, weld location and type, corrosion level, burial depth and pipeline protection measures; the operating parameters include pipeline pressure, temperature, medium type and historical pressure fluctuation records.

3. The method according to claim 1, characterized in that, The dynamic segmentation described in S2 specifically includes the following: S2.

1. The first segmentation is carried out using pipeline pressure change points, weld locations, corrosion anomaly points, and surface deformation data as initial segmentation nodes. S2.

2. The first segmentation results are further subdivided by combining topographic data and InSAR surface deformation monitoring data. S2.3 Set minimum / maximum segment length constraints during the segmentation process to balance recognition accuracy and engineering feasibility.

4. The method according to claim 1, characterized in that, The geological hazards described in S3 include landslides, collapses, debris flows, ground subsidence, and karst collapses; the multi-model collaborative mechanism uses at least two models, including statistical models, random forests, XGBoost, convolutional neural networks, or long short-term memory networks, for prediction, and generates the final hazard susceptibility index through weighted fusion or stacking methods; wherein, the weighting coefficients of the weighted fusion are determined through cross-validation or historical case backtracking.

5. The method according to claim 1, characterized in that, In S6, the comprehensive indicators are all calculated using a unified weighted calculation framework, specifically as follows: For the normalized index vector With weight vector Calculated using a weighted linear combination: In the formula, Y This indicates the comprehensive index to be calculated, which may be a disaster susceptibility index, vulnerability index, or consequence index. Represents the normalized i-th j Item; Indicates the first j The weight of each indicator; The weight vector w is calculated by combining the subjective weight vector and the objective weight vector, and the calculation formula is as follows: In the formula, This represents the active weight vector, which is derived through the analytic hierarchy process and has undergone a consistency test. This represents the objective weight vector, which is calculated using the entropy weighting method. Represents the coefficient of difference. It is determined through methods such as historical accident retrospection, cross-validation, or expert consultation.

6. The method according to claim 5, characterized in that, In calculating comprehensive indicators Y When an individual indicator exceeds a preset threshold, a nonlinear penalty mapping function is used to amplify its risk contribution; wherein, the nonlinear penalty mapping function is a quadratic penalty function or an exponential amplification function.

7. The method according to claim 1, characterized in that, The contribution analysis described in S8 uses Shapley values, feature importance ranking, or feature contribution methods based on locally interpretable models to identify the dominant risk factors for each pipeline segment.

8. The method according to claim 7, characterized in that, The monitoring triggering conditions described in S8 include: InSAR deformation rate exceeding a preset deformation threshold, cumulative short-term heavy rainfall exceeding a preset rainfall threshold, abnormal pipeline pressure monitoring value, or pipeline strain monitoring value exceeding a preset strain threshold.

9. A computer device, characterized in that, The computer device includes a processor and a memory, wherein the memory stores at least one instruction, at least one program, code set, or instruction set, and the instruction, program, code set, or instruction set is loaded and executed by the processor to implement the intelligent risk assessment method for long-distance oil and gas pipelines oriented towards geological hazards as described in any one of claims 1-8.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores at least one instruction, at least one program, code set, or instruction set, which is loaded and executed by a processor to implement the intelligent risk assessment method for long-distance oil and gas pipelines oriented towards geological hazards as described in any one of claims 1-8.

Images