A multi-source data fusion gas pipe network risk intelligent evaluation method

Abstract

The present application relates to gas pipe network risk management and control technical field, especially to a kind of multi-source data fusion's gas pipe network risk intelligent evaluation method, comprising: gas pipe network is divided into several actual evaluation sections;Information collection is carried out to each actual evaluation section, detects the geographic information data of each actual evaluation section, and obtains all historical maintenance records of gas pipe network, analyzes the direct risk type data of each actual evaluation section, and is parsed, in combination with the geographic information data of corresponding actual evaluation section Determine the risk tendency discrete variable of the actual evaluation section Determine the risk tendency category of corresponding actual evaluation section;According to the risk tendency category of actual evaluation section, determine the parameter selection mode of subsequent evaluation, according to the evaluation result, determine whether the corresponding actual evaluation section needs to be early warned and send early warning information to the corresponding actual evaluation section.The present application improves the comprehensiveness of pipe network risk identification, the accuracy of early warning signal and the pertinence of preventive maintenance.

Description

Technical Field

[0001] This invention relates to the field of gas pipeline network risk management technology, and in particular to a method for intelligent risk assessment of gas pipeline networks using multi-source data fusion. Background Technology

[0002] As a core component of urban lifeline engineering, the safe and stable operation of gas pipeline networks is directly related to public safety, economic development, and social stability. With the continuous expansion of urban scale, gas pipeline systems are becoming increasingly complex, and their operating environment is becoming more volatile, facing traditional challenges from multiple risks such as pipeline corrosion, third-party damage, geological disasters, equipment aging, and lagging management. Currently, the industry's risk management of gas pipeline networks still largely relies on traditional models such as regular manual inspections, reactive alarms from isolated monitoring systems, and planned maintenance based on fixed cycles.

[0003] Traditional methods have significant limitations. They typically rely on data from a single or limited dimension, such as focusing only on pressure or leakage concentration, failing to systematically integrate operating parameters with macro-geographical information about the pipeline network. This data silo phenomenon leads to a one-sided perspective in risk assessment, making it difficult to comprehensively and accurately capture early hidden dangers under complex risk coupling effects. Secondly, existing methods mostly focus on real-time alarms for anomalies or malfunctions, lacking the ability to predict risk evolution trends. Essentially, they are passive responses and cannot achieve early warning and preventive intervention. Furthermore, relying on human experience to judge risk levels and formulate maintenance strategies also suffers from strong subjectivity, inconsistent standards, and low efficiency, making it difficult to meet the needs of refined and intelligent management of modern large-scale pipeline networks. Summary of the Invention

[0004] To address this, the present invention provides a multi-source data fusion-based intelligent risk assessment method for gas pipeline networks, which overcomes the problems of delayed early warning, one-sided risk identification, and insufficient prevention capabilities caused by isolated data and reliance on human experience in the prior art.

[0005] To achieve the above objectives, this invention provides a method for intelligent risk assessment of gas pipeline networks based on multi-source data fusion, comprising:

[0006] Step S1: Determine the overall structure of the gas pipeline network, and divide the gas pipeline network into several actual evaluation sections based on the overall structure.

[0007] Step S2: Information is collected for each actual assessment section, the geographic information data of each actual assessment section is detected, and all historical maintenance records of the gas pipeline network are obtained. The collected pipeline network detection data includes direct risk type data and indirect risk type data.

[0008] Step S3: Analyze and parse the direct risk type data of each actual assessment segment, combine it with the geographic information data of the corresponding actual assessment segment to determine the risk tendency discrete parameter of the actual assessment segment, and determine the risk tendency category of the corresponding actual assessment segment based on the risk tendency discrete parameter.

[0009] Step S4: Determine the parameter selection method for subsequent assessments based on the risk tendency category of the actual assessment segment, including:

[0010] Obtain the latest data on the aforementioned indirect risk types and conduct inspections of the gas pipeline network in the current actual assessment section;

[0011] Alternatively, by combining the historical maintenance records and unused geographic information data, the comprehensive risk value of the current actual assessment section can be calculated;

[0012] Step S5: Determine whether an early warning is needed for the corresponding actual assessment section based on the evaluation results of the parameter selection methods, and send early warning information to the corresponding actual assessment section.

[0013] As a preferred technical solution for the intelligent risk assessment method of gas pipeline network based on multi-source data fusion, the direct risk type data includes: pressure data, flow data, and leak detection data;

[0014] The indirect risk type data includes: cathodic protection data, temperature data, and vibration information;

[0015] The geographic information data includes: distribution of geological hazards, soil pH, and traffic load.

[0016] As a preferred technical solution for the intelligent risk assessment method of gas pipeline networks based on multi-source data fusion, in step S1, the gas pipeline network is divided into several actual assessment sections based on the overall structure, including:

[0017] Construct a topology model of the gas pipeline network, which includes all pipe segments as well as valves and pressure regulating stations connecting each pipe segment;

[0018] Based on the geographic information system, gas pipeline networks within the same administrative division boundary are divided into initial assessment sections; gas pipeline networks within the same geological unit boundary are divided into second assessment sections; and gas pipeline networks within the same road network unit are divided into actual assessment sections.

[0019] The initial evaluation segment includes several second evaluation segments, and the second evaluation segment includes several actual evaluation segments.

[0020] As a preferred technical solution for the intelligent risk assessment method of gas pipeline networks based on multi-source data fusion, step S3, which involves parsing the direct risk type data, includes:

[0021] Calculate the average pressure and average pressure deviation of the actual assessment section, and compare the average pressure deviation with the average pressure to obtain the pressure risk dispersion factor;

[0022] Calculate the average flow rate and average flow rate deviation of the actual assessment section, and compare the average flow rate deviation with the average flow rate to obtain the flow rate risk dispersion factor;

[0023] Obtain the integral value of the leakage concentration in the actual assessment section, compare the integral value of the leakage concentration with the leakage concentration threshold and round down to obtain the leakage risk dispersion factor.

[0024] As a preferred technical solution for the intelligent risk assessment method of gas pipeline networks based on multi-source data fusion, in step S3, the risk tendency discrete parameter of the actual assessment section is determined by combining the parsing results of the direct risk type data of the actual assessment section with the geographic information data of the corresponding actual assessment section, including:

[0025] The geological impact weights are determined based on the distribution of the geological hazards.

[0026] Calculate the product of the geological influence weight and the sum of the pressure risk discrete factor and the flow risk discrete factor, and add the leakage risk discrete factor. The result is denoted as the risk propensity discrete parameter.

[0027] The geological influence weight is set to 1, 2, or 3.

[0028] As a preferred technical solution for the intelligent risk assessment method of gas pipeline networks based on multi-source data fusion, the risk tendency category of the corresponding actual assessment section is determined according to the discrete parameter of risk tendency, including:

[0029] If the discrete parameter of risk propensity is greater than the standard risk propensity parameter, then the risk propensity category of the current actual assessment segment is determined to be the high-risk propensity category;

[0030] If the discrete parameter of risk propensity is less than or equal to the standard risk propensity parameter, then the risk propensity category of the current actual assessment segment is determined to be the low risk propensity category.

[0031] The standard risk propensity parameter has a value less than 1.

[0032] As a preferred technical solution for the intelligent risk assessment method of gas pipeline networks based on multi-source data fusion, in step S4, the parameter selection method for subsequent assessment is determined according to the risk tendency category of the actual assessment section, including:

[0033] If the actual assessment section is classified as a high-risk tendency category, the parameter selection method is to obtain the latest indirect risk type data and conduct an inspection of the gas pipeline network in the current actual assessment section.

[0034] If the actual assessment section is classified as low-high risk, the parameter selection method is to combine the historical maintenance records and unused geographic information data to calculate the comprehensive risk value of the current actual assessment section.

[0035] As a preferred technical solution for the intelligent risk assessment method of gas pipeline networks based on multi-source data fusion, in step S5, determining whether an early warning is needed for the corresponding actual assessment section based on the assessment results of indirect risk type data includes:

[0036] Obtain the latest cathodic protection data, temperature data, and vibration information for the current actual evaluation section;

[0037] If the cathodic protection data is lower than the effective protection threshold, or the temperature data exceeds the normal operating range, or the vibration information exceeds the safe vibration threshold, then an early warning is required.

[0038] As a preferred technical solution for the intelligent risk assessment method of gas pipeline networks based on multi-source data fusion, in step S5, determining whether an early warning is needed for the corresponding actual assessment section based on the assessment results of historical maintenance records and unused geographic information data includes:

[0039] The comprehensive risk value is calculated using a weighted model based on the maintenance frequency in the historical maintenance records, the soil pH, and the traffic load.

[0040] If the overall risk value exceeds the preset risk threshold, an early warning is required.

[0041] Compared with existing technologies, the advantages of this invention lie in its ability to achieve a breakthrough shift from passive response to proactive early warning by constructing a multi-source data fusion-based intelligent risk assessment system for gas pipeline networks. Based on a geographic information system, a multi-level zoning assessment model is established, dividing the pipeline network into three assessment segments: administrative, geological, and road network levels, providing a precise spatial basis for risk assessment. Furthermore, by introducing a risk dispersion factor mechanism, direct risk data such as pressure, flow rate, and leakage concentration are relativized, effectively overcoming the interference caused by pressure level differences and load fluctuations in traditional absolute threshold judgments. On this basis, combined with geological influence weighting and objectively weighted comprehensive risk value calculation, deep coupling between operational data and geographic environmental information is achieved. In addition, a dynamic assessment path selection mechanism based on risk tendency categories is established, enabling real-time precise diagnosis of high-risk tendency segments and long-term trend prediction of low-risk tendency segments, thereby optimizing the allocation of inspection resources. This invention improves the comprehensiveness of pipeline network risk identification, the accuracy of early warning signals, and the targeted nature of preventative maintenance. Attached Figure Description

[0042] Figure 1 This is a flowchart of the intelligent risk assessment method for gas pipeline networks based on multi-source data fusion, as described in an embodiment of the present invention.

[0043] Figure 2 This is a flowchart illustrating the parsing of direct risk type data according to an embodiment of the present invention;

[0044] Figure 3 A flowchart for determining the risk tendency category of the corresponding actual assessment section in an embodiment of the present invention. Detailed Implementation

[0045] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0046] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0047] It should be noted that in the description of this invention, the terms "upper", "lower", "left", "right", "inner", "outer", etc., which indicate directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings. This is only for the convenience of description and is not intended to indicate or imply that the device or element must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation of this invention.

[0048] Furthermore, it should be noted that, in the description of this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0049] Please see Figure 1 The diagram shows a flowchart of a multi-source data fusion-based intelligent risk assessment method for gas pipeline networks according to an embodiment of the present invention. The present invention provides a multi-source data fusion-based intelligent risk assessment method for gas pipeline networks, comprising:

[0050] Step S1: Determine the overall structure of the gas pipeline network, and divide the gas pipeline network into several actual evaluation sections based on the overall structure.

[0051] Step S2: Information is collected for each actual assessment section, the geographic information data of each actual assessment section is detected, and all historical maintenance records of the gas pipeline network are obtained. The collected pipeline network detection data includes direct risk type data and indirect risk type data.

[0052] Step S3: Analyze and parse the direct risk type data of each actual assessment segment, combine it with the geographic information data of the corresponding actual assessment segment to determine the risk tendency discrete parameter of the actual assessment segment, and determine the risk tendency category of the corresponding actual assessment segment based on the risk tendency discrete parameter.

[0053] Step S4: Determine the parameter selection method for subsequent assessments based on the risk tendency category of the actual assessment segment, including:

[0054] Obtain the latest data on the aforementioned indirect risk types and conduct inspections of the gas pipeline network in the current actual assessment section;

[0055] Alternatively, by combining the historical maintenance records and unused geographic information data, the comprehensive risk value of the current actual assessment section can be calculated;

[0056] Step S5: Determine whether an early warning is needed for the corresponding actual assessment section based on the evaluation results of the parameter selection methods, and send early warning information to the corresponding actual assessment section.

[0057] Specifically, the direct risk type data includes: pressure data, flow data, and leak detection data;

[0058] The indirect risk type data includes: cathodic protection data, temperature data, and vibration information;

[0059] The geographic information data includes: distribution of geological hazards, soil pH, and traffic load.

[0060] In implementation, direct risk type data is used for real-time pipeline status diagnosis, directly reflecting whether the pipeline network is experiencing anomalies or malfunctions. Pressure data is collected in real time by pressure transmitters installed at key pipeline nodes, such as pressure regulating stations and both sides of valves. A sudden drop in pressure indicates a core characteristic of pipeline leakage; an abnormal surge in pressure may indicate a malfunction or blockage of pressure regulating equipment. Pressure data is one of the most direct and fastest indicators for monitoring pipeline network integrity. Flow data is continuously monitored by installing flow meters at the source, end, and major branch points of the pipeline network, preferably ultrasonic flow meters. By analyzing the flow difference between the pipeline inlet and outlet, or identifying abnormal increases that do not conform to historical gas usage patterns, medium to large-scale leaks can be effectively detected. Leak detection data is obtained by installing fixed combustible gas detectors in locations prone to gas accumulation, such as valve wells, underground spaces, and integrated pipe corridors. Leak detection data provides direct evidence of gas leaks and is crucial for early leak warning and preventing gas accumulation to the explosive limit.

[0061] Indirect risk data is used for predictive assessment, reflecting the health status of the pipeline network and external chronic threats, and providing early warnings of long-term risks. Cathodic protection data is obtained through buried reference electrodes and test piles, periodically or in real-time measuring the pipeline's potential to ground. Cathodic protection data is used to determine the effectiveness of the cathodic protection system applied to suppress pipeline electrochemical corrosion. A substandard potential indicates that the pipeline is at risk of corrosion, which may lead to thinning of the pipe wall or even perforation in the long term. Temperature data is collected by temperature sensors installed on the pipeline or in the environment to compensate for errors in pressure / flow data caused by temperature changes, improving monitoring accuracy. Abnormal temperature changes may also indicate third-party thermal interference or equipment malfunction. Vibration information is monitored by installing vibration acceleration sensors on pipelines near areas with active third-party construction or busy traffic sections. This is used to identify abnormal activities that may cause physical damage to the pipeline, such as illegal mechanical construction or heavy vehicle traffic, and is a cutting-edge means of achieving early warning of third-party damage.

[0062] Geographic information data is used for risk assessment and consequence analysis, identifying high-risk sections. Geological hazard distribution maps (such as landslides, subsidence, and earthquake fault zones) are used to identify high-risk areas for pipeline failure due to natural disasters, thereby assigning higher basic risk levels to these sections and guiding differentiated inspections. Soil pH values ​​are obtained through soil sampling and laboratory analysis along the pipeline route, or by utilizing soil chemistry databases in Geographic Information Systems (GIS), and are used to assess soil corrosivity. Low pH soil significantly accelerates the corrosion rate of pipeline metals. Traffic load is spatially overlaid with traffic flow data (especially heavy vehicle data) from traffic management departments and pipeline routing maps. Traffic load is used to assess the risk of pipeline material fatigue due to long-term traffic vibration and to identify sections with a high probability of direct damage due to accidents such as vehicle collisions.

[0063] Specifically, in step S1, the gas pipeline network is divided into several actual evaluation sections based on the overall structure, including:

[0064] Construct a topology model of the gas pipeline network, which includes all pipe segments as well as valves and pressure regulating stations connecting each pipe segment;

[0065] Based on the geographic information system, gas pipeline networks within the same administrative division boundary are divided into initial assessment sections; gas pipeline networks within the same geological unit boundary are divided into second assessment sections; and gas pipeline networks within the same road network unit are divided into actual assessment sections.

[0066] The initial evaluation segment includes several second evaluation segments, and the second evaluation segment includes several actual evaluation segments.

[0067] In implementation, administrative division boundaries are based on publicly released district and county-level administrative map data. For example, the boundary between District A and District B of a city is used as the initial segment dividing line. This boundary is directly derived from official GIS data and does not require additional calibration. Geological unit boundaries are based on fault zone distribution maps in geological exploration reports. For example, pipe segments within the influence range of the same fault zone are designated as the second assessment segment. The fault zone boundary distance threshold is determined through correlation analysis between historical earthquake data and pipeline damage records. For example, a 500-meter range on each side of the fault is used as the geological unit boundary. Road network units are based on road grade maps published by urban planning departments. Pipe segments within the coverage area of ​​roads of the same grade, such as urban arterial roads, are designated as actual assessment segments. The road boundary threshold is determined through superimposed analysis of road red line width and pipeline burial depth data. For example, a 300-meter range on each side of an arterial road forms a road network unit.

[0068] Understandably, the initial assessment section is based on administrative divisions to ensure compatibility between the assessment unit and the administrative management system, facilitating cross-departmental coordination. The second section is based on geological units to focus on common geological risks, such as pipe sections within the same fault zone potentially facing similar seismic wave impact risks. The final section is based on road network levels, which refines the assessment to the engineering implementation level. Pipe sections within the same road unit typically have the same burial standards, maintenance frequencies, and traffic load characteristics. This avoids risk assessment distortion due to overly large sections or analytical redundancy due to overly small sections in risk prediction, ultimately improving the targeting of preventative maintenance strategies and the efficiency of resource allocation.

[0069] Please see Figure 2 The diagram shows a flowchart of the process for parsing direct risk type data according to an embodiment of the present invention. In step S3, parsing the direct risk type data includes:

[0070] Calculate the average pressure and average pressure deviation of the actual assessment section, and compare the average pressure deviation with the average pressure to obtain the pressure risk dispersion factor;

[0071] Calculate the average flow rate and average flow rate deviation of the actual assessment section, and compare the average flow rate deviation with the average flow rate to obtain the flow rate risk dispersion factor;

[0072] Obtain the integral value of the leakage concentration in the actual assessment section, compare the integral value of the leakage concentration with the leakage concentration threshold and round down to obtain the leakage risk dispersion factor.

[0073] Understandably, the actual operating pressure of the heating network, whether static or dynamic, fluctuates within a small range around a set pressure value, determined by design calculations. There will be no drastic, large fluctuations or continuous drops / rises.

[0074] For pressure data analysis, the pressure risk dispersion factor is obtained by calculating the ratio of the average pressure deviation to the average pressure. This allows for the sensitive detection of abnormal pressure fluctuations in the system, unaffected by the absolute operating pressure of the pipeline. For example, a medium-pressure pipeline section with a lower design pressure and a sub-high-pressure pipeline section with a higher design pressure will experience drastically different absolute pressure drops even if they experience the same proportion of leakage. However, by comparing relative values, they can exhibit similar risk factors. This enables the system to uniformly identify abnormal operating conditions of the same severity in pipeline networks of different pressure levels. This allows the monitoring strategy to adapt to the characteristics of different pressure level sections within the pipeline network, thereby achieving effective monitoring of the operating status of various pipeline sections.

[0075] For flow data analysis, the same relative value principle as for pressure is used to calculate the flow risk dispersion factor, which can effectively avoid fluctuations in normal gas consumption peaks and troughs caused by seasons, holidays, or even different times of day. The system focuses on the relative deviation between real-time flow and the average flow of the pipeline segment over similar historical periods, such as rest days and workdays, rather than the absolute value of the flow. This allows it to accurately identify abnormal increments within the overall gas consumption pattern. Such abnormal increments may be strong correlation signals of pipeline leaks, thus avoiding unnecessary warnings triggered by normal and predictable load changes, and greatly improving the signal-to-noise ratio of leak detection.

[0076] For leak detection data, a leak risk discrete factor is generated by rounding down the ratio of the leak concentration integral value to a set threshold. This allows for a quantitative assessment of the persistence and severity of leak events and provides an integer-discrete output of the risk level. Simple instantaneous concentration sampling may be affected by environmental factors such as gusts of wind, while the integral value reflects the cumulative trend of hazardous gases, providing a more reliable assessment of whether a continuous leak exists. By comparing with a threshold and rounding, continuous concentration signals can be transformed into clear, graded risk alarm levels. For example, a ratio of 0.3 rounded down to level 0 has no impact. This facilitates subsequent risk classification and provides a clear triggering basis for the emergency response procedures corresponding to different warning levels.

[0077] In this invention, a risk dispersion factor mechanism is introduced to relativize and normalize three key data points in the operation of heating pipeline networks: pressure, flow rate, and leakage concentration. This effectively overcomes the problems of pressure level differences, load fluctuation interference, and instantaneous environmental disturbances caused by traditional absolute threshold judgment. By constructing pressure risk dispersion factors and flow rate risk dispersion factors, abnormal fluctuation trends can be sensitively captured, unaffected by changes in the design pressure of pipeline sections or normal gas consumption rhythms. This enables consistent monitoring of pipeline networks with different pressure levels and operating states at different times, significantly improving the consistency and accuracy of abnormal condition identification. Simultaneously, by rounding down the ratio of the leakage concentration integral value to the threshold, a leakage risk dispersion factor is constructed, effectively suppressing instantaneous environmental interference, quantifying the persistence and severity of leakage, and achieving integer and graded output of risk levels. This provides a clear and reliable triggering basis for subsequent classification and precise emergency response.

[0078] Please see Figure 3 The flowchart shown is a process for determining the risk tendency category of a corresponding actual assessment section according to an embodiment of the present invention. In step S3, the risk tendency discrete parameter of the actual assessment section is determined based on the parsing results of the direct risk type data of the actual assessment section and the geographic information data of the corresponding actual assessment section, including:

[0079] The geological impact weights are determined based on the distribution of the geological hazards.

[0080] Calculate the product of the geological influence weight and the sum of the pressure risk discrete factor and the flow risk discrete factor, and add the leakage risk discrete factor. The result is denoted as the risk propensity discrete parameter.

[0081] The geological influence weight is set to 1, 2, or 3.

[0082] In practice, the classification of geological impact levels is mainly based on two categories of data: historical geological disaster records and regional geological structural characteristics. Specifically, the first step is to obtain the "Geological Disaster Distribution Map" and its basic survey data from national or local geological survey institutions. These data typically include information on the location, scale, and activity level of historical landslides, collapses, debris flows, ground subsidence, ground fissures, and active fault zones.

[0083] Based on this data, the following three-level classification system is established: High geological hazard distribution level: This refers to the presence of active geological hazard points in the actual assessment section, or its location within a major geological hazard-prone area. For example, the area has experienced multiple small landslides in the past decade, or is situated on a large, active landslide, or crosses an active fault zone with the potential for moderate-intensity earthquakes; Medium geological hazard distribution level: This refers to the presence of potential or intermittently active geological hazard threats in the actual assessment section. For example, the area has a complex geological structure, contains ancient landslides but is currently basically stable, or is located within a ground subsidence monitoring area with a moderate annual subsidence rate; Low geological hazard distribution level: This refers to the relatively stable geological conditions in the actual assessment section, where no obvious geological hazard points or major hidden dangers have been found in the available historical records and existing geological models.

[0084] Understandably, different levels of geological hazard risks pose significant differences in their potential destructive power and probability of occurrence for pipelines. This classification allows the system to transform macroscopic, spatially distributed geological risks into comparable discrete parameters (e.g., quantifiable as high=3, medium=2, low=1) tied to a specific actual assessment section. Incorporating this weight into the weighted calculation of the risk propensity discrete parameter ensures that pipeline sections traversing high-risk areas such as landslide zones along major rivers and earthquake fault zones receive higher basic risk scores due to their challenging geological environments, thus gaining greater priority and stricter monitoring in subsequent assessments and early warning systems.

[0085] Therefore, in this embodiment, the geological impact weight for geological disasters with a high geological disaster distribution level is 3, the geological impact weight for geological disasters with a medium geological disaster distribution level is 2, and the geological impact weight for geological disasters with a low geological disaster distribution level is 1.

[0086] In this invention, by introducing geological impact weights based on historical geological disaster records and regional geological structural characteristics, high, medium, and low-level geological disaster hazard zones are discretized and assigned values. Combined with various risk discretization factors, a comprehensive risk tendency discrete parameter is finally generated. This transforms macroscopic spatial geological risk into comparable and calculable discrete weights, enabling pipe sections traversing high-risk geological areas such as landslide zones, fault zones, and subsidence zones to obtain higher basic risk scores due to their severe geological background, even when operational data fluctuations are not yet significant. This results in priority attention and resource allocation in early warning ranking, inspection frequency, and emergency response, significantly improving the spatial sensitivity and forward-looking early warning capabilities of risk assessment. It effectively avoids the hidden danger of underestimating geologically high-risk but stable data sections, achieving collaborative perception, hierarchical control, and precise governance of geological and operational risks.

[0087] Specifically, determining the risk tendency category for the corresponding actual assessment segment based on the aforementioned discrete risk tendency parameter includes:

[0088] If the discrete parameter of risk propensity is greater than the standard risk propensity parameter, then the risk propensity category of the current actual assessment segment is determined to be the high-risk propensity category;

[0089] If the discrete parameter of risk propensity is less than or equal to the standard risk propensity parameter, then the risk propensity category of the current actual assessment segment is determined to be the low risk propensity category.

[0090] The standard risk propensity parameter has a value less than 1.

[0091] In implementation, the standard risk propensity parameter is set to less than 1. This is based on statistical analysis of the distribution characteristics of the discrete risk propensity parameter of the pipeline system under long-term normal operation. Under ideal safe operating conditions, the calculation result composed of pressure, flow rate, and leakage risk discrete factors should approach zero. A leakage risk discrete factor of 1 indicates that the cumulative amount of hazardous gas exceeds the standard, suggesting a problem in a portion of the pipeline network. Therefore, a threshold less than 1 can effectively distinguish between normal fluctuations in background noise levels and abnormal risk signals that require attention. Specifically, the standard risk propensity parameter is calculated by examining a large number of actual assessment sections operating safely in historical data to obtain the distribution range of their discrete risk propensity parameter. Ideally, this distribution range should be 95% of its upper limit. Referring to the method for determining confidence intervals, this threshold can maximally distinguish the critical value between normal and abnormal states.

[0092] In this invention, the standard risk propensity parameter ensures that the risk assessment system has good specificity, that is, it can correctly classify the vast majority of normal operating states into the low-risk propensity category, thereby avoiding resource waste and warning fatigue caused by excessive warnings. It also makes the entire classification decision-making process objective and repeatable, effectively ensuring the consistency and comparability of risk assessment results between different periods and different sections, and providing a stable and reliable logical basis for the subsequent activation of different levels of inspection or prediction plans.

[0093] In this invention, a standard risk propensity parameter is set as a rigid threshold to divide massive sections into two categories: high-risk and low-risk. This not only suppresses background noise fluctuations into the low-risk zone but also ensures that any leakage anomalies or geological-operating condition superposition anomalies are detected sensitively. The classification results directly drive differentiated subsequent strategies. For high-risk sections, the latest indirect risk data is immediately retrieved to initiate on-site inspections, enabling early problem detection and precise resource allocation. For low-risk sections, historical maintenance records and remaining geographic information are automatically retrieved for offline reassessment of comprehensive risk values, ensuring the continuous accumulation of data throughout the entire lifecycle and the self-iteration of the risk model.

[0094] Specifically, in step S4, determining the parameter selection method for subsequent assessments based on the risk tendency category of the actual assessment segment includes:

[0095] If the actual assessment section is classified as a high-risk tendency category, the parameter selection method is to obtain the latest indirect risk type data and conduct an inspection of the gas pipeline network in the current actual assessment section.

[0096] If the actual assessment section is classified as low-high risk, the parameter selection method is to combine the historical maintenance records and unused geographic information data to calculate the comprehensive risk value of the current actual assessment section.

[0097] Specifically, in step S5, determining whether a warning is needed for the corresponding actual assessment section based on the assessment results of indirect risk type data includes:

[0098] Obtain the latest cathodic protection data, temperature data, and vibration information for the current actual evaluation section;

[0099] If the cathodic protection data is lower than the effective protection threshold, or the temperature data exceeds the normal operating range, or the vibration information exceeds the safe vibration threshold, then an early warning is required.

[0100] In practice, the effective protection threshold for cathodic protection is directly adopted from the minimum protection potential value specified in the national mandatory standard "Technical Specification for Cathodic Protection of Buried Steel Pipelines", such as -0.85V.

[0101] The normal operating temperature range is determined based on the design operating temperature of the pipe material and historical operating data. The upper limit of the range is taken from the highest withstand temperature of the material provided by the pipe manufacturer, while the lower limit is determined based on the local historical extreme minimum temperature and the antifreeze requirements of the medium. For example, the operating temperature range of PE pipes may be set to -20℃ to 40℃. Outside this range, it is considered an abnormal situation that may cause a decline in material performance or equipment failure.

[0102] The safe vibration threshold is calibrated by collecting background vibration data of the pipeline under known safe conditions and combining it with a limited number of simulated impact tests. Specifically, in sections with frequent third-party construction activities, the environmental vibration amplitude during periods of no construction is monitored and recorded as a background value; simultaneously, the vibration amplitude transmitted to the pipeline by typical construction machinery (such as a rammer) operating at a safe distance is recorded. The safe vibration threshold is then calculated as 95% of the maximum value of the latter's vibration acceleration or frequency characteristic value, which can distinguish between normal traffic vibration and potentially destructive construction vibration.

[0103] Specifically, in step S5, determining whether a warning is needed for the corresponding actual assessment section based on the evaluation results of historical maintenance records and unused geographic information data includes:

[0104] The comprehensive risk value is calculated using a weighted model based on the maintenance frequency in the historical maintenance records, the soil pH, and the traffic load.

[0105] If the overall risk value exceeds the preset risk threshold, an early warning is required.

[0106] In implementation, a data-driven objective weighting method was used to construct a weighted model. Weights were determined by analyzing the statistical correlation between soil pH and traffic load and historical pipeline maintenance records. Objective weighting was based on the degree of variability and conflict among indicators, specifically using the entropy weighting method. A certain amount of historical sample data was collected, with each sample containing the maintenance frequency, soil pH, and traffic load values ​​for a specific pipeline section, along with their corresponding historical states. Data standardization was performed, transforming the raw data of each indicator to a 0-1 range using range normalization. Based on this, the entropy value of each indicator was calculated. This entropy value reflects the dispersion of the indicator data; the greater the dispersion, the smaller the entropy value, indicating that the indicator carries more information in distinguishing risk levels, and thus receives a higher weight. Finally, the weight coefficients of each indicator were calculated based on their entropy values, forming a fixed weight set for subsequent risk calculations of all sections.

[0107] After obtaining the objective weights, the calculation of the comprehensive risk value follows a linear weighted model. For any actual section to be assessed, its standardized maintenance frequency value, soil pH, and traffic load index are multiplied by the corresponding weights calculated by the entropy weight method, and then summed to obtain the comprehensive risk value of that section.

[0108] Understandably, the weight allocation is entirely determined by the inherent patterns of historical data, enabling the model to accurately reflect the impact of each risk factor on pipeline safety in actual operation. This ensures the objectivity and reproducibility of the risk assessment results and provides stable and reliable data support for automated early warning decision-making.

[0109] For example, data were collected from five representative sample sections of the pipeline system in the previous assessment year. Sample section A had 2 maintenance visits, a soil pH of 5.5, and a heavy vehicle traffic volume of 300 vehicles per day. Sample section B had 1 maintenance visit, 6.8 vehicles per day, and 150 vehicles per day. Sample section C had 4 maintenance visits, 4.5 vehicles per day, and 500 vehicles per day. Sample section D had 1 maintenance visit, 7.0 vehicles per day, and 100 vehicles per day. Sample section E had 3 maintenance visits, 5.0 vehicles per day, and 400 vehicles per day.

[0110] Data normalization is performed. For positive indicators such as maintenance frequency and heavy vehicle traffic volume, where higher values ​​indicate higher risk, the calculation formula is: subtract the minimum value from the value of a certain segment, then divide by the difference between the maximum and minimum values. For negative indicators such as soil pH value, where lower values ​​indicate higher risk, the formula is: subtract the value of a certain segment from the maximum value, then divide by the difference between the maximum and minimum values.

[0111] The calculations yielded the following results: For sample segment A, the normalized values ​​for maintenance frequency were 0.333, soil pH were 0.600, and heavy vehicle traffic volume was 0.500. For sample segment B, the normalized values ​​for the three indicators were 0.000, 0.080, and 0.125, respectively. For sample segment C, all normalized values ​​were 1.000. For sample segment D, the normalized values ​​were 0.000, 0.000, and 0.000, respectively. For sample segment E, the normalized values ​​were 0.667, 0.800, and 0.750, respectively.

[0112] Calculate the entropy value of each indicator, and calculate the proportion of each normalized value under the corresponding indicator. The proportion of maintenance frequency in section A is 0.333 divided by the sum of the normalized values ​​of the indicator in all sections. Then, according to the information entropy formula, calculate the information entropy of each indicator. The information entropy of the maintenance frequency indicator is 0.686, the information entropy of the soil pH value indicator is 0.619, and the information entropy of the heavy vehicle flow indicator is 0.660.

[0113] Then, the weights are calculated based on the entropy values. The weight of the maintenance frequency indicator is 1 minus its entropy value of 0.686, divided by the sum of the redundancies of all indicators, resulting in a final weight of 0.350. The weight of the soil pH value indicator is calculated to be 0.429, and the weight of the heavy vehicle flow indicator is 0.221.

[0114] Through the aforementioned purely data-driven objective process, the final weights of the three risk factors were determined: maintenance frequency (0.350), soil corrosivity (0.429), and traffic load (0.221). This set of weights will be written into the model as fixed parameters for calculating the comprehensive risk value of all actual assessment sections.

[0115] The preset risk threshold is determined by identifying all unmaintained nodes in all historical sample segments and calculating the comprehensive risk value of each actual assessment area, and then taking the 75% to 95% percentile of the high quantile of the risk value as the preset risk threshold.

[0116] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

[0117] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for intelligent risk assessment of gas pipeline networks using multi-source data fusion, characterized in that, include: Step S1: Determine the overall structure of the gas pipeline network, and divide the gas pipeline network into several actual evaluation sections based on the overall structure. Step S2: Information is collected for each actual assessment section, the geographic information data of each actual assessment section is detected, and all historical maintenance records of the gas pipeline network are obtained. The collected pipeline network detection data includes direct risk type data and indirect risk type data. The direct risk type data includes: pressure data, flow data, and leak detection data; The indirect risk type data includes: cathodic protection data, temperature data, and vibration information; The geographic information data includes: geological hazard distribution, soil pH, and traffic load; Step S3: Analyze and parse the direct risk type data of each actual assessment segment, combine it with the geographic information data of the corresponding actual assessment segment to determine the risk tendency discrete parameter of the actual assessment segment, and determine the risk tendency category of the corresponding actual assessment segment based on the risk tendency discrete parameter. In step S3, parsing the direct risk type data includes: Calculate the average pressure and average pressure deviation of the actual assessment section, and compare the average pressure deviation with the average pressure to obtain the pressure risk dispersion factor; Calculate the average flow rate and average flow rate deviation of the actual assessment section, and compare the average flow rate deviation with the average flow rate to obtain the flow rate risk dispersion factor; Obtain the integral value of the leakage concentration in the actual assessment section, compare the integral value of the leakage concentration with the leakage concentration threshold, and round down to obtain the leakage risk dispersion factor; In step S3, based on the parsing results of the direct risk type data of the actual assessment segment and combined with the geographic information data of the corresponding actual assessment segment, the risk tendency discrete parameter of the actual assessment segment is determined, including: The geological impact weights are determined based on the distribution of the geological hazards. Calculate the product of the geological influence weight and the sum of the pressure risk discrete factor and the flow risk discrete factor, and add the leakage risk discrete factor. The result is denoted as the risk propensity discrete parameter. The geological influence weight is set to 1, 2, or 3. Step S4: Determine the parameter selection method for subsequent assessments based on the risk tendency category of the actual assessment segment, including: Obtain the latest data on the aforementioned indirect risk types and conduct inspections of the gas pipeline network in the current actual assessment section; Alternatively, by combining the historical maintenance records and unused geographic information data, the comprehensive risk value of the current actual assessment section can be calculated; Step S5: Determine whether an early warning is needed for the corresponding actual assessment section based on the evaluation results of the parameter selection methods, and send early warning information to the corresponding actual assessment section.

2. The intelligent risk assessment method for gas pipeline networks based on multi-source data fusion according to claim 1, characterized in that, In step S1, the gas pipeline network is divided into several actual assessment sections based on the overall structure, including: Construct a topology model of the gas pipeline network, which includes all pipe segments as well as valves and pressure regulating stations connecting each pipe segment; Based on the geographic information system, gas pipeline networks within the same administrative division boundary are divided into initial assessment sections; gas pipeline networks within the same geological unit boundary are divided into second assessment sections; and gas pipeline networks within the same road network unit are divided into actual assessment sections. The initial evaluation segment includes several second evaluation segments, and the second evaluation segment includes several actual evaluation segments.

3. The intelligent risk assessment method for gas pipeline networks based on multi-source data fusion according to claim 2, characterized in that, Based on the aforementioned discrete parameters of risk propensity, the risk propensity category corresponding to the actual assessment segment is determined, including: If the discrete parameter of risk propensity is greater than the standard risk propensity parameter, then the risk propensity category of the current actual assessment segment is determined to be the high-risk propensity category; If the discrete parameter of risk propensity is less than or equal to the standard risk propensity parameter, then the risk propensity category of the current actual assessment segment is determined to be the low risk propensity category. The standard risk propensity parameter has a value less than 1.

4. The intelligent risk assessment method for gas pipeline networks based on multi-source data fusion according to claim 3, characterized in that, In step S4, the parameter selection method for subsequent assessments is determined based on the risk propensity category of the actual assessment segment, including: If the actual assessment section is classified as a high-risk tendency category, the parameter selection method is to obtain the latest indirect risk type data and conduct an inspection of the gas pipeline network in the current actual assessment section. If the actual assessment section is classified as low-high risk, the parameter selection method is to combine the historical maintenance records and unused geographic information data to calculate the comprehensive risk value of the current actual assessment section.

5. The intelligent risk assessment method for gas pipeline networks based on multi-source data fusion according to claim 4, characterized in that, In step S5, determining whether an early warning is needed for the corresponding actual assessment section based on the assessment results of indirect risk type data includes: Obtain the latest cathodic protection data, temperature data, and vibration information for the current actual evaluation section; If the cathodic protection data is lower than the effective protection threshold, or the temperature data exceeds the normal operating range, or the vibration information exceeds the safe vibration threshold, then an early warning is required.

6. The intelligent risk assessment method for gas pipeline networks based on multi-source data fusion according to claim 5, characterized in that, In step S5, based on the evaluation results of historical maintenance records and unused geographic information data, it is determined whether an early warning is needed for the corresponding actual evaluation section, including: The comprehensive risk value is calculated using a weighted model based on the maintenance frequency in the historical maintenance records, the soil pH, and the traffic load. If the overall risk value exceeds the preset risk threshold, an early warning is required.

Images