Digital twin pipe network leakage tracing method fusing GIS three-dimensional topology

Abstract

The invention relates to the technical field of pipe network leakage detection, and discloses a GIS three-dimensional topology-fused digital twin pipe network leakage tracing method, which comprises the following steps of: 1, performing parametric modeling on a water supply pipeline, a valve and a connecting piece in a target area through a building information module and a GIS platform, and constructing a digital twin body; and 2, installing various types of sensors at the water supply pipelines, the valves and the connecting pieces, marking the sensors as acquisition nodes in the three-dimensional underground pipe network model, periodically acquiring various data of the water supply pipelines, and transmitting the acquired various data of the water supply pipelines to a digital twinborn body. Multiple types of sensors are installed in a water supply pipe network to collect pressure, flow and vibration data in real time, leakage characteristic comprehensive indexes are obtained according to the data, whether real leakage exists or not is judged, and abnormal nodes are marked; and calculating a leakage suspicious index and positioning a suspicious pipeline according to the time difference and the intensity difference between the abnormal node and other acquisition nodes and the shortest water flow path.

Description

Technical Field

[0001] This invention relates to the field of pipeline network leakage detection technology, specifically a digital twin pipeline network leakage tracing method that integrates GIS three-dimensional topology. Background Technology

[0002] Pipeline networks are network systems composed of pipes, valves, and auxiliary facilities. They are mainly used to transport and distribute fluids such as water, gas, and heat media. They are a core component of urban and industrial infrastructure. Their core functions include water transport, pressure regulation, water quality assurance, and distribution scheduling, ensuring that various users receive a safe and reliable fluid supply. When a leak occurs in the urban water supply network, in most cases, maintenance personnel need to carry portable detection equipment to inspect the network one by one or conduct targeted inspections after the user reports the problem. This not only consumes a lot of manpower and time, but is also easily affected by the depth of the network, the surrounding environmental noise, and the difference in pipe materials, making it difficult to quickly locate early small leaks. Often, the leak is not discovered until the amount of leakage increases and causes obvious problems such as road flooding and a sudden drop in water pressure, resulting in a large waste of water resources and increased maintenance costs. To address the shortcomings of existing technologies, this invention proposes a digital twin pipeline leakage tracing method that integrates GIS three-dimensional topology. Summary of the Invention

[0003] The purpose of this invention is to provide a digital twin pipeline network leakage tracing method that integrates GIS three-dimensional topology, in order to solve the problems mentioned in the background art.

[0004] The objective of this invention can be achieved through the following technical solutions: A method for tracing leaks in a digital twin pipeline network that integrates GIS 3D topology includes the following steps: Step 1: Parametrically model the water supply pipes, valves, and connectors in the target area using the building information module and GIS platform to construct a digital twin; Step 2: Install various types of sensors at the water supply pipes, valves, and connectors, and mark them as acquisition nodes in the three-dimensional underground pipe network model. Periodically acquire various data of the water supply pipes and transmit the acquired data of the water supply pipes to the digital twin. Step 3: Based on the multi-sensor data acquired by each acquisition node, calculate and obtain the leakage characteristics of each acquisition node, determine whether the water supply pipeline is leaking based on the acquired leakage characteristics, trigger a leakage warning based on the judgment result, and mark the abnormal acquisition node; Step 4: After the leak warning is triggered, obtain the water supply pipeline segments that are directly or indirectly connected to the abnormal node, obtain the leakage suspicion index of each water supply pipeline segment, and select the water supply pipeline segments where multiple candidate points with the smallest leakage suspicion index are located as suspected leaking pipelines. Step 5: Locate the suspected leaking channel in the digital twin and set a virtual leak point in the digital twin to verify whether the suspected leaking channel has a real leak; Step 6: Based on the verification results, locate the leaking pipeline and notify maintenance personnel.

[0005] Preferably, the multiple types of sensors include pressure sensors, flow sensors, and vibration sensors, and the method for performing step three is as follows: The static pressure value of the acquisition node is periodically obtained through the pressure sensor. The instantaneous flow rate value of the acquisition node is periodically obtained through the flow sensor. The vibration sensor periodically acquires the vibration signal energy of the acquisition node within a preset frequency band. ; Based on the obtained static pressure value Instantaneous flow rate Vibration signal energy within the preset frequency band Calculate and obtain the data leakage characteristics during the acquisition period: Based on the currently obtained static pressure value The pressure deviation value is calculated by comparing it with the historical baseline pressure value. Based on the current instantaneous flow rate value The flow rate mutation value is obtained by calculating the instantaneous flow rate value of the adjacent period. Obtain the vibration signal energy within the current preset frequency band. Vibration ratio to total vibration signal energy ; For each data collection node, a comprehensive leakage characteristic index is calculated based on various data leakage characteristics. The comprehensive leakage characteristic index of each data collection node is compared with a set dynamic threshold. If the comprehensive leakage characteristic index of a data collection node is greater than the dynamic threshold for multiple consecutive periods, the node is determined to have a water supply pipeline leak and is marked as an abnormal data collection node.

[0006] Preferably, the method for performing step four is as follows: The average of the time point when the pressure deviation value of the abnormal acquisition node exceeds the normal pressure deviation fluctuation threshold and the time point when the vibration ratio exceeds the normal vibration ratio threshold is set as the arrival time point of the abnormal signal of the acquisition node. The time difference is calculated based on the arrival time points of the abnormal signals of adjacent acquisition nodes or adjacent abnormal acquisition nodes. The comprehensive intensity is obtained by weighted summing of the maximum pressure deviation value and the maximum vibration ratio value within multiple cycles after the abnormal acquisition node becomes abnormal. The intensity difference is then calculated based on the comprehensive intensity of adjacent acquisition nodes or adjacent abnormal nodes. A directed graph model is constructed based on all data acquisition nodes and the water supply pipelines connecting them. The nodes in the directed graph model represent all data acquisition nodes, and are recorded as follows: ; In the directed graph model, the edges represent the water supply pipe segments between each data acquisition node, and are recorded as follows: The weight of each edge is the actual physical length of the water supply pipeline segment between the data collection nodes; Instantaneous flow rate values ​​at each data collection node To determine the direction of water flow, if the data collection node... and between Then the water flows from arrive ,side This is positive, and vice versa; Traverse all water supply pipeline segments that are directly or indirectly connected to the abnormal node, divide each water supply pipeline segment into multiple candidate points, calculate the leakage point suspicion index for each candidate point based on the shortest water flow path between the abnormal acquisition node and the acquisition node, the speed of vibration wave propagation along the pipeline, and the leakage signal intensity attenuation coefficient, and select the water supply pipeline segments where the multiple candidate points with the smallest leakage suspicion index are located and mark them as suspected leakage pipelines.

[0007] Preferably, the calculation process for the leakage point suspicion index is as follows: Based on the material information of each water supply pipeline section, the propagation speed of the leakage signal along the pipeline and the attenuation coefficient of the signal strength are obtained. The adjacent collection nodes of the abnormal collection node are selected as comparison nodes, based on the shortest water flow path length, time difference, and intensity difference from the abnormal node to each comparison node.

[0008] The suspicion index is obtained by combining the ratio of the path length difference and propagation speed from the candidate point to the abnormal node and from the candidate point to the comparison node, the signal strength ratio between the abnormal node and the comparison node, the shortest water flow path length from the abnormal node to each comparison node, the time difference, and the intensity difference.

[0009] Preferably, the method for performing step five is as follows: Locate the suspected leaking pipeline in the three-dimensional pipeline network model of the digital twin, and obtain the real data of all collection nodes contained in the suspected pipeline during the leak warning period; Multiple virtual leak points are set at even intervals on each suspected leaking pipeline, and simulation parameters are set for each virtual leak point; Initiate a virtual leak simulation in the digital twin to obtain parameter changes at each acquisition node in the water supply pipeline when the virtual leak point exists, and generate simulation operation data; Compare the simulated running data generated when the virtual leak point exists with the real running data; Based on the comparison results, it is determined whether there is a real leak in the suspected leaking pipeline. If the simulated data of a virtual leak point on a suspected pipeline matches the real data in all dimensions, the matching degree exceeds the set matching threshold, and it is determined that there is a real leak at the physical location corresponding to the virtual leak point. If the simulated data and real data of all virtual leak points on the suspected pipeline have a matching degree lower than the matching threshold, it indicates that there is no real leak in the pipeline. For pipelines where leaks are confirmed, mark the precise location of the actual leak point in the digital twin.

[0010] Preferably, the process of comparing the simulated operating data generated when the virtual leak point exists with the real operating data is as follows: Based on the leak warning trigger time in the real data, the timestamp of the synchronous simulation data is used to filter the static pressure value, instantaneous flow value, and vibration signal energy in the preset frequency band in the two types of data, and the units are unified to replace the error data caused by sensor interference in the real data. For abnormal leakage nodes and multiple adjacent core nodes, compare the static pressure values, instantaneous flow values, and the direction and amplitude of vibration signal energy changes within the preset frequency band with real and simulated data. Set a qualified threshold for comparing the real data and simulated data of static pressure value, instantaneous flow value, and vibration signal energy within a preset frequency band.

[0011] Preferably, the process of replacing the error data caused by sensor interference in the real data is as follows: The real data is normalized, and the local data variance is calculated for windows through multiple collection cycles. Data within windows where the variance suddenly increases by more than three times is marked as suspicious points. Calculate the absolute value of the difference between the suspicious point and the two adjacent data points. If the absolute difference is greater than the average value of the normal period and the subsequent data drops in the next period, it is determined to be error data. Weighting coefficients are assigned to multiple normal data points adjacent to the error data to generate smooth data that replaces the error data.

[0012] Preferably, the method for performing step six is ​​as follows: generate a precise location and repair report based on the water supply pipeline leak coordinates, material, and valve parameters; match qualified repair personnel and dispatch spare parts; plan the route to the site based on road conditions; dispatch the order to the repair team and confirm the response; simultaneously inform the affected users of the water outage period and water intake point; collect data to verify the repair effect after the repair is completed; if no abnormalities are found, the repair is deemed qualified; and archive the repair record.

[0013] The beneficial effects of this invention are: 1. This invention constructs a digital twin of a water supply network by combining a building information module (BIM) with a GIS platform. Multiple sensors are installed in the water supply network to collect pressure, flow, and vibration data in real time. A comprehensive leakage characteristic index is obtained based on these data points to determine the existence of a real leak and mark abnormal nodes. A leakage suspicion index is calculated based on the time difference, intensity difference, and shortest water flow path between the abnormal node and other data collection nodes to locate suspicious pipes. Finally, multi-dimensional matching between the virtual leak point in the digital twin and the real data from the data collection nodes verifies the existence of a real leak. This invention achieves accurate source tracing, rapid location, and efficient handling of leaks in the water supply network, effectively solving the shortcomings of existing technologies that rely on manual inspection for leak detection, resulting in low efficiency and large leak location errors.

[0014] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0015] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments 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.

[0016] Figure 1 This is a flowchart illustrating the steps of a digital twin pipeline leakage tracing method integrating GIS three-dimensional topology according to the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Please see Figure 1As shown, this invention is a digital twin pipeline leakage tracing method that integrates GIS 3D topology. It integrates Building Information Modeling (BIM) to provide millimeter-level refined parameters of water supply pipelines in the target area, such as pipe diameter, material, pressure rating, valve and connector model specifications, and installation process details such as sealing type and interface method to generate a geometric model. It provides macro-geographic information such as community topography elevation, underground pipeline burial depth, and plot boundaries through the GIS platform. It establishes a spatial framework through a coordinate system and achieves centimeter-level registration between the BIM model and the GIS framework through seven-parameter transformation. In the modeling process, not only is the 3D geometric reconstruction of pipes, valves, and connectors completed, but also the pipeline network topology relationship network is constructed, the water flow branch direction and valve control range are marked, and the full life cycle attributes such as pipeline installation time and historical maintenance records are embedded to form a digital twin.

[0019] After the digital twin was constructed, high-precision pressure sensors, flow sensors, and vibration sensors were deployed and installed on the main sections and branch nodes of the water supply pipeline, before and after valves, at the boundaries of pipeline zones, and at leak-prone connections such as tees and elbows. Pressure sensors were used to acquire pressure fluctuations caused by leaks in real time, flow sensors were used to monitor abnormal flow losses, and vibration sensors were used to capture vibration signals generated by leak impacts. Subsequently, in the GIS 3D underground pipeline network model, each sensor was bound to its corresponding physical component, assigned a unique device identifier, and its spatial coordinates were accurately marked. A visual acquisition node layer with sensor type and acquisition parameters was generated. Different acquisition cycles were set according to the characteristics of the signals to be acquired. Pressure and flow data were acquired periodically every 1 minute, while vibration signals were acquired once every 10 seconds to capture instantaneous leak impacts. Before transmitting data, data cleaning and format unification were completed, and then high real-time data transmission of each acquisition node was achieved through the 5G network and pushed to the digital twin database to update the dynamic data of the acquisition nodes in the twin in real time.

[0020] Based on the data acquired from each data collection node, the leakage characteristics of each node can be obtained. These leakage characteristics mainly include pressure deviation values. Flow mutation value Vibration ratio The static pressure value obtained periodically by the pressure sensor The obtained static pressure value The pressure deviation value is calculated by comparing it with the historical baseline pressure value. , The instantaneous flow rate value of the data acquisition node is periodically obtained through a flow sensor. The instantaneous flow value will be obtained Instantaneous flow rate values ​​in adjacent periods Calculations were performed to obtain the flow rate mutation value. The vibration sensor periodically acquires the vibration signal energy of the acquisition node within a preset frequency band. The acquired vibration signal energy within the preset frequency band The vibration ratio is obtained by comparing it with the total vibration signal capability. The preset frequency band is 50-500Hz; By acquiring each leakage characteristic, the comprehensive leakage characteristic index of the acquisition node can be obtained. Through the formula:

[0021] in, This represents the standard deviation of the normal pressure at this node over the past 72 hours. This represents the average flow rate of this node under normal operating conditions over the past 72 hours. This represents the percentage of average vibration energy at this node under normal operating conditions over the past 72 hours. , , For feature weights, , , , , , This is a non-linear correction coefficient, dynamically adjusted according to the pipe material; for example, for metal pipes... , , PE pipes , , ; For example, consider the data collected by a certain data acquisition node of a community's water supply network in period t: , , Metal pipes , , Historical baseline data: , Average flow , Real-time data for period t: static pressure value Instantaneous flow rate Adjacent cycle flow The proportion of vibration energy in the preset frequency band 5; but , , but ; Set the threshold for the data collection nodes Then the current Greater than the threshold Continuous monitoring is required; if the value exceeds the threshold for three consecutive cycles... If it is, it will be marked as an abnormal data collection node, triggering a leak warning.

[0022] After a leak warning is triggered, the average of the time point when the pressure deviation value of the abnormal acquisition node exceeds the normal pressure deviation fluctuation threshold and the time point when the vibration ratio exceeds the normal vibration ratio threshold is set as the arrival time point of the abnormal signal of the acquisition node. The time difference is calculated based on the arrival time points of the abnormal signals of adjacent acquisition nodes or adjacent abnormal acquisition nodes. The time difference is the time difference between the arrival of the leakage anomaly signal at the node. With nodes The time interval is standardized, and the timestamps of all acquisition nodes are unified to ensure data timeline consistency. This is used to detect pressure signals from the acquisition nodes. The moment when the normal fluctuation threshold is first exceeded For vibration signals, detection The moment when the normal energy percentage threshold is first exceeded Take the mean of the two time points as the node. The arrival time of the abnormal signal Then abnormal nodes With any surrounding nodes The time difference is ; The comprehensive intensity is obtained by weighted summing of the maximum pressure deviation value and the maximum vibration ratio value within multiple cycles after the abnormal acquisition node becomes abnormal. The intensity difference is then calculated based on the comprehensive intensity of adjacent acquisition nodes or adjacent abnormal nodes. Intensity difference is defined as the leakage abnormal signal at the node. With nodes The intensity attenuation difference is determined by taking the maximum pressure deviation within three cycles after the abnormal moment. , The vibration intensity is taken as the maximum value of the vibration ratio within 3 cycles after the abnormal moment. = The overall strength is Then abnormal nodes With any surrounding nodes The strength difference is ; A directed graph model is constructed based on all data acquisition nodes and the water supply pipelines connecting them. The nodes in the directed graph model represent all data acquisition nodes, and are recorded as follows: ; In the directed graph model, the edges represent the water supply pipe segments between each data acquisition node, and are recorded as follows: The weight of each edge is the actual physical length of the water supply pipeline segment between the data collection nodes; Instantaneous flow rate values ​​at each data collection node To determine the direction of water flow, if the data collection node... and between Then the water flows from arrive ,side This is positive, and vice versa; Traverse all water supply pipeline segments directly or indirectly connected to the abnormal node, and divide each segment into multiple candidate points. Calculate the leakage point suspicion index for each candidate point based on the shortest water flow path between the abnormal acquisition node and the acquisition node, the velocity of the vibration wave propagating along the pipeline, and the leakage signal intensity attenuation coefficient. The leakage point suspicion index calculation process is as follows: Based on the material information of each water supply pipeline section, the propagation speed of the leakage signal along the pipeline and the attenuation coefficient of the signal strength were obtained. The propagation speed of the metal pipeline was set to 1200±50 m / s, and the propagation speed of the PE pipeline was set to 800±50 m / s. The signal strength attenuation coefficient of the metal pipeline was 0.002 m. -1 PE pipe λ=0.003m -1 ; Adjacent acquisition nodes to the abnormal acquisition node are selected as comparison nodes. The shortest water flow path length, time difference, and intensity difference from the abnormal node to each comparison node are obtained. The suspicion index is calculated by combining the ratio of the path length difference and propagation speed from the candidate point to the abnormal node to the candidate point to the comparison node, the signal strength ratio between the abnormal node and the comparison node, and the shortest water flow path length, time difference, and intensity difference from the abnormal node to each comparison node. The leakage suspicion index is selected using the formula: ,

[0023] in, Leakage point To the node The shortest water flow path length, Leakage point To the node The shortest water flow path length, The speed at which the leakage signal propagates along the pipeline. This is the attenuation coefficient of the signal strength; For example, an abnormal node is identified by its number in the water supply network of a certain residential community. With adjacent acquisition nodes For example, if the water supply pipe is a metal pipe, the signal propagation speed... attenuation coefficient m -1 Abnormal signal time difference The overall strengths are respectively , Abnormal nodes With adjacent acquisition nodes The shortest water flow path length is 300m, at the abnormal node. With adjacent acquisition nodes Three candidate points are evenly spaced on the pipe section between them: Candidate points Distance from abnormal node 100m, then , Candidate points Distance from abnormal node 150m, then , Candidate points Distance from abnormal node 200m, then , ; but , , , In conclusion, ,Right now , The leakage suspicion index is the lowest, and , All are located at abnormal nodes With adjacent acquisition nodes The pipe section between them, therefore, will be an abnormal node. With adjacent acquisition nodes The pipe segments between them are marked as suspicious pipe segments.

[0024] After marking suspected leaking pipelines, the marked pipelines are located in the 3D pipeline network model of the digital twin. Information such as pipeline ID, start and end nodes, and pipe diameter is aligned with the GIS topology data to ensure that the spatial location of the virtual verification is completely consistent with the physical pipeline network. Subsequently, based on the collected real operating data, including the key parameters of all collection nodes associated with the suspected pipeline during the leak warning period, pressure fluctuation curves, flow change trends, vibration signal characteristics, and the overall operating conditions of the pipeline network during that period, multiple virtual leak points are evenly spaced along the length of each suspected leaking pipeline. The interval is adjusted according to the pipeline length, such as setting one virtual leak point every 5 or 10 meters to ensure coverage of the entire pipeline section and avoid missing real leak points. Each virtual leak point needs to be set with simulation parameters that conform to the actual scenario: referring to the common leak types of this type of pipeline, three typical leak diameters are set, such as small leak, medium leak, and large leak. After completing the parameter settings, the virtual leak simulation is started in the digital twin to simulate the parameter changes of each collection node in the pipeline network when the virtual leak point exists, generating simulated operating data that must be completely consistent with the collection cycle and parameter type of the real data.

[0025] The simulated operating data generated from the virtual leak point is compared with the real operating data. The comparison process includes: using the leak warning trigger time in the real data as a benchmark, synchronizing the timestamp of the simulated data, filtering the static pressure value, instantaneous flow value, and vibration signal energy within a preset frequency band from both types of data, and unifying the units; replacing the error data caused by sensor interference in the real data. The process of replacing the error data caused by sensor interference in the real data is as follows: The real data is normalized to the [0,1] interval. The local data variance is calculated using five adjacent acquisition cycles as windows. Data within the window whose variance increases by more than three times is marked as suspicious points. After marking suspicious points, the absolute value of the difference between the suspicious point and the two adjacent data points is calculated. If the absolute value of the difference is five times greater than the average value of the normal period and the subsequent data falls back in an adjacent cycle, it is determined that the data in that cycle is affected by sensor interference, resulting in error data. After determining the error data, weights are assigned to the three normal data points adjacent to the error data. The closer the collected data is to the error data, the higher the weight. Smooth data is generated to replace the error data, and the mean square error between the processed data and the normal time period data is calculated.

[0026] For the abnormal leakage node and multiple adjacent core nodes, compare the static pressure values, instantaneous flow values, and the direction and amplitude of vibration signal energy changes within the preset frequency band with real and simulated data: Trend Comparison: If the pressure of an abnormal node in the real data decreases by 8% within 10 minutes after the warning and then stabilizes at a low pressure value, check whether the node in the simulated data shows the same decrease duration and magnitude. If the pressure of surrounding nodes in the real data fluctuates slightly due to leakage, the simulated data should reflect this fluctuation range synchronously, rather than showing no fluctuation or excessive fluctuation. If the upstream flow of an abnormal node in the real data increases by 4% due to leakage replenishment and the downstream flow decreases by 3% due to leakage loss, the simulated data should match the change in flow distribution, and the increase or decrease should not deviate from the real data by more than 1%. If the simulated change magnitude of a certain parameter deviates from the real change magnitude by less than or equal to 5%, the change direction is consistent, and the stabilization time deviation is less than or equal to 10%, then the trend matching degree of that parameter is scored as 100 points. For each deviation threshold exceeding the threshold, 20 points are deducted. Finally, the average score of all parameters is taken as the node trend matching degree.

[0027] Feature Comparison: If over 80% of the energy of the vibration signal corresponding to leakage in the real data is concentrated in the 120–180 Hz frequency band, then the proportion of vibration amplitude in this frequency band in the simulated data must be consistent with the real data, and the amplitude of the non-leakage frequency band must be synchronously at a low level. If the vibration signal in the real data detects an anomaly 1–2 seconds before the pressure signal, the simulated data must show the same timing relationship of vibration triggering first and pressure responding later, rather than the timing being reversed or synchronously triggered. Regarding the proportion of vibration frequency bands, if the energy proportion in the 50–500 Hz frequency band in the real data is 65%, and the simulated data is 60% or 70%, then the frequency band feature matches. If the vibration in the real data is 1.5 seconds earlier than the pressure, and the deviation in the simulated data is less than or equal to 0.3 seconds, then the timing feature matches. The frequency band concentration and trigger timing features are allocated with weights of 0.4 and 0.6, respectively. Each matching feature is scored 100 points, and each non-matching feature is scored 0 points. The feature matching degree is obtained by weighted summation.

[0028] Operating condition coordination comparison: If the total water supply flow of the water plant increases by 3.5% due to leakage in the real data, the total flow in the simulation data must reflect this increase, and the time of the increase must be consistent with the real data. If the pressure of the node 100 meters away from the suspected pipeline does not change significantly in the real data, the pressure of that node in the simulation data must remain stable to avoid unreasonable situations such as abnormal pressure at distant nodes. The total flow coordination and the coordination of distant node pressure are assigned weight coefficients of 0.6 and 0.4 respectively. Each indicator is scored out of 100 points, and 10 points are deducted for every 1% deviation. The weighted result is the operating condition coordination degree.

[0029] Qualification thresholds are set for trend matching degree, feature matching degree, and operating condition coordination degree: when the trend matching degree is greater than or equal to 90%, the feature matching degree is greater than or equal to 85%, and the operating condition coordination degree is greater than or equal to 80%, the virtual leak point is qualified. Then, the comprehensive matching degree of the virtual leak point is calculated according to the weight allocation of trend matching degree 0.4, feature matching degree 0.4, and operating condition coordination degree 0.2. If the comprehensive matching degree exceeds 85%, the virtual leak point has a real leak. For example, if the trend matching degree of a virtual leak point is 92%, the feature matching degree is 88%, and the operating condition coordination degree is 82%, then the comprehensive matching degree = 92%×0.4 + 88%×0.4 + 82%×0.2 = 89.2%. 89.2% is greater than 85%, so the virtual leak point has a real leak.

[0030] After identifying the virtual leak point, the actual leak point is located in the digital twin. The 3D GIS coordinates are extracted and linked to identifiable reference information on site. At the same time, core maintenance parameters such as pipe material, upstream and downstream valve numbers, and current pipe pressure are integrated to generate a precise leak location report. The report includes screenshots of the digital twin 3D model and photos of the site environment. Qualified maintenance personnel are matched, spare parts are dispatched, and a route to the site is planned based on road conditions. The work order is assigned to the maintenance team and the response is confirmed. Affected users are informed of the water outage period and water intake point. After the maintenance is completed, data is collected to verify the repair effect. If no abnormalities are found, the repair is deemed qualified and the maintenance record is archived.

[0031] The above description is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined in the claims, they should all fall within the protection scope of the present invention.

Claims

1. A method for tracing leaks in a digital twin pipeline network integrating GIS 3D topology, characterized in that, Includes the following steps: Step 1: Parametrically model the water supply pipes, valves, and connectors in the target area using the building information module and GIS platform to construct a digital twin; Step 2: Install various types of sensors at the water supply pipes, valves, and connectors, and mark them as acquisition nodes in the three-dimensional underground pipe network model. Periodically acquire various data of the water supply pipes and transmit the acquired data of the water supply pipes to the digital twin. Step 3: Based on the multi-sensor data acquired by each acquisition node, calculate and obtain the leakage characteristics of each acquisition node, determine whether the water supply pipeline is leaking based on the acquired leakage characteristics, trigger a leakage warning based on the judgment result, and mark the abnormal acquisition node; Step 4: After the leak warning is triggered, obtain the water supply pipeline segments that are directly or indirectly connected to the abnormal node, obtain the leakage suspicion index of each water supply pipeline segment, and select the water supply pipeline segments where multiple candidate points with the smallest leakage suspicion index are located as suspected leaking pipelines. Step 5: Locate the suspected leaking channel in the digital twin and set a virtual leak point in the digital twin to verify whether the suspected leaking channel has a real leak; Step 6: Based on the verification results, locate the leaking pipeline and notify maintenance personnel.

2. The method for tracing leaks in a digital twin pipeline network integrating GIS three-dimensional topology as described in claim 1, characterized in that, The various types of sensors include pressure sensors, flow sensors, and vibration sensors. The method for performing step three is as follows: The static pressure value of the acquisition node is periodically obtained through the pressure sensor. The instantaneous flow rate value of the acquisition node is periodically obtained through the flow sensor. The vibration sensor periodically acquires the vibration signal energy of the acquisition node within a preset frequency band. ; Based on the obtained static pressure value Instantaneous flow rate Vibration signal energy within the preset frequency band Calculate and obtain the data leakage characteristics during the acquisition period: Based on the currently obtained static pressure value The pressure deviation value is calculated by comparing it with the historical baseline pressure value. Based on the current instantaneous flow rate value The flow rate mutation value is obtained by calculating the instantaneous flow rate value of the adjacent period. Obtain the vibration signal energy within the current preset frequency band. Vibration ratio to total vibration signal energy ; For each data collection node, a comprehensive leakage characteristic index is calculated based on various data leakage characteristics. The comprehensive leakage characteristic index of each data collection node is compared with a set dynamic threshold. If the comprehensive leakage characteristic index of a data collection node is greater than the dynamic threshold for multiple consecutive periods, the node is determined to have a water supply pipeline leak and is marked as an abnormal data collection node.

3. The method for tracing leaks in a digital twin pipeline network integrating GIS three-dimensional topology according to claim 1, characterized in that, The method for performing step four is as follows: The average of the time point when the pressure deviation value of the abnormal acquisition node exceeds the normal pressure deviation fluctuation threshold and the time point when the vibration ratio exceeds the normal vibration ratio threshold is set as the arrival time point of the abnormal signal of the acquisition node. The time difference is calculated based on the arrival time points of the abnormal signals of adjacent acquisition nodes or adjacent abnormal acquisition nodes. The comprehensive intensity is obtained by weighted summing of the maximum pressure deviation value and the maximum vibration ratio value within multiple cycles after the abnormal acquisition node becomes abnormal. The intensity difference is then calculated based on the comprehensive intensity of adjacent acquisition nodes or adjacent abnormal nodes. A directed graph model is constructed based on all data acquisition nodes and the water supply pipelines connecting them. The nodes in the directed graph model represent all data acquisition nodes, and are recorded as follows: ; In the directed graph model, the edges represent the water supply pipe segments between each data acquisition node, and are recorded as follows: The weight of each edge is the actual physical length of the water supply pipeline segment between the data collection nodes; Instantaneous flow rate values ​​at each data collection node To determine the direction of water flow, if the data collection node... and between Then the water flows from arrive ,side This is positive, and vice versa; Traverse all water supply pipeline segments that are directly or indirectly connected to the abnormal node, divide each water supply pipeline segment into multiple candidate points, calculate the leakage point suspicion index for each candidate point based on the shortest water flow path between the abnormal acquisition node and the acquisition node, the speed of vibration wave propagation along the pipeline, and the leakage signal intensity attenuation coefficient, and select the water supply pipeline segments where the multiple candidate points with the smallest leakage suspicion index are located and mark them as suspected leakage pipelines.

4. The method for tracing leaks in a digital twin pipeline network integrating GIS three-dimensional topology according to claim 3, characterized in that, The calculation process for the suspicion index of the leakage point is as follows: Based on the material information of each water supply pipeline section, the propagation speed of the leakage signal along the pipeline and the attenuation coefficient of the signal strength are obtained. Select adjacent collection nodes of abnormal collection nodes as comparison nodes, and obtain the shortest water flow path length, time difference, and intensity difference from the abnormal node to each comparison node; The suspicion index is obtained by combining the ratio of the path length difference and propagation speed from the candidate point to the abnormal node and from the candidate point to the comparison node, the signal strength ratio between the abnormal node and the comparison node, the shortest water flow path length from the abnormal node to each comparison node, the time difference, and the intensity difference.

5. The method for tracing leaks in a digital twin pipeline network integrating GIS three-dimensional topology according to claim 1, characterized in that, The method for performing step five is as follows: Locate the suspected leaking pipeline in the three-dimensional pipeline network model of the digital twin, and obtain the real data of all collection nodes contained in the suspected pipeline during the leak warning period; Multiple virtual leak points are set at even intervals on each suspected leaking pipeline, and simulation parameters are set for each virtual leak point; Initiate a virtual leak simulation in the digital twin to obtain parameter changes at each acquisition node in the water supply pipeline when the virtual leak point exists, and generate simulation operation data; Compare the simulated running data generated when the virtual leak point exists with the real running data; Based on the comparison results, it is determined whether there is a real leak in the suspected leaking pipeline. If the simulated data of a virtual leak point on a suspected pipeline matches the real data in all dimensions, the matching degree exceeds the set matching threshold, and it is determined that there is a real leak at the physical location corresponding to the virtual leak point. If the simulated data and real data of all virtual leak points on the suspected pipeline have a matching degree lower than the matching threshold, it indicates that there is no real leak in the pipeline. For pipelines where leaks are confirmed, mark the precise location of the actual leak point in the digital twin.

6. The method for tracing leaks in a digital twin pipeline network integrating GIS three-dimensional topology according to claim 5, characterized in that, The process of comparing the simulated running data generated when the virtual leak point exists with the real running data is as follows: Based on the leak warning trigger time in the real data, the timestamp of the synchronous simulation data is used to filter the static pressure value, instantaneous flow value, and vibration signal energy in the preset frequency band in the two types of data, and the units are unified to replace the error data caused by sensor interference in the real data. For abnormal leakage nodes and multiple adjacent core nodes, compare the static pressure values, instantaneous flow values, and the direction and amplitude of vibration signal energy changes within the preset frequency band with real and simulated data. Set a qualified threshold for comparing the real data and simulated data of static pressure value, instantaneous flow value, and vibration signal energy within a preset frequency band.

7. The method for tracing leaks in a digital twin pipeline network integrating GIS three-dimensional topology according to claim 6, characterized in that, The process of replacing the error data caused by sensor interference in the real data is as follows: The real data is normalized, and the local data variance is calculated for windows through multiple collection cycles. Data within windows where the variance suddenly increases by more than three times is marked as suspicious points. Calculate the absolute value of the difference between the suspicious point and the two adjacent data points. If the absolute difference is greater than the average value of the normal period and the subsequent data drops in the next period, it is determined to be error data. Weighting coefficients are assigned to multiple normal data points adjacent to the error data to generate smooth data that replaces the error data.

8. The method for tracing leaks in a digital twin pipeline network integrating GIS three-dimensional topology according to claim 1, characterized in that, The method for performing step six is ​​as follows: generate a precise location and repair report based on the coordinates, material, and valve parameters of the water supply pipeline leak; match qualified repair personnel and dispatch spare parts; plan the route to the site based on road conditions; dispatch the order to the repair team and confirm the response; simultaneously inform the affected users of the water outage period and water intake point; collect data to verify the repair effect after the repair is completed; if no abnormalities are found, it is deemed qualified; and archive the repair record.

Images