Water supply network leakage positioning system and method based on sound wave-pressure cooperation

Abstract

The invention discloses a water supply network leakage positioning system and method based on sound wave-pressure cooperation, and belongs to the technical field of water supply network monitoring and leakage control. The system comprises a pressure sensing layer, a cooperative control unit, a sound wave sensing layer, an underwater sound cooperative positioning solver and a result output module. The pressure sensing layer monitors the pressure of a pipe network in real time through pressure sensors which are sparsely deployed, and identifies an abnormal area; the cooperative control unit dynamically activates the acoustic sensors of the minimum set; the underwater sound cooperative positioning resolver corrects sound wave parameters through a hydraulic model, and a leakage point is accurately positioned through sound wave time difference inversion in a pressure abnormal area; the result output module pushes the positioning result, the leakage amount and the maintenance scheme to the operation and maintenance platform; and pressure early warning, sound wave accurate measurement and joint inversion are carried out.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of urban water supply network monitoring and leakage control, and particularly relates to a water supply network leakage positioning system and method based on sound wave sensing and pressure sensing cooperation. BACKGROUND

[0002] The urban water supply network is an important part of the urban infrastructure, and its safe operation is directly related to people's livelihood protection and social stability. At present, the leakage monitoring of the water supply network mainly relies on a single technical means, such as the sound wave monitoring method or the pressure monitoring method. There are three major bottlenecks of high monitoring cost, insufficient positioning accuracy and delayed decision feedback. Specifically, the sound wave monitoring method has high positioning accuracy, but needs to densely deploy sensors, and the system deployment and maintenance cost is high, which is difficult to achieve full coverage of the network. The pressure monitoring method has low cost and is easy to deploy widely, but can only achieve rough positioning of the leakage area and cannot accurately identify the leakage point position, and is easily disturbed by the working condition of the network.

[0003] At the technical architecture level, the existing solutions mostly adopt the architecture of separating sensing and decision-making, and cannot realize the deep fusion and intelligent correlation analysis of multi-source information. This leads to the fact that the monitoring data cannot be effectively converted into accurate leakage positioning instructions, the decision lacks real-time and reliable data support, and the management measures are seriously out of touch with the actual leakage situation. In addition, the traditional system completely lacks a multi-source data cooperative verification mechanism, cannot exclude environmental interference and signal aliasing, has low diagnosis reliability, and the system is difficult to adapt to the dynamic changes of the network.

[0004] Therefore, the existing technical system cannot meet the high-level requirements of intelligent water management for leakage positioning accuracy, cost controllability and decision real-time, and an innovative system solution is needed to realize the integration of pressure wide-area early warning and sound wave accurate positioning. SUMMARY

[0005] The present application aims to provide a water supply network leakage positioning system and method based on sound wave-pressure cooperation to overcome the defects of high deployment cost of the sound wave monitoring method and insufficient positioning accuracy of the pressure monitoring method in the prior art. Through the complementary advantages of the two technologies, a cooperative mechanism of wide-area early warning and accurate positioning is realized. The sparse pressure sensor network is used to realize early identification and regional rough positioning of leakage events in a large range at low cost, and then the sound wave sensor in a specific area is intelligently activated for accurate detection and data acquisition. Finally, through data fusion and joint inversion, the accurate coordinates of the leakage point are output with the optimal system cost and energy consumption.

[0006] The present application can be implemented by the following technical scheme: a water supply network leakage positioning system based on sound wave-pressure cooperation, comprising a pressure sensing layer, a cooperative control unit, a sound wave sensing layer, a water sound cooperative positioning solver and a result output module.

[0007] Pressure sensing layer: Composed of pressure sensors deployed at key hydraulic sensitivity nodes of the pipeline network, used to acquire pipeline network pressure data in real time; This layer performs real-time data analysis through a built-in hydraulic model, and once an abnormal pressure event is identified, it can output a candidate range of the abnormal pressure area, realizing the preliminary identification and coarse location of leakage.

[0008] Furthermore, the pressure sensing layer includes: a pressure monitoring and data acquisition unit, and a hydraulic analysis and zone generation unit;

[0009] The pressure monitoring and data acquisition unit consists of pressure sensors sparsely deployed at key nodes of the pipeline network. The key nodes are determined based on offline hydraulic sensitivity analysis and selected using the longest path principle of graph theory and a comprehensive evaluation algorithm for node hydraulic sensitivity. This unit is configured to monitor and acquire pipeline network pressure data in real time at a set sampling frequency and upload the data to the data center via a communication module.

[0010] The hydraulic analysis and region generation unit has a built-in hydraulic model based on the pipeline network topology. It is configured to automatically call the hydraulic model when pressure anomalies are detected, trace the source of the anomaly by analyzing the pressure transmission path, and output the candidate regions of pressure anomalies.

[0011] Collaborative Control Unit: Communicatively connected to the pressure sensing layer, this unit serves as the intelligent scheduling center of the system. It is pre-stored with a regional node mapping table generated through offline hydraulic sensitivity analysis. When the pressure sensing layer responds, this unit can automatically convert the pressure anomaly area into a list of node IDs of the minimum set of acoustic sensors that need to be activated, and dynamically schedule the sensors in the list to enter the working state, while keeping other sensors in a low-power standby state, thereby realizing on-demand allocation of energy and communication bandwidth.

[0012] The collaborative control unit achieves its function in the following ways:

[0013] By using a pre-generated region-node mapping relationship, the received pressure anomaly region is converted into a corresponding set of acoustic sensor nodes;

[0014] The minimum vertex cover algorithm based on graph theory calculates the minimum list of nodes that can cover the entire abnormal region from the set of nodes.

[0015] Based on the generated list of minimum nodes, a wake-up command is sent to the sensors in the list, while maintaining the low-power standby state of the sensors outside the list.

[0016] The system monitors the operating status of all sensors in real time and automatically activates backup sensors when a device malfunction is detected, ensuring the continuous execution of monitoring tasks.

[0017] Acoustic sensing layer: Communicates with the cooperative control unit and consists of a mobile array of acoustic sensors for precise positioning; the sensors in this layer are usually in a low-power standby state and are only activated in a specific suspected area after receiving a system command to collect high-quality acoustic signals.

[0018] The acoustic wave sensing layer includes an acoustic wave sensor array, a signal preprocessing unit, and a partitioned asynchronous monitoring unit;

[0019] An acoustic sensor array with edge computing capabilities supports mobile deployment and is configured to acquire acoustic signals in candidate areas of pressure anomalies.

[0020] The signal preprocessing unit is configured to filter, reduce noise, and extract features from the raw acoustic signal locally, and upload the compressed feature data.

[0021] The partitioned asynchronous monitoring unit is configured to virtually divide the acoustic sensor into several monitoring sub-regions according to the pipeline topology, and activate the sensor sequentially according to the hydraulic correlation priority.

[0022] By acquiring signals through rapidly deployable mobile sensors and utilizing local edge computing for data preprocessing and feature extraction, data quality and transmission efficiency are improved. Combined with real-time corrected acoustic parameters and intelligent inversion algorithms, the final output is a high-precision positioning result with credibility assessment.

[0023] Underwater acoustic collaborative positioning solver: It communicates with the acoustic wave sensing layer and dynamically corrects the propagation parameters of sound waves in the pipeline based on real-time collected pipeline operation data; under the spatial constraints of the pressure anomaly area, it uses the corrected acoustic parameters to perform joint inversion calculation; after the leakage event is confirmed, it collects on-site data to perform self-evolution training on the positioning model.

[0024] The underwater acoustic co-localization solver also includes:

[0025] The hydraulic correction module is configured to dynamically calculate and correct the actual propagation speed and attenuation coefficient of sound waves in the pipeline based on real-time collected data on pipeline pressure, water temperature, pipeline material and pipe diameter.

[0026] The joint inversion module is configured to perform joint inversion calculations using an algorithm based on the time difference of arrival of sound waves, with the pressure anomaly area as the spatial constraint and the corrected acoustic parameters as the basis, and output the leakage point location results.

[0027] The self-evolution module is configured to collect the pressure change curves and acoustic feature vectors of the entire network after each leakage event is confirmed by on-site maintenance, form sample pairs and store them in the historical sample library, and perform online incremental training on the positioning model to achieve continuous self-optimization of the system's positioning accuracy and event recognition capabilities.

[0028] Furthermore, the joint inversion module specifically employs a maximum a posteriori probability estimation method based on Bayesian theory to solve for the coordinates of the leakage points, including:

[0029] A probabilistic computation framework is constructed to transform the physical constraints of sound wave propagation, pipeline topology characteristics, and spatial prior information of pressure anomaly areas into probabilistic computational elements.

[0030] The Markov chain Monte Carlo random sampling algorithm is used to efficiently solve the posterior probability distribution of the location of the leakage point;

[0031] The output includes comprehensive positioning information containing confidence intervals and uncertainty ranges.

[0032] By real-time correction of acoustic parameters, combined with pressure zone locking and intelligent algorithms, the system can calculate accurate leak locations with high reliability, and can learn and optimize itself after each maintenance confirmation.

[0033] Results output module: Communicatively connected to the underwater acoustic co-location solver, used to output the precise positioning results, leakage estimation and recommended valve shut-off scheme to the operation and maintenance platform and the pipeline network digital twin model for visualization annotation and work order dispatch;

[0034] The results output module also includes:

[0035] The intelligent push engine is configured to adaptively select push channels and information detail levels based on the urgency and scope of an event.

[0036] The visualization enhancement module is configured to combine the digital twin model to generate a comprehensive visualization view that includes the location of the leakage point, the affected watershed, and a schematic diagram of the valve closure scheme.

[0037] The closed-loop feedback module is configured to track the status and progress of work orders, realizing a closed-loop business process from location to handling.

[0038] It achieves a complete business closed loop from precise location to repair and handling, can automatically push key information according to the urgency of the leakage, intuitively display the location of the leak and the valve shut-off plan on the digital twin model, and track the repair progress throughout the process to ensure that every leakage event can be handled quickly and effectively.

[0039] The pressure sensing layer, collaborative control unit, acoustic sensing layer, underwater acoustic collaborative positioning solver, and result output module together constitute a closed-loop control system from pressure early warning to precise acoustic positioning.

[0040] Furthermore, the method for locating leakage in water supply networks based on acoustic wave-pressure coordination is characterized by comprising the following steps:

[0041] S1. Real-time pressure data of the pipeline network is acquired by pressure sensors sparsely deployed at key nodes of the pipeline network. The pressure data is analyzed based on a real-time hydraulic model to identify pressure anomalies and output the candidate range of pressure anomalies.

[0042] S2. Based on the pressure anomaly area output in step S1, query the pre-generated area-node mapping table, convert it into a list of minimum nodes of acoustic sensors that need to be activated, dynamically schedule the corresponding sensors to enter the working state, and at the same time control the sensors outside the list to enter the low-power standby state.

[0043] S3. The acoustic wave sensor activated after being scheduled in step S2 collects acoustic wave signals in the pressure anomaly area, performs preprocessing such as filtering, noise reduction and feature extraction locally, and uploads the compressed feature data.

[0044] S4. Receive the pressure anomaly area data from step S1 and the acoustic characteristic data uploaded in step S3. Correct the acoustic propagation parameters using the hydraulic model. Based on the corrected acoustic parameters and with the pressure anomaly area as the spatial constraint, perform joint inversion calculation based on the acoustic time difference algorithm to output the accurate location result of the leak point.

[0045] S5. Based on the precise positioning results obtained in step S4, combined with the hydraulic model, estimate the leakage and generate a recommended valve shut-off scheme. Output the scheme to the operation and maintenance platform and the pipeline network digital twin model through a standardized interface for visualization annotation and work order dispatch.

[0046] Steps S1 to S5 are sequentially linked to form a closed-loop control process from pressure early warning and sound wave precision measurement to decision output.

[0047] Beneficial effects

[0048] This invention forms a control loop for early warning, scheduling, positioning, and decision-making by combining a pressure sensing layer, a collaborative control unit, an acoustic sensing layer, a hydroacoustic collaborative positioning solver, and a result output module. It solves the problems of high cost and insufficient accuracy in traditional leakage location due to single monitoring methods, limited positioning mechanisms, and delayed decision feedback. By utilizing the wide-area monitoring of the pressure sensing layer, the dynamic scheduling of the collaborative control unit, the precise acquisition of the acoustic sensing layer, the joint inversion of the hydroacoustic collaborative positioning solver, and the intelligent push of the result output module, it achieves comprehensive diagnosis from coarse positioning of pressure anomaly areas to precise acoustic positioning, meeting the needs of smart water management for early detection, accuracy, and real-time decision-making in leakage location.

[0049] By sparsely deploying the pressure sensing layer and coordinating the minimum set scheduling of the collaborative control unit, the system cost and energy consumption are significantly reduced. Combined with the linkage of hydraulic model correction and Bayesian inversion algorithm, a two-way feedback mechanism of multi-source information collaboration and decision optimization is formed, which solves the problems of wide-area monitoring and precise positioning, and separation of state diagnosis and control decision in traditional positioning systems, and significantly improves the accuracy of leak point identification and the precision of maintenance operations.

[0050] The pressure sensing layer enables wide-area monitoring of pipeline pressure and identification of abnormal areas. Under the scheduling of the collaborative control unit, the acoustic sensing layer is activated for precise signal acquisition. The underwater acoustic collaborative positioning solver completes the joint inversion based on hydraulic model correction. Finally, the result output module generates positioning results and maintenance plans, realizing a closed-loop application of diagnostic decision-making. Through the collaborative work of the modules, the shortcomings of traditional leakage positioning, such as single information dimension, low positioning reliability, and delayed decision feedback, are effectively overcome. This provides an integrated intelligent equipment solution for smart water management, from perception to decision-making. Attached Figure Description

[0051] To better understand and implement this application, the technical solution is described in detail below with reference to the accompanying drawings.

[0052] Fig. 1 This is a schematic diagram of the structure of a water supply network leakage location system based on acoustic wave-pressure coordination according to the present invention;

[0053] Fig. 2 This is a flowchart illustrating a method for locating leaks in a water supply network based on acoustic wave-pressure coordination, according to the present invention. Detailed Implementation

[0054] To further illustrate the technical means and effects adopted by the present invention to achieve the intended purpose, exemplary embodiments will be described in detail below, examples of which are illustrated in the accompanying drawings; when the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements; the embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application.

[0055] The terminology used in this application is for descriptive purposes only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used herein are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed purposes.

[0056] The following detailed description of the specific implementation methods, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided in detail.

[0057] Example 1

[0058] Please see Figs. 1-2 This embodiment provides a water supply network leakage location system based on acoustic wave-pressure coordination, including a pressure sensing layer, a coordination control unit, an acoustic wave sensing layer, a hydroacoustic coordination location solver, and a result output module;

[0059] The pressure sensing layer, collaborative control unit, acoustic sensing layer, underwater acoustic collaborative positioning solver, and result output module together constitute a closed-loop control system from pressure early warning to precise acoustic positioning.

[0060] The pressure sensing layer is used to monitor pipeline pressure data in real time through pressure sensors sparsely deployed at key nodes of the pipeline network's hydraulic sensitivity. It has a built-in hydraulic model based on the pipeline network topology, which can quickly locate abnormal areas when leakage occurs, achieving early warning and coarse location at low cost, and providing clear targets for precise acoustic positioning.

[0061] Furthermore, the pressure sensing layer includes: a pressure monitoring and data acquisition unit, and a hydraulic analysis and zone generation unit;

[0062] The pressure monitoring and data acquisition unit consists of pressure sensors sparsely deployed at key nodes of the pipeline network. The key nodes are determined based on offline hydraulic sensitivity analysis and selected using the longest path principle of graph theory and a comprehensive evaluation algorithm for node hydraulic sensitivity. This unit is configured to monitor and acquire pipeline network pressure data in real time at a set sampling frequency and upload the data to the data center via a communication module.

[0063] Among them, the longest path principle of graph theory and the node hydraulic sensitivity comprehensive evaluation algorithm were adopted for selection.

[0064] Specifically, it includes:

[0065] Based on the actual pipeline distribution and node connection relationships of the water supply network, such as valves, water meters, and pipe interfaces, a network topology model containing pipe segment parameters and node attributes is constructed. This model is then transformed into an undirected weighted graph in graph theory. Specifically, physical nodes in the network are used as vertices of the graph, including pipe junctions and equipment connection points. Pipe segments are used as edges connecting vertices, and the pipe segment length and hydraulic resistance coefficient are used as edge weights.

[0066] Using the longest path algorithm in graph theory, such as a variation of Dijkstra's algorithm or dynamic programming, the longest path covering the main branches of the pipeline network and reflecting large-scale hydraulic changes is calculated in the constructed pipeline network topology. Nodes on this longest path are selected as initial candidate nodes. These nodes can monitor the pressure signal propagating along the longest path, covering a wider pipeline area with the fewest nodes. This avoids dense sensor deployment, meets the cost control requirements of sparse deployment, and ensures effective capture of pressure changes in the main pipeline.

[0067] For the selected candidate nodes, a hydraulic sensitivity evaluation index system is constructed, including the pressure change amplitude, pressure response speed, and signal transmission efficiency to surrounding pipe sections during leakage. Offline hydraulic simulation is used to model leakage scenarios at different locations and scales, calculating the sensitivity values ​​of each candidate node under various scenarios. A weighted comprehensive evaluation algorithm is then used to quantify and rank the node sensitivity, selecting the nodes with the highest sensitivity ranking to significantly reflect pressure anomalies caused by leakage and ensure the effectiveness of pressure monitoring.

[0068] The hydraulic analysis and region generation unit has a built-in hydraulic model based on the pipeline network topology. It is configured to automatically call the hydraulic model when pressure anomalies are detected, trace the source of the anomaly by analyzing the pressure transmission path, and output the candidate regions of pressure anomalies.

[0069] The candidate region for abnormal output pressure specifically includes:

[0070] By setting a sampling frequency, pressure sensor data sparsely deployed at key nodes is uploaded to the data center in real time. Based on preset anomaly judgment rules, such as pressure values ​​exceeding the normal threshold range or pressure mutation amplitude reaching a set standard, pressure data changes are monitored in real time. When an anomaly event is detected, the hydraulic model built into the hydraulic analysis and region generation unit, which is based on the pipeline network topology, is automatically triggered to start the anomaly tracing process. The pipeline network topology includes node connection relationships, pipe segment parameters, etc.

[0071] The activated hydraulic model calls upon the pipeline topology data and real-time pressure data to simulate the pressure transmission pattern within the pipeline under normal operating conditions. By reverse-engineering the propagation process of pressure anomalies, it compares the timing of pressure changes and the degree of pressure attenuation between the abnormal node and its surrounding related nodes. Combining the influence of hydraulic characteristics such as pipe diameter and material on pressure transmission, it eliminates interference from normal hydraulic fluctuations, accurately identifies the main transmission path of pressure anomalies spreading from the source to the surrounding areas, and initially locks in the approximate direction of the anomaly source.

[0072] Based on the direction of the anomaly source and the transmission path obtained through reverse tracing, the hydraulic model, combined with the hydraulic sensitivity distribution of the pipe network and the response sensitivity of key nodes to leakage, defines the range of areas where pressure anomalies occur. The model comprehensively considers the pipe segment length on the anomaly transmission path and the pressure coordination characteristics of adjacent key nodes, and delineates the area where leakage may occur as a specific pressure anomaly candidate area, such as a continuous pipe segment or a sub-region of the pipe network surrounded by multiple key nodes. This ensures that the candidate area covers the potential leakage range while avoiding excessive expansion that would waste subsequent monitoring resources, and provides a clear target area for the collaborative control unit.

[0073] Specifically, the pressure sensing layer, through pressure sensors sparsely deployed at key nodes and combined with real-time analysis of pressure data using a built-in hydraulic model, can quickly identify and trace back to areas of abnormal pressure when leakage occurs. This enables early warning and intelligent coarse localization of leakage under low-cost, wide-area monitoring, providing clear detection targets and trigger signals for the precise activation and collaborative localization of subsequent acoustic sensors, and laying the foundation for the pressure-guided acoustic collaborative working mechanism of the entire system.

[0074] The collaborative control unit communicates with the pressure sensing layer to receive the pressure anomaly area defined by the pressure sensing layer and convert it into a minimal set of acoustic sensor node list. Using spatial analysis and minimum vertex coverage algorithm, it accurately calculates the minimum set of sensors to be activated and dynamically schedules the sensors in the list to enter the working state. It achieves efficient scheduling with on-demand wake-up and low-power standby, and automatically activates backup nodes when equipment fails, ensuring continuous monitoring and optimal system energy efficiency.

[0075] Furthermore, dynamic scheduling is achieved through a coordination mechanism. First, the GIS polygon coordinates of the pressure anomaly area output by the pressure sensing layer are received. A pre-generated region-node mapping table is queried, defining the correspondence between each virtual monitoring sub-region and the deployed sensor node IDs. Through GIS spatial overlay analysis, all sub-regions intersecting with the anomaly area are identified. Using the minimum vertex cover algorithm of graph theory, a minimum set of nodes covering the entire anomaly area is calculated from the sensor nodes in the intersecting sub-regions. An encrypted wake-up command is sent to the sensors in the list, switching them from low-power sleep mode to high-power operating state, while maintaining the low-power standby state of sensors outside the list. Sensor status is monitored in real time, and backup nodes are automatically activated in case of equipment failure to ensure monitoring continuity.

[0076] Furthermore, when the pressure sensing layer detects an abnormal area, the unit immediately queries the pre-stored map, converts this area into a list of acoustic sensors that need to operate, and uses mathematical algorithms to calculate the optimal combination of sensors that can cover the entire area with the fewest sensors. It then wakes up the sensors on the list to work intensively, while allowing other sensors to remain dormant to conserve power. Through this on-demand activation and precise deployment method, it ensures sufficient and efficient monitoring power to capture acoustic signals in suspected leakage areas, while also preventing the entire system from consuming excessive energy due to all devices operating continuously. The collaborative control unit uses an intelligent scheduling mechanism to monitor and capture leakage acoustic signals.

[0077] Specifically, it achieves efficient scheduling that accurately wakes up the minimum number of sensors covering abnormal areas, minimizing system energy consumption while ensuring continuous monitoring.

[0078] The acoustic sensing layer, connected to the collaborative control unit, collects acoustic signals in a designated area through a mobile, deployable array of intelligent sensors. It then utilizes edge computing capabilities to perform signal preprocessing and feature extraction on-site. Simultaneously, it optimizes the scheduling of sensor resources through a partitioned monitoring mechanism. The underwater acoustic collaborative positioning solver, by real-time correction of acoustic parameters, performs joint calculations using a Bayesian inversion algorithm under the spatial constraints of anomaly pressure areas. It outputs leakage point location results with confidence intervals and continuously improves the system's positioning accuracy through a self-evolution mechanism, forming a complete technology chain from signal acquisition to positioning solution.

[0079] The mobile smart sensor array uses a magnetic base and standardized interface, which can be carried by maintenance personnel to areas with abnormal pressure and quickly attached to pipe accessories such as fire hydrants and valves. During deployment, the system automatically updates the sensor coordinates by scanning and binding the geographical location through a mobile terminal, forming a mobile monitoring mode in which the personnel carry the equipment and the equipment follows the leak point, so as to achieve flexible and accurate coverage of the entire network by a limited number of sensors.

[0080] Furthermore, the acoustic wave sensing layer includes an acoustic wave sensor array, a signal preprocessing unit, and a partitioned asynchronous monitoring unit;

[0081] An acoustic sensor array with edge computing capabilities supports mobile deployment and is configured to acquire acoustic signals in candidate areas of pressure anomalies.

[0082] The signal preprocessing unit is configured to filter, reduce noise, and extract features from the raw acoustic signal locally, and upload the compressed feature data.

[0083] The partitioned asynchronous monitoring unit is configured to virtually divide the acoustic sensor into several monitoring sub-regions according to the pipeline topology, and activate the sensor sequentially according to the hydraulic correlation priority.

[0084] Specifically, the system collects signals through rapidly deployable mobile sensors, performs data preprocessing and feature extraction using local edge computing, thereby improving data quality and transmission efficiency. Combined with real-time corrected acoustic parameters and intelligent inversion algorithms, it ultimately outputs high-precision positioning results with credibility assessment. Furthermore, the system can continuously learn and optimize through a self-evolution mechanism, achieving continuous improvement in accuracy with use.

[0085] The underwater acoustic collaborative positioning solver communicates with the acoustic wave sensing layer and dynamically corrects the propagation parameters of sound waves in the pipeline based on real-time collected pipeline operation data. Under the spatial constraints of the pressure anomaly area, it uses the corrected acoustic parameters to perform joint inversion calculation. After the leakage event is confirmed, it collects on-site data to perform self-evolution training on the positioning model.

[0086] Furthermore, the underwater acoustic co-localization solver includes a hydraulic correction module, a joint inversion module, and a self-evolution module;

[0087] The hydraulic correction module is configured to dynamically calculate and correct the actual propagation speed and attenuation coefficient of sound waves in the pipeline based on real-time collected data on pipeline pressure, water temperature, pipe material, and pipe diameter.

[0088] The joint inversion module integrates multi-source information through intelligent algorithms and outputs accurate positioning results with credibility assessment; the results output module transforms this professional data into multi-level decision support information, building a system from problem discovery to resolution.

[0089] The joint inversion module is configured to perform joint inversion calculations using an algorithm based on the time difference of arrival of sound waves, with the pressure anomaly area as the spatial constraint and the corrected acoustic parameters as the basis, and output the leakage point location results.

[0090] The self-evolution module is configured to collect the pressure change curves and acoustic feature vectors of the entire network after each leakage event is confirmed by on-site maintenance, form sample pairs and store them in the historical sample library, and perform online incremental training on the positioning model to achieve continuous self-optimization of the system's positioning accuracy and event recognition capabilities.

[0091] The method employs a maximum a posteriori probability estimation method based on Bayesian theory to solve for the coordinates of the leak point. This includes: constructing a probabilistic computational framework to transform the physical constraints of sound wave propagation, pipeline topology characteristics, and spatial prior information of pressure anomaly areas into probabilistic computational elements; using a Markov chain Monte Carlo random sampling algorithm to efficiently solve for the posterior probability distribution of the leak point location; and outputting comprehensive location information including confidence intervals and uncertainty ranges.

[0092] Based on Bayesian theory, a probabilistic framework for joint inversion is constructed. The corrected sound wave propagation speed, pipeline topology, and pressure anomaly region boundary are used as spatial prior information, and the sound wave arrival time difference data are used as observational evidence. Together, they constitute the basic input for inversion calculation.

[0093] The Markov chain Monte Carlo method is used to efficiently sample the posterior probability distribution of the leak location. Candidate leak points are randomly generated within the pressure anomaly area. The theoretical time difference is calculated by combining the sound wave propagation model and compared with the measured time difference. The candidate locations are iteratively optimized to gradually approximate the most likely leak point coordinates.

[0094] Furthermore, the optimal estimated coordinates of the leak points are generated based on the sampling results, and their confidence interval and uncertainty range are calculated; the output results include the precise location of the leak points, the location confidence level, and the error range.

[0095] Specifically, by real-time correction of acoustic parameters, combined with pressure zone locking and intelligent algorithms, the system can calculate accurate leak locations with high reliability. Furthermore, it can learn and optimize itself after each maintenance confirmation, making the entire system more and more accurate with use.

[0096] The results output module communicates with the underwater acoustic co-positioning solver and is used to output the accurate positioning results, leakage estimation and recommended valve shut-off scheme to the operation and maintenance platform and the pipeline network digital twin model for visualization annotation and work order dispatch.

[0097] Furthermore, the results output module also includes:

[0098] The intelligent push engine is configured to adaptively select push channels and information detail levels based on the urgency and scope of an event.

[0099] The visualization enhancement module is configured to combine the digital twin model to generate a comprehensive visualization view that includes the location of the leakage point, the affected watershed, and a schematic diagram of the valve closure scheme.

[0100] The closed-loop feedback module is configured to track the status and progress of work orders, realizing a closed-loop business process from location to handling.

[0101] The precise leakage location results include:

[0102] The system receives precise location results of leak points, leakage estimation data, and recommended valve shut-off schemes from the hydroacoustic collaborative positioning solver; it transmits these data through a standardized RESTful API to ensure the integrity and real-time nature of the information; and it integrates this data with the pipeline topology and hydraulic model parameters to provide a basis for subsequent decision-making.

[0103] The intelligent push engine within the module adaptively selects push channels such as SMS, email, or operation and maintenance platform notifications and information detail levels based on the urgency and scope of the event; the visualization enhancement module utilizes the pipeline network digital twin model to generate a comprehensive visualization view that includes the location of leak points, affected watersheds, and valve shut-off scheme diagrams, which are intuitively marked on a 3D map for easy understanding by operation and maintenance personnel.

[0104] The processed information is automatically generated into a maintenance work order, which is pushed to the on-site operation and maintenance terminal through the Internet of Things gateway. The closed-loop feedback module tracks the status of the work order and the progress of the handling in real time, and collects maintenance confirmation feedback to ensure a closed loop of business from location to handling.

[0105] Specifically, it achieves a complete business closed loop from precise location to repair and handling. It can automatically push key information according to the urgency of the leakage, intuitively display the location of the leak and the valve shut-off plan on the digital twin model, track the repair progress throughout the process, and ensure that every leakage event can be handled quickly and effectively.

[0106] This embodiment provides an efficient and intelligent solution for locating leaks in water supply networks. The system achieves wide-area early warning through the sparse deployment of the pressure sensing layer, and the collaborative control unit uses graph theory algorithms to dynamically activate the minimum sensor set, achieving precise energy-saving scheduling. The acoustic sensing layer performs high-precision signal acquisition and preprocessing in the target area, providing a data foundation for accurate positioning. The core underwater acoustic collaborative positioning solver innovatively integrates a hydraulic model to dynamically correct acoustic parameters and performs joint inversion calculations under the spatial constraints of anomaly pressure zones. Its self-evolutionary capability ensures continuous improvement in system performance. Finally, the result output module completes the multi-dimensional push and visualization of decision information.

[0107] Example 2

[0108] This embodiment provides a method for locating leaks in a water supply network based on acoustic wave-pressure coordination, including the following steps:

[0109] S1. Pressure data of the pipeline network is acquired in real time by pressure sensors sparsely deployed at key nodes of the pipeline network hydraulic sensitivity. The pressure data is analyzed based on a real-time hydraulic model. A change detection algorithm based on wavelet transform is used to identify pressure anomaly events. A reverse tracing algorithm of the pressure transmission path is used to output the candidate range of pressure anomalies. Based on the pressure anomaly region, a pre-generated region-node mapping table is queried. The minimum vertex cover algorithm in graph theory is used to calculate the minimum set of acoustic sensor nodes to be activated. The corresponding sensor is controlled to enter the working state through a dynamic scheduling mechanism. At the same time, the sensors outside the list are controlled to enter the low-power standby state based on the energy-saving strategy.

[0110] S2 and the collaborative control unit, based on the pressure anomaly area output by S1, query the pre-generated region-node mapping table, use the minimum vertex cover algorithm in graph theory to calculate the minimum set of acoustic sensor nodes that need to be activated, and send wake-up commands to the sensors in the list through a dynamic scheduling mechanism to enable them to enter the working state, while controlling the sensors outside the list to enter the low-power standby state.

[0111] S3. Acoustic wave signals in the pressure anomaly area are collected by a scheduled and activated acoustic wave sensor array. The preprocessing of filtering, noise reduction and feature extraction is performed locally using an end-to-end signal processing model based on deep learning. The processed feature data is then uploaded using a data compression algorithm.

[0112] S4. Receive data on pressure anomaly areas and uploaded acoustic characteristic data. Dynamically correct the propagation parameters of acoustic waves in complex pipe network environments through hydraulic models. Based on the corrected acoustic parameters and with pressure anomaly areas as spatial constraints, use a joint inversion algorithm based on Bayesian theory to calculate the time difference of arrival of acoustic waves. Output the accurate location results and confidence intervals of leakage points through Markov chain Monte Carlo sampling method.

[0113] S5. Based on the obtained accurate positioning results, combined with the real-time hydraulic model, the leakage amount is estimated and a recommended valve shut-off scheme is generated. The results are output to the operation and maintenance platform and the pipeline network digital twin model through a standardized RESTful interface. A WebGL-based visualization engine is used to mark the leakage points in the three-dimensional scene and dispatch intelligent work orders. After each maintenance confirmation, on-site verification data is collected to incrementally train the positioning model, so as to continuously optimize the system's diagnostic capabilities.

[0114] This method achieves a leapfrog diagnosis from abnormal area identification to precise leak location through deep collaboration between wide-area pressure monitoring and precise acoustic positioning. It establishes an intelligent decision-making mechanism based on multi-source data fusion, effectively solving the problems of high cost or insufficient accuracy of single monitoring methods, and providing a complete technical solution for smart water management that integrates perception, decision-making, and optimization.

[0115] Example 3

[0116] Application scenario: The water supply network leakage location system based on sound wave-pressure coordination is used to monitor water supply network leakage in old residential communities;

[0117] This scenario addresses the problem of water supply network leakage in old residential areas, solving the technical challenge of quickly and accurately locating leakage points in the complex network environment of these areas. The following explanation combines specific application scenarios and implementation details.

[0118] In a certain city, an old residential community built in 2000 experienced abnormally high minimum water flow rates at night, suggesting a possible leak. The community's water supply network has a mixed ring and branch structure, buried at a depth of 1.2 meters, with the main pipes being DN200-DN100 cast iron pipes. Traditional sound-listening methods are difficult to pinpoint accurately due to environmental noise interference. Traditional methods for locating leaks in water supply networks rely on manual labor, are inefficient, and lack sufficient accuracy.

[0119] System Implementation and Operation Process: After hydraulic model analysis of the pipeline network, fixed pressure sensors are installed at four key nodes: pump station outlets, two ring network connection points, and the end of the pipeline network. The sensors are fixed to the pressure testing ports of the pipeline via threaded interfaces and externally fitted with protective manhole covers. Movable acoustic sensors are pre-installed in the 15 fire hydrants and valve wells of the community. The bottom of the sensor has a magnetic base, which can be quickly attached to the fire hydrant or valve. It has a built-in 4G communication module and battery. Software modules such as the collaborative control unit and the underwater acoustic collaborative fusion positioning processor are deployed in the form of microservice containers on the edge server of the community property management center.

[0120] At 2:15 AM one day, a pressure sensor deployed at the end of the pipeline recorded a continuous drop in pressure from 0.28 MPa to 0.22 MPa, below the system's set response threshold of 0.25 MPa. The data analysis algorithm of the pressure sensing layer identified this abnormal pattern and, through reverse tracing using a hydraulic model, pinpointed the abnormal area to a 50-meter-long pipe section north of Building 3 in the residential complex. The collaborative control unit received the GIS polygon coordinates of the abnormal pressure area.

[0121] The system immediately initiates the scheduling algorithm, which internally executes the following key processes to achieve precise sensor deployment: The system queries a pre-generated region-node mapping table based on offline hydraulic sensitivity analysis; this table defines the correspondence between each virtual monitoring sub-region and the IDs of all acoustic sensor nodes deployed within that sub-region; the system performs GIS spatial overlay and intersection analysis on the received pressure anomaly region polygons and all monitoring sub-regions in the mapping table to quickly identify all sub-regions covered or intersecting by the polygons; the system uses the minimum vertex cover algorithm from graph theory to calculate a minimum set of sensors that can effectively monitor the entire anomaly region and has the fewest number of nodes from all the sensor nodes contained in the intersecting sub-regions, ultimately generating an optimal node ID list: [S02, S05, S08, S09, S11, [S13] The system sends encrypted wake-up commands to the six sensors in the list via the 4G network. After the commands take effect, the sensors in the list are woken up from low-power sleep mode, and start the acoustic wave acquisition module, signal processor, and high-speed communication module, entering a high-power working state. At the same time, the system strictly controls the other nine sensors outside the list to remain in low-power standby mode. In this state, they only maintain the most basic communication monitoring function, and the core acquisition and processing unit is powered off, thereby greatly reducing the overall power consumption of the system. After the activated sensors are started, they return an acknowledgment signal to the collaborative control unit. The unit updates the real-time status table of all sensors in the network accordingly, laying the foundation for subsequent data acquisition. The unit queries the built-in region-node mapping table to generate a minimum activation list ID containing six acoustic wave sensors: S02, S05, S08, S09, S11, S13. The system sends activation commands to these six sensors via the 4G network. After the commands take effect: the sensors in the list are woken up from sleep mode and start the acoustic wave acquisition module; the other nine sensors outside the list remain in low-power standby mode.

[0122] The six activated acoustic sensors attach their magnetic bases to the designated fire hydrant and begin synchronously collecting acoustic signals from the pipeline for 30 seconds. The edge computing module built into the sensors filters and reduces noise in the raw signals, extracts effective leakage characteristic frequency band data, and uploads the compressed data packet. The hydraulic model calculates the actual propagation speed of the sound wave under the current operating conditions as 1120 m / s (instead of the default 1200 m / s) based on the real-time water temperature (15°C), pipeline material (cast iron), and pipe diameter (DN150). The joint inversion module uses the pressure anomaly area as a spatial constraint, utilizes the arrival time difference of the sound waves recorded by the six sensors, and combines it with the corrected sound velocity to perform the solution using a Bayesian inversion algorithm.

[0123] The output of the underwater acoustic fusion positioning processor shows that the leak point is located on the DN150 water supply branch pipe under the lawn on the north side of Building 3, with coordinates (X=285473.25, Y=4638921.60), confidence level 92%, and estimated leakage of about 1.8 cubic meters per hour.

[0124] Using the RESTful API, the location results, leakage amount, and the system-recommended solution to shut down valve V-103 are pushed to the property management work order system with one click: an emergency repair work order is generated; the leakage point is highlighted on the 3D map; a response message containing map coordinates and on-site photos is pushed; the maintenance team, following the work order instructions, precisely excavated the site early that morning and found a hole of about 3mm caused by pipe corrosion, which was completely consistent with the system's location.

[0125] After maintenance confirmation, technicians marked the event as "repaired" in the system. The self-evolution module automatically stored the pressure curve and acoustic feature vector of the entire event process into the sample library for subsequent incremental training of the graph neural network, optimizing the identification accuracy of similar cast iron pipe corrosion and leakage in the future. This embodiment demonstrates the complete workflow of the system in a real-world scenario. Through the synergy of pressure and acoustic waves, a seamless connection from macro-area early warning to precise location is achieved. A dynamic scheduling strategy significantly reduces system energy consumption and data redundancy. Finally, standardized output achieves a rapid closed loop from discovery to handling, effectively reducing water resource losses and economic losses. The self-evolution module is the core software component for continuous learning and performance optimization. Deployed as a microservice alongside the hydraulic character correction module, it is logically tightly integrated with the underwater acoustic collaborative positioning solver. Its working mechanism is as follows: When the result output module pushes... Once the leak location information is confirmed by the on-site maintenance team, the confirmation signal triggers the self-evolution module to start. The module then retrieves the spatiotemporal pressure change curve data covering the entire network throughout the leak event from the data center, as well as the pre-processed acoustic feature vector uploaded by the acoustic sensing layer. It precisely pairs these two types of data to form a high-quality pressure-sound leak location sample pair, which is then encrypted and stored in a dedicated historical sample library. As the sample library continues to expand, the system periodically initiates an online incremental training process. This process uses these continuously accumulated real-world case samples to retrain the built-in graph neural network model, optimizing the model's ability to map leak locations from complex pressure patterns and acoustic features. Through this mechanism, the system can gradually adapt to the dynamic changes in the pipeline network and develop increasingly accurate identification capabilities for various leak patterns, improving the success rate and confidence of location in complex scenarios.

[0126] Example 4

[0127] The application scenario of this embodiment is: acoustic-pressure collaborative identification and localization of multiple concurrent leaks in urban trunk pipelines; it solves the technical problem that traditional methods cannot separate and locate each leak point due to signal aliasing when multiple concurrent leaks occur in urban trunk pipelines.

[0128] This embodiment differs from Embodiment 3. In order to better solve the problem that urban main pipelines are prone to multiple concurrent leaks due to pressure fluctuations, and that traditional methods cannot distinguish them due to signal aliasing, this embodiment differs from Embodiment 3 in that it also includes and fully utilizes the deep collaboration between the SCADA closed-loop interface unit and the self-evolution module. Through innovative means of active hydraulic excitation, feature separation and recognition, and continuous optimization based on experience, it better achieves effective separation, accurate identification, and precise location of multiple adjacent leak points.

[0129] In a city center, a DN600 ductile iron main water supply pipeline built in 1985 experienced a water hammer event caused by a pump station switch. The dispatch center monitored a continuous abnormality in the regional water supply pressure with a complex fluctuation pattern. The pressure dropped from 0.65 MPa to 0.52 MPa and then continued to fluctuate. Since the pipeline is buried at a depth of more than 3 meters and is located under the city's main road, it is subject to severe traffic noise interference. It was initially determined that there may be multiple leakage points, and the sound wave signals generated by these leakage points superimposed and interfered with each other in the pipeline network, forming a serious aliasing effect.

[0130] System-specific operations and deep integration processes:

[0131] The pressure sensing layer detected a broad and unstable pressure drop pattern. Preliminary analysis of the hydraulic model indicated an anomaly in a 280-meter-long pipe section, but it was impossible to accurately determine whether it was caused by a single major leak or the combined effect of multiple scattered leaks.

[0132] The collaborative control unit activated 15 acoustic sensors in the suspected area according to the standard procedure; the hydraulic-acoustic underwater acoustic collaborative fusion positioning processor performed the first joint inversion calculation, but the results showed that the confidence level was only 28%, and the positioning points were discretely distributed on the map, indicating that the traditional algorithm could not extract effective source location information from the aliased signal.

[0133] The system decision invokes the SCADA closed-loop interface unit to initiate a special handling procedure. This unit sends precise instructions to the SCADA system to perform a series of preset, rapid, step-by-step micro-opening and closing operations on the upstream main control valve V-201 and the downstream regulating valve V-205 of the suspected pipe section (specifically: close 25% → hold for 30 seconds → close to 50% → hold for 20 seconds → return to full open). This precisely controlled hydraulic transient process artificially creates unique pressure fluctuation characteristics in the pipeline network. Each real physical leak point, after being subjected to this specific hydraulic impact, generates an enhanced acoustic signal with unique transient characteristics, the response of which is closely related to the type, size, and location of the leak.

[0134] The acoustic sensing layer synchronously recorded the full-frequency acoustic signals during this active excitation process. Utilizing the synchronicity of the excitation signals and the different locations of each leakage point, the system successfully separated the characteristics of three independent acoustic source signals: high-frequency transient signal (L1: porosity leakage), mid-frequency continuous signal (L2: crack leakage), and low-frequency fluctuation signal (L3: loose interface).

[0135] Based on the separated pure signals and combined with dynamically corrected acoustic parameters, the underwater acoustic collaborative fusion positioning processor successfully calculated the precise coordinates of three leakage points using an improved Bayesian inversion algorithm: Point L1: located 15 meters southeast of the intersection, a corrosion hole at the top of the DN600 main pipeline; Point L2: located 82 meters downstream of Point L1, a crack in the pipeline joint; Point L3: located 43 meters downstream of Point L2, a failure of the flange connection seal. The maintenance team conducted precise excavation verification based on the positioning results, and the actual location of the three leakage points was less than 1.5 meters from the system positioning error.

[0136] The self-evolution module stores the entire process data of this successful pressure excitation mode, acoustic response characteristics, and precise positioning results as a complete multi-point leakage treatment sample in the special working condition sample library; the system uses this sample to incrementally train the graph neural network, especially enhancing the ability to identify the aliasing pattern of multi-point leakage signals.

[0137] When the system detects a similar complex pressure fluctuation pattern after water hammer again, it can immediately identify it as a "suspected multi-point leakage" scenario and directly optimize and call the "active hydraulic activation" process, which can improve the handling efficiency of such problems by more than 60%.

[0138] Compared to Example 3, this example demonstrates the system's unique technical advantages under extremely complex operating conditions:

[0139] The problem of signal aliasing has been solved: by active hydraulic excitation, the technical bottleneck of traditional acoustic positioning in the case of multi-source signals has been broken.

[0140] Multi-point precise positioning was achieved: three different types and locations of leakage points were successfully separated and located in a single event, and the positioning accuracy met the engineering requirements;

[0141] A closed-loop intelligent handling system has been formed: successful experiences are transformed into system intelligence, realizing a complete technical closed loop from passive response to proactive diagnosis and then to experience accumulation;

[0142] The system's adaptive capability has been enhanced: through a self-evolution mechanism, the system has the ability to continuously learn and handle special and complex working conditions.

[0143] This embodiment fully demonstrates that the system is not only suitable for conventional leakage location, but also shows unique technical value in complex scenarios where traditional methods fail, providing a more reliable technical guarantee for the safe operation of urban water supply networks.

[0144] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A water supply gateway leakage location system based on acoustic wave-pressure coordination, characterized in that, It includes a pressure sensing layer, a collaborative control unit, an acoustic sensing layer, an underwater acoustic collaborative positioning solver, and a result output module; The pressure sensing layer is used to monitor the pressure data of the pipeline network in real time through pressure sensors sparsely deployed at key nodes of the pipeline network's hydraulic sensitivity. It also incorporates a hydraulic model based on the pipeline network topology to obtain areas of pressure anomalies. The collaborative control unit is communicatively connected to the pressure sensing layer. It receives the pressure anomaly area defined by the pressure sensing layer and converts the pressure anomaly area into a minimal set of acoustic sensor node list. It uses a machine learning-based intelligent dynamic scheduling algorithm to dynamically schedule the sensors in the list to enter the working state to monitor and capture the acoustic signals generated by leakage, while controlling the sensors outside the list to enter a low-power standby state. The acoustic wave sensing layer is communicatively connected to the collaborative control unit and is used to be activated in the target area by a movable acoustic wave sensor array with edge computing capabilities, to collect the acoustic wave vibration signal of the pipeline, and to complete the signal preprocessing and feature data upload locally. The underwater acoustic collaborative positioning solver is communicatively connected to the acoustic wave sensing layer. Based on real-time collected pipeline operation data, it dynamically corrects the propagation parameters of sound waves in the pipeline. Under the spatial constraints of the pressure anomaly area, it uses the corrected acoustic parameters to perform joint inversion calculations to obtain the location of the leakage point. The result output module is communicatively connected to the underwater acoustic co-positioner and is used to output the location results of the leakage point, the leakage amount estimation, and the recommended valve shut-off scheme to the operation and maintenance platform and the pipeline network digital twin model for visualization annotation and work order dispatch.

2. The water supply network leakage location system based on acoustic wave-pressure coordination according to claim 1, characterized in that, The pressure sensing layer includes: The pressure monitoring and data acquisition unit consists of pressure sensors sparsely deployed at key nodes of the pipeline network. These key nodes are determined based on offline hydraulic sensitivity analysis and selected using the longest path principle of graph theory and a comprehensive evaluation algorithm for node hydraulic sensitivity. The unit is configured to monitor and acquire pipeline network pressure data in real time at a set sampling frequency and upload the data to the data center. The hydraulic analysis and region generation unit has a built-in hydraulic model based on the pipeline network topology. It is configured to automatically call the hydraulic model when an anomaly is detected, trace the source of the anomaly by analyzing the pressure transmission path, and output the candidate region of the pressure anomaly.

3. The system according to claim 1, characterized in that, The collaborative control unit includes: By using a pre-generated region-node mapping relationship, the received pressure anomaly region is converted into a corresponding set of acoustic sensor nodes; The minimum vertex cover algorithm based on graph theory calculates the minimum list of nodes that covers the entire abnormal region from the set of nodes; Based on the generated list of minimum nodes, a wake-up command is sent to the sensors in the list, while maintaining the low-power standby state of the sensors outside the list. It monitors the operating status of all sensors in real time and automatically activates backup sensors when a device failure is detected, so as to ensure the continuous execution of the monitoring task.

4. The water supply network leakage location system based on acoustic wave-pressure coordination according to claim 1, characterized in that, The acoustic sensing layer includes: An array of acoustic sensors configured to acquire acoustic signals within candidate regions of pressure anomalies; The signal preprocessing unit is configured to filter, reduce noise, and extract features from the raw acoustic signal locally, and upload the compressed feature data. The partitioned asynchronous monitoring unit is configured to virtually divide the acoustic sensor into several monitoring sub-regions according to the pipeline topology, and activate the sensor sequentially according to the hydraulic association priority.

5. A water supply network leakage location system based on acoustic wave-pressure coordination according to claim 1, characterized in that, The underwater acoustic cooperative localization solver includes: The hydraulic correction module is configured to dynamically calculate and correct the actual propagation speed and attenuation coefficient of sound waves in the pipeline based on real-time collected data on pipeline pressure, water temperature, pipeline material and pipe diameter. The joint inversion module is configured to perform joint inversion calculations using the pressure anomaly region as a spatial constraint, the corrected acoustic parameters as a basis, and the maximum a posteriori probability estimation method based on Bayesian theory, and output the leakage point location results. The self-evolution module is configured to collect the pressure change curves and acoustic feature vectors of the entire network after each leakage event is confirmed by on-site maintenance, form sample pairs and store them in the historical sample library, and perform online incremental training on the positioning model to achieve continuous self-optimization of the system's positioning accuracy and event recognition capabilities.

6. A water supply network leakage location system based on acoustic wave-pressure coordination according to claim 5, characterized in that, The joint inversion module specifically employs a maximum a posteriori probability estimation method based on Bayesian theory to solve for the coordinates of the leakage points, including: A probabilistic computation framework is constructed to transform the physical constraints of sound wave propagation, pipeline topology characteristics, and spatial prior information of pressure anomaly areas into probabilistic computational elements. The Markov chain Monte Carlo random sampling algorithm is used to efficiently solve the posterior probability distribution of the location of the leakage point; The output includes comprehensive positioning information containing confidence intervals and uncertainty ranges.

7. A water supply network leakage location system based on acoustic wave-pressure coordination according to claim 1, characterized in that, The result output module includes: The intelligent push engine is configured to adaptively select push channels and information detail levels based on the urgency and scope of an event. The visualization enhancement module is configured to combine the digital twin model to generate a comprehensive visualization view that includes the location of the leakage point, the affected watershed, and a schematic diagram of the valve closure scheme. The closed-loop feedback module is configured to track the status and progress of work orders, realizing a closed-loop business process from location to handling.

8. A method for locating leakage in a water supply network based on acoustic wave-pressure coordination for implementing the system described in claims 1-7, characterized in that, Includes the following steps: S1. Real-time pressure data of the pipeline network is acquired by pressure sensors sparsely deployed at key nodes of the pipeline network. The pressure data is analyzed based on a real-time hydraulic model to identify pressure anomalies and output the candidate range of pressure anomalies. S2. Based on the pressure anomaly area output in step S1, query the pre-generated area-node mapping table, convert it into a list of minimum nodes of acoustic sensors that need to be activated, dynamically schedule the corresponding sensors to enter the working state, and at the same time control the sensors outside the list to enter the low-power standby state. S3. The acoustic wave sensor activated after being scheduled in step S2 collects acoustic wave signals in the pressure anomaly area, performs preprocessing such as filtering, noise reduction and feature extraction locally, and uploads the compressed feature data. S4. Receive the pressure anomaly area data from step S1 and the acoustic characteristic data uploaded in step S3. Correct the acoustic propagation parameters using the hydraulic model. Based on the corrected acoustic parameters and with the pressure anomaly area as the spatial constraint, perform joint inversion calculation based on the acoustic time difference algorithm to output the accurate location result of the leak point. S5. Based on the accurate positioning results obtained in step S4, combined with the hydraulic model, estimate the leakage and generate a recommended valve shut-off scheme. Output the scheme to the operation and maintenance platform and the pipeline network digital twin model through a standardized interface for visualization annotation and work order dispatch.

9. A method for locating leakage in a water supply network based on acoustic wave-pressure coordination, as described in claim 8, is characterized in that... The active detection step is included after step S4: After obtaining the preliminary location of the leak, a command is sent to the system to control the relevant valves to perform micro-opening and closing operations. The acoustic signal at the leak is enhanced by controlled hydraulic changes, and the valve action feedback signal is received to verify and refine the location results.

Images