Urban drainage pipe network intelligent health analysis system based on multi-mode twinborn deduction

Abstract

The invention discloses an urban drainage pipe network intelligent health analysis system based on multi-modal twin deduction, relates to the technical field of drainage pipe network monitoring, and constructs an analysis system composed of a three-dimensional modeling module, a topological relation construction module, a characteristic parameter extraction module, a coupling accuracy evaluation module, a flow state misjudgment risk prediction module, a dynamic adjustment module and the like. By extracting a wellhead terrain residual height gradient and a historical waterlogging point height difference offset index and combining local hydraulic simulation and error sensitivity index analysis, classified recognition and dynamic correction of wellhead modeling deviation are achieved, the elevation coupling precision and hydraulic simulation reliability of a drainage pipe network model are effectively improved, and the method is suitable for large-scale popularization and application. Therefore, the intelligent monitoring, risk early warning and decision support capabilities of the urban drainage system are enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of drainage network monitoring, and in particular to an intelligent health analysis system for urban drainage networks based on multimodal twin deduction. Background Art

[0002] Intelligent health analysis of urban drainage networks utilizes technologies such as the Internet of Things, big data, and artificial intelligence to comprehensively analyze the structure, operational status, and historical data of urban drainage systems. This allows for real-time monitoring of the network's operational efficiency, blockage risks, structural damage, and other issues, assessing its "health status," and assisting urban drainage management departments in making maintenance decisions and issuing risk warnings, thereby improving the safety and management efficiency of the drainage system. However, in the 3D modeling of urban drainage networks, if the wellhead elevation is not accurately coupled with the terrain elevation (DEM), it can easily lead to misjudgment of the wellhead height in the model, affecting the assessment of free surface flow and pressure flow states. This can especially lead to misjudgment of backflow, overtopping, or water hammer risks in low-lying areas. Summary of the Invention

[0003] The purpose of the present invention is to provide an intelligent health analysis system for urban drainage pipe networks based on multimodal twin deduction to address the shortcomings of the background technology.

[0004] To achieve the above objectives, the present invention provides the following technical solutions: an intelligent health analysis system for urban drainage pipe networks based on multimodal twinning, comprising a three-dimensional modeling module, a topological relationship construction module, a characteristic parameter extraction module, a coupling accuracy assessment module, a flow pattern misjudgment risk prediction module, and a dynamic adjustment module;

[0005] 3D modeling module, used to build 3D structural models of drainage pipe networks based on GIS and BIM data;

[0006] A topological relationship building module, used to build a topological network in the three-dimensional structural model of the drainage pipe network for hydraulic simulation and event propagation analysis;

[0007] A feature parameter extraction module is used to extract the residual height gradient of the wellhead terrain and the height difference offset index of the historical waterlogging point based on the three-dimensional model and the DEM data;

[0008] A coupling accuracy assessment module is used to classify and assess the coupling between the wellhead elevation and the DEM based on the wellhead terrain residual height gradient and the historical waterlogging point height difference offset index, and divide the coupling into three categories: accurate, inaccurate, and incompletely accurate;

[0009] The flow pattern misjudgment risk prediction module is used to construct a local hydraulic model for inaccurate wellhead locations, simulate flow pattern changes under original and adjusted elevations, calculate the error sensitivity index, and evaluate the simulation error risk based on the error sensitivity index;

[0010] The dynamic adjustment module is used to adjust the elevation parameters of the identified sensitive wellhead position based on the error sensitivity index to form a proposed corrected elevation for subsequent simulation of the hydraulic model.

[0011] Preferably, the three-dimensional modeling module includes: obtaining two-dimensional or 2.5D geographic spatial data of the urban drainage network, including pipeline directions, node locations and buried depth information;

[0012] BIM data was introduced to construct 3D models of pipes, manholes, and gully components, with component attribute information embedded in the models, including pipe type, laying year, flow direction, and maintenance history.

[0013] Spatially align GIS and BIM data and fuse them to generate a high-precision 3D drainage network model with spatial coordinates, geometric forms, and semantic attributes.

[0014] Preferably, the topology relationship building module specifically includes:

[0015] Extract the connected components in the 3D structural model, identify them as nodes or edges, and establish unique identifiers;

[0016] Generate a directed topological graph structure based on the spatial position relationship and flow direction information, and record the starting point, end point, pipe diameter, length, material properties, flow direction and operation status of each edge;

[0017] Perform consistency checks on the topology network to detect whether there are isolated nodes, closed loop anomalies, or direction conflict topology errors. If any, perform automatic repairs or prompt manual intervention.

[0018] Preferably, the method for obtaining the residual height gradient of the wellhead terrain is as follows: obtaining the three-dimensional coordinates and elevation of the wellhead point: P0 = (x0, y0, z 井 ); where z 井 Indicates the elevation value of the wellhead in the three-dimensional model; x0 is the horizontal coordinate of the wellhead in the geographic space; y0 is the vertical coordinate of the wellhead; in the DEM, with the wellhead as the center, select the terrain elevation sampling point set within the radius R {P i}, each point is: P i =(x i ,y i ,z i ), i=1,2,...,n; n is the total number of points; calculate the horizontal distance d between each sampling point and the wellhead i , the expression is: Calculate the elevation residual ΔH for each point i , the expression is: ΔH i =z 井 -z i ;z iis the DEM elevation of the surrounding terrain sampling points; calculate the residual height gradient of the wellhead terrain, the expression is: Where SD is the residual height gradient of the wellhead terrain.

[0019] Preferably, the method for obtaining the historical waterlogging point height difference offset index is: obtaining the wellhead position coordinates and elevation z 井 , find the historical waterlogging point set within the buffer radius {Q j}, each waterlogging point is: Q j =(x j ,y j ,z j ),j=1,2,...,m; where z j is the measured water surface elevation of the historical waterlogging point; x j is the horizontal coordinate of the jth historical waterlogging point, y j is the ordinate of the jth historical waterlogging point, and the height difference offset δ from the wellhead is calculated. j , the expression is: δ j =z 井 -z j ; Calculate the historical waterlogging point height difference offset index EK, the expression is: m is the total number of waterlogging points.

[0020] Preferably, the residual height gradient of the wellhead terrain and the height difference offset index of the historical waterlogging point are converted into a comprehensive feature vector, and the comprehensive feature vector is used as the input of the machine learning model. The machine learning model uses each group of comprehensive feature vectors to predict the wellhead elevation and DEM coupling accuracy score value label as the prediction target, and takes minimizing the sum of the prediction errors of all wellhead elevation and DEM coupling accuracy score value labels as the training target. The machine learning model is trained until the sum of the prediction errors reaches convergence, and the model training is stopped. The wellhead elevation and DEM coupling accuracy score value is determined according to the model output results, wherein the machine learning model is a polynomial regression model.

[0021] Preferably, the obtained wellhead elevation and DEM coupling accuracy score value is compared with the gradient standard threshold, the gradient standard threshold includes a first standard threshold and a second standard threshold, and the first standard threshold is less than the second standard threshold, and the wellhead elevation and DEM coupling accuracy score value is compared with the first standard threshold and the second standard threshold respectively;

[0022] If the accuracy score of the coupling between the wellhead elevation and the DEM is greater than the second standard threshold, the accuracy of the coupling between the wellhead and the DEM is considered high and no adjustment is required, and it is marked as accurate;

[0023] If the accuracy score of the coupling between the wellhead elevation and the DEM is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, the accuracy of the coupling between the wellhead and the DEM is considered medium and marked as inaccurate, requiring the subsequent hydraulic simulation sensitivity analysis and proposed correction process;

[0024] If the accuracy score of the coupling between the wellhead elevation and the DEM is less than the first standard threshold, it is considered that the coupling accuracy between the wellhead and the DEM is poor, and there is a serious modeling deviation. It is marked as inaccurate and manually reviewed or directly marked as an incorrect input.

[0025] Preferably, the flow state misjudgment risk prediction module includes:

[0026] For wellheads that are judged to be inaccurate, the topological structure and boundary conditions of the local subnet are extracted; hydraulic simulations are run under the original wellhead elevation and the proposed adjusted elevation, and the elevation is corrected proportionally; the water level, flow type and overflow state under different scenarios are compared, and the error sensitivity index S is calculated. error , the expression is: W B 、W A It represents the local maximum water level or average water depth in the simulation before and after the adjustment. orig is the original wellhead elevation, H adj The adjusted wellhead elevation is used to divide the risk into three levels: high, medium, and low according to the index value, and the high-risk wellheads are recommended to enter the proposed correction process.

[0027] Preferably, according to the error sensitivity index S error The size of the corresponding elevation correction coefficient α is determined; based on the expression H adj =H orig -α·ΔH, calculate the proposed corrected elevation; use the corrected elevation as input for subsequent simulation models, and mark the wellhead as the proposed corrected state in the system.

[0028] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0029] 1. This invention integrates GIS and BIM to build a high-precision three-dimensional model, introduces a topological relationship construction module to form a complete network structure, and combines two uncommon but key elevation accuracy assessment indicators: the residual height gradient of the wellhead terrain and the elevation offset index of the historical waterlogging point. A machine learning model is constructed to intelligently score the accuracy of wellhead modeling, realizing the entire health analysis process from data-driven identification to model prediction and early warning.

[0030] 2. This invention introduces a flow pattern misjudgment risk prediction mechanism based on an error sensitivity index and automatically corrects elevation parameters for highly sensitive wellheads through a dynamic adjustment module, making the simulation more accurate to actual drainage conditions and significantly improving the accuracy, stability, and intelligence of the urban drainage system digital twin model. This invention has broad application in drainage system maintenance decision-making, risk warning, and simulation optimization, and is particularly suitable for modeling drainage networks in complex, aging, or high-risk urban areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments described in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0032] Figure 1 This is a mind map of the system modules of the present invention. DETAILED DESCRIPTION

[0033] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0034] For examples, see Figure 1 As shown, the urban drainage network intelligent health analysis system based on multimodal twin deduction described in this embodiment includes a three-dimensional modeling module, a topological relationship construction module, a characteristic parameter extraction module, a coupling accuracy assessment module, a flow state misjudgment risk prediction module, and a dynamic adjustment module;

[0035] 3D modeling module, used to build 3D structural models of drainage pipe networks based on GIS and BIM data;

[0036] A topological relationship building module, used to build a topological network in the three-dimensional structural model of the drainage pipe network for hydraulic simulation and event propagation analysis;

[0037] A feature parameter extraction module is used to extract the residual height gradient of the wellhead terrain and the height difference offset index of the historical waterlogging point based on the three-dimensional model and the DEM data;

[0038] A coupling accuracy assessment module is used to classify and assess the coupling between the wellhead elevation and the DEM based on the wellhead terrain residual height gradient and the historical waterlogging point height difference offset index, and divide the coupling into three categories: accurate, inaccurate, and incompletely accurate;

[0039] The flow pattern misjudgment risk prediction module is used to construct a local hydraulic model for inaccurate wellhead locations, simulate flow pattern changes under original and adjusted elevations, calculate the error sensitivity index, and evaluate the simulation error risk based on the error sensitivity index;

[0040] The dynamic adjustment module is used to adjust the elevation parameters of the identified sensitive wellhead position based on the error sensitivity index to form a proposed corrected elevation for subsequent simulation of the hydraulic model.

[0041] GIS systems capture the spatial distribution of urban drainage networks, including the geometric location, path direction, burial depth, and connection nodes of drainage pipes. This information typically comes from urban pipeline databases, as-built drawings, underground exploration data (such as underground radar (GPR), and historical construction records. GIS systems provide a two-dimensional or two-and-a-half-two-dimensional (2.5D) representation of geographic space, with precise geolocation capabilities. This ensures the precise location of each drainage facility on the map, serving as the geographic coordinate reference for 3D modeling.

[0042] BIM data is introduced to perform component-level 3D modeling of drainage facilities. BIM models focus on the detailed expression of physical structures and have rich component attribute definition capabilities. In the present invention, BIM is used to model various structural units in the drainage network, including but not limited to main pipes, branch pipes, inspection wells, rainwater inlets, gates, pump station outlets, etc. Each component not only has geometric information (such as pipe diameter, length, wall thickness, elbow angle, well depth, etc.), but also comes with a series of attribute information. These attributes are embedded in the model in a parameterized manner, specifically including:

[0043] Pipe type: such as ductile iron, reinforced concrete, polyethylene, polyvinyl chloride, etc. Different materials correspond to different corrosion characteristics, hydraulic roughness and life cycle.

[0044] Laying age: reflects the service time of the pipeline and is a key input for subsequent structural aging analysis and risk assessment.

[0045] Construction unit and construction batch number: used to trace construction quality and section maintenance responsibilities.

[0046] Access node information: includes the unique identification, relative height difference, flow direction, etc. of upstream and downstream connection components, which is used to build a topological relationship network.

[0047] Historical operation and maintenance data: such as recent dredging records, accident records, video inspection results, structural repair status, blockage frequency, etc., for health analysis and predictive maintenance.

[0048] Functional type: For example, sanitary sewer, storm sewer, or combined sewer, each with different flow characteristics and regulatory requirements.

[0049] Operation status identification: whether it is in use, whether it is a backup channel, whether it is in a blocked or renovated state.

[0050] The modeling process also requires consideration of the network's 3D topological logic, including node connectivity, upstream and downstream relationships within each pipe segment, flow direction definitions, and the intersection and vertical distribution of pipes within 3D space. By integrating GIS and BIM data, a highly accurate 3D drainage network model can be constructed, encompassing precise locations, structural attributes, and engineering semantics.

[0051] This 3D model not only enables visualization and spatial query but also serves as the core foundation for subsequent hydraulic simulation analysis, structural health assessment, accident simulation, maintenance scheduling, and digital twin interaction. It represents a significant leap from static 2D drawings to a dynamic, structured, data-driven 3D pipe network system, offering significant engineering value and potential for technological expansion.

[0052] The topology building module aims to reveal and quantify the connectivity, directionality, dependencies, and propagation paths between the fundamental elements of a drainage network. Its core task is to transform drainage components (such as pipes, manholes, and stormwater inlets) into nodes and edges in a network diagram, building a directed or weighted graph model within the spatial structure to achieve logical pathways for information flow, water flow, or event propagation in the simulation.

[0053] Extract components with connectivity attributes (e.g., pipes, wells, gates, etc.) from GIS and BIM models. Identify each inspection well, node well, stormwater inlet, pump station, etc. as a "node" and assign a unique ID to it as a NodeID. Each pipe segment connecting two nodes is considered an "edge," assigned a unique EdgeID, and the starting and ending points of the connection are recorded.

[0054] Based on 3D spatial coordinates, the spatial adjacency relationship between all nodes is calculated. The system determines whether connections exist between components and whether the connection sequence is reasonable (e.g., whether connections are made to illegal endpoints or whether dead zones are formed). For special structures such as elbows, diversion wells, and confluence wells, the spatial bifurcation or confluence locations are accurately calculated.

[0055] Flow direction attributes for each edge are automatically or semi-automatically generated based on the pipeline's longitudinal slope (elevation difference), designed flow direction, and historical flow data. If IoT flow monitoring points are available, real-time flow direction data is preferred for mapping. In complex confluences, a weighted function is used to determine the relationship between main and tributary streams, for example, based on maximum flow or conventional water level difference.

[0056] A complete directed graph of the drainage system is constructed, with nodes as vertices and edges as lines. Each edge records the following topological attributes: starting NodeID, ending NodeID, length, pipe diameter, material properties, hydraulic resistance coefficient (Manning coefficient), design flow direction flag, and current operating status (e.g., plugged, maintenance, or interrupted). If bridge structures exist (e.g., wells within wells, or switches between risers and laterals), these hierarchical connections are represented as virtual edges or child nodes.

[0057] A consistency check is performed on the constructed topology network to determine whether there are: isolated nodes (no connections), abnormal closed-loop paths (undesigned return flows), multiple edge overlaps (repeated mapping), and physically unreasonable connections where downstream nodes are higher than upstream nodes. If topology errors are found, they are automatically repaired through system rules or prompted for manual adjustments.

[0058] The constructed topological network not only serves structural modeling, but is also used for a variety of dynamic analysis tasks: topological relationships serve as boundary inputs for hydraulic models (such as SWMM and hydrodynamic simulation engines); support dynamic simulation of multiple flow regimes such as pressure flow, free surface flow, and tidal backflow; and calculate the continuous transmission and response delay of water level, water pressure, and water volume based on the sequence of pipe sections.

[0059] Topology maps support fault propagation path analysis, such as the impact range of a blockage from a manhole downstream; the diffusion path of a pollution source within a pipe network; and the time it takes for rainwater intrusion or waterlogging to spread from the source to the impact point. By combining weights (such as node importance, flow velocity, and response time), a propagation path map is constructed for emergency response decision-making.

[0060] The system can perform drainage path analysis, redundant path identification, and breakpoint analysis based on the topology map; it supports network robustness assessment and identifies key nodes and potential single points of failure.

[0061] The topological relationship is not isolated from the 3D model, but is fully synchronized with the 3D structure through component binding and attribute indexing mechanisms: each NodeID is bound to the manhole component in the model; each EdgeID is associated with the pipe segment component and can be highlighted in the 3D view; the topological relationship diagram can be superimposed on the 3D space as a layer in the platform to achieve image-model linkage.

[0062] The residual height gradient of the wellhead terrain indicates the rate of change of the elevation residual between the wellhead and its surrounding terrain, and reflects whether there are unreasonable high convex points or depressions at the wellhead.

[0063] The acquisition method is: Get the three-dimensional coordinates and elevation of the wellhead point: P0 = (x0, y0, z 井 ); where z 井 Indicates the elevation value of the wellhead in the three-dimensional model; x0 is the horizontal coordinate of the wellhead in the geographic space (usually the longitude direction or the X coordinate in the plane coordinate system); y0 is the vertical coordinate of the wellhead (usually the latitude direction or the Y coordinate); in the DEM, a set of terrain elevation sampling points within a certain radius R (such as 5 meters) is selected with the wellhead as the center. i}, each point is: P i =(x i ,y i ,z i ), i=1,2,...,n; n is the total number of points; calculate the horizontal distance d between each sampling point and the wellhead i , the expression is: Calculate the elevation residual ΔH for each point i , the expression is: ΔH i =z 井 -z i ;z i is the DEM elevation of the surrounding terrain sampling points; calculate the residual height gradient of the wellhead terrain, the expression is: Where SD is the residual height gradient of the wellhead terrain, which indicates the average rate of change of the elevation difference in the wellhead area with distance. The larger the gradient, the higher the degree of elevation anomaly of the wellhead in the area.

[0064] The historical waterlogging point elevation deviation index is used to measure whether the wellhead elevation is overestimated, causing the system to fail to correctly identify waterlogging risk. If the wellhead is significantly higher than the historical waterlogging point, there is an overestimation risk.

[0065] The acquisition method is: Get the wellhead position coordinates and elevation z 井 , find the historical waterlogging point set within a certain buffer radius (such as 30 meters) {Q j}, each waterlogging point is: Q j =(x j ,y j ,z j ),j=1,2,...,m; where z j is the measured water surface elevation of the historical waterlogging point; x j is the horizontal coordinate of the jth historical waterlogging point, which usually represents the east-west coordinate of the point on the map (for example, the X value in a plane coordinate system); jis the ordinate of the jth historical waterlogging point, representing the north-south coordinate of the point on the map (i.e., Y value); calculate the height difference offset δ from the wellhead j , the expression is: δ j =z 井 -z j ; Calculate the historical waterlogging point height difference offset index EK, the expression is: m is the total number of waterlogging points. The larger the elevation difference offset index value of historical waterlogging points, the higher the degree of wellhead overestimation in the model and the greater the possibility of misjudging waterlogging risk.

[0066] The residual height gradient of the wellhead terrain and the height difference offset index of the historical waterlogging point are converted into a comprehensive feature vector, which is used as the input of the machine learning model. The machine learning model uses the predicted wellhead elevation and DEM coupling accuracy score value label for each group of comprehensive feature vectors as the prediction target, and minimizes the sum of the prediction errors of all wellhead elevation and DEM coupling accuracy score value labels as the training target. The machine learning model is trained until the sum of the prediction errors reaches convergence, and the model training is stopped. The wellhead elevation and DEM coupling accuracy score value is determined according to the model output results. Among them, the machine learning model is a polynomial regression model.

[0067] The obtained wellhead elevation and DEM coupling accuracy score is compared with the gradient standard threshold. The gradient standard threshold includes a first standard threshold and a second standard threshold, and the first standard threshold is smaller than the second standard threshold. The wellhead elevation and DEM coupling accuracy score is compared with the first standard threshold and the second standard threshold respectively.

[0068] If the accuracy score of the coupling between the wellhead elevation and the DEM is greater than the second standard threshold, the accuracy of the coupling between the wellhead and the DEM is considered high and no adjustment is required, and it is marked as accurate;

[0069] If the accuracy score of the coupling between the wellhead elevation and the DEM is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, the accuracy of the coupling between the wellhead and the DEM is considered medium and marked as inaccurate, requiring the subsequent hydraulic simulation sensitivity analysis and proposed correction process;

[0070] If the accuracy score of the coupling between the wellhead elevation and the DEM is less than the first standard threshold, it is considered that the coupling accuracy between the wellhead and the DEM is poor, and there may be serious modeling deviations. It is marked as inaccurate and manual review or direct marking as an incorrect input is recommended.

[0071] The flow pattern misjudgment risk prediction module only processes wellheads that have been previously determined to be inaccurate through the scoring mechanism. These wellhead elevations may have slight deviations, but not enough to be directly classified as serious errors. Therefore, further evaluation is required to determine the extent of their impact on the flow pattern simulation results.

[0072] Extract the local subnetwork topology centered on the target wellhead (such as upstream and downstream pipe sections and nodes within a 50-meter radius); reference the hydraulic parameters of the entire network (such as flow, water depth, flow velocity, and boundary conditions); and form a local independent solution domain for efficient simulation.

[0073] Run a hydraulic simulation engine (such as SWMM, IFAS, or a custom model) based on the following two wellhead elevation settings:

[0074] Scenario A: Original wellhead elevation H orig ;

[0075] Scenario B: Adjusting the wellhead elevation H adj =H orig -α·ΔH, where α∈[0.5,1] is moderately adjusted downward according to the scoring error; ΔH is the wellhead elevation deviation, which represents the elevation difference between the wellhead elevation and the surrounding DEM terrain data;

[0076] Key flow parameters were recorded in two scenarios:

[0077] Node water level change W A 、W B ;

[0078] The water depth at the node exceeds the overflow state mark at the wellhead;

[0079] Whether the flow classification (free flow / pressure flow) conversion point has changed;

[0080] Calculate the error sensitivity index S error , which quantifies the response strength of the elevation error to changes in the simulated output:

[0081] According to the calculated S error , divide the risk level of flow state misjudgment into:

[0082] If S error >γ2: indicates extremely high sensitivity and high risk, requiring immediate correction or manual confirmation;

[0083] If γ1 error ≤γ2: indicates medium sensitivity, medium risk, and is included in the key monitoring list;

[0084] If S error ≤γ1: indicates a small impact, low risk, and retains the original elevation setting

[0085] Here, γ1 and γ2 are risk sensitivity thresholds preset by the system (eg, γ1 = 0.2, γ2 = 0.5).

[0086] The module outputs S for each incomplete wellhead error ​Value and risk level; high-risk wellheads will be marked as simulation distortion sources, automatically recommended to enter the proposed correction state, and recommended to enter the field retest; synchronously highlighted with color / icon in the digital twin system for decision makers to deal with first.

[0087] The dynamic adjustment module is used to dynamically adjust the elevation parameters of the wellhead based on the pre-set correction mechanism when there is an incomplete deviation in the coupling accuracy between the wellhead and the DEM, and when the error sensitivity analysis determines that the hydraulic simulation result is highly sensitive, to generate a proposed corrected elevation value (H aej ) and uses this value in subsequent hydraulic simulations and system feedback. The module's primary function is to implement intelligent, hierarchical, and adaptive adjustment of wellhead elevations, avoiding errors that can lead to misjudgment of flow patterns, deviations in water level calculations, and distortion of simulation results.

[0088] The dynamic adjustment module is activated only for wellhead locations that meet the following two conditions:

[0089] The coupling is determined to be incompletely accurate by the characteristic parameter extraction module;

[0090] The error sensitivity index calculated by the flow state misjudgment risk prediction module exceeds the sensitivity threshold γ2.

[0091] For each target wellhead, the dynamic adjustment module performs the following calculation steps:

[0092] Read the original elevation value H of the wellhead orig ;

[0093] Calculate or call the elevation deviation ΔH of the wellhead;

[0094] Determine the correction coefficient α: According to the error sensitivity index S error The adjustment coefficient α is adaptively determined according to the system's preset sensitivity classification rules, and its value range is: α=f(S error ),α∈[0.5,1.0]; general rule examples are as follows:

[0095] If S error >0.8, then set α = 1.0;

[0096] If 0.5 error ≤0.8, then set α=0.75;

[0097] If S error ≤0.5, then set α=0.5;

[0098] Calculate the proposed corrected elevation value H adj , the expression is: H adj =H orig -α·ΔH; the adjustment results are used as new input boundary conditions for local simulation. ​

[0099] The wellhead is set to the proposed correction state and marked synchronously in the digital twin model; metadata such as adjustment time, adjustment amplitude, original and proposed correction values ​​are recorded for subsequent tracking and manual verification.

[0100] After the adjustment is completed, the system will adj Injection is then performed into the subsequent local or global hydraulic simulation model; the simulation results will be used to determine whether the flow state judgment should be improved. If the simulation effect is significantly improved, the proposed correction value can be further submitted as a recommended correction value; manual review and intervention are supported. If the on-site measurement results are consistent with the proposed correction value, the wellhead basic model can be automatically updated; if there are a large number of concentrated highly sensitive wellheads, the system can output a "model abnormal hot zone" warning for management departments to investigate the source of data errors.

[0101] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0102] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer program are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center via a wired (e.g., infrared, wireless, microwave, etc.) method. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0103] The above is only a specific implementation method of the present application, but the scope of protection of the present application is not limited thereto. Any technician familiar with this technical field can easily think of changes or replacements within the technical scope disclosed in this application, which should be covered by the scope of protection of the present application.

Claims

1. An intelligent health analysis system for urban drainage networks based on multimodal twinning, featuring: It includes a three-dimensional modeling module, a topological relationship construction module, a characteristic parameter extraction module, a coupling accuracy evaluation module, a flow state misjudgment risk prediction module, and a dynamic adjustment module; 3D modeling module, used to build 3D structural models of drainage pipe networks based on GIS and BIM data; A topological relationship building module, used to build a topological network in the three-dimensional structural model of the drainage pipe network for hydraulic simulation and event propagation analysis; A feature parameter extraction module is used to extract the residual height gradient of the wellhead terrain and the height difference offset index of the historical waterlogging point based on the three-dimensional model and the DEM data; A coupling accuracy assessment module is used to classify and assess the coupling between the wellhead elevation and the DEM based on the wellhead terrain residual height gradient and the historical waterlogging point height difference offset index, and divide the coupling into three categories: accurate, inaccurate, and incompletely accurate; The flow pattern misjudgment risk prediction module is used to construct a local hydraulic model for inaccurate wellhead locations, simulate flow pattern changes under original and adjusted elevations, calculate the error sensitivity index, and evaluate the simulation error risk based on the error sensitivity index; The dynamic adjustment module is used to adjust the elevation parameters of the identified sensitive wellhead position based on the error sensitivity index to form a proposed corrected elevation for subsequent simulation of the hydraulic model.

2. The urban drainage network intelligent health analysis system based on multimodal twinning according to claim 1 is characterized by: The three-dimensional modeling module includes: obtaining two-dimensional or 2.5D geographic spatial data of the urban drainage network, including pipeline direction, node location and burial depth information; BIM data was introduced to construct 3D models of pipes, manholes, and gully components, with component attribute information embedded in the models, including pipe type, laying year, flow direction, and maintenance history. Spatially align GIS and BIM data and fuse them to generate a high-precision 3D drainage network model with spatial coordinates, geometric forms, and semantic attributes.

3. The urban drainage network intelligent health analysis system based on multimodal twinning according to claim 1 is characterized by: The topology relationship building module specifically includes: Extract the connected components in the 3D structural model, identify them as nodes or edges, and establish unique identifiers; Generate a directed topological graph structure based on the spatial position relationship and flow direction information, and record the starting point, end point, pipe diameter, length, material properties, flow direction and operation status of each edge; Perform consistency checks on the topology network to detect whether there are isolated nodes, closed loop anomalies, or direction conflict topology errors. If any, perform automatic repairs or prompt manual intervention.

4. The urban drainage network intelligent health analysis system based on multimodal twinning according to claim 1 is characterized by: The method for obtaining the residual height gradient of the wellhead terrain is as follows: obtain the three-dimensional coordinates and elevation of the wellhead point: P0 = (x0, y0, z 井 ); where z 井 Indicates the elevation value of the wellhead in the three-dimensional model; x0 is the horizontal coordinate of the wellhead in geographic space; y0 is the vertical coordinate of the wellhead; In DEM, with the wellhead as the center, select the terrain elevation sampling point set within the radius R {P i }, each point is: P i =(x i ,y i ,z i ),i=1,2,...,n; n is the total number of points; Calculate the horizontal distance d between each sampling point and the wellhead i , the expression is: Calculate the elevation residual ΔH for each point i , the expression is: ΔH i =z 井 -z i ;z i is the DEM elevation of the surrounding terrain sampling points; Calculate the residual height gradient of the wellhead terrain, the expression is: Where SD is the residual height gradient of the wellhead terrain.

5. The urban drainage network intelligent health analysis system based on multimodal twinning according to claim 4 is characterized by: The method for obtaining the height difference offset index of historical waterlogging points is as follows: obtain the wellhead location coordinates and elevation z 井 , find the historical waterlogging point set within the buffer radius {Q j }, each waterlogging point is: Q j =(x j ,y j ,z j ),j=1,2,...,m; where z j is the measured water surface elevation of the historical waterlogging point; x j is the horizontal coordinate of the jth historical waterlogging point, y j is the ordinate of the jth historical waterlogging point, and the height difference offset δ from the wellhead is calculated. j , the expression is: δ j =z 井 -z j ; Calculate the historical waterlogging point height difference offset index EK, the expression is: m is the total number of waterlogging points.

6. The urban drainage network intelligent health analysis system based on multimodal twinning according to claim 5 is characterized by: The residual height gradient of the wellhead terrain and the height difference offset index of the historical waterlogging point are converted into a comprehensive feature vector, which is used as the input of the machine learning model. The machine learning model uses the predicted wellhead elevation and DEM coupling accuracy score value label for each group of comprehensive feature vectors as the prediction target, and minimizes the sum of the prediction errors of all wellhead elevation and DEM coupling accuracy score value labels as the training target. The machine learning model is trained until the sum of the prediction errors reaches convergence, and the model training is stopped. The wellhead elevation and DEM coupling accuracy score value is determined according to the model output results. Among them, the machine learning model is a polynomial regression model.

7. The urban drainage network intelligent health analysis system based on multimodal twinning according to claim 6 is characterized by: The obtained wellhead elevation and DEM coupling accuracy score is compared with the gradient standard threshold. The gradient standard threshold includes a first standard threshold and a second standard threshold, and the first standard threshold is smaller than the second standard threshold. The wellhead elevation and DEM coupling accuracy score is compared with the first standard threshold and the second standard threshold respectively. If the accuracy score of the coupling between the wellhead elevation and the DEM is greater than the second standard threshold, the accuracy of the coupling between the wellhead and the DEM is considered high and no adjustment is required, and it is marked as accurate; If the accuracy score of the coupling between the wellhead elevation and the DEM is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, the accuracy of the coupling between the wellhead and the DEM is considered medium and marked as inaccurate, requiring the subsequent hydraulic simulation sensitivity analysis and proposed correction process; If the accuracy score of the coupling between the wellhead elevation and the DEM is less than the first standard threshold, it is considered that the coupling accuracy between the wellhead and the DEM is poor, and there is a serious modeling deviation. It is marked as inaccurate and manually reviewed or directly marked as an incorrect input.

8. The urban drainage network intelligent health analysis system based on multimodal twinning according to claim 7 is characterized by: The flow state misjudgment risk prediction module includes: For wellheads that are judged to be inaccurate, the topological structure and boundary conditions of the local subnet are extracted; hydraulic simulations are run under the original wellhead elevation and the proposed adjusted elevation, and the elevation is corrected proportionally; the water level, flow type and overflow state under different scenarios are compared, and the error sensitivity index S is calculated. error , the expression is: W B 、W A It represents the local maximum water level or average water depth in the simulation before and after the adjustment. orig is the original wellhead elevation, H adj The adjusted wellhead elevation is used to divide the risk into three levels: high, medium, and low according to the index value, and the high-risk wellheads are recommended to enter the proposed correction process.

9. The urban drainage network intelligent health analysis system based on multimodal twinning according to claim 8 is characterized by: According to the error sensitivity index S error The size of the corresponding elevation correction coefficient α is determined; based on the expression H adj =H orig -α·ΔH, calculate the proposed corrected elevation; The corrected elevation is used as input for subsequent simulation models, and the wellhead is marked as being in a proposed correction state in the system.

Images