A sewer network monitoring method and system

Abstract

The application provides a sewer network monitoring method and system, the method comprising: collecting multi-source monitoring data of the sewer network in real time; constructing a directed weighted graph topology model based on basic geographic information data, basic structure data and a graph theory algorithm; identifying a topology structure problem of the sewer network according to the directed weighted graph topology model and correcting the topology structure problem; diagnosing a functional defect state of the sewer network and an internal physical defect state of the sewer network; extracting a liquid level sequence in the multi-source monitoring data and predicting a water level change trend of a key node of the sewer network; mapping the corrected standard topology structure file, the functional defect state, the internal physical defect state and the water level change trend to a hydraulic model to construct a digital twin body that is real-time synchronized with the sewer network. The application effectively monitors the sewer network by fusing multi-source monitoring data, identifying a pipe network background structure problem, diagnosing a functional defect and predicting a water level change trend.

Description

Technical Field

[0001] This invention belongs to the technical field of pipeline network monitoring, and specifically relates to a method and system for monitoring drainage pipeline networks. Background Technology

[0002] As the "lifeline" of urban infrastructure, urban drainage networks play a vital role in rainwater collection, sewage transportation, and flood control. However, with the acceleration of urbanization and the frequent occurrence of extreme weather events, drainage networks face increasingly severe challenges. Statistics show that the total length of urban drainage pipelines in my country exceeds 900,000 kilometers, but a considerable portion of these pipelines suffer from varying degrees of topological problems and physical defects due to their age, low design standards, and inadequate maintenance. This leads to reduced drainage capacity, increased risk of urban flooding, and frequent overflow pollution, seriously threatening urban operational safety and the ecological environment.

[0003] Currently, the monitoring and management of drainage pipe networks mainly rely on the following technical means: Firstly, there is the defect detection technology based on manual inspection and closed-circuit television (CCTV) testing. This technology involves workers going down into the well to inspect or robots equipped with cameras entering the pipeline to acquire image data. Professionals then analyze the images to identify defects such as damage, blockages, and deformation in the pipeline. However, this method has significant shortcomings: manual inspection is inefficient, dangerous, and requires highly experienced technicians; CCTV testing equipment is expensive and complex to operate, requiring pre-treatment such as sealing and pumping water before testing, resulting in a large workload and long cycle, making it difficult to achieve real-time monitoring and rapid response to pipeline network conditions.

[0004] Secondly, there is early warning technology based on single-point liquid level monitoring. This technology deploys liquid level gauges at key nodes in the pipeline network to collect water level data in real time, triggering an alarm when the water level exceeds a preset threshold. However, this single-dimensional monitoring method cannot reveal the root cause of water level anomalies, making it difficult to distinguish whether the cause is pipeline blockage, pump station failure, or excessive rainfall intensity, and it cannot effectively diagnose topological problems and internal physical defects in the pipeline network.

[0005] Thirdly, simulation technology based on hydraulic models. This technology constructs hydraulic models of drainage pipe networks to simulate the distribution of water level and flow velocity under different rainfall scenarios, providing a reference for planning, design, and management scheduling. However, traditional hydraulic model construction relies on static pipe network geographic information system data, which cannot dynamically reflect the actual operating status and defect evolution of the pipe network; the parameter calibration of the model requires a large amount of manual intervention and is difficult to effectively integrate with real-time monitoring data, resulting in significant deviations between simulation results and actual conditions.

[0006] Fourth, prediction and scheduling technologies based on single algorithms. Existing research attempts to use machine learning algorithms for water level prediction or genetic algorithms to optimize pump station operating parameters. However, these methods are often isolated and fail to integrate topology identification, defect diagnosis, water level prediction, and optimized scheduling into a unified framework. At the same time, existing prediction models struggle to simultaneously capture the long-distance dependencies and local temporal characteristics of water level data, resulting in lower prediction accuracy. Optimization of scheduling schemes typically considers only a single flood control objective, neglecting the synergy between pollution control and ecological benefits.

[0007] In summary, existing technologies lack a drainage network monitoring system capable of integrating multi-source monitoring data, identifying underlying structural problems in the pipe network, diagnosing functional defects, and predicting water level change trends. Summary of the Invention

[0008] To address the aforementioned technical problems, this invention provides a drainage network monitoring method and system to overcome the shortcomings of the prior art.

[0009] In a first aspect, the present invention provides a method for monitoring drainage pipe networks, the method comprising: Real-time acquisition of multi-source monitoring data of the drainage pipe network, and processing of the multi-source monitoring data; Obtain the basic structural data and basic geographic information data of the drainage pipe network, and construct a directed weighted graph topology model with inspection wells as nodes and pipes as directed edges based on the basic geographic information data, the basic structural data and graph theory algorithm; The topological problems of the drainage network are identified based on the directed weighted graph topology model, and the topological problems are identified and corrected by calculating the parameters of each node to obtain a standard topology file. Based on the processed multi-source monitoring data, and by using a trained machine learning classification model and a deep learning object detection algorithm respectively, the functional defect status and internal physical defect status of the drainage network are diagnosed. The liquid level sequence is extracted from the multi-source monitoring data, and the liquid level sequence is fused with the meteorological rainfall data of the same period to obtain fused data. Based on the hybrid neural network and the fused data, the water level change trend of the key nodes of the drainage network is predicted. The revised standard topology file, the functional defect status, the internal physical defect status, and the water level change trend are mapped to the hydraulic model to construct a digital twin that is synchronized with the drainage network in real time.

[0010] Compared with the prior art, the beneficial effects of the present invention are as follows: by constructing a directed weighted graph topology model, the topological problems of the drainage pipe network are identified and corrected, thereby obtaining a standard topological structure file of the drainage pipe network. Furthermore, by diagnosing the functional defect status, internal physical defect status, standard topological structure file, and water level change trend of the drainage pipe network and mapping them to the hydraulic model, a real-time synchronized digital twin of the drainage pipe network is obtained. In this way, the drainage pipe network can be effectively monitored by integrating multi-source monitoring data, identifying the underlying structural problems of the pipe network, diagnosing functional defects, and predicting water level change trends through the digital twin.

[0011] Furthermore, the step of real-time acquisition of multi-source monitoring data of the drainage pipe network and processing of the multi-source monitoring data includes: Collect multi-source monitoring data from intelligent monitoring terminals deployed at key nodes of the drainage network. The multi-source monitoring data includes upstream liquid level data, downstream liquid level data, internal image data, manhole cover opening and closing status, methane concentration, equipment power parameters, and positioning information. The multi-source monitoring data are sequentially cleaned, normalized, and filled with missing values.

[0012] Furthermore, the step of acquiring the basic structural data and basic geographic information data of the drainage pipe network, and constructing a directed weighted graph topology model with inspection wells as nodes and pipes as directed edges based on the basic geographic information data, the basic structural data, and graph theory algorithms includes: A node attribute table is established based on the basic structural data and an edge attribute table is established based on the basic geographic information data. An adjacency matrix is ​​constructed based on the node attribute table and the edge attribute table to represent the directed connection relationship between each node. A directed weighted graph topology model is constructed based on adjacency matrices and graph theory algorithms, with inspection wells as nodes and pipes as directed edges. The construction expression of the directed weighted graph topology model is as follows: ; ; ; ; In the formula, This represents a directed weighted graph topology model. Represents a set of nodes. Represents a set of directed edges. Represents the flow direction state function, Represents the adjacency matrix, Indicates a directed pipe, , Representing nodes respectively ,node The inner bottom elevation of the pipe, Represents the set of nodes The total number of nodes in the middle, In the adjacency matrix, the first... Line 1 The elements of the column.

[0013] Furthermore, the step of identifying the topological structure problem of the drainage network based on the directed weighted graph topology model, and identifying and correcting the topological structure problem by calculating the parameters of each node, includes: Based on the directed weighted graph topology model, the in-degree, out-degree, and betweenness of each node are calculated. Based on the in-degree, out-degree, and betweenness, nodes with broken connections and abnormal node pairs are identified. The expressions for calculating the in-degree, out-degree, and betweenness are as follows: ; ; ; In the formula, Represents a node in-degree, Represents a set of nodes The total number of nodes in the middle, In the adjacency matrix, the first... Line 1 Column elements, Represents a node The degree of exit, Represents a node betweenness, , These represent two different nodes. Indicates from node To the node All shortest paths passing through nodes The number of paths, From node To the node The total number of all shortest paths; Based on geometric constraints, the connection interruption nodes are automatically extended, and the abnormal node pairs are transposed to obtain a corrected topology. The corrected topology is then updated to the directed weighted graph topology model. A connectivity analysis model is constructed based on the depth-first traversal algorithm. The connectivity of all non-terminal nodes is analyzed, and a set of nodes that are not connected to the terminal node is selected. Each node in the set of unconnected nodes is replaced by the shortest path connectivity relationship between adjacent nodes.

[0014] Furthermore, the steps of diagnosing the functional defect status and internal physical defect status of the drainage network based on the processed multi-source monitoring data and using a trained machine learning classification model and a deep learning object detection algorithm, respectively, include: The upstream and downstream liquid level data of the processed multi-source monitoring data are extracted as input feature vectors, and the input feature vectors are input into the trained K-nearest neighbor algorithm classification model; Based on the trained K-nearest neighbor algorithm classification model, the feature distance between the input feature vector and the training samples is calculated to diagnose the functional defect status of the drainage pipe network. The expression for calculating the feature distance between the input feature vector and the training samples is as follows: ; In the formula, This represents the feature distance between the input feature vector and the training samples. This represents the input feature vector for real-time monitoring. This represents the feature vector in the training samples. The dimension of the feature vector. Index representing the feature dimension , These represent the input feature vector at the th... The values ​​in each dimension, and the feature vector of the training sample in the th dimension. Values ​​in each dimension Indicates the distance parameter; Extract the internal image data from the processed multi-source monitoring data, and identify the internal physical defects of the drainage network based on the improved YOLO-v4 deep learning object detection algorithm model and the internal image data; The diagnostic results of the functional defect state and the identification results of the internal physical defect state are spatiotemporally correlated and fused to generate a comprehensive defect diagnosis report.

[0015] Furthermore, the step of mapping the modified standard topology file, the functional defect status, the internal physical defect status, and the water level change trend to the hydraulic model to construct a digital twin synchronized with the drainage network in real time includes: The corrected standard topology file is parsed into recognizable node objects and pipe objects. Based on the node objects and pipe objects, the functional defect state is mapped to the initial water level boundary conditions of the pipe, and the internal physical defect state is mapped to the corresponding hydraulic parameter correction values. The multi-source monitoring data is used as a dynamic boundary condition, and the dynamic boundary condition is input into the hydraulic model at a preset time frequency to simulate the hydraulic distribution state of the drainage network and obtain simulation results. The simulation results are compared with the multi-source monitoring data in real time to obtain the comparison results. If the comparison results exceed the preset threshold, the key parameters of the hydraulic model are automatically calibrated, and the automatically calibrated hydraulic model is fused and rendered with the basic geographic information data to generate an interactive three-dimensional visual digital twin.

[0016] Furthermore, following the step of constructing a digital twin synchronized in real time with the drainage network, the method further includes: Using the digital twin as the computing engine, a non-dominated sorting genetic algorithm with an elite strategy as the optimization engine, and multiple objective functions such as minimizing the total amount of waterlogging, minimizing the total amount of combined sewer overflow pollution, and maximizing the overall scheduling benefits, the decision variables for adjusting the operating parameters of the drainage network are iteratively optimized to obtain the optimal collaborative scheduling scheme.

[0017] Secondly, the present invention also provides a drainage pipe network monitoring system, the system comprising: The data acquisition and processing module is used to acquire multi-source monitoring data of the drainage pipe network in real time and process the multi-source monitoring data. The module is used to acquire the basic structural data and basic geographic information data of the drainage network, and to construct a directed weighted graph topology model with inspection wells as nodes and pipes as directed edges based on the basic geographic information data, the basic structural data and graph theory algorithms. The identification and calculation module is used to identify the topological problems of the drainage network based on the directed weighted graph topology model, and to identify and correct the topological problems by calculating the parameters of each node to obtain a standard topology file. The diagnostic module is used to diagnose the functional defect status and internal physical defect status of the drainage network based on the processed multi-source monitoring data and the trained machine learning classification model and deep learning object detection algorithm, respectively. The extraction and fusion module is used to extract the liquid level sequence from the multi-source monitoring data, fuse the liquid level sequence with the meteorological rainfall data of the same period to obtain fused data, and predict the water level change trend of key nodes of the drainage network based on the hybrid neural network and the fused data. The mapping construction module is used to map the corrected standard topology file, the functional defect status, the internal physical defect status, and the water level change trend to the hydraulic model, so as to construct a digital twin that is synchronized with the drainage network in real time.

[0018] Thirdly, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described drainage network monitoring method.

[0019] Fourthly, the present invention also provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described drainage network monitoring method. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a flowchart of the drainage network monitoring method in the first embodiment of the present invention; Figure 2 This is a structural block diagram of the drainage network monitoring system in the second embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of the electronic device in the third embodiment of the present invention.

[0022] Explanation of key component symbols: 10. Acquisition and Processing Module; 20. Acquisition and Construction Module; 30. Recognition and Calculation Module; 40. Diagnosis Module; 50. Extraction and Fusion Module; 60. Mapping and Construction Module; 70. Bus; 71. Processor; 72. Memory; 73. Communication interface.

[0023] The embodiments of the present invention will be further described below with reference to the accompanying drawings. Detailed Implementation

[0024] To facilitate understanding of the present invention, a more complete description will be given below with reference to the accompanying drawings. Several embodiments of the invention are illustrated in the drawings. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete.

[0025] It should be noted that when a component is said to be "fixed to" another component, it can be directly on the other component or there may be an intervening component. When a component is said to be "connected to" another component, it can be directly connected to the other component or there may be an intervening component. The terms "vertical," "horizontal," "left," "right," and similar expressions used in this document are for illustrative purposes only.

[0026] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0027] Example 1 Please see Figure 1 The diagram illustrates a drainage network monitoring method according to a first embodiment of the present invention, the method comprising steps S1 to S6: S1, collect multi-source monitoring data of the drainage pipe network in real time, and process the multi-source monitoring data; Specifically, step S1 includes steps S11 to S12: S11, Collect multi-source monitoring data from intelligent monitoring terminals deployed at key nodes of the drainage network. The multi-source monitoring data includes upstream liquid level data, downstream liquid level data, internal image data, manhole cover opening and closing status, methane concentration, equipment power parameters, and positioning information. S12, the multi-source monitoring data are cleaned, normalized and filled with missing values ​​in sequence; It should be noted that modular intelligent monitoring terminals are deployed at key nodes of the drainage network, and periscopes are installed within the network to detect image data. Multi-source monitoring data is collected in real time through these terminals. The terminals integrate a main control board, a power management board, and a sensor expansion board. The power management board achieves a microampere-level low-power mode through independent power supply control via external chips. The sensor expansion board integrates a Hall effect magnetic switch, an electrode-type water level detection circuit, a submersible hydrostatic level gauge, a methane sensor, and a BeiDou positioning module. The real-time collected data includes at least upstream and downstream liquid level data, internal image data, manhole cover opening / closing status, methane concentration, equipment voltage and current, and positioning information. The collected data is then sequentially cleaned, normalized, and filled with missing values ​​to improve data quality.

[0028] S2, acquire the basic structural data and basic geographic information data of the drainage pipe network, and construct a directed weighted graph topology model with inspection wells as nodes and pipes as directed edges based on the basic geographic information data, the basic structural data and graph theory algorithm; Specifically, step S2 includes steps S21 to S22: S21. Establish a node attribute table based on the basic structure data and an edge attribute table based on the basic geographic information data, and construct an adjacency matrix based on the node attribute table and the edge attribute table to represent the directed connection relationship between each node. It should be noted that, based on the aforementioned infrastructure data, a node attribute table and an edge attribute table are established. The node attribute table includes at least the node name, node type, in-degree, out-degree, inner bottom elevation, planar coordinates, and connectivity identifier. The edge attribute table includes at least the edge name, edge type, pressure state, pipe geometry, start node identifier, and end node identifier. An adjacency matrix is ​​constructed based on the node attribute table and edge attribute table to represent the directed connections between nodes. In addition, the data in the node attribute table and edge attribute table undergoes availability verification, including non-emptiness verification, uniqueness verification, and theoretical constraint verification of parameter values. Data that does not conform to the verification rules is marked with corresponding error codes or warning codes. S22, Based on the adjacency matrix and graph theory algorithm, a directed weighted graph topology model is constructed with inspection wells as nodes and pipelines as directed edges; It should be noted that a directed weighted graph topology model is constructed using graph theory computation and adjacency matrices. In this embodiment, the construction expression of the directed weighted graph topology model is: ; ; ; ; In the formula, This represents a directed weighted graph topology model. Represents a set of nodes. Represents a set of directed edges. Represents the flow direction state function, Represents the adjacency matrix, Indicates a directed pipe, , Representing nodes respectively ,node The inner bottom elevation of the pipe, Represents the set of nodes The total number of nodes in the middle, In the adjacency matrix, the first... Line 1 The elements of the column.

[0029] S3. Identify the topological problems of the drainage network based on the directed weighted graph topology model, and identify and correct the topological problems by calculating the parameters of each node to obtain a standard topology file; Specifically, step S3 includes steps S31 to S33: S31, calculate the in-degree, out-degree, and betweenness of each node based on the directed weighted graph topology model, and identify disconnected nodes and abnormal node pairs based on the in-degree, out-degree, and betweenness. It should be noted that, based on the directed weighted graph topology model, the in-degree, out-degree, and betweenness of each node are calculated; nodes that satisfy the condition of in-degree greater than 0 and out-degree equal to 0 and whose node type does not belong to discharge outlet or sewage treatment plant are identified as nodes with disconnected connections; node pairs that simultaneously satisfy the condition of in-degree equal to 0 and out-degree greater than 1, and in-degree greater than 1 and out-degree equal to 0 are identified as nodes with abnormal flow direction; for confluence nodes with in-degree greater than 1 and out-degree equal to 1, their flow area ratio is calculated, which is the ratio of the sum of the cross-sectional areas of the inflow pipes to the cross-sectional area of ​​the outflow pipe. When the flow area ratio is greater than a preset threshold, it is determined that there is an abnormal pipe diameter in the downstream pipe of the node; nodes whose upstream pipe types simultaneously include rainwater pipes and sewage pipes are identified as nodes with mixed rainwater and sewage connections. In this embodiment, the calculation expressions for the in-degree, the out-degree, and the betweenness are as follows: ; ; ; In the formula, Represents a node in-degree, Represents a set of nodes The total number of nodes in the middle, In the adjacency matrix, the first... Line 1 Column elements, Represents a node The degree of exit, Represents a node betweenness, , These represent two different nodes. Indicates from node To the node All shortest paths passing through nodes The number of paths, From node To the node The total number of all shortest paths; S32, based on geometric constraints, the connection interruption nodes are automatically extended, and the abnormal node pairs are transposed to obtain the corrected topology, and the corrected topology is updated to the directed weighted graph topology model. It should be noted that for identified nodes with interrupted connections, automatic extension is performed based on geometric constraints. These constraints include that the pipe diameter of the downstream candidate node is not less than the pipe diameter of the current node, the flow direction angle is less than 90 degrees, and the extension distance is minimized. For identified pairs of nodes with abnormal flow directions, the pipe flow direction between the node pairs is transposed and corrected. The corrected topology is then updated in the node attribute table and adjacency matrix of the directed weighted graph topology model. S33. Based on the depth-first traversal algorithm, a connected path analysis model is constructed. Connectivity analysis is performed on all non-terminal nodes, and a set of nodes that are not connected to the terminal node is selected. Each node in the set of nodes that are not connected is replaced by the shortest path connectivity relationship between adjacent nodes. It should be noted that a connectivity analysis model is constructed based on the depth-first traversal algorithm. Connectivity analysis is performed on all non-terminal nodes to filter out the set of nodes that are not connected to the terminal node. For each node in the set of non-connected nodes, all related nodes flowing into the node are found. The connectivity relationship of the node is replaced with the shortest path connectivity relationship of the adjacent nodes and the flow direction is modified until all non-terminal nodes have at least one path connected to the terminal node.

[0030] S4. Based on the processed multi-source monitoring data, and by using a trained machine learning classification model and a deep learning target detection algorithm respectively, diagnose the functional defect status of the drainage network and the internal physical defect status of the drainage network. Specifically, step S4 includes steps S41 to S44: S41, extract the upstream liquid level data and downstream liquid level data of the processed multi-source monitoring data as input feature vectors, and input the input feature vectors into the trained K-nearest neighbor algorithm classification model; S42, calculate the feature distance between the input feature vector and the training sample based on the trained K-nearest neighbor algorithm classification model, so as to diagnose and output the functional defect status of the drainage network; It should be noted that upstream and downstream liquid level values ​​are extracted from multi-source monitoring data as input feature vectors, which are then input into a pre-trained K-nearest neighbor (KNN) classification model. The KNN classification model uses upstream and downstream water level feature vectors at different blockage levels and locations obtained from hydraulic model experiments of physical pipelines as training samples, and optimizes and determines the nearest neighbor number k and distance metric p value through a grid cross-search method. The KNN classification model calculates the feature distance between the real-time input feature vector and the training samples, and outputs the equivalent blockage area ratio, blockage location, corresponding defect level, and treatment suggestions for the current pipeline. The expression for calculating the feature distance between the input feature vector and the training samples is as follows: ; In the formula, This represents the feature distance between the input feature vector and the training samples. This represents the input feature vector for real-time monitoring. This represents the feature vector in the training samples. The dimension of the feature vector. Index representing the feature dimension , These represent the input feature vector at the th... The values ​​in each dimension, and the feature vector of the training sample in the th dimension. Values ​​in each dimension This represents the distance parameter.

[0031] S43, extract the internal image data of the processed multi-source monitoring data, and identify the internal physical defects of the drainage network based on the improved YOLO-v4 deep learning target detection algorithm model and the internal image data; It should be noted that the process involves acquiring closed-circuit television or periscope inspection image data of the drainage pipeline, and then inputting the image data into an improved YOLO-v4 deep learning object detection algorithm model. The improved YOLO-v4 algorithm model introduces the SENet channel attention mechanism into the backbone network to enhance feature extraction capabilities, employs a bidirectional feature pyramid network in the neck network for multi-scale feature fusion, and optimizes the loss function to SIoU to improve positioning accuracy. The improved YOLO-v4 algorithm model performs frame-by-frame analysis of the image data and outputs the physical defect types, location coordinates, and identification confidence scores of internal pipe damage, interface misalignment, branch pipe unauthorized connections, foreign object occupation, and pipe deposition. S44, the diagnostic results of the functional defect state and the identification results of the internal physical defect state are spatiotemporally correlated and fused to generate a comprehensive defect diagnosis report.

[0032] It should be noted that the diagnostic results of functional defects are spatiotemporally correlated and fused with the identification results of internal physical defects to generate a comprehensive defect diagnosis report that includes defect type, severity, location coordinates, and confidence level, providing data support for subsequent operation and maintenance scheduling plans.

[0033] S5, extract the liquid level sequence from the multi-source monitoring data, fuse the liquid level sequence with the meteorological rainfall data of the same period to obtain fused data, and predict the water level change trend of the key nodes of the drainage network based on the hybrid neural network and the fused data; It should be noted that by aligning and concatenating the liquid level sequence with concurrent meteorological rainfall data along the time dimension, a multi-dimensional feature input vector is constructed. The fusion method can be represented as follows: , To fuse feature vectors, for The liquid level monitoring value at any given time. for The rainfall at any given moment; for inputs that include historical time windows, this forms a two-dimensional matrix. ,in The time window length, This is the fused feature vector at the first moment within the time window, i.e., the fused feature vector at the earliest moment. The feature vector is fused at the second time point within the time window. This represents the matrix transpose. This fusion method enables the hybrid neural network to simultaneously learn the coupling effect between the temporal patterns of water levels and rainfall driving factors, thereby more accurately capturing sudden changes and nonlinear variations in water levels caused by rainfall events. This significantly improves the prediction accuracy of peak water levels and trends at key nodes, providing more reliable boundary conditions for subsequent optimized scheduling.

[0034] S6, map the revised standard topology file, the functional defect status, the internal physical defect status, and the water level change trend to the hydraulic model to construct a digital twin that is synchronized with the drainage network in real time; Specifically, step S6 includes steps S61 to S64: S61, the modified standard topology file is parsed into identifiable node objects and pipe objects, and based on the node objects and pipe objects, the functional defect state is mapped to the initial water level boundary conditions of the pipe, and the internal physical defect state is mapped to the corresponding hydraulic parameter correction values. It should be noted that the corrected standard topology file is parsed into nodes and pipe objects recognizable by the hydraulic model. Node objects include at least manholes, discharge outlets, sewage treatment plants, and pumping stations, while pipe objects include at least rainwater pipes, sewage pipes, and combined sewer pipes. The diagnostic results of internal physical defects are mapped to corresponding hydraulic parameter correction values ​​according to defect type and confidence level. The hydraulic parameter correction values ​​include at least the increment of pipe roughness coefficient, leakage loss flow, and reduction in pipe cross-sectional area. The diagnostic results of functional defects are mapped to the initial water level boundary conditions of the pipe. S62, the multi-source monitoring data is used as a dynamic boundary condition, and the dynamic boundary condition is input into the hydraulic model at a preset time frequency to simulate the hydraulic distribution state of the drainage network and obtain the simulation results; It should be noted that the real-time liquid level, flow rate, and manhole cover status from multi-source monitoring data are used as dynamic boundary conditions and input into the hydraulic model at a preset time frequency; the water level change trend is used as a predictive boundary condition and set within the future time window of the hydraulic model; the hydraulic model is based on the Saint-Venant equations and Manning formulas to perform hydrodynamic solutions on the mapped nodes and pipe objects, simulating the distribution of water level, flow velocity, and fullness of the drainage network at the current and future times; S63, compare the simulation results with the multi-source monitoring data in real time to obtain the comparison results. If the comparison results exceed the preset threshold, automatically calibrate the key parameters of the hydraulic model, and fuse and render the automatically calibrated hydraulic model with the basic geographic information data to generate an interactive three-dimensional visual digital twin. It should be noted that the simulation results of the hydraulic model are compared with multi-source monitoring data in real time. When the deviation between the two exceeds a preset threshold, the parameter calibration module is triggered to automatically calibrate the key parameters of the hydraulic model. The calibrated hydraulic model is then fused and rendered with geographic information system data to generate an interactive 3D visualized digital twin. The digital twin includes at least the topology of the drainage network, real-time operating status, defect distribution information, and future change trends.

[0035] S64, using the digital twin as the computing engine, the non-dominated sorting genetic algorithm with elite strategy as the optimization engine, and the multi-objective function of minimizing the total amount of waterlogging, minimizing the total amount of combined sewer overflow pollution, and maximizing the comprehensive scheduling benefits, the decision variables for adjusting the operating parameters of the drainage network are iteratively optimized to obtain the optimal collaborative scheduling scheme. It should be noted that, using a digital twin as the computing engine and a non-dominated sorting genetic algorithm with an elite strategy as the optimization engine, and with the multiple objective functions of minimizing the total amount of waterlogging, minimizing the total amount of pollution from combined sewer overflows, and maximizing the comprehensive scheduling benefits, the decision variables of pump station start-up and shutdown liquid levels, gate opening and closing conditions, water inlet and outlet rules of regulating reservoirs, and sewage treatment plant operating parameters are iteratively optimized. From the Pareto optimal solution set, the optimal "plant-network-river-green" collaborative scheduling scheme is selected according to the weight preference of the current rainfall scenario, and the scheme is output to the monitoring platform for execution. It is worth noting that, in this embodiment, the expressions for the continuity equation, momentum equation, and friction slope of the hydraulic model are as follows: ; ; ; In the formula, Indicates the sign of the partial derivative. This indicates the cross-sectional area of ​​the pipe through which water flows. Indicates time, Indicates the instantaneous flow rate of the pipeline. Represents the ordinate along the direction of the pipeline. Represents gravitational acceleration. Indicates the water head in the pipe. Indicates the friction slope. This represents the Manning roughness coefficient. Indicates the hydraulic radius.

[0036] In summary, the drainage network monitoring method in the above embodiments of the present invention identifies and corrects topological problems in the drainage network by constructing a directed weighted graph topology model, thereby obtaining a standard topology file for the drainage network. Furthermore, by diagnosing the functional defects, internal physical defects, standard topology file, and water level change trends of the drainage network and mapping them to a hydraulic model, a real-time synchronized digital twin of the drainage network is obtained. This digital twin can then effectively monitor the drainage network by fusing multi-source monitoring data, identifying underlying structural problems, diagnosing functional defects, and predicting water level change trends.

[0037] Example 2 Please see Figure 2 The image shows a drainage network monitoring system according to a second embodiment of the present invention. The system includes: The data acquisition and processing module 10 is used to acquire multi-source monitoring data of the drainage pipe network in real time and process the multi-source monitoring data. The acquisition module 20 is used to acquire the basic structural data and basic geographic information data of the drainage pipe network, and to construct a directed weighted graph topology model with inspection wells as nodes and pipes as directed edges based on the basic geographic information data, the basic structural data and graph theory algorithm. The identification and calculation module 30 is used to identify the topological structure problems of the drainage network according to the directed weighted graph topology model, and to identify and correct the topological structure problems by calculating the parameters of each node to obtain a standard topology file. The diagnostic module 40 is used to diagnose the functional defect status of the drainage network and the internal physical defect status of the drainage network based on the processed multi-source monitoring data and the trained machine learning classification model and deep learning target detection algorithm, respectively. The extraction and fusion module 50 is used to extract the liquid level sequence from the multi-source monitoring data, fuse the liquid level sequence with the meteorological rainfall data of the same period to obtain fused data, and predict the water level change trend of key nodes of the drainage network based on the hybrid neural network and the fused data. The mapping construction module 60 is used to map the corrected standard topology file, the functional defect status, the internal physical defect status, and the water level change trend to the hydraulic model, so as to construct a digital twin that is synchronized with the drainage network in real time.

[0038] In some optional embodiments, the acquisition and processing module 10 includes: The data acquisition unit is used to collect multi-source monitoring data from intelligent monitoring terminals deployed at key nodes of the drainage network. The multi-source monitoring data includes upstream liquid level data, downstream liquid level data, internal image data, manhole cover opening and closing status, methane concentration, equipment power parameters, and positioning information. The processing unit is used to sequentially clean, normalize, and fill in missing values ​​for the multi-source monitoring data.

[0039] In some alternative embodiments, the acquisition building module 20 includes: The establishment unit is used to establish a node attribute table based on the basic structure data and an edge attribute table based on the basic geographic information data, and to construct an adjacency matrix based on the node attribute table and the edge attribute table to represent the directed connection relationship between each node. The construction unit is used to construct a directed weighted graph topology model with inspection wells as nodes and pipes as directed edges based on the adjacency matrix and graph theory algorithms. The construction expression for the directed weighted graph topology model is as follows: ; ; ; ; In the formula, This represents a directed weighted graph topology model. Represents a set of nodes. Represents a set of directed edges. Represents the flow direction state function, Represents the adjacency matrix, Indicates a directed pipe, , Representing nodes respectively ,node The inner bottom elevation of the pipe, Represents the set of nodes The total number of nodes in the middle, In the adjacency matrix, the first... Line 1 The elements of the column.

[0040] In some alternative embodiments, the identification calculation module 30 includes: The calculation unit is used to calculate the in-degree, out-degree, and betweenness of each node based on the directed weighted graph topology model, and to identify nodes with broken connections and abnormal node pairs based on the in-degree, out-degree, and betweenness. The calculation expressions for the in-degree, out-degree, and betweenness are as follows: ; ; ; In the formula, Represents a node in-degree, Represents a set of nodes The total number of nodes in the middle, In the adjacency matrix, the first... Line 1 Column elements, Represents a node The degree of exit, Represents a node betweenness, , These represent two different nodes. Indicates from node To the node All shortest paths passing through nodes The number of paths, From node To the node The total number of all shortest paths; The correction and update unit is used to automatically extend the connection of the interrupted nodes based on geometric constraints, and to perform transpose correction between the abnormal node pairs to obtain the corrected topology, and to update the corrected topology to the directed weighted graph topology model. The analysis unit is used to construct a connectivity analysis model based on the depth-first traversal algorithm, perform connectivity analysis on all non-terminal nodes, filter out the set of nodes that are not connected to the terminal node, and replace each node in the set of unconnected nodes with the shortest path connectivity relationship between adjacent nodes.

[0041] In some alternative embodiments, the diagnostic module 40 includes: An input extraction unit is used to extract the upstream and downstream liquid level data of the processed multi-source monitoring data as input feature vectors, and input the input feature vectors into the trained K-nearest neighbor algorithm classification model; The diagnostic unit is used to calculate the feature distance between the input feature vector and the training samples based on the trained K-nearest neighbor algorithm classification model, so as to diagnose the functional defect status of the drainage pipe network. The expression for calculating the feature distance between the input feature vector and the training samples is as follows: ; In the formula, This represents the feature distance between the input feature vector and the training samples. This represents the input feature vector for real-time monitoring. This represents the feature vector in the training samples. The dimension of the feature vector. Index representing the feature dimension , These represent the input feature vector at the th... The values ​​in each dimension, and the feature vector of the training sample in the th dimension. Values ​​in each dimension Indicates the distance parameter; An extraction and identification unit is used to extract internal image data from the processed multi-source monitoring data, and to identify internal physical defects of the drainage network based on the improved YOLO-v4 deep learning target detection algorithm model and the internal image data. The generation unit is used to perform spatiotemporal correlation fusion of the diagnostic results of the functional defect state and the identification results of the internal physical defect state to generate a comprehensive defect diagnosis report.

[0042] In some alternative embodiments, the mapping construction module 60 includes: The parsing unit is used to parse the modified standard topology file into recognizable node objects and pipe objects, and based on the node objects and pipe objects, map the functional defect state to the initial water level boundary conditions of the pipe, and map the internal physical defect state to the corresponding hydraulic parameter correction values. The input unit is used to take the multi-source monitoring data as dynamic boundary conditions and input the dynamic boundary conditions into the hydraulic model at a preset time frequency to simulate the hydraulic distribution state of the drainage network and obtain simulation results. The comparison unit is used to compare the simulation results with the multi-source monitoring data in real time to obtain the comparison results. If the comparison results exceed the preset threshold, the key parameters of the hydraulic model are automatically calibrated, and the automatically calibrated hydraulic model is fused and rendered with the basic geographic information data to generate an interactive three-dimensional visual digital twin. The optimization unit is used to iteratively optimize the decision variables of the operation parameters of the drainage network by using the digital twin as the computing engine, the non-dominated sorting genetic algorithm with elitist strategy as the optimization engine, and the multi-objective functions of minimizing the total amount of waterlogging, minimizing the total amount of combined sewer overflow pollution, and maximizing the overall scheduling benefits, so as to obtain the optimal coordinated scheduling scheme.

[0043] The functions or operation steps implemented by the above modules and units are largely the same as those in the above method embodiments, and will not be repeated here.

[0044] The drainage network monitoring system provided in this embodiment of the invention has the same implementation principle and technical effects as the aforementioned method embodiment. For the sake of brevity, any parts not mentioned in the system embodiment can be referred to the corresponding content in the aforementioned method embodiment.

[0045] Example 3 The third embodiment of the present invention also proposes an electronic device, please refer to [link / reference]. Figure 3 The image shows an electronic device according to a third embodiment of the present invention.

[0046] The electronic device may include a processor 71 and a memory 72 storing computer program instructions.

[0047] Specifically, the processor 71 may include a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement this application.

[0048] The memory 72 may include a mass storage device for data or instructions. For example, and not limitingly, the memory 72 may include a hard disk drive (HDD), a floppy disk drive, a solid-state drive (SSD), flash memory, an optical disk drive, a magneto-optical disk drive, magnetic tape, or a Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, the memory 72 may include removable or non-removable (or fixed) media. Where appropriate, the memory 72 may be internal or external to a data processing device. In a particular embodiment, the memory 72 is non-volatile memory. In a particular embodiment, the memory 72 includes read-only memory (ROM) and random access memory (RAM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), an electrically alterable read-only memory (EAROM), or flash memory, or a combination of two or more of these. Where appropriate, the RAM can be Static Random-Access Memory (SRAM) or Dynamic Random-Access Memory (DRAM). DRAM can be Fast Page Mode Dynamic Random-Access Memory (FPMDRAM), Extended Data Out Dynamic Random-Access Memory (EDODRAM), Synchronous Dynamic Random-Access Memory (SDRAM), etc.

[0049] The memory 72 can be used to store or cache various data files that need to be processed and / or communicated, as well as possible computer program instructions executed by the processor 71.

[0050] The processor 71 reads and executes the computer program instructions stored in the memory 72 to implement the drainage network monitoring method of the above embodiment 1.

[0051] In some embodiments, the electronic device may further include a communication interface 73 and a bus 70. For example, Figure 3 As shown, the processor 71, memory 72, and communication interface 73 are connected through bus 70 and complete communication with each other.

[0052] The communication interface 73 is used to enable communication between the various modules, devices, units, and / or equipment in this application. The communication interface 73 can also enable data communication with other components such as external devices, image / data acquisition devices, databases, external storage, and image / data processing workstations.

[0053] Bus 70 includes hardware, software, or both, that couples components of a device together. Bus 70 includes, but is not limited to, at least one of the following: data bus, address bus, control bus, expansion bus, and local bus. For example, and not as a limitation, bus 70 may include an Accelerated Graphics Port (AGP) or other graphics bus, an Extended Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a Hyper Transport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an InfiniBand interconnect, a Low Pin Count (LPC) bus, a memory bus, a Micro Channel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local Bus (VLB) bus, or other suitable buses, or a combination of two or more of these. Where appropriate, bus 70 may include one or more buses. Although this application describes and illustrates a specific bus, this application considers any suitable bus or interconnection.

[0054] The electronic device can acquire the drainage network monitoring system and execute the drainage network monitoring method of this embodiment.

[0055] In addition, in conjunction with the drainage network monitoring method in Embodiment 1 above, this application can provide a storage medium for implementation. This storage medium stores computer program instructions; when these computer program instructions are executed by a processor, they implement the drainage network monitoring method of Embodiment 1 above.

[0056] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0057] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this invention patent should be determined by the appended claims.

Claims

1. A method for monitoring drainage pipe networks, characterized in that, The method includes: Real-time acquisition of multi-source monitoring data of the drainage pipe network, and processing of the multi-source monitoring data; Obtain the basic structural data and basic geographic information data of the drainage pipe network, and construct a directed weighted graph topology model with inspection wells as nodes and pipes as directed edges based on the basic geographic information data, the basic structural data and graph theory algorithm; The topological problems of the drainage network are identified based on the directed weighted graph topology model, and the topological problems are identified and corrected by calculating the parameters of each node to obtain a standard topology file. Based on the processed multi-source monitoring data, and by using a trained machine learning classification model and a deep learning object detection algorithm respectively, the functional defect status and internal physical defect status of the drainage network are diagnosed. The liquid level sequence is extracted from the multi-source monitoring data, and the liquid level sequence is fused with the meteorological rainfall data of the same period to obtain fused data. Based on the hybrid neural network and the fused data, the water level change trend of the key nodes of the drainage network is predicted. The revised standard topology file, the functional defect status, the internal physical defect status, and the water level change trend are mapped to the hydraulic model to construct a digital twin that is synchronized with the drainage network in real time.

2. The drainage pipe network monitoring method according to claim 1, characterized in that, The steps of real-time acquisition of multi-source monitoring data of the drainage pipe network and processing of the multi-source monitoring data include: Collect multi-source monitoring data from intelligent monitoring terminals deployed at key nodes of the drainage network. The multi-source monitoring data includes upstream liquid level data, downstream liquid level data, internal image data, manhole cover opening and closing status, methane concentration, equipment power parameters, and positioning information. The multi-source monitoring data are sequentially cleaned, normalized, and filled with missing values.

3. The drainage pipe network monitoring method according to claim 1, characterized in that, The steps of acquiring the basic structural data and basic geographic information data of the drainage pipe network, and constructing a directed weighted graph topology model with inspection wells as nodes and pipes as directed edges based on the basic geographic information data, the basic structural data, and graph theory algorithms include: A node attribute table is established based on the basic structural data and an edge attribute table is established based on the basic geographic information data. An adjacency matrix is ​​constructed based on the node attribute table and the edge attribute table to represent the directed connection relationship between each node. A directed weighted graph topology model is constructed based on adjacency matrices and graph theory algorithms, with inspection wells as nodes and pipes as directed edges. The construction expression of the directed weighted graph topology model is as follows: ； ； ； ； In the formula, This represents a directed weighted graph topology model. Represents a set of nodes. Represents a set of directed edges. Represents the flow direction state function, Represents the adjacency matrix, Indicates a directed pipe, , Representing nodes respectively ,node The inner bottom elevation of the pipe, Represents the set of nodes The total number of nodes in the middle, In the adjacency matrix, the first... Line number The elements of the column.

4. The drainage pipe network monitoring method according to claim 1, characterized in that, The steps of identifying the topological structure problem of the drainage network based on the directed weighted graph topology model, and identifying and correcting the topological structure problem by calculating the parameters of each node, include: Based on the directed weighted graph topology model, the in-degree, out-degree, and betweenness of each node are calculated. Based on the in-degree, out-degree, and betweenness, nodes with broken connections and abnormal node pairs are identified. The expressions for calculating the in-degree, out-degree, and betweenness are as follows: ； ； ； In the formula, Represents a node in-degree, Represents a set of nodes The total number of nodes in the middle, In the adjacency matrix, the first... Line number Column elements, Represents a node The degree of exit, Represents a node betweenness, , These represent two different nodes. Indicates from node To the node All shortest paths passing through nodes The number of paths, From node To the node The total number of all shortest paths; Based on geometric constraints, the connection interruption nodes are automatically extended, and the abnormal node pairs are transposed to obtain a corrected topology. The corrected topology is then updated to the directed weighted graph topology model. A connectivity analysis model is constructed based on the depth-first traversal algorithm. The connectivity of all non-terminal nodes is analyzed, and a set of nodes that are not connected to the terminal node is selected. Each node in the set of unconnected nodes is replaced by the shortest path connectivity relationship between adjacent nodes.

5. The drainage pipe network monitoring method according to claim 1, characterized in that, The steps of diagnosing the functional defect status and internal physical defect status of the drainage network based on the processed multi-source monitoring data and using a trained machine learning classification model and a deep learning object detection algorithm, respectively, include: The upstream and downstream liquid level data of the processed multi-source monitoring data are extracted as input feature vectors, and the input feature vectors are input into the trained K-nearest neighbor algorithm classification model; Based on the trained K-nearest neighbor algorithm classification model, the feature distance between the input feature vector and the training samples is calculated to diagnose the functional defect status of the drainage pipe network. The expression for calculating the feature distance between the input feature vector and the training samples is as follows: ； In the formula, This represents the feature distance between the input feature vector and the training samples. This represents the input feature vector for real-time monitoring. This represents the feature vector in the training samples. The dimension of the feature vector. Index representing the feature dimension , These represent the input feature vector at the th... The values ​​in each dimension, and the feature vector of the training sample in the th dimension. Values ​​in each dimension Indicates the distance parameter; Extract the internal image data from the processed multi-source monitoring data, and identify the internal physical defects of the drainage network based on the improved YOLO-v4 deep learning object detection algorithm model and the internal image data; The diagnostic results of the functional defect state and the identification results of the internal physical defect state are spatiotemporally correlated and fused to generate a comprehensive defect diagnosis report.

6. The drainage pipe network monitoring method according to claim 1, characterized in that, The step of mapping the modified standard topology file, the functional defect status, the internal physical defect status, and the water level change trend to the hydraulic model to construct a digital twin synchronized with the drainage network in real time includes: The corrected standard topology file is parsed into recognizable node objects and pipe objects. Based on the node objects and pipe objects, the functional defect state is mapped to the initial water level boundary conditions of the pipe, and the internal physical defect state is mapped to the corresponding hydraulic parameter correction values. The multi-source monitoring data is used as a dynamic boundary condition, and the dynamic boundary condition is input into the hydraulic model at a preset time frequency to simulate the hydraulic distribution state of the drainage network and obtain simulation results. The simulation results are compared with the multi-source monitoring data in real time to obtain the comparison results. If the comparison results exceed the preset threshold, the key parameters of the hydraulic model are automatically calibrated, and the automatically calibrated hydraulic model is fused and rendered with the basic geographic information data to generate an interactive three-dimensional visual digital twin.

7. The drainage pipe network monitoring method according to claim 1, characterized in that, Following the step of constructing a digital twin synchronized in real time with the drainage network, the method further includes: Using the digital twin as the computing engine, a non-dominated sorting genetic algorithm with an elite strategy as the optimization engine, and multiple objective functions such as minimizing the total amount of waterlogging, minimizing the total amount of combined sewer overflow pollution, and maximizing the overall scheduling benefits, the decision variables for adjusting the operating parameters of the drainage network are iteratively optimized to obtain the optimal collaborative scheduling scheme.

8. A drainage pipe network monitoring system, characterized in that, The system includes: The data acquisition and processing module is used to acquire multi-source monitoring data of the drainage pipe network in real time and process the multi-source monitoring data. The module is used to acquire the basic structural data and basic geographic information data of the drainage network, and to construct a directed weighted graph topology model with inspection wells as nodes and pipes as directed edges based on the basic geographic information data, the basic structural data and graph theory algorithms. The identification and calculation module is used to identify the topological problems of the drainage network based on the directed weighted graph topology model, and to identify and correct the topological problems by calculating the parameters of each node to obtain a standard topology file. The diagnostic module is used to diagnose the functional defect status and internal physical defect status of the drainage network based on the processed multi-source monitoring data and the trained machine learning classification model and deep learning object detection algorithm, respectively. The extraction and fusion module is used to extract the liquid level sequence from the multi-source monitoring data, fuse the liquid level sequence with the meteorological rainfall data of the same period to obtain fused data, and predict the water level change trend of key nodes of the drainage network based on the hybrid neural network and the fused data. The mapping construction module is used to map the corrected standard topology file, the functional defect status, the internal physical defect status, and the water level change trend to the hydraulic model, so as to construct a digital twin that is synchronized with the drainage network in real time.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the drainage network monitoring method as described in any one of claims 1 to 7.

10. A storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the drainage network monitoring method as described in any one of claims 1 to 7.

Images