Geological disaster-oriented long-distance pipeline risk intelligent assessment system

Abstract

The application discloses a geological disaster-oriented long-distance pipeline risk intelligent assessment system and belongs to the technical field of infrastructure safety and geological disaster prevention and treatment. The system comprises a dynamic segmentation module, a geological disaster-prone area identification module, a pipeline vulnerability assessment module, a vulnerability and failure consequence weight fusion module, a comprehensive risk calculation module and a risk explanation and updating module. The system can realize fine identification of the risk along the line, interpretable risk contribution analysis and dynamic updating based on monitoring data, can output key risk sections, dominant factors and sectional risk indexes, and is suitable for geological disaster risk assessment and management of oil, natural gas, communication, water supply and drainage and other shallowly buried pipelines.

Description

Technical Field

[0001] This invention relates to the field of infrastructure safety and geological disaster prevention technology, and in particular to an intelligent risk assessment system for long-distance pipelines oriented towards geological disasters. It can be used for risk identification, assessment and dynamic management of various types of long-distance shallow buried pipelines such as oil, natural gas, communication cables, water supply and drainage pipelines, and hydrogen pipelines under the influence of geological disasters. Background Technology

[0002] Long-distance pipelines, as an important component of energy transmission, urban operation, and infrastructure security, are widely distributed in complex environments such as mountainous areas, hills, river valleys, and urban underground spaces. Affected by factors such as topography, geological structure, environmental exposure, human activities, and climate change, long-distance pipelines are susceptible to various types of geological disasters, including landslides, collapses, debris flows, ground subsidence, karst collapses, and earthquake-induced secondary disasters.

[0003] Once geological disasters interact with pipelines, they may cause pipeline deformation, breakage, rupture, leakage, power outage, or cable insulation damage, leading to significant economic losses, environmental pollution, energy disruptions, or risks to urban operations.

[0004] The existing technologies for geological hazard risk assessment of long-distance pipelines have the following main shortcomings:

[0005] (1) The segmentation method is highly subjective. The use of fixed lengths or manual experience in segmentation makes it difficult to reflect the spatial differences in geological hazards and cannot adapt to the complexity of different regions.

[0006] (2) The disaster susceptibility prediction model is too simple. They rely heavily on a single machine learning model or empirical model, have weak generalization ability, and are difficult to adapt to complex scenarios with multiple disasters, multiple regions, and multiple data sources.

[0007] (3) The weighting system is imperfect. Subjective weights are heavily influenced by expert experience, while objective weights ignore the correlation between indicators and are difficult to balance engineering experience and data characteristics.

[0008] (4) Risk outcomes lack interpretability and dynamic updating capability. Existing methods are mostly static assessments, which cannot be updated in real time based on monitoring data, disaster events or environmental changes, nor can they identify the dominant risk factors.

[0009] Therefore, there is an urgent need for a comprehensive intelligent assessment framework for geological hazard risks of long-distance pipelines that can integrate multi-source data, support multi-model collaboration, have dynamic segmentation and dynamic updating capabilities, and have good interpretability. Summary of the Invention

[0010] The purpose of this invention is to propose an intelligent risk assessment system for long-distance pipelines oriented towards geological hazards, in order to solve the problems of unscientific segmentation, single model, imperfect weight system, lack of interpretability and dynamic updating capability of risk results in the existing technology; this invention can realize intelligent, refined, interpretable and updatable assessment of geological hazard risks along the pipeline.

[0011] To achieve the above objectives, the present invention adopts the following technical solution: A smart risk assessment system for long-distance pipelines oriented towards geological hazards, the system comprising: The data acquisition and preprocessing module is used to acquire and uniformly process multi-source heterogeneous data. The dynamic segmentation module is used to dynamically segment long-distance pipelines based on pipeline attributes, topography, geological structure, environmental exposure and monitoring data, using a data-driven approach, and to optimize the segmentation results by applying constraints. The geological hazard susceptibility identification module is used to predict the probability of geological hazards occurring in different sections along the pipeline based on a multi-model collaborative mechanism, and generate a geological hazard susceptibility map. The pipeline vulnerability assessment module is used to construct a vulnerability evaluation system and calculate the vulnerability index of each segment based on pipeline structural parameters, operating parameters and their spatial relationship with geological hazards. The indicator weight fusion module is used to determine the comprehensive weight of vulnerability indicators and failure consequence indicators through a subjective and objective fusion mechanism. The comprehensive risk calculation module is used to integrate geological disaster susceptibility, pipeline vulnerability, and disaster consequences to calculate the risk value of each segment and classify the risk level. The risk interpretation and update module is used to output the dominant risk factors, key risk segments and early warning information, and dynamically update the risk assessment results based on new data.

[0012] Preferably, the dynamic segmentation module includes: Multi-source data clustering units are used to generate initial segments using density-based or spectral clustering methods; The segment length constraint unit is used to optimize the initial segmentation by merging or splitting based on the preset minimum and maximum segment lengths. The deformation monitoring constraint unit is used to automatically subdivide the segments of the deformation anomaly area based on the gradient or rate threshold of the surface deformation monitoring data.

[0013] Preferably, the geological hazard susceptibility identification module includes: The statistical model unit is used for prediction based on statistical analysis of historical disaster data and disaster-prone factors; A machine learning model unit for performing predictions using at least one of the algorithms, including random forest, support vector machine, and XGBoost. Deep learning model unit, used to perform time series predictions using algorithms including convolutional neural networks or long short-term memory networks; The model fusion unit is used to integrate or weightedly fuse the outputs of at least two types of statistical model units, machine learning model units, and deep learning model units to generate the final susceptibility result.

[0014] Preferably, in the pipeline vulnerability assessment module, the indicators on which the vulnerability evaluation system is based include: pipeline material, wall thickness, burial depth, corrosion status, insulation class or sheath material, operating pressure, medium type, joint type, and the angle between the pipeline axis and the direction of geological disaster movement.

[0015] Preferably, the indicator weight fusion module achieves adaptive weight fusion based on subjective weight, objective weight, and difference coefficient, wherein the subjective weight is determined by the analytic hierarchy process and the objective weight is determined by the entropy weight method.

[0016] Preferably, the risk results output by the comprehensive risk calculation module include a risk level map, a risk index spatial distribution map, and a risk change trend map.

[0017] Preferably, the risk interpretation and update module includes: The risk barrier analysis unit is used to identify the key dominant factors affecting the risk value and their contribution ranking; and The dynamic update unit is used to trigger the recalculation of relevant modules to update the risk results when new geological disaster event data, surface deformation monitoring data, or environmental change data are input.

[0018] Preferably, the data acquisition and preprocessing module supports the processing of multi-source heterogeneous data including: GIS data, remote sensing images, InSAR deformation monitoring data, GNSS monitoring data, pipeline attribute data, historical disaster records, population density data, and building distribution data.

[0019] Preferably, the system is suitable for conducting geological hazard risk assessments on at least one of oil pipelines, natural gas pipelines, water pipelines, optical fiber cables, and hydrogen pipelines.

[0020] Preferably, the geological hazards include at least one of landslides, collapses, debris flows, ground subsidence, karst collapses, and earthquake-induced secondary hazards.

[0021] Compared with the prior art, the present invention has the following beneficial effects: (1) Scientific segmentation: Based on the dynamic segmentation method driven by multi-source data, it overcomes the problem of strong subjectivity in traditional segmentation.

[0022] (2) Intelligent prediction: adopt a multi-model collaborative mechanism to improve the accuracy and adaptability of geological disaster susceptibility prediction.

[0023] (3) Adaptive weighting: The subjective and objective weighting system takes into account both engineering experience and data characteristics.

[0024] (4) Risk interpretability: By analyzing risk barriers, identify the dominant factors and improve the transparency of the results.

[0025] (5) Dynamic assessment: Risk results can be automatically updated based on new data, which is suitable for long-term operation and management.

[0026] (6) Strong applicability: It can be used in multiple types of pipelines, multiple disasters, and multiple regional scenarios. Attached Figure Description

[0027] 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.

[0028] Figure 1 is an overall architecture diagram of the intelligent risk assessment system for long-distance pipelines oriented towards geological disasters proposed in this invention; Figure 2 is a flowchart of the intelligent risk assessment method for long-distance pipelines to geological disasters based on the system proposed in this invention; Figure 3 is a schematic diagram of the risk interpretation and update mechanism proposed in this invention; Figure 4 is a schematic diagram of the geological disaster risk assessment results of a natural gas long-distance pipeline proposed in this invention. Detailed Implementation

[0029] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0030] The following description, in conjunction with relevant accompanying drawings and specific examples, illustrates the intelligent risk assessment system for long-distance pipelines oriented towards geological disasters proposed in this invention.

[0031] Example 1: Please see Figure 1-3 This invention proposes an intelligent risk assessment system for long-distance pipelines oriented towards geological disasters, specifically including the following: (1) Data acquisition and preprocessing module Acquire DEM, geological maps, remote sensing images, InSAR deformation data, pipeline attributes, historical disaster records, population and building data, etc.

[0032] (2) Dynamic segmentation module Based on multi-source data such as pipeline attributes, topography, geological structure, environmental exposure, and surface deformation monitoring, the pipeline is dynamically segmented through clustering or other data-driven methods, and the segmentation results can be optimized by combining minimum segment length constraints, maximum segment length constraints, or monitoring data.

[0033] (3) Geological disaster susceptibility identification module Inputting factors such as topography, geology, environment, and human activities, and combining them with historical data on geological disasters, a multi-model collaborative mechanism is used to generate a disaster susceptibility map, which predicts the probability of occurrence of geological disasters such as landslides, collapses, debris flows, and ground subsidence. The models can include statistical models, machine learning models, or deep learning models, and the final susceptibility results are generated through integration or weighting.

[0034] (4) Pipeline vulnerability and failure consequences assessment module A vulnerability assessment system is constructed based on pipeline structural parameters (including pipeline material, wall thickness, burial depth, corrosion status, operating pressure, insulation class, and the relationship between the pipeline and the disaster direction) and their spatial relationship with the disaster, and a vulnerability index is calculated to quantify the possibility of pipeline damage under the action of geological disasters.

[0035] (5) Indicator weight fusion module A subjective-objective fusion mechanism is adopted, which combines subjective weights, objective weights, and difference coefficients to achieve adaptive fusion of vulnerability and failure consequence index weights.

[0036] (6) Comprehensive Risk Calculation Module A comprehensive risk model is constructed based on the susceptibility to geological disasters, the vulnerability of pipelines, and the consequences of failure, and outputs the risk level, risk index, or spatial risk distribution.

[0037] (7) Risk Explanation and Update Module like Figure 3 As shown, an analytical method for determining dominant risk factors is employed by identifying key influencing factors, blocking pathways, and their contributions in the risk formation process. Key risk factors are identified based on the risk barrier analysis method, and the risk results are dynamically updated according to new geological disaster events, monitoring data, or environmental changes.

[0038] Example 2: Based on Embodiment 1, but with a difference, this invention uses the geological hazard risk assessment of a long-distance natural gas pipeline in a mountainous area as an example, combined with the system proposed in this invention, and specifically includes the following: This embodiment takes a long-distance natural gas pipeline in a mountainous area as an example, and uses the intelligent risk assessment system proposed in this invention to identify and assess geological disaster risks, including: (1) Data acquisition and preprocessing Collect the following multi-source data: 1.1) Topographic and geomorphological data (DEM, slope, aspect, surface roughness, etc.); 1.2) Geological structure data (lithology, faults, stratigraphic occurrence, etc.); 1.3) Environmental exposure data (population density, building distribution, road distribution, etc.); 1.4) Surface deformation monitoring data (InSAR time-series deformation); 1.5) Pipeline attribute data (material, wall thickness, burial depth, corrosion status, operating pressure, etc.); 1.6) Historical geological disaster data (landslides, collapses, debris flows, etc.); The above data is then processed by coordinate unification, rasterization, and normalization.

[0039] (2) Dynamic segmentation Data such as pipeline attributes, topography, geological structure, and environmental exposure are input into the dynamic segmentation module, and an initial segmentation result is generated using a data-driven clustering method.

[0040] Then, optimization is performed based on the following constraints: 2.1) Minimum segment length constraint (to avoid excessively short segments); 2.2) Maximum segment length constraint (to avoid excessively long segments); 2.3) Surface deformation monitoring constraints (automatic subdivision of abnormal deformation areas); The final product consists of several pipeline segments of varying lengths, each sensitive to geological hazards.

[0041] (3) Identification of geological hazard susceptibility Inputting factors such as topography, geology, environment, and human activities, and combining them with historical geological disaster data along the natural gas pipeline, a multi-model collaborative mechanism is used for prediction, including but not limited to: 3.1) Statistical model; 3.2) Machine learning models; 3.3) Deep learning models; The final geological hazard susceptibility map is generated through model integration.

[0042] (4) Pipeline vulnerability assessment A vulnerability assessment system is constructed based on the following indicators: 4.1) Pipeline structural parameters (material, wall thickness, burial depth, corrosion status); 4.2) Pipeline operating parameters (pressure, medium type); 4.3) Relationship between pipeline and disaster direction (e.g., the angle between the landslide movement direction and the pipeline axis); 4.4) Topographic features around the pipeline (slope, aspect, etc.); Output the vulnerability index for each segment.

[0043] (5) Integration of indicator weights Adopting a subjective and objective integration mechanism: Subjective weighting: Constructing a hierarchical model based on expert knowledge; Objective weights: calculated based on data differences and correlations; Difference coefficient: used to adjust the ratio of subjective to objective weights; The final comprehensive weighting of vulnerability and failure consequences indicators is generated.

[0044] (6) Comprehensive risk calculation The susceptibility to geological disasters, pipeline vulnerability, and disaster consequences are input into a comprehensive risk model to calculate the risk value for each segment and classify the risk level (e.g., Figure 4 (As shown).

[0045] (7) Risk explanation and output System output: 7.1) Risk Level Chart; 7.2) Key Risk Section 7.3) Dominant Risk Factors 7.4) Ranking of Risk Contribution It provides a basis for pipeline inspection, maintenance and planning.

[0046] Example 3: Based on Embodiment 1, but with a difference, this invention uses the geological hazard risk assessment of a water diversion pipeline as an example, combined with the system proposed in this invention, and specifically includes the following: This embodiment takes a water transmission pipeline (pressure pipeline) of a large-scale water diversion project as the object, and uses the risk intelligent assessment system proposed in this invention to assess its risk under the influence of ground subsidence, karst collapse and earthquake-induced disasters, including: (1) Data acquisition and preprocessing Collect the following multi-source data: 1.1) Topographic and geomorphological data: DEM, slope, valley topography index; 1.2) Geological structure data: karst development zone, fault zone, weak interlayer; 1.3) Hydrological data: changes in groundwater level, rainfall, and seepage field; 1.4) Surface deformation monitoring data: InSAR, ground subsidence monitoring wells; 1.5) Water pipeline attributes: pipe diameter, wall thickness, material, weld type, operating pressure; 1.6) Historical disaster data: karst collapse points, subsidence centers, and ground fissures; The above data was preprocessed, including coordinate unification, rasterization (20 m), and normalization.

[0047] (2) Dynamic segmentation The segmentation was performed using a method of "spectral clustering + ground settlement gradient constraint".

[0048] The constraints include: 2.1) Minimum segment length 100 m; 2.2) Maximum segment length 3000 m; 2.3) Automatic subdivision when settlement gradient > 3 mm / m; The final result is 52 segments.

[0049] (3) Identification of geological hazard susceptibility A synergistic mechanism of RF + SVM + LSTM (time series settlement prediction) is used to predict the following hazards: 3.1) Susceptibility to ground subsidence; 3.2) Susceptibility to karst collapse; 3.3) Susceptibility to secondary disasters induced by earthquakes; Output a comprehensive susceptibility index.

[0050] (4) Vulnerability assessment of water pipelines Vulnerability indicators include: 4.1) Pipe wall thickness and yield strength; 4.2) Joint type and quality grade; 4.3) Operating pressure (0.6–1.2 MPa); 4.4) The angle θ between the pipeline and the settlement direction; 4.5) Degree of development of karst cavities; Output the vulnerability index.

[0051] (5) Integration of indicator weights Adopting a subjective and objective integration mechanism: Subjective weighting: AHP Objective weighting: Entropy weighting method Coefficient of difference λ = 0.37 (6) Comprehensive risk calculation Consequence indicators include: 6.1) Impact on water supply; 6.2) Size of the affected population; 6.3) Dependence of critical facilities on water supply; Output a risk level map (level 5).

[0052] (7) Risk explanation and output System output: 7.1) High-risk segment 5; 7.2) Dominant factors: subsidence gradient, karst development, and operational pressure; 7.3) Risk change trend chart; 7.4) Suggestions for scheduling and inspection.

[0053] Example 4: Based on Embodiment 1, but with a difference, this invention uses the geological hazard risk assessment of shallow-buried optical fiber cable laying in mountainous areas as an example, combined with the system proposed in this invention, to illustrate the following: This embodiment takes a shallowly buried long-distance optical fiber (fiber) communication cable in a mountainous area as an example. The risk intelligent assessment system proposed in this invention is used to identify, assess, and dynamically update the risks of the cable under the influence of landslides, collapses, debris flows, and surface deformation, including: (1) Data acquisition and preprocessing Collect the following multi-source data: 1.1) Topographic data: DEM (5 m resolution), slope, aspect, curvature; 1.2) Geological structure data: lithology, faults, joint density, stratigraphic attitude; 1.3) Environmental exposure data: distribution of communication sites, density of optical fiber nodes, and location of important communication facilities; 1.4) Surface deformation monitoring data: InSAR time-series deformation (12-day revisit), GNSS monitoring points; 1.5) Optical cable attribute data: optical cable type (ADSS / ordinary optical cable), sheath material, burial depth (0.6–1.2 m), laying method; 1.6) Historical disaster data: landslide sites, collapse sites, debris flow channels; The above data was preprocessed, including coordinate unification, rasterization (10 m), normalization, and missing value imputation.

[0054] (2) Dynamic segmentation Data such as fiber optic cable attributes, topography, geological structure, and environmental exposure are input into the dynamic segmentation module, and the initial segments are generated using the method of "density-based clustering (DBSCAN) + minimum segment length constraint (30 m)".

[0055] Then, optimization is performed based on the following constraints: 2.1) Minimum segment length constraint: Avoid excessively short segments; 2.2) Maximum segment length constraint: not exceeding 1200 m; 2.3) InSAR deformation rate threshold: Automatic subdivision occurs when the deformation rate > 6 mm / yr; Ultimately, 112 optical cable segments with geological hazard sensitivity were generated.

[0056] (3) Identification of geological hazard susceptibility Input factors such as terrain, geology, environment, and human activities, and use a multi-model collaborative mechanism of "Random Forest (RF) + XGBoost + CNN" for prediction.

[0057] The model output includes: 3.1) Landslide susceptibility index; 3.2) Collapse susceptibility index; 3.3) Debris flow susceptibility index; 3.4) Surface deformation anomaly index; The final susceptibility map is generated by weighted fusion.

[0058] (4) Vulnerability assessment of optical cables Vulnerability indicators include: 4.1) Durability of the sheath material; 4.2) Fiber optic cable burial depth (shallow sections are more vulnerable); 4.3) Laying method (direct burial / pipe laying); 4.4) The angle θ between the optical cable and the direction of landslide movement; 4.5) Surface deformation gradient; 4.6) Fiber optic cable node density (the denser the nodes, the more serious the consequences); Output the vulnerability index for each segment.

[0059] (5) Integration of indicator weights Adopting a subjective and objective integration mechanism: Subjective weighting: AHP (5 experts); Objective weighting: Entropy weighting method; The coefficient of difference λ = 0.38; Generate a comprehensive weighted index of final vulnerability and failure consequences.

[0060] (6) Comprehensive risk calculation Consequence indicators include: 6.1) Impact level of communication interruption; 6.3) Number of important communication nodes; 6.3) Communication traffic density; Output a risk level map (level 5).

[0061] (7) Risk explanation and output System output: 7.1) Key Risk Sections (9 sections in total); 7.2) Dominant risk factors (deformation rate, slope, fiber optic cable burial depth); 7.3) Ranking of risk contribution; 7.4) Inspection priority recommendations.

[0062] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A smart risk assessment system for long-distance pipelines oriented towards geological disasters, characterized in that, The system includes: The data acquisition and preprocessing module is used to acquire and uniformly process multi-source heterogeneous data. The dynamic segmentation module is used to dynamically segment long-distance pipelines based on pipeline attributes, topography, geological structure, environmental exposure and monitoring data, using a data-driven approach, and to optimize the segmentation results by applying constraints. The geological hazard susceptibility identification module is used to predict the probability of geological hazards occurring in different sections along the pipeline based on a multi-model collaborative mechanism, and generate a geological hazard susceptibility map. The pipeline vulnerability assessment module is used to construct a vulnerability evaluation system and calculate the vulnerability index of each segment based on pipeline structural parameters, operating parameters and their spatial relationship with geological hazards. The indicator weight fusion module is used to determine the comprehensive weight of vulnerability indicators and failure consequence indicators through a subjective and objective fusion mechanism. The comprehensive risk calculation module is used to integrate geological disaster susceptibility, pipeline vulnerability, and disaster consequences to calculate the risk value of each segment and classify the risk level. The risk interpretation and update module is used to output the dominant risk factors, key risk segments and early warning information, and dynamically update the risk assessment results based on new data.

2. The system according to claim 1, characterized in that, The dynamic segmentation module includes: Multi-source data clustering units are used to generate initial segments using density-based or spectral clustering methods; The segment length constraint unit is used to optimize the initial segmentation by merging or splitting based on the preset minimum and maximum segment lengths. The deformation monitoring constraint unit is used to automatically subdivide the segments of the deformation anomaly area based on the gradient or rate threshold of the surface deformation monitoring data.

3. The system according to claim 1, characterized in that, The geological hazard susceptibility identification module includes: The statistical model unit is used for prediction based on statistical analysis of historical disaster data and disaster-prone factors; A machine learning model unit for performing predictions using at least one of the algorithms, including random forest, support vector machine, and XGBoost. Deep learning model unit, used to perform time series predictions using algorithms including convolutional neural networks or long short-term memory networks; The model fusion unit is used to integrate or weightedly fuse the outputs of at least two types of statistical model units, machine learning model units, and deep learning model units to generate the final susceptibility result.

4. The system according to claim 1, characterized in that, The vulnerability assessment module for pipelines includes the following indicators: pipeline material, wall thickness, burial depth, corrosion status, insulation class or sheath material, operating pressure, medium type, joint type, and the angle between the pipeline axis and the direction of geological hazard movement.

5. The system according to claim 1, characterized in that, The indicator weight fusion module achieves adaptive weight fusion based on subjective weight, objective weight, and difference coefficient. The subjective weight is determined by the analytic hierarchy process, and the objective weight is determined by the entropy weight method.

6. The system according to claim 1, characterized in that, The risk results output by the comprehensive risk calculation module include a risk level map, a risk index spatial distribution map, and a risk change trend map.

7. The system according to claim 1, characterized in that, The risk interpretation and update module includes: The risk barrier analysis unit is used to identify the key dominant factors affecting the risk value and their contribution ranking; and The dynamic update unit is used to trigger the recalculation of relevant modules to update the risk results when new geological disaster event data, surface deformation monitoring data, or environmental change data are input.

8. The system according to claim 1, characterized in that, The data acquisition and preprocessing module supports processing multi-source heterogeneous data, including: GIS data, remote sensing images, InSAR deformation monitoring data, GNSS monitoring data, pipeline attribute data, historical disaster records, population density data, and building distribution data.

9. The system according to claim 1, characterized in that, The system is applicable to geological hazard risk assessment of at least one of oil pipelines, natural gas pipelines, water pipelines, optical fiber cables, and hydrogen pipelines.

10. The system according to claim 9, characterized in that, The geological hazards mentioned include at least one of the following: landslides, collapses, debris flows, ground subsidence, karst collapses, and earthquake-induced secondary disasters.

Images