Intelligent Detection Method and System for Underground Pipeline Network Data Based on Multi-Source Data Fusion

Abstract

This invention discloses an intelligent detection method and system for underground pipeline network data based on multi-source data fusion, specifically relating to the field of underground pipeline network data processing technology. It addresses the problems of low accuracy and poor reliability in existing technologies where pipeline connection relationship determination relies on manual experience. The method involves constructing an original dataset containing spatial location information and pipeline attribute information, then sequentially performing connection information integrity analysis, connection relationship analysis based on engineering and spatial rules, graph signal processing evaluation combining dynamic response correlation and network power flow distribution compatibility, network structure robustness and cascading failure risk analysis, multi-agent game simulation, and constraint satisfaction analysis. Finally, it automatically generates an accurate and reliable pipeline network topology diagram.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of underground pipe network data processing, and particularly relates to a multi-source data fusion underground pipe network data intelligent detection method and system. BACKGROUND

[0002] In the field of underground pipe network surveying and mapping, in order to obtain comprehensive and accurate pipeline information, various technical means are generally used for collaborative work. The existing technology usually determines the spatial position of the pipeline point by using global navigation satellite system and total station instrument, and uses pipeline detection instrument to explore the buried depth and attributes of the pipeline. On this basis, the operation personnel need to integrate the above multi-source data, and manually determine the connection relationship between each pipeline point according to the point information, exploration records and field experience in the indoor stage, and finally draw a comprehensive underground pipeline map.

[0003] However, when the existing technical method faces complex working conditions such as signal interference, manhole cover burying or dense pipeline, the obtained pipeline connection information often presents fragmentation and uncertainty, which leads to that in the indoor mapping process, the determination of the pipeline network topological relationship highly depends on the personal experience and subjective inference of the operation personnel, and it is difficult to systematically guarantee the logical correctness, thereby introducing potential risks for subsequent engineering design and construction. SUMMARY

[0004] The present application provides a multi-source data fusion underground pipe network data intelligent detection method and system to solve the technical problems in the prior art.

[0005] The technical solution of the present application to solve the above technical problems is as follows:

[0006] The multi-source data fusion underground pipe network data intelligent detection method comprises:

[0007] S1, acquiring pipeline point spatial position information and pipeline attribute information collected by multi-source detection equipment, and constructing an original data set;

[0008] S2, performing pipeline point connection information integrity analysis based on the original data set, and generating a connection information integrity judgment result;

[0009] S3, when the judgment result is incomplete, performing connection relationship analysis on the pipeline points with missing connection information based on engineering rules and spatial rules to generate a preliminary connection relationship analysis result;

[0010] S4, based on the preliminary connection relationship analysis result, evaluating the pipeline connection possibility by analyzing the dynamic response correlation between pipelines and combining with network flow distribution compatibility, and generating a pipeline connection possibility evaluation result by using a graph signal processing algorithm;

[0011] S5. Based on the pipeline connection possibility evaluation result, network structure robustness and cascade failure risk analysis is performed to obtain a robust path subset, and based on the robust path subset, an optimal connection path set is generated through multi-agent game simulation of the group consensus formation process;

[0012] S6. Constraint satisfaction analysis is performed according to the optimal connection path set to generate a pipeline network topology relationship diagram.

[0013] Further, the pipeline point spatial position information and pipeline attribute information collected by the multi-source detection equipment are obtained to construct an original data set, including:

[0014] The longitude and latitude coordinates included in the pipeline point spatial position information are collected by a global navigation satellite system receiving device;

[0015] The elevation coordinates included in the pipeline point spatial position information are obtained by total station measurement;

[0016] The pipeline material and pipeline burial depth included in the pipeline attribute information are obtained by pipeline detection instrument exploration;

[0017] The longitude and latitude coordinates, elevation coordinates, pipeline material, and pipeline burial depth are integrated to construct the original data set.

[0018] Further, based on the original data set, pipeline point connection information integrity analysis is performed to generate a connection information integrity judgment result, including:

[0019] Based on the pipeline point spatial position information included in the original data set, an initial connection graph representing the adjacency relationship between pipeline points is constructed;

[0020] According to the connection degree distribution characteristics of each pipeline point in the initial connection graph, pipeline points with missing connection information are identified;

[0021] By analyzing the spatial distance and pipeline attribute information matching relationship between the pipeline points with missing connection information and their adjacent pipeline points, the connection information missing state is confirmed;

[0022] It is determined whether there is at least one pipeline point confirmed to be in a connection information missing state;

[0023] If so, a connection information incomplete judgment result is generated; otherwise, a connection information complete judgment result is generated.

[0024] Further, when the judgment result is incomplete, the pipeline points with missing connection information are analyzed based on engineering rules and spatial rules to generate a preliminary connection relationship analysis result, including:

[0025] Based on the spatial position information between the pipeline points with missing connection information and their adjacent pipeline points, candidate connection relationships are constructed according to the pipeline direction continuity constraint;

[0026] Based on the pipeline attribute information contained in the original dataset, candidate connection relationships are filtered according to the priority connection rule for pipelines of the same type.

[0027] By analyzing the pipeline burial depth data between adjacent pipeline points, the rationality of the candidate connection relationship is verified based on the principle of burial depth consistency.

[0028] Preliminary connection analysis results are generated by combining the screened and validated candidate connections.

[0029] Furthermore, based on the preliminary connectivity analysis results, by analyzing the dynamic response correlation between pipelines and combining it with the network power flow distribution compatibility, a graph signal processing algorithm is used to evaluate the pipeline connectivity probability, generating pipeline connectivity probability evaluation results, including:

[0030] Based on the candidate connectivity relationships from the preliminary connectivity analysis results, construct a graph structure to be evaluated that contains potential connectivity paths;

[0031] Define a graph signal on the graph structure to be evaluated, using pipeline pressure or flow monitoring data as the graph vertex signal;

[0032] Calculate the dynamic response correlation coefficient between the vertex signals of each candidate connection relationship in the graph structure to be evaluated;

[0033] Simulate network power flow distribution based on pipeline attribute information in the original dataset and calculate the compatibility index between power flow distribution and pipeline attribute information;

[0034] The dynamic response correlation coefficient and the network power flow distribution compatibility index are input into the graph signal processing algorithm for fusion calculation to obtain the pipeline connection probability score;

[0035] The pipeline connection probability assessment results are generated based on the pipeline connection probability score.

[0036] Furthermore, the pipeline connection probability score is obtained by fusing the dynamic response correlation coefficient and the network power flow distribution compatibility index into a graph signal processing algorithm. This process includes: mapping the dynamic response correlation coefficient and the network power flow distribution compatibility index to the graph spectrum domain using the graph Fourier transform in the graph signal processing algorithm based on the graph vertex signals of each candidate connection relationship in the graph structure to be evaluated; performing weighted fusion calculation on the dynamic response correlation coefficient and the network power flow distribution compatibility index after normalization in the graph spectrum domain; and converting the weighted fusion result back to the vertex domain using the inverse graph Fourier transform to obtain the pipeline connection probability score.

[0037] Furthermore, based on the pipeline connection probability assessment results, a robust path subset is obtained through network structure robustness and cascading failure risk analysis. Based on this robust path subset, an optimal connection path set is generated through multi-agent game simulation of the group consensus formation process, including:

[0038] An initial path set containing each candidate connection path is constructed based on the pipeline connection probability assessment results;

[0039] The robustness of network structure is evaluated by calculating the degree of network connectivity preservation after removing non-critical connection paths.

[0040] The risk of cascading failure is assessed by simulating the cascading effect caused by the failure of a single connection path.

[0041] A robust path subset was obtained based on the network structure robustness assessment results and the cascade failure risk assessment results;

[0042] Establish a multi-agent negotiation model on a robust path subset, with path attributes as the interest demand;

[0043] Through multiple rounds of negotiation and iteration, the interests of each intelligent agent are brought into agreement, forming a group consensus;

[0044] The optimal set of connection paths is generated based on the connection path selection that achieves group consensus.

[0045] Furthermore, establishing a multi-agent negotiation model based on path attributes as the interest demands on a robust path subset includes: defining each candidate connection path in the robust path subset as an independent agent; extracting path attributes, including pipeline material and pipeline burial depth, from the pipeline attribute information in the original dataset as the interest demands of each agent; and establishing multi-agent negotiation rules with path attribute similarity and network structure robustness as negotiation objectives.

[0046] Furthermore, constraint satisfaction is analyzed based on the optimal connection path set to generate a pipeline network topology graph, including:

[0047] A constraint satisfaction problem model is established based on the optimal set of connection paths, which includes spatial geometric constraints and engineering logical constraints.

[0048] The coordinate data from the spatial location information of pipeline points and the material type and burial depth data from the pipeline attribute information are used as constraint variables to input the constraint satisfaction problem model;

[0049] Eliminate contradictory connection paths that violate both spatial geometric constraints and engineering logic constraints in the optimal connection path set by using a constraint propagation algorithm;

[0050] The backtracking search algorithm is used to determine the final pipeline connection relationship in the solution space that satisfies all constraints;

[0051] A pipeline network topology diagram is generated based on the final pipeline connection relationships and the spatial location information of pipeline points.

[0052] On the other hand, the present invention provides an intelligent detection system for underground pipeline network data based on multi-source data fusion, comprising:

[0053] The information acquisition module is used to acquire the spatial location information and pipeline attribute information of pipeline points collected by multi-source detection equipment, and to construct the original dataset;

[0054] The complete judgment module is used to perform integrity analysis of pipeline point connection information based on the original dataset and generate connection information integrity judgment results.

[0055] The preliminary analysis module is used to generate preliminary connection relationship analysis results by performing connection relationship analysis on pipeline points with missing connection information based on engineering rules and spatial rules when the judgment result is incomplete.

[0056] The probability assessment module is used to evaluate the probability of pipeline connections based on the preliminary connection relationship analysis results, by analyzing the dynamic response correlation between pipelines and combining the network power flow distribution compatibility, and by using graph signal processing algorithms to generate pipeline connection probability assessment results.

[0057] The set generation module is used to analyze the robustness of network structure and the risk of cascading failure based on the pipeline connection probability assessment results to obtain a robust path subset. Based on the robust path subset, the optimal connection path set is generated by simulating the group consensus formation process through multi-agent game theory.

[0058] The relationship generation module is used to perform constraint satisfaction analysis based on the optimal connection path set and generate a pipeline network topology diagram.

[0059] The beneficial effects of this invention are:

[0060] 1. By constructing a systematic data processing workflow, intelligent fusion and analysis of multi-source pipeline detection data are realized. First, missing data links are automatically identified through connection information integrity analysis. Then, preliminary connection relationships are constructed based on engineering rules and spatial rules, effectively solving the logical inconsistency problem caused by the reliance on human experience in traditional methods. By introducing graph signal processing algorithms to analyze pipeline dynamic response and network power flow compatibility, the objectivity and accuracy of connection probability assessment are significantly improved. The determination of pipeline connection relationships is transformed into a calculable and verifiable automated process, greatly reducing the error risk introduced by human factors.

[0061] 2. Through multi-agent game simulation and constraint satisfaction analysis, intelligent generation of pipeline network topology was achieved. Based on network structure robustness analysis and cascade failure risk assessment, robust paths were selected. Then, a group consensus was formed through a multi-agent negotiation mechanism. Finally, the optimal connection scheme was generated under the condition of satisfying spatial geometric constraints and engineering logic constraints. This not only ensured the logical consistency of pipeline network topology, but also significantly improved the reliability and engineering applicability of pipeline network data products, providing a solid data foundation for subsequent pipeline network planning, design, operation and maintenance. Attached Figure Description

[0062] Figure 1 This is a flowchart of the intelligent detection method for underground pipe network data based on multi-source data fusion according to the present invention;

[0063] Figure 2 This is a schematic diagram of the structure of the intelligent detection system for underground pipe network data based on multi-source data fusion according to the present invention. Detailed Implementation

[0064] 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.

[0065] Example 1: Figure 1 This invention presents an intelligent detection method for underground pipe network data based on multi-source data fusion, comprising:

[0066] S1. Obtain the spatial location information and pipeline attribute information of pipeline points collected by multi-source detection equipment, and construct the original dataset;

[0067] S2. Perform pipeline point connection information integrity analysis based on the original dataset and generate connection information integrity judgment results;

[0068] S3. When the judgment result is incomplete, perform connection relationship analysis on pipeline points with missing connection information based on engineering rules and spatial rules to generate preliminary connection relationship analysis results.

[0069] S4. Based on the preliminary connection relationship analysis results, by analyzing the dynamic response correlation between pipelines and combining the network power flow distribution compatibility, a graph signal processing algorithm is used to evaluate the pipeline connection probability and generate pipeline connection probability evaluation results.

[0070] S5. Based on the pipeline connection probability assessment results, conduct network structure robustness and cascade failure risk analysis to obtain a robust path subset. Based on the robust path subset, generate the optimal connection path set by simulating the group consensus formation process through multi-agent game.

[0071] S6. Perform constraint satisfaction analysis based on the optimal connection path set to generate a pipeline network topology diagram.

[0072] S1. Obtain the spatial location information and pipeline attribute information of pipeline points collected by multi-source detection equipment, and construct the original dataset. The specific implementation is as follows:

[0073] In the process of acquiring spatial location and attribute information of pipeline points from multi-source detection equipment and constructing the original dataset, the spatial location information of the pipeline points, including latitude and longitude coordinates, is first collected using a Global Navigation Satellite System (GNSS) receiver. Specifically, a high-precision GNSS receiver, a multi-frequency, multi-satellite receiver, is used, capable of simultaneously receiving signals from multiple satellite navigation systems, such as GPS, GLONASS, and BeiDou, to improve signal coverage and positioning stability. The sampling frequency is set to once per second to ensure the real-time and continuous nature of data acquisition. Real-time dynamic differential positioning is employed, which uses correction data provided by a reference station to eliminate satellite orbit errors and atmospheric delay errors, thereby achieving a planar positioning accuracy better than 0.05 meters. When placing the receiver at the actual location of the pipeline point, it is ensured that the center of the receiver antenna is strictly aligned with the ground marker point to avoid installation deviations affecting data accuracy. The receiving equipment continuously receives satellite signals and calculates latitude and longitude coordinate data through its built-in processing unit. The coordinate data is output in decimal degree format, for example, longitude values ​​range from 110.500000 degrees to 120.800000 degrees, and latitude values ​​range from 20.300000 degrees to 30.600000 degrees. A timestamp and point identifier are appended to the data recording. The point identifier uses a unique numerical sequence, for example, starting from 1 and incrementing, to ensure that the latitude and longitude coordinates of each pipeline point are unique and traceable. Data quality is checked through the equipment's built-in software, with indicators including signal-to-noise ratio (SNR), multipath error, and positioning calculation residual. For example, if the SNR is below 30 dB or the residual exceeds 0.1 meters, the data point is automatically discarded and a re-acquisition process is triggered to ensure the reliability and consistency of the coordinate data. Furthermore, environmental factors are considered during field operations, such as selecting open areas to avoid signal obstruction by buildings or trees, and conducting multiple measurements at different time periods to balance the influence of satellite geometric distribution.

[0074] The spatial location information of the pipeline point, including its elevation coordinates, is obtained through total station measurements. An electronic total station is used, equipped with automatic distance and angle measurement functions, achieving a distance measurement accuracy of 1 mm + 1 ppm and an angle measurement accuracy of 1 second. During total station installation, a known control point is selected as the station. The elevation coordinates of the control point have been pre-determined through precise leveling, for example, a control point elevation of 50.000 meters. Both the instrument height and the target height are set to 1.5 meters. The instrument height refers to the vertical distance from the center of the total station telescope to the station point, and the target height refers to the vertical distance from the center of the reflecting prism to the pipeline point. The reflecting prism is used as the target point and placed at the pipeline point location, ensuring the prism is vertically installed without wobbling. The measurement process includes aiming at the prism and reading the horizontal angle, vertical angle, and slope distance data. The horizontal angle is used to determine the azimuth, the vertical angle is used to calculate the elevation difference, and the slope distance is used to derive the horizontal distance and elevation values. Elevation coordinates are calculated using the principles of trigonometric leveling. The specific calculation steps involve first converting the slope distance and vertical angle into elevation differences, then summing these differences by combining the station elevation, instrument height, and target height. Earth curvature and atmospheric refraction corrections are applied; for example, the average Earth radius of 6371 km is used for Earth curvature correction, and a standard refractive index of 0.13 is used for atmospheric refraction correction. Elevation coordinates are expressed in meters, with the local mean sea level as the reference surface. For example, the elevation range is from 0 meters to 100 meters. Multiple independent observations are performed during measurement, such as taking the arithmetic mean of three measurements to reduce random errors. Environmental parameters such as temperature and air pressure are also recorded, for example, with a temperature range of 10 degrees Celsius to 35 degrees Celsius and an air pressure range of 1000 hPa to 1020 hPa, for real-time data correction to ensure the accuracy of the elevation coordinates is within 0.01 meters. After data output, it is stored in association with pipeline point identifiers; for example, the point identifier, elevation value, and measurement timestamp are recorded in the data file.

[0075] Pipeline detectors are used to obtain pipeline attribute information, including pipeline material and burial depth. These detectors employ electromagnetic induction, consisting of a transmitter and receiver. The transmitter generates an alternating electromagnetic field, and the receiver detects the induced signals from underground pipelines. The detection frequency is set to a low 33 kHz for metallic pipelines and a high 100 kHz for non-metallic pipelines. Low-frequency signals have strong penetrating power but lower resolution, while high-frequency signals offer high resolution but limited penetration depth. During detection, the detector moves at a constant speed along the pipeline path, controlled within 0.5 meters per second, receiving the electromagnetic signals generated by the underground pipeline. Pipeline location is identified by signal strength changes and phase characteristics; for example, signal strength peaks correspond to the pipeline centerline, and phase reversal points indicate pipeline boundaries. Pipeline material is distinguished based on signal response characteristics. For example, metallic pipelines typically exhibit high signal amplitude and low phase delay, while non-metallic pipelines exhibit low signal amplitude and high phase delay. For non-metallic pipelines, ground-penetrating radar (GPR) is used for auxiliary identification. GPR infers material type by emitting high-frequency electromagnetic waves and analyzing the reflected signals.

[0076] The pipeline burial depth is calculated using the double-coil method, and the formula is: ;in, Indicates the burial depth of the pipeline, in meters; Indicates the coil spacing in meters, for example, 0.5 meters; This represents the signal strength ratio; it is dimensionless and read from the device's display screen. This represents a dimensionless correction factor for soil electrical conductivity, based on field-tested soil electrical conductivity. Confirmed, soil electrical conductivity The unit is Siemens per meter, ranging from 0.01 Siemens per meter to 0.1 Siemens per meter.

[0077] Burial depth data is recorded in meters, for example, the burial depth range is from 0.5 meters to 3 meters. During the exploration process, the material type, such as cast iron or polyethylene, and the burial depth value are recorded for each pipeline point. When storing data, it is matched with the spatial location information of the pipeline point through point identifiers. For example, material and burial depth fields are added to the attribute table and associated with point identifiers.

[0078] The latitude and longitude coordinates, elevation coordinates, pipeline material, and pipeline burial depth are integrated to construct the original dataset. The integration process uses Geographic Information System (GIS) software as the data processing platform, such as commonly used commercial or open-source GIS tools. First, the latitude and longitude coordinates and elevation coordinates are merged into three-dimensional spatial coordinate points. The coordinate system adopts a plane projection coordinate system and an elevation system, such as the universal transverse Mercator projection zone 50N. The elevation system is based on the mean sea level datum, ensuring the consistency and convertibility of spatial location information. Then, pipeline material and burial depth attribute data are associated with spatial coordinate points through point identifiers. The data format is set to tabular form, with table fields including point identifier, longitude, latitude, elevation, material, and burial depth. For example, the point identifier is a numerical sequence number, such as 1, 2, 3; the material field stores a text description, such as "steel pipe"; and the burial depth field stores a numerical value, such as 1.2 meters. During integration, data validation is performed to check for missing or outlier values. For example, if longitude or latitude values ​​exceed a reasonable geographical range, they are marked as anomalous data and a review process is initiated. Values ​​with a burial depth less than 0 or exceeding a preset maximum depth are considered invalid and require re-exploration. The final raw dataset is stored as a database file or spreadsheet, such as an SQLite database or Excel file, for easy use in subsequent analysis. Furthermore, data cleaning steps are implemented during integration, such as removing duplicate records, standardizing data units, and standardizing text formats to ensure the integrity and consistency of the data quality. The entire data collection and integration process follows standardized operating procedures to ensure the repeatability and verifiability of each step.

[0079] S2. Based on the original dataset, perform integrity analysis of pipeline point connection information and generate connection information integrity judgment results. The specific implementation is as follows:

[0080] In the process of performing integrity analysis of pipeline point connection information based on the original dataset and generating integrity judgment results, an initial connection graph representing the adjacency relationships between pipeline points is first constructed based on the spatial location information of the pipeline points contained in the original dataset. Specifically, an undirected graph model from graph theory is used as the data structure, where each pipeline point is a vertex. The connection relationship between vertices is determined by calculating the three-dimensional Euclidean distance based on the latitude, longitude, and elevation coordinates in the spatial location information of the pipeline points. The input parameters for calculating the Euclidean distance are the differences in longitude, latitude, and elevation between adjacent pipeline points. The actual spatial distance value is obtained by summing the squares of each difference and then taking the square root, for example, the distance value is output in meters. To define the adjacency relationship, a distance threshold is set. This threshold is based on the actual layout characteristics of the pipeline network and engineering experience. For example, the typical spacing range of common underground pipelines is from 5 meters to 20 meters, so the distance threshold is set to 10 meters. When the Euclidean distance between two pipeline points is less than or equal to 10 meters, an undirected edge is added to the initial connection graph to indicate that they have an adjacency relationship. The initial connectivity graph is constructed using a computer program. The program reads the spatial location information of pipeline points from the original dataset, iterates through all pipeline point pairs, calculates the Euclidean distance between each pair and compares it to a distance threshold, ultimately generating a graph structure containing a set of vertices and edges. This graph is stored in an adjacency list format for easy retrieval. Data preprocessing is also performed during the construction process, such as checking for missing or out-of-range coordinate data, and removing obvious outliers such as latitude and longitude values ​​exceeding geographical boundaries or negative elevation values, ensuring the accuracy and reliability of the initial connectivity graph.

[0081] Pipeline points with missing connection information are identified based on the connectivity distribution characteristics of each pipeline point in the initial connectivity graph. Connectivity is defined as the number of edges connected to each vertex in the graph. To calculate connectivity, each vertex in the initial connectivity graph is traversed, and the number of its adjacent vertices is counted as the connectivity value. The connectivity distribution characteristics are quantified using statistical analysis methods, such as calculating the arithmetic mean, standard deviation, and distribution histogram of the connectivity of all pipeline points, identifying pipeline points with connectivity values ​​significantly lower than normal levels. A connectivity threshold is set during the identification process. This threshold is dynamically set based on the characteristics of the connectivity distribution. For example, when the connectivity distribution approximates a normal distribution, the connectivity threshold is set to the mean minus twice the standard deviation. If the connectivity value of a pipeline point is lower than this threshold, it is marked as a pipeline point that may have missing connection information. The basis for setting the connectivity threshold includes historical pipeline network data analysis and domain expert knowledge. For example, in a typical pipeline network, the normal connectivity range is usually from 1 to 5; therefore, the connectivity threshold may be set to 1, indicating that pipeline points with a connectivity value less than 1 are considered abnormal candidate points. The identification algorithm is implemented by a program. The program reads the adjacency list structure of the initial connected graph, calculates the connectivity of each vertex, analyzes the overall distribution characteristics, applies a connectivity threshold to filter out candidate missing points, and outputs a list of identifiers for the candidate missing points.

[0082] By analyzing the spatial distances between pipeline points with missing connection information and their adjacent pipeline points, and comparing these distances with pipeline attribute information, the missing connection information status is confirmed. Spatial distance analysis, based on the spatial location information of the pipeline points, calculates the Euclidean distances between the missing pipeline point and all its adjacent pipeline points. The distance calculation method is the same as when constructing the initial connection map, with coordinate differences as the input parameter. Simultaneously, pipeline attribute information matching analysis is performed. Pipeline attribute information includes pipeline material and burial depth. The matching relationship is determined by comparing the attribute values ​​of the missing pipeline point with its adjacent pipeline points. For example, material matching checks if the material type strings are completely identical, and burial depth matching checks if the difference in burial depth values ​​is within the allowable range. Matching rules set similarity thresholds; for example, a matching score of 1 is assigned when the materials are completely identical, otherwise 0; a matching score of 1 is assigned when the burial depth difference is less than 0.5 meters, otherwise 0. A weighted scoring method is used to confirm the missing connection information state, considering both spatial distance and attribute matching relationships. The weights for spatial distance and attribute matching are each set to 0.5, based on engineering experience principles indicating that spatial and attribute factors are equally important. The scoring process involves first normalizing the spatial distance values ​​to a score between 0 and 1 (e.g., smaller distances receive higher scores), then summing and normalizing the attribute matching scores. The final score equals the spatial distance score multiplied by its weight plus the attribute matching score multiplied by its weight. If the total score is below a set threshold, such as 0.5, the pipeline point is confirmed to be in a missing connection information state. The confirmation process is implemented programmatically. The program iterates through each candidate missing point, calculates the distance and attribute matching degree with all adjacent points, applies the weighted score and the total score threshold for judgment, and finally outputs the confirmation status result for each point.

[0083] The system determines whether at least one pipeline point is confirmed to be in a state of missing connection information. This determination process is based on the previous step's confirmation of missing connection information. The program checks the status records of all pipeline points and counts the number of points in a missing connection information state. If the count is greater than or equal to 1, it is determined that at least one pipeline point is confirmed to be in a state of missing connection information; otherwise, it is determined that no such point exists. The determination logic is implemented through conditional statements. For example, a counter variable is set in the program to iterate through the status list. If at least one status is found to be missing, the existence condition is triggered. This step ensures that the determination is based on objective data statistics, avoiding subjective judgment, and also handles boundary cases. For example, when all pipeline point statuses are complete, it is directly determined that no such point exists, and the determination timestamp is recorded for auditing purposes.

[0084] If a connection exists, a judgment result indicating incomplete connection information is generated; otherwise, a judgment result indicating complete connection information is generated. The generation process is based on the judgment result, and the program outputs the judgment result in text form. For example, if at least one missing point exists, the output string "Connection information is incomplete" is output; if none exists, the output string "Connection information is complete" is output. The judgment result is stored as a separate data file or integrated into the system log to ensure traceability and verifiability. The generation logic is simple and direct, relying on the judgment output of the previous step, requiring no additional calculations, but includes data post-processing, such as recording the generation timestamp and associating it with the original dataset identifier, to ensure the consistency and repeatability of the entire analysis process. The final judgment result is used to guide subsequent processing flows, such as triggering connection relationship analysis or marking a data quality report.

[0085] S3. When the judgment result is incomplete, based on engineering rules and spatial rules, perform connection relationship analysis on pipeline points with missing connection information to generate preliminary connection relationship analysis results. The specific implementation is as follows:

[0086] In the process of generating preliminary connection relationship analysis results for pipeline points with missing connection information based on engineering and spatial rules when the judgment result is incomplete, candidate connection relationships are first constructed based on the spatial location information between the pipeline point with missing connection information and its adjacent pipeline points, according to pipeline continuity constraints. Specifically, the construction process obtains the list of pipeline point identifiers with missing connection information and the spatial location information of these points and all adjacent pipeline points from the output of step S2. The spatial location information includes latitude and longitude coordinates and elevation coordinates, stored in decimal degrees and meters. Pipeline continuity constraints are implemented through direction vector analysis. The direction vector is derived from the coordinate difference between the pipeline point with missing connection information and each adjacent pipeline point. The input parameters for vector calculation are the longitude difference, latitude difference, and elevation difference. For example, the longitude difference is the longitude of the adjacent point minus the longitude of the missing point; the latitude difference is the latitude of the adjacent point minus the latitude of the missing point; and the elevation difference is the elevation of the adjacent point minus the elevation of the missing point.

[0087] Continuity of direction is evaluated by calculating the angle between two adjacent direction vectors, as shown in the formula: ;in, Indicates the included angle, in degrees; This represents the first direction vector, whose components are... , , ; This represents the second direction vector, whose components are... , , ; The vector dot product is calculated as follows: ; Representing vectors The modulus is calculated as ; Representing vectors The modulus is calculated as ; This represents the longitude difference of the first vector, in degrees. This represents the difference in latitude between the first vectors, in degrees. This represents the elevation difference of the first vector, in meters. This represents the longitude difference between the second vectors, in degrees. This represents the difference in latitude between the second vectors, in degrees. This represents the elevation difference of the second vector, in meters.

[0088] The angle threshold is set to 30 degrees. This threshold is based on historical pipeline layout data and engineering experience, as pipeline routes are typically smooth and continuous, and angle changes exceeding 30 degrees may indicate inflection points or measurement errors. For each pipeline point with missing connection information, the program iterates through all its adjacent pipeline points, calculates the direction vector between the current point and its neighbors, and calculates the angle between this vector and the reference vector of the line connecting the neighbors. If the angle is less than or equal to 30 degrees, the neighboring point is added to the candidate connection relationships. Candidate connection relationships are recorded in point-pair format, including a missing point identifier and a candidate point identifier, and stored as a list structure. During the construction process, the program performs data preprocessing, such as checking whether the coordinate values ​​are within a reasonable range, removing obvious outliers such as latitude and longitude exceeding geographical boundaries or negative elevation values, to ensure the correctness of vector calculations.

[0089] Based on the pipeline attribute information contained in the original dataset, candidate connection relationships are filtered according to the rule of prioritizing connections between pipelines of the same type. Pipeline attribute information is read from the original dataset, including a pipeline material type field, where the material type is a string data such as cast iron or polyethylene. The rule of prioritizing connections between pipelines of the same type requires prioritizing connections between pipeline points with the same material. The filtering process is achieved by comparing whether the material strings of two pipeline points in a candidate connection relationship are completely identical. A priority mechanism is set in the rule implementation: a priority score of 1 is assigned when the materials are completely identical, and a priority score of 0 is assigned when the materials are inconsistent. Then, the candidate connection relationships are sorted according to the priority scores, with higher-scoring relationships being retained first. The filtering algorithm is implemented programmatically. The program reads the list of candidate connection relationships and the corresponding pipeline attribute information. For each point pair, it retrieves the material values ​​of the two points, compares whether the strings are completely identical, calculates the priority score, and then sorts the entire list in descending order of score. The filtered candidate connection relationship list is output, retaining only relationships with a score of 1 or, as needed, the top N relationships (e.g., N set to 10). The priority setting is based on domain knowledge; for example, pipelines of the same material have similar physical characteristics and durability, making the connection more reliable and reducing the risk of system leakage or failure. During the filtering process, the program handles situations where material data is missing. For example, when the material value is empty, a priority score of 0 is assigned by default to ensure the robustness of rule application.

[0090] The rationality of candidate connection relationships is verified by analyzing the burial depth data between adjacent pipeline points according to the principle of burial depth consistency. The burial depth data is obtained from the original dataset and recorded in meters. The burial depth consistency principle requires that the difference in burial depth between connection points be within an allowable range. The verification process calculates the absolute value of the difference in burial depth between the two pipeline points in the candidate connection relationship. A burial depth difference threshold of 0.5 meters is set, based on common pipeline construction practices and allowable error ranges; for example, burial depth errors are usually controlled within 0.5 meters. For each candidate connection relationship, the program reads the burial depth values ​​of the two points, calculates the absolute value of the difference, and marks the rationality as true if the difference is less than or equal to 0.5 meters; otherwise, it is marked as false. The verification process iterates through the filtered list of candidate connection relationships, performs burial depth calculations and threshold comparisons, and outputs the rationality flag for each relationship. The program also handles abnormal situations, such as using default values ​​or skipping verification when burial depth data is missing, and records verification logs for subsequent auditing. The implementation of the principle of consistent burial depth is based on engineering practice. For example, abrupt changes in burial depth may indicate different pipeline layers or incorrect connections. Therefore, a threshold of 0.5 meters ensures a smooth transition in connections.

[0091] The preliminary connection analysis results are generated by synthesizing the screened and validated candidate connections. The synthesis process, based on the outputs of screening and validation, retains only candidate connections that have a priority score of 1 in screening and a reasonableness flag of true in validation. The program merges the two lists, removing duplicate or conflicting entries, for example, by comparing point identifier pairs to ensure the uniqueness of each connection, forming the final preliminary connection analysis results. The analysis results are output in a structured data format, including a list of point pairs, priority scores, and reasonableness flags, and stored as a JSON file or database table. The generation logic ensures data consistency, for example, by associating all information through point identifiers and recording generation timestamps and reference to the judgment result identifier in S2. The final output is used for subsequent processing steps, such as the connection probability assessment in S4, and provides auditable logs of the input and output parameters of the entire analysis process. The synthesis step also includes post-processing, such as counting the number of valid connections and generating a summary report, ensuring the readability and traceability of the results.

[0092] S4. Based on the preliminary connection relationship analysis results, by analyzing the dynamic response correlation between pipelines and combining the network power flow distribution compatibility, a graph signal processing algorithm is used to evaluate the pipeline connection probability and generate pipeline connection probability evaluation results. The specific implementation is as follows:

[0093] In the process of evaluating pipeline connection probability based on the preliminary connectivity analysis results, analyzing the correlation of dynamic responses between pipelines and combining network power flow distribution compatibility with graph signal processing algorithms, the evaluation of pipeline connection probability is first carried out by constructing a graph structure containing potential connection paths based on the candidate connectivity relationships in the preliminary connectivity analysis results. Specifically, the construction process obtains the candidate connectivity relationship list from the output of step S3, which contains point pairs between pipeline points with missing connection information and candidate connection points. The graph structure to be evaluated is represented by an undirected graph model, where each pipeline point is a vertex of the graph, with vertex identifiers consistent with those in the original dataset. Candidate connectivity relationships are represented as edges of the graph, and an edge connecting two vertices represents a potential connection path. During construction, the program reads the candidate connectivity relationship list, initializes an empty graph structure, and adds vertices and edges one by one, ensuring no duplicate vertices or edges. The graph structure is stored in adjacency list format for easy subsequent graph operations and signal processing. Data validation is performed during construction, such as checking if vertices exist in the original dataset and removing invalid vertices to ensure the integrity and consistency of the graph structure. The graph structure to be evaluated also includes vertex attributes, such as spatial location information, to aid in analysis.

[0094] A graph signal is defined on the graph structure to be evaluated, using pipeline pressure or flow monitoring data as the vertex signals. The pipeline pressure or flow monitoring data is obtained from an external monitoring system or historical database; for example, pressure data is in kilopascals (kPa) and flow data is in cubic meters per second (m³ / s). The graph signal is defined as the monitoring data value associated with each vertex. The data format can be scalar or time series; for example, each vertex corresponds to a pressure value or a flow series at a set of time points. The signal assignment process is implemented programmatically. The program iterates through all vertices of the graph structure to be evaluated, retrieves the corresponding data from the monitoring data source based on the vertex identifier, and maps the data to the vertex as the signal value. If monitoring data is missing, interpolation methods or default values ​​are used to fill the gaps, such as interpolating using the average value of adjacent vertices. The graph signal is stored in vector form, consistent with the vertex order, for easy operation by graph signal processing algorithms. The signal definition also includes data preprocessing, such as noise and outlier removal, to ensure signal quality.

[0095] The program calculates the dynamic response correlation coefficient between the vertex signals corresponding to each candidate connectivity in the graph structure to be evaluated. The dynamic response correlation coefficient quantifies the degree of association between two vertex signals and is calculated based on signal time series data. For each candidate connectivity, the program retrieves their graph vertex signals, such as stress time series, and then calculates the Pearson correlation coefficient as the dynamic response correlation coefficient. The input parameters for the Pearson correlation coefficient calculation are the numerical values ​​of the two signal sequences. The calculation steps include first calculating the mean of the two sequences, then calculating the covariance and their respective standard deviations, and finally, the correlation coefficient is equal to the covariance divided by the product of the two standard deviations. The correlation coefficient value ranges from -1 to +1, with a value closer to +1 indicating a stronger positive correlation. During the calculation, the program handles cases of inconsistent signal lengths, such as by truncating or interpolating to make the sequences the same length, ensuring the accuracy of the calculation. The dynamic response correlation coefficient is output as a numerical value and stored in association with the candidate connectivity.

[0096] The simulation of network power flow distribution is based on centralized pipeline attribute information from the original dataset, and the compatibility index between the power flow distribution and the pipeline attribute information is calculated. The network power flow distribution simulation uses a hydraulic model. Input parameters include pipeline attribute information such as pipeline material, diameter, length, and burial depth, as well as boundary conditions such as source pressure and sink flow. The simulation process is achieved by solving the pipeline network flow equations, for example, using iterative methods to calculate flow and pressure distribution. The simulation output is the theoretical flow or pressure value for each pipeline segment. The compatibility index is calculated by comparing the degree of matching between the simulated power flow distribution and actual monitoring data or attribute constraints. For example, the root mean square error (RMSE) between the simulated flow and the expected flow derived from the attributes is used as the compatibility index. The input for the RMSE calculation is a sequence of simulated and expected values. The calculation steps include first calculating the square of the difference at each point, then calculating the average, and finally taking the square root. A smaller compatibility index value indicates better compatibility. The program automatically performs the simulation and calculation, handling cases where attribute data is missing, such as using default parameters.

[0097] The dynamic response correlation coefficient and the network power flow distribution compatibility index are input into a graph signal processing algorithm for fusion calculation to obtain a pipeline connection probability score. The fusion calculation, based on the graph signal processing algorithm, first maps the dynamic response correlation coefficient and the network power flow distribution compatibility index to the graph spectral domain using a graph Fourier transform. The inputs to the graph Fourier transform are the Laplace matrix of the graph structure to be evaluated and the graph signal. The transformation process includes calculating the eigenvalues ​​and eigenvectors of the Laplace matrix, and then projecting the signal onto the eigenvector space to obtain a spectral representation. In the graph spectral domain, the dynamic response correlation coefficient and the network power flow distribution compatibility index are normalized separately, for example, using min-max normalization to scale the values ​​to the range of 0 to 1. After normalization, a weighted fusion calculation is performed. The weights are set based on historical data analysis; for example, the weight of the dynamic response correlation coefficient is 0.6, and the weight of the network power flow distribution compatibility index is 0.4, indicating that the dynamic response correlation is given more importance. The weighted fusion calculation involves multiplying the normalized values ​​by their respective weights and then summing them. Finally, the weighted fusion result is transformed back to the vertex domain using an inverse graph Fourier transform. The inverse transform process uses the same eigenvectors to reconstruct the spectral signal into a vertex domain signal, yielding a pipeline connection probability score for each candidate connection. The score ranges from 0 to 1, with higher values ​​indicating a greater connection probability.

[0098] The pipeline connection probability assessment results are generated based on the pipeline connection probability score. The generation process is based on the pipeline connection probability score; the program sorts all candidate connections in descending order of score and sets a scoring threshold, for example, 0.7. Candidate connections with scores higher than 0.7 are marked as high-probability connections. The assessment results are output as structured data, including a list of candidate connections, score values, and probability levels, and stored as a database table or file. The generation logic includes data post-processing, such as removing connections with excessively low scores, and generating a summary report.

[0099] S5. Based on the pipeline connection probability assessment results, conduct network structure robustness and cascade failure risk analysis to obtain a robust path subset. Based on the robust path subset, generate the optimal connection path set through multi-agent game simulation of the group consensus formation process. The specific implementation is as follows:

[0100] In the process of implementing network structure robustness and cascading failure risk analysis based on pipeline connection probability assessment results to obtain a robust path subset and generating an optimal connection path set through multi-agent game simulation of the group consensus formation process, the initial path set containing each candidate connection path is first constructed based on the pipeline connection probability assessment results. Specifically, the construction process obtains the pipeline connection probability assessment results from the output of step S4, which includes a list of candidate connection paths and their corresponding connection probability scores. The construction of the initial path set is implemented through a computer program. The program reads the candidate connection path information from the assessment results, including pipeline point identifiers at the path's start and end points and connection probability scores. All candidate connection paths with scores higher than a set threshold are included in the initial path set. The threshold is determined based on historical data analysis; for example, by analyzing the connection success rate of similar pipeline network projects in the past, the threshold is set to 0.6, filtering out paths with scores higher than 0.6. The initial path set is stored in the form of a data structure containing path identifiers, endpoint information, and score values ​​for easy subsequent analysis and processing. During the construction process, the program performs data validation, such as checking whether the path endpoints exist in the original dataset, removing invalid paths, and ensuring the integrity of the set.

[0101] Network structural robustness is assessed by calculating network connectivity retention after removing non-critical connection paths. Non-critical connection paths are identified based on their importance within the network topology, quantified by their betweenness centrality, defined as the proportion of shortest paths that pass through that path. The program iterates through each path in the initial path set, temporarily removing it from the network topology, and then calculates network connectivity retention. Connectivity retention is measured as the ratio of the number of connected components in the remaining network to the original network. For example, if the original network has 10 connected components and removing a path leaves 9, the connectivity retention is 0.9. A connectivity threshold of 0.8 is set, based on engineering experience, representing the minimum acceptable level of network connectivity. Paths with a connectivity retention below 0.8 after removal are marked as critical paths. The network structural robustness assessment results are output as path importance scores, calculated based on connectivity retention; higher scores indicate a smaller impact of the path on network connectivity. During the evaluation process, the program handles boundary conditions, such as when the network is completely disconnected, the connectivity level is 0, and an exception log is recorded.

[0102] Cascade failure risk is assessed by simulating the cascading effect of a single connection path failure. The simulation is based on a network hydraulic model, with input parameters including pipeline attributes and initial flow distribution. For each path in the initial path set, the program sets its failure state and then simulates the propagation of the failure within the network. The propagation simulation uses an iterative algorithm, checking the load changes of adjacent paths in each iteration. Load changes are calculated based on flow redistribution, and a path is marked as failed when its load exceeds its capacity threshold. The capacity threshold is calculated based on pipe diameter and material from the pipeline attributes; for example, a larger pipe diameter results in a higher capacity threshold. The cascading effect range is quantified by the ratio of the number of failed paths to the total number of paths. For example, if there are 100 paths in total and failure propagates to the point that 10 paths fail, the cascading effect range is 0.1. The cascading failure risk assessment result is output as a risk coefficient, which is equal to the cascading effect range value; a higher value indicates a greater risk. The program handles abnormal situations during the simulation, such as terminating the simulation when the number of iterations exceeds a set limit to avoid infinite loops.

[0103] A robust path subset was selected based on the network structure robustness assessment results and cascading failure risk assessment results. The selection process combined path importance scores and risk coefficients, employing a weighted scoring method. In the weighted scoring, the path importance score had a weight of 0.6, and the risk coefficient had a weight of 0.4. These weights were set based on domain expert knowledge and historical project data analysis, indicating that network connectivity maintenance is slightly more important than risk control. The weighted score was calculated by multiplying the path importance score by its corresponding weight, adding the risk coefficient by its corresponding weight, and then negatively assigning a risk coefficient, as a higher risk coefficient indicates a less stable path. A weighted scoring threshold of 0.7 was set, determined through experimental testing, such as adjusting the threshold in a simulated environment to observe the selection effect. Paths with scores higher than 0.7 were included in the robust path subset. The selection algorithm was implemented programmatically. The program read the assessment results, calculated the weighted scores, applied the threshold for selection, and output a list of robust path subsets. During the selection process, the program handled data inconsistencies, such as using default values ​​when scores were missing.

[0104] A multi-agent negotiation model is established on a robust path subset, with path attributes as the interest demands. Specifically, each candidate connection path in the robust path subset is defined as an independent agent, with the agent identifier matching the path identifier. Path attributes, including pipeline material and burial depth, are extracted from pipeline attribute information in the original dataset as the interest demands of each agent. Material attributes are represented as strings, such as cast iron or polyethylene, and burial depth is represented in meters, such as 1.5 meters. Multi-agent negotiation rules are established with path attribute similarity and network structure robustness as negotiation objectives. These rules define the interaction logic between agents. Path attribute similarity is calculated by comparing the material and burial depth attributes of two agents. Material similarity is 1 if the materials are the same, and 0 otherwise. Burial depth similarity is calculated based on the burial depth difference, with 1 if the difference is less than 0.3 meters, and 0 otherwise. The difference threshold is set based on construction accuracy standards. Network structure robustness is represented by a path importance score based on the aforementioned evaluation results. The negotiation rules require agents to adjust their demands in each round of negotiation, moving closer to other agents with high attribute similarity and good network structure robustness.

[0105] Through multiple rounds of negotiation and iteration, the interests of each agent are aligned to form a group consensus. The negotiation and iteration process employs a distributed algorithm, where each agent updates its own attribute demands according to negotiation rules in each iteration. These update rules are calculated based on the average demand value of neighboring agents; for example, an agent's new burial depth demand is equal to the average of the burial depth demands of all agents in its neighborhood. The neighborhood is defined as other agents with an attribute similarity higher than 0.8, and the similarity threshold is set based on experimental data analysis. The iteration termination condition is set so that the change in the attribute demands of all agents in two consecutive iterations is less than a set threshold, such as 0.01 meters, which is determined through convergence testing. The negotiation process is simulated by a program. The program initializes the demands of each agent, then repeatedly executes updates and judgments until the termination condition is met, outputting the final consensus-reaching agent state. The program handles exceptions during iteration; for example, if consensus cannot be reached, it forcibly terminates and records the reason.

[0106] The optimal set of connection paths is generated based on consensus reached among the stakeholders. The generation process selects paths based on the states of the agents after negotiation, considering both path attribute consistency and network structural robustness. The program filters paths from the consensus-reaching agents whose attribute demands most closely resemble the overall characteristics of the pipeline network; for example, it selects the path with the smallest difference between the requested burial depth and the average burial depth of the pipeline. Simultaneously, it considers path importance scores, prioritizing paths with higher scores. The optimal set of connection paths is output as a path list, including path identifiers and explanations of the selection rationale, for subsequent processing steps. The generation logic ensures that path selection satisfies both attribute coordination and network stability, and the negotiation process is logged for auditing purposes. Post-processing is performed during the generation process, such as removing duplicate paths and generating a summary report, ensuring readability and traceability of the results.

[0107] S6. Based on the optimal connection path set, perform constraint satisfaction analysis to generate a pipeline network topology diagram. The specific implementation is as follows:

[0108] In the process of generating pipeline network topology diagrams based on constraint satisfaction analysis using the optimal connection path set, a constraint satisfaction problem model is first established, incorporating both spatial geometric constraints and engineering logical constraints. Specifically, the optimal connection path set is obtained from the output of step S5, containing multiple candidate connection paths and their attribute information. The constraint satisfaction problem model adopts the standard constraint satisfaction problem framework, defining each candidate connection path as a decision variable, with the variable's value domain being the connection state, such as connected or not connected. Spatial geometric constraints are defined based on the coordinate data in the spatial location information of pipeline points. Constraints include minimum distance requirements between connection paths and requirements to avoid intersections. For example, a minimum distance threshold of 2 meters is set; when the spatial distance between two connection paths is less than 2 meters, it is considered a constraint violation. This threshold is set based on historical pipeline layout data and engineering practice, determined by analyzing the safety spacing requirements of similar past pipeline network projects. Engineering logical constraints are defined based on the material type and burial depth data in the pipeline attribute information. Constraints include material compatibility requirements and burial depth consistency requirements. For example, connection paths are required to have the same or compatible material types, and the burial depth difference must be less than 0.5 meters. These thresholds are determined by analyzing historical pipeline fault data and engineering experience. Model construction is achieved through a computer program. The program reads the optimal set of connection paths, initializes variables and constraints, and establishes a constraint network for subsequent solving. During the construction process, the program performs data validation, such as checking the existence of path endpoints and eliminating invalid paths, to ensure the integrity of the model.

[0109] The coordinate data from the spatial location information of pipeline points, along with the material type and burial depth data from the pipeline attribute information, are used as constraint variables in the constraint satisfaction problem model. The input process is implemented through data mapping. The program extracts the spatial location information of pipeline points from the original dataset, including latitude and longitude coordinates and elevation coordinates, with coordinate data in decimal degrees and meters. Simultaneously, it extracts pipeline attribute information, including a material type string and a burial depth value, with material types such as cast iron or polyethylene, and burial depth in meters. This data is converted into attribute values ​​for constraint variables; for example, each connection path variable is associated with its endpoint coordinates, material, and burial depth. During the input process, the program performs data preprocessing, such as checking whether the coordinate data is within a reasonable geographical range and whether the material data is complete. Missing values ​​are filled using interpolation or default values, for example, using the material of adjacent pipelines as the default value, ensuring the validity of the input data. Constraint variables are stored as data structures and associated with the constraint satisfaction problem model for easy constraint checking and processing. The input logic ensures data consistency, for example, by associating all information through point identifiers.

[0110] The constraint propagation algorithm eliminates contradictory connection paths that violate both spatial geometric and engineering logical constraints from the optimal connection path set. The algorithm employs an arc consistency check method, traversing all variables and constraints in the constraint network. For each connection path variable, the algorithm checks its constraint satisfaction with adjacent variables, such as calculating the spatial distance between the endpoints of two paths and comparing material type and burial depth differences. Spatial distance calculation is based on coordinate data, using the Euclidean distance formula, with input parameters being longitude difference, latitude difference, and elevation difference, and output in meters. If a constraint violation is found, such as a spatial distance less than 2 meters or material incompatibility, the path is marked as a contradictory path and removed from the current solution candidate set. The algorithm iterates until no new contradictions are found; for example, a maximum of 100 iterations is set to avoid infinite loops. After contradictory paths are eliminated, the remaining path set is output as a temporary solution. The algorithm implementation includes exception handling; for example, when multiple paths simultaneously violate constraints, the path with the greater impact is removed first. Impact assessment is based on the path's connectivity importance in the network, calculated through the path's degree centrality.

[0111] A backtracking search algorithm is used to determine the final pipeline connections within the solution space that satisfies all constraints. The algorithm starts with a provisional solution set and recursively attempts to assign path connection states. It first selects an unassigned path variable, attempts to set it as a connection state, and then checks if all constraints are satisfied. If satisfied, it continues to the next variable; otherwise, it backtracks to the previous variable and tries other assignments. A pruning strategy is used to optimize efficiency during the search, such as terminating a branch early when a partial solution violates constraints. The solution space is defined as a subset of all possible combinations of connection paths. The algorithm ensures that it finds either the first feasible solution or all feasible solutions, for example, by setting the search depth limit to twice the number of paths. The final output is a set of pipeline connections that satisfy all spatial geometric and engineering logic constraints, stored as a list of paths containing path identifiers and connection states. The algorithm handles boundary cases, such as returning an empty set and logging an error when no solution is found.

[0112] A pipeline network topology graph is generated based on the final pipeline connection relationships and the spatial location information of pipeline points. The generation process uses a graphical modeling tool; the program reads the final set of pipeline connections and the spatial location information of pipeline points. The topology graph is constructed as an undirected graph, where pipeline points are the vertices and connections are the edges. Vertex attributes include coordinate data, and edge attributes include connection type and path information. The graph generation algorithm traverses all connections, adds vertices and edges, and ensures topological consistency, such as avoiding isolated vertices or duplicate edges. The topology graph is output in a standard graphical format, such as a vector graphics file or a GIS-compatible format, facilitating visualization and further analysis. Post-processing is performed during the generation process, such as verifying the connectivity and integrity of the graph and adding metadata such as generation timestamps and data sources to ensure the traceability and availability of the results. The generation logic includes exception handling, such as triggering warnings and logging the reasons when the graph is not connected.

[0113] Example 2: Figure 2 A schematic diagram of the intelligent detection system for underground pipe network data based on multi-source data fusion of the present invention is provided. The intelligent detection system for underground pipe network data based on multi-source data fusion includes:

[0114] The information acquisition module is used to acquire the spatial location information and pipeline attribute information of pipeline points collected by multi-source detection equipment, and to construct the original dataset;

[0115] The complete judgment module is used to perform integrity analysis of pipeline point connection information based on the original dataset and generate connection information integrity judgment results.

[0116] The preliminary analysis module is used to generate preliminary connection relationship analysis results by performing connection relationship analysis on pipeline points with missing connection information based on engineering rules and spatial rules when the judgment result is incomplete.

[0117] The probability assessment module is used to evaluate the probability of pipeline connections based on the preliminary connection relationship analysis results, by analyzing the dynamic response correlation between pipelines and combining the network power flow distribution compatibility, and by using graph signal processing algorithms to generate pipeline connection probability assessment results.

[0118] The set generation module is used to analyze the robustness of network structure and the risk of cascading failure based on the pipeline connection probability assessment results to obtain a robust path subset. Based on the robust path subset, the optimal connection path set is generated by simulating the group consensus formation process through multi-agent game theory.

[0119] The relationship generation module is used to perform constraint satisfaction analysis based on the optimal connection path set and generate a pipeline network topology diagram.

[0120] The calculations involved in the embodiments are all dimensionless numerical calculations, and the preset parameters and thresholds in the calculations are set by those skilled in the art according to the actual situation.

[0121] It should be noted that this invention can be deployed on the device itself to realize embedded applications, or it can run on a PC or other terminal with a user interface, thereby meeting various hardware environments and usage requirements.

[0122] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions according to the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. Computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wireless or wired transmission; wired transmission methods include optical fiber, twisted pair, coaxial cable, etc.; wireless transmission includes infrared, microwave, etc. Computer-readable storage media can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more sets of available media. Available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media. Semiconductor media can be solid-state drives.

[0123] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and modules described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0124] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0125] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0126] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0127] If a function is implemented as a software module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0128] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0129] In conclusion, the above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for intelligent detection of underground pipe network data through multi-source data fusion, characterized in that, The method comprises the following steps: S1, acquiring pipeline point spatial position information and pipeline attribute information collected by a multi-source detection device, and constructing an original data set; S2, performing pipeline point connection information integrity analysis based on the original data set to generate a connection information integrity judgment result; S3, when the judgment result is incomplete, performing connection relationship analysis on the pipeline points with missing connection information based on engineering rules and spatial rules to generate a preliminary connection relationship analysis result; S4, based on the preliminary connection relationship analysis result, the pipeline connection possibility is evaluated by analyzing the dynamic response correlation between pipelines and combining network power flow distribution compatibility to generate a pipeline connection possibility evaluation result, including: constructing a to-be-evaluated graph structure containing potential connection paths based on the candidate connection relationships in the preliminary connection relationship analysis result; defining a graph signal on the to-be-evaluated graph structure, taking pipeline pressure or flow monitoring data as the graph vertex signal; calculating the dynamic response correlation coefficient between the graph vertex signals corresponding to each candidate connection relationship in the to-be-evaluated graph structure; simulating network power flow distribution based on the pipeline attribute information in the original data set and calculating the compatibility index of power flow distribution and pipeline attribute information; inputting the dynamic response correlation coefficient and the network power flow distribution compatibility index into the graph signal processing algorithm for fusion calculation to obtain a pipeline connection possibility score; generating a pipeline connection possibility evaluation result according to the pipeline connection possibility score; S5, based on the pipeline connection possibility evaluation result, network structure robustness and cascading failure risk analysis are performed to obtain a robust path subset, and based on the robust path subset, an optimal connection path set is generated through a multi-agent game simulation of the group consensus formation process, including: constructing an initial path set containing each candidate connection path based on the pipeline connection possibility evaluation result; evaluating the network structure robustness by calculating the network connectivity retention degree after removing non-critical connection paths; evaluating the cascading failure risk by simulating the range of chain reactions triggered by the failure of a single connection path; based on the network structure robustness evaluation result and the cascading failure risk evaluation result, a robust path subset is obtained; establishing a multi-agent negotiation model on the robust path subset, taking path attributes as the interest demands; forming a group consensus by making the interest demands of each agent consistent through multiple rounds of negotiation iteration; based on the connection path selected by reaching the group consensus, an optimal connection path set is generated; S6, constraint satisfaction analysis is performed according to the optimal connection path set to generate a pipeline network topology relationship graph.

2. The intelligent probing of underground pipe network data method of multi-source data fusion according to claim 1, characterized in that, Acquire pipeline point spatial position information and pipeline attribute information collected by a multi-source detection device, and construct an original data set, including: acquire the longitude and latitude coordinates included in the pipeline point spatial position information through a global navigation satellite system receiving device; acquire the elevation coordinates included in the pipeline point spatial position information through a total station instrument; acquire the pipeline material and pipeline burial depth included in the pipeline attribute information through pipeline detection instrument exploration; integrate the longitude and latitude coordinates, elevation coordinates, pipeline material and pipeline burial depth to construct the original data set.

3. The intelligent probing of underground pipe network data from multi-source data fusion method according to claim 1, characterized in that, Perform pipeline point connection information integrity analysis based on the original data set to generate a connection information integrity judgment result, including: Based on the spatial position information of the pipeline points contained in the original data set, an initial connection graph representing the adjacency relationship between the pipeline points is constructed; According to the connection degree distribution characteristics of each pipeline point in the initial connection graph, pipeline points with missing connection information are identified; By analyzing the spatial distance and pipeline attribute information matching relationship between the pipeline points with missing connection information and their adjacent pipeline points, the connection information missing state is confirmed; Determine whether there is at least one pipeline point confirmed to be in the connection information missing state; If so, a connection information incomplete judgment result is generated; otherwise, a connection information complete judgment result is generated.

4. The intelligent probing of underground pipe network data from multi-source data fusion method according to claim 1, characterized in that, When the judgment result is incomplete, the missing connection information of the pipeline points is analyzed based on the engineering rules and spatial rules to generate a preliminary connection relationship analysis result, including: Based on the spatial position information between the pipeline points with missing connection information and their adjacent pipeline points, candidate connection relationships are constructed according to the pipeline direction continuity constraint; Based on the pipeline attribute information contained in the original data set, the candidate connection relationships are screened according to the same type pipeline priority connection rule; By analyzing the pipeline burial depth data between adjacent pipeline points, the rationality of the candidate connection relationships is verified according to the burial depth consistency principle; The preliminary connection relationship analysis result is generated by comprehensively screening and verifying the candidate connection relationships.

5. The intelligent probing of underground pipe network data using multi-source data fusion method as claimed in claim 1, wherein, The pipeline connection possibility score is obtained by fusing and calculating the dynamic response correlation coefficient and the network power flow distribution compatibility index into the graph signal processing algorithm, including: based on the graph vertex signal of each candidate connection relationship in the to-be-evaluated graph structure, the dynamic response correlation coefficient and the network power flow distribution compatibility index are mapped to the graph frequency domain through the graph Fourier transform in the graph signal processing algorithm; After normalization processing, the dynamic response correlation coefficient and the network power flow distribution compatibility index are weighted and fused in the graph frequency domain; The weighted fusion result is converted back to the vertex domain by inverse graph Fourier transform to obtain the pipeline connection possibility score.

6. The intelligent probing of underground pipe network data using multi-source data fusion method as claimed in claim 1, wherein, A multi-agent negotiation model based on path attributes is established on the robust path subset, including: each candidate connection path in the robust path subset is defined as an independent agent; The path attributes including pipeline material and pipeline burial depth are extracted as the benefit pursuit of each agent based on the pipeline attribute information in the original data set; A multi-agent negotiation rule is established with path attribute similarity and network structure robustness as the negotiation target.

7. The intelligent probing of underground pipe network data from multi-source data fusion method according to claim 1, characterized in that, According to the constraint satisfaction analysis of the optimal connection path set, a pipeline network topology relationship graph is generated, including: Based on the optimal connection path set, a constraint satisfaction problem model is established containing spatial geometric constraints and engineering logic constraints; The coordinate data in the pipeline point spatial position information and the material type and burial depth data in the pipeline attribute information are input as constraint variables into the constraint satisfaction problem model; By using the constraint propagation algorithm, the contradictory connection paths in the optimal connection path set that violate the spatial geometric constraints and engineering logic constraints are eliminated; The backtracking search algorithm is used to determine the final pipeline connection relationship in the solution space that satisfies all the constraints; Based on the final pipeline connection relationship and the pipeline point spatial position information, a pipeline network topology relationship graph is generated.

8. The underground pipe network data intelligent detection system of multi-source data fusion, used for realizing the underground pipe network data intelligent detection method of multi-source data fusion according to any one of claims 1-7, characterized in that, ​ An information acquisition module is configured to acquire pipeline point spatial position information and pipeline attribute information collected by a multi-source detection device and construct an original data set; A completeness judgment module is configured to perform pipeline point connection information completeness analysis based on the original data set and generate a connection information completeness judgment result; A preliminary analysis module is configured to, when the judgment result is incomplete, perform connection relationship analysis on pipeline points with missing connection information based on engineering rules and spatial rules and generate a preliminary connection relationship analysis result; A possibility evaluation module is configured to, based on the preliminary connection relationship analysis result, perform pipeline connection possibility evaluation by analyzing dynamic response correlation between pipelines and combining network power flow distribution compatibility, and generate a pipeline connection possibility evaluation result by using a graph signal processing algorithm; A set generation module is configured to perform network structure robustness and cascading failure risk analysis based on the pipeline connection possibility evaluation result, obtain a robust path subset, and generate an optimal connection path set by simulating a group consensus formation process through multi-agent game based on the robust path subset; A relationship generation module is configured to perform constraint satisfaction analysis based on the optimal connection path set and generate a pipeline network topology relationship graph.

Images