Water network monitoring method and system based on intelligent water supply and drainage

Abstract

The invention provides a water network monitoring method and system based on intelligent water supply and drainage, and the method comprises the steps: obtaining multi-source heterogeneous monitoring data of a water network in real time, carrying out the unified format conversion and time-space synchronization processing of the multi-source heterogeneous monitoring data based on a protocol adaptive fusion mechanism, and obtaining a standardized multi-source time series data set; performing exception identification processing on the standardized multi-source time sequence data set based on the time-varying statistical characteristics and the dynamic control limit to obtain a candidate exception event set; performing multi-sensor cross validation processing on the candidate abnormal event set based on a pipe network topological space association knowledge graph to obtain a high-confidence abnormal event set; performing reverse scene matching analysis on the high-confidence abnormal event set based on a hydraulic numerical simulation model to obtain a preliminary positioning area; performing positioning optimization processing on the preliminary positioning area based on a graph theory pressure wave propagation algorithm to obtain an abnormal position; and performing multi-dimensional fusion display processing on the abnormal position to generate a monitoring report. According to the method, accurate positioning of the abnormal position is realized.

Description

Technical Field

[0001] This application relates to the field of equipment monitoring technology, and in particular to a water network monitoring method and system based on smart water supply and drainage. Background Technology

[0002] With the acceleration of urbanization and the continuous growth of water supply demand, urban water supply networks are becoming increasingly large and complex. To ensure water supply safety and service quality, various monitoring devices are widely deployed in the water supply network, including smart water meters, pressure sensors, flow meters, and water quality monitors. These devices collect network operation data in real time, providing data support for network operation and management. Meanwhile, with the development of Internet of Things (IoT) technology, the types and quantities of water supply network monitoring devices have formed a complex data acquisition network.

[0003] However, due to the large scale, deep burial, and complex structure of water supply networks, existing monitoring equipment often can only determine the general area where the anomaly occurs when it is detected, making it difficult to accurately pinpoint the specific location of the anomaly. This situation forces maintenance personnel to conduct investigations over a large area, which not only prolongs the troubleshooting time and increases maintenance costs but also affects the quality of water supply services. Summary of the Invention

[0004] This application provides a water network monitoring method and system based on intelligent water supply and drainage, in order to solve the problems mentioned in the background art.

[0005] Firstly, this application provides a water network monitoring method based on smart water supply and drainage, including: Real-time acquisition of multi-source heterogeneous monitoring data of water network, and unified format conversion and spatiotemporal synchronization processing of the multi-source heterogeneous monitoring data based on protocol adaptive fusion mechanism to obtain standardized multi-source time series dataset; Anomaly identification processing is performed on the standardized multi-source time-series dataset based on time-varying statistical features and dynamic control limits to obtain a candidate set of abnormal events. The candidate abnormal event set is subjected to multi-sensor cross-validation processing based on the knowledge graph of pipeline topology spatial association to obtain a high-confidence abnormal event set. Based on the hydraulic numerical simulation model, a reverse scenario matching analysis was performed on the set of high-confidence abnormal events to obtain the preliminary location area; The preliminary location area is optimized based on the graph theory pressure wave propagation algorithm to obtain the abnormal location. The abnormal locations are processed through multi-dimensional fusion display to generate a monitoring report.

[0006] In one possible implementation, the anomaly identification processing of the standardized multi-source time-series dataset based on time-varying statistical features and dynamic control limits to obtain a candidate set of anomaly events includes: The standardized multi-source time series dataset is divided into sliding time windows to obtain time window data; The time window data is processed by statistical analysis methods to obtain a time-varying statistical feature vector. Based on the historical data distribution, the time-varying statistical feature vector is processed to calculate the control limit, thereby obtaining the dynamic control limit parameters; An anomaly detection result is obtained by performing anomaly detection processing on the dynamic control limit parameters based on a multi-level detection strategy. The anomaly detection results are subjected to confidence scoring to obtain the anomaly confidence level; The set of candidate abnormal events is obtained by filtering the anomaly confidence based on the confidence threshold.

[0007] In one possible implementation, the candidate anomaly event set is subjected to multi-sensor cross-validation processing based on the pipeline network topology spatial association knowledge graph to obtain a high-confidence anomaly event set, including: Based on the candidate abnormal event set and the pipeline network topology, a pipeline network topology spatial association knowledge graph is constructed to obtain node attribute data and edge attribute data. Based on the graph theory path search algorithm, the node attribute data and edge attribute data are processed by topological distance calculation to obtain the correlation strength coefficient between monitoring points; Based on the correlation strength coefficient, a neighboring point search process is performed on the abnormal monitoring points to obtain neighboring monitoring data; Based on the principles of pipeline hydraulics, the abnormal behavior verification processing of the adjacent monitoring data was performed to obtain the verification results. The verification results are subjected to consistency scoring to obtain a verification score; The verification scores are filtered based on the verification score threshold to obtain the set of high-confidence abnormal events.

[0008] In one possible implementation, the inverse scenario matching analysis of the high-confidence anomaly event set based on the hydraulic numerical simulation model to obtain the preliminary location area includes: Based on the characteristics of the high-confidence anomaly event set, the anomaly type and boundary conditions are determined to obtain an anomaly scenario template; The abnormal scenario template is subjected to search range division processing to obtain a candidate location set; The candidate location set is simulated point by point using a hydraulic simulation model to obtain theoretical response data. The theoretical response data is compared and analyzed using a multi-dimensional similarity evaluation method to obtain a matching score; The matching scores are then sorted to obtain the optimal matching position. Based on the optimal matching location, the area is delineated to obtain the preliminary positioning area.

[0009] In one possible implementation, the graph-based pressure wave propagation algorithm is used to optimize the initial location area to obtain the abnormal location, including: The preliminary positioning area is divided into grids to obtain candidate location grids; The pressure wave propagation velocity of the candidate location grid is calculated based on the pipeline network physical parameters to obtain propagation velocity data. Based on the propagation speed data, the theoretical propagation time of each monitoring node is calculated to obtain time series data; The time series data is subjected to consistency evaluation processing to obtain a time consistency score; Based on the intensity of the anomaly's impact, the temporal consistency score is subjected to spatial constraint verification to obtain a spatial consistency score. The anomaly location is obtained by performing multi-objective optimization based on the time consistency score and spatial consistency score.

[0010] In one possible implementation, the multi-dimensional fusion display processing of the abnormal location to generate a monitoring report includes: The abnormal locations are labeled using a 3D geographic information system to obtain location labeling data. The influence range data is calculated based on the location annotation data. Based on the data of the area of ​​influence, a heat map is drawn to obtain spatial distribution data; Based on the spatial distribution data, a report template is selected to obtain the report template data; The monitoring report is obtained by populating the data in the report template with information.

[0011] Secondly, this application provides a water network monitoring system based on intelligent water supply and drainage, comprising: The acquisition module is used to acquire multi-source heterogeneous monitoring data of the water network in real time, and perform unified format conversion and spatiotemporal synchronization processing on the multi-source heterogeneous monitoring data based on the protocol adaptive fusion mechanism to obtain a standardized multi-source time series dataset. An anomaly identification module is used to perform anomaly identification processing on the standardized multi-source time series dataset based on time-varying statistical features and dynamic control limits to obtain a candidate anomaly event set; The cross-validation processing module is used to perform multi-sensor cross-validation processing on the candidate abnormal event set based on the pipeline topology spatial association knowledge graph to obtain a high-confidence abnormal event set. The matching analysis module is used to perform reverse scenario matching analysis on the high-confidence abnormal event set based on the hydraulic numerical simulation model to obtain the preliminary location area; The positioning module is used to perform positioning optimization processing on the preliminary positioning area based on the graph theory pressure wave propagation algorithm to obtain the abnormal location; The generation module is used to perform multi-dimensional fusion display processing on the abnormal locations and generate a monitoring report.

[0012] This application provides a water network monitoring method and system based on smart water supply and drainage. The method includes: acquiring multi-source heterogeneous monitoring data of the water network in real time; performing unified format conversion and spatiotemporal synchronization processing on the multi-source heterogeneous monitoring data based on a protocol adaptive fusion mechanism to obtain a standardized multi-source time-series dataset; performing anomaly identification processing on the standardized multi-source time-series dataset based on time-varying statistical features and dynamic control limits to obtain a candidate anomaly event set; performing multi-sensor cross-validation processing on the candidate anomaly event set based on a pipeline topology spatial association knowledge graph to obtain a high-confidence anomaly event set; performing reverse scenario matching analysis on the high-confidence anomaly event set based on a hydraulic numerical simulation model to obtain a preliminary location area; performing location optimization processing on the preliminary location area based on a graph theory pressure wave propagation algorithm to obtain the anomaly location; and performing multi-dimensional fusion display processing on the anomaly location to generate a monitoring report. On the one hand, by introducing an adaptive fusion mechanism based on protocols and an anomaly identification method based on time-varying statistical features, unified processing of multi-source heterogeneous monitoring data and rapid discovery of abnormal events are achieved, improving data processing efficiency and anomaly detection accuracy. On the other hand, by establishing a knowledge graph of pipeline topology spatial association and a hydraulic numerical simulation model, multi-sensor cross-validation and reverse scenario matching analysis of abnormal events are achieved, reducing false alarm rate and narrowing the location range of abnormal events. Furthermore, by combining graph theory pressure wave propagation algorithm and multi-dimensional fusion display processing, accurate location and visualization of abnormal locations are achieved, shortening fault location time and reducing maintenance costs. Attached Figure Description

[0013] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 A flowchart illustrating the water network monitoring method based on smart water supply and drainage provided in this application embodiment; Figure 2 A schematic block diagram of the structure of a water network monitoring system based on smart water supply and drainage provided in the embodiments of this application; Figure 3A schematic block diagram of the structure of a terminal device provided in an embodiment of this application. Detailed Implementation

[0015] 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, not all, of the embodiments of the present invention. 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.

[0016] The flowchart shown in the attached diagram is for illustrative purposes only and does not necessarily include all content and operations / steps, nor does it necessarily have to be performed in the order described. For example, some operations / steps can be broken down, combined, or partially merged, so the actual execution order may change based on the actual situation.

[0017] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the scope of the application. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0018] It should also be further understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the relevant listed items and all possible combinations, and includes such combinations.

[0019] The following detailed description of some embodiments of this application is provided in conjunction with the accompanying drawings. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0020] Please see Figure 1 , Figure 1 A flowchart illustrating the water network monitoring method based on smart water supply and drainage provided in this application embodiment is shown below. Figure 1 As shown, the water network monitoring method based on smart water supply and drainage provided in this application includes steps S1 to S6.

[0021] Step S1: Acquire multi-source heterogeneous monitoring data of the water network in real time, and perform unified format conversion and spatiotemporal synchronization processing on the multi-source heterogeneous monitoring data based on the protocol adaptive fusion mechanism to obtain a standardized multi-source time series dataset.

[0022] It should be noted that the multi-source heterogeneous monitoring data includes water consumption data from smart water meters, pressure value data from pressure sensors, instantaneous flow data from flow meters, water quality parameter data from water quality monitors, and equipment status data from SCADA control systems. The protocol adaptive fusion mechanism refers to a processing mechanism that automatically determines the communication protocol type of the access device through a device type identification module, and calls the corresponding protocol parsing engine to decode the data, uniformly converting the private data formats of different devices into a predefined standard format.

[0023] In step S1, firstly, a multi-protocol compatible data acquisition and access gateway is established. Multi-source heterogeneous monitoring data is acquired in real time through a protocol adaptive fusion mechanism. Then, time synchronization processing is performed on the collected raw data. A network time protocol server is used as the standard time source, and the timestamps of all data are uniformly calibrated to millisecond-level accuracy. Secondly, data format standardization conversion is performed, converting data of different formats into a unified JSON structured format, including unified fields such as device identifier, timestamp, numerical type, and measurement unit. Finally, data integrity verification is performed, classifying problematic data to form a standardized multi-source time-series dataset. This processing method effectively solves the problem of fusion of multi-source heterogeneous data, ensuring the timeliness and consistency of the data.

[0024] For example, a city's water supply network is equipped with various monitoring devices. Pressure sensors using the MODBUS-RTU protocol collect network pressure data every second. Smart water meters based on the LoRaWAN protocol upload water consumption data every 5 minutes, and flow meters connected via Ethernet record flow data every 30 seconds. Upon receiving the pressure sensor data packet, the data acquisition gateway identifies it as part of the MODBUS-RTU protocol, calls the corresponding parser to decode it, and obtains the original pressure value of 0.45 MPa, device number PS001, and acquisition time 2024-01-01 10:00:00.123. Upon receiving data from the smart water meter, it identifies it as part of the LoRaWAN protocol, decodes it to obtain water consumption of 2.5 m³, water meter number WM023, and time 2024-01-01 10:00:05.000. Flow meter data is transmitted via TCP / IP protocol. Parsing yields a flow rate of 80 m³ / h, flow meter number FM056, and time 2024-01-01 10:00:15.500. The timestamps of this data are then uniformly calibrated to the network time server time and converted to standard JSON format: {"device_id":"PS001", "type":"pressure", "value":0.45, "unit":"MPa", "timestamp":"2024-01-01 10:00:00.123"}, {"device_id":"WM023", "type":"waterconsumption", "value":2.5, "unit":"m³", "timestamp":"2024-01-01 10:00:05.000"}, {"device_id":"FM056", "type":"water flow", "value":80, "unit":"m³ / h", The data standardization process is completed using the format "timestamp":"2024-01-01 10:00:00.123".

[0025] Step S2: Based on time-varying statistical features and dynamic control limits, perform anomaly identification processing on the standardized multi-source time series dataset to obtain a candidate set of abnormal events.

[0026] It should be noted that the time-varying statistical characteristics refer to statistical measures such as the mean, variance, skewness, kurtosis, and quantiles of the data calculated within a sliding time window, used to describe the distribution characteristics of the data. The dynamic control limits refer to the reasonable range boundaries of the data that are dynamically adjusted based on the statistical characteristics and seasonal variation patterns of historical data, used to determine whether the data is abnormal.

[0027] In step S2, firstly, a sliding time window is set for the standardized data, with an appropriate window length determined according to different parameter types. The time-varying statistical characteristics of the data within the window are calculated. Then, a dynamic control limit calculation model is established based on the historical data distribution, considering the periodic changes in the data, and the control limit parameters are dynamically updated. Secondly, a multi-level detection strategy is used for anomaly identification, including single-point anomaly detection, trend anomaly detection, and correlation anomaly detection. Finally, the detection results are scored with confidence. When the confidence of an anomaly event is greater than 0.6, it is identified as a candidate anomaly event, generating a candidate anomaly event set. This method avoids the misjudgment problem caused by fixed thresholds and improves the accuracy of anomaly detection.

[0028] For example, the normal operating pressure range of a pressure monitoring node in a water supply network is between 0.4 and 0.5 MPa. The system sets a 1-hour sliding time window and calculates the mean pressure data within the window as 0.45 MPa, standard deviation as 0.02 MPa, skewness as 0.1, and kurtosis as 3.0. Based on historical data from the same period, the system calculates the dynamic control upper limit as 0.52 MPa and the lower limit as 0.38 MPa. When the pressure value drops to 0.35 MPa at a certain moment, a single-point anomaly detection is triggered. If the pressure continues to drop within the next 15 minutes, a trend anomaly detection is triggered. At the same time, an abnormal increase in flow rate is observed at an adjacent flow monitoring node, triggering a correlation anomaly detection. Combining the results of the three layers of detection, the confidence level of the anomaly event is calculated to be 0.85, and it is included in the candidate anomaly event set as a highly suspected pipeline leakage event.

[0029] Step S3: Perform multi-sensor cross-validation processing on the candidate abnormal event set based on the pipeline network topology spatial association knowledge graph to obtain a high-confidence abnormal event set.

[0030] It should be noted that the aforementioned pipeline topology spatial association knowledge graph refers to the knowledge representation of elements such as monitoring nodes, pipeline connection relationships, and valve control points in the pipeline network as a graph theory structure. Among them, the attributes of monitoring nodes include geographical coordinates, equipment type, pipe diameter information, etc., and the edge attributes include physical characteristic parameters such as pipe length, resistance coefficient, and connectivity status.

[0031] In step S3, firstly, the topological distance between any two monitoring nodes is calculated based on the knowledge graph of pipeline topology spatial association, establishing a spatial association strength evaluation mechanism. Then, when an abnormal event is detected, its topologically associated neighboring monitoring nodes are automatically searched, and the presence of corresponding abnormal behavior at neighboring points is verified according to the principles of pipeline hydraulics. Secondly, a verification consistency scoring mechanism is established, comprehensively considering factors such as the synchronicity of the anomaly occurrence time, the rationality of the anomaly magnitude, and the matching of the anomaly type to calculate the verification score. Finally, the abnormal events are screened based on the verification score. When the score of an abnormal event is greater than 0.7, it is identified as a high-confidence abnormal event, forming a high-confidence abnormal event set. This method based on multi-sensor cross-validation helps reduce the false alarm rate and improve the reliability of anomaly identification.

[0032] For example, a suspected pipe rupture occurred in area A of a city's water supply network. Monitoring node P1 detected a sharp drop in pressure from 0.45 MPa to 0.32 MPa. Based on the network topology spatial association knowledge graph, three monitoring nodes directly connected to P1 were identified: upstream pressure monitoring node P2 (200 meters away), downstream pressure monitoring node P3 (150 meters away), and downstream flow monitoring node F1 (180 meters away). Analysis showed that the pressure at point P2 rose slightly to 0.48 MPa, and the pressure at point P3 subsequently dropped to 0.35 MPa. The flow rate at point F1 increased by 20%, which is consistent with the hydraulic response characteristics caused by the pipe rupture. The upstream pressure at the rupture point rose slightly, the downstream pressure dropped significantly, and the downstream flow rate increased significantly. The consistency score of the calculation verification is as follows: time synchronization score 0.9 (the time difference of pressure change is consistent with the propagation delay), amplitude reasonableness score 0.85 (the amplitude of pressure change is consistent with hydraulic calculation), and type matching score 0.95 (the changes of multiple parameters are consistent with the characteristics of pipe rupture). The comprehensive verification score is 0.9, which is greater than 0.7. Therefore, this abnormal event is included in the high-confidence abnormal event set.

[0033] Step S4: Perform reverse scenario matching analysis on the high-confidence abnormal event set based on the hydraulic numerical simulation model to obtain the preliminary location area.

[0034] It should be noted that the hydraulic numerical simulation model refers to a mathematical model established based on the pipeline network topology and physical parameters. It includes information such as pipeline geometry, material properties, monitoring node elevations, and boundary conditions, and is capable of simulating the hydraulic response characteristics of the pipeline network under different abnormal conditions. The inverse scenario matching analysis refers to the process of setting different types of abnormal boundary conditions at possible fault locations, performing hydraulic simulation calculations, and comparing the simulation results with actual monitoring data to determine the most likely fault area.

[0035] In step S4, firstly, based on the characteristics of high-confidence anomaly events, corresponding anomaly scenario templates are established in the hydraulic simulation model, including boundary condition setting schemes for typical anomaly types such as pipe bursts, pipe blockages, and illegal water extraction. Then, candidate locations are selected within the suspected fault area, and hydraulic simulation calculations are performed sequentially to obtain the theoretical response data for each monitoring node. Secondly, a multi-dimensional similarity evaluation method is used to compare the numerical deviation, trend, and peak characteristics of the simulation data with the measured data. Finally, the preliminary location area is determined based on the comprehensive matching score. This method, through comparison and verification with simulation and measured data, helps to narrow down the fault location range and improve the reliability of the location.

[0036] For example, for the high-confidence pipeline rupture event in area A identified in step S3, a circular area with a radius of 300 meters centered on monitoring node P1 is selected as the initial search range in the hydraulic simulation model. Within this initial search range, a candidate location is selected every 20 meters, for a total of 90 simulation monitoring nodes. For each monitoring node, the system sets the same scale of pressure leakage boundary conditions (setting the monitoring node pressure to 0.1 MPa and the flow rate to 80 m³ / h), and performs steady-state hydraulic calculations. Taking point P1 as an example, when the leakage condition is set at the actual rupture point (approximately 100 meters from P1), the simulation shows that the pressure at point P1 is 0.3 MPa. The simulated pressure at point P1 is 3 MPa, close to the measured value of 0.32 MPa. The simulated pressure at point P2 is 0.47 MPa, close to the measured value of 0.48 MPa. The simulated pressure at point P3 is 0.34 MPa, close to the measured value of 0.35 MPa. The simulated flow rate at point F1 increases by 18%, which is basically consistent with the measured increase of 20%. The comprehensive matching score is calculated as follows: numerical difference score 0.92 (average deviation between simulated and measured values ​​is less than 5%), trend consistency score 0.88 (pressure change direction is completely consistent), and peak value correspondence score 0.85 (pressure drop time conforms to the propagation law). The overall matching degree at this location is 0.88, the highest among all candidate locations. Considering model error, an 80-meter radius around this point is defined as the preliminary positioning area, providing spatial constraints for subsequent precise positioning.

[0037] Step S5: Based on the graph theory pressure wave propagation algorithm, perform positioning optimization processing on the preliminary positioning area to obtain the abnormal location.

[0038] It should be noted that the graph theory pressure wave propagation algorithm refers to the method of abstracting the pressure disturbance propagation process in the pipeline network into a signal propagation problem in graph theory. By analyzing the propagation speed and attenuation characteristics of pressure waves in pipes of different diameters and materials, and combining the time series of anomalies detected by monitoring nodes, the location of the anomaly can be deduced.

[0039] In step S5, firstly, the preliminary location area is meshed to establish a dense grid of candidate locations. Then, based on the pipeline network's physical parameters, the propagation velocity of pressure waves in various types of pipelines is calculated, considering the influence of factors such as pipe diameter, pipe wall thickness, and material elastic modulus on propagation characteristics. Secondly, time inversion technology is used to analyze the temporal characteristics of anomalies detected by each monitoring node, calculating the consistency between theoretical propagation time and actual detection time. Simultaneously, a spatial constraint verification mechanism is introduced, optimizing based on the spatial distribution law of anomaly influence intensity. Finally, a multi-objective optimization method is used to comprehensively consider temporal and spatial consistency to determine the optimal anomaly location. This location method based on pressure wave propagation characteristics helps improve the accuracy of anomaly location.

[0040] For example, for the preliminary positioning area determined in step S4, a uniform grid of 2m × 2m is first established to form multiple candidate location points. For each candidate location point, the theoretical propagation time of the pressure wave to each monitoring node is calculated: the propagation speed of the pressure wave to point P1 is 1200m / s, the propagation distance is 100m, and the theoretical time is 0.083 seconds; the propagation speed of the pressure wave to point P2 is 1000m / s, the distance is 200m, and the theoretical time is 0.200 seconds; the propagation speed of the pressure wave to point P3 is 1400m / s, the distance is 150m, and the theoretical time is 0.107 seconds. Comparing the theoretical propagation time with the actual detected time series: Point P1 detected the pressure decrease first (10:15:30.000), followed by point P3 (10:15:30.110), and finally point P2 (10:15:30.195). The time consistency score was calculated: the propagation time sequence matching score was 0.95, and the time interval close to the theoretical value scored 0.90. Simultaneously, spatial constraint characteristics were analyzed: point P1 experienced the largest pressure decrease (-0.13 MPa), followed by P3 (-0.10 MPa), and P2 experienced the smallest (-0.07 MPa), conforming to the distance decay law, with a spatial consistency score of 0.92. After comprehensive optimization calculations, the highest comprehensive score of 0.93 was obtained at coordinates (X=520m, Y=320m), and this point was identified as the anomaly location.

[0041] Step S6: Perform multi-dimensional fusion display processing on the abnormal locations and generate a monitoring report.

[0042] It should be noted that the multi-dimensional fusion display processing refers to the unified display of multi-level information such as the geographic information, topology, monitoring data, and abnormal locations of the pipeline network on a three-dimensional visualization platform, and the generation of standardized monitoring reports by combining report templates for different types of abnormal events.

[0043] In step S6, firstly, the abnormal location information is imported into a 3D geographic information system and precisely marked on an electronic map. Different colors and symbols are used to identify different types of abnormal events. Then, based on the anomaly type and its impact level, the impact range is calculated, and a heat map is generated to show the distribution of the abnormal event's impact on the surrounding pipeline network. Secondly, a corresponding report template is selected, and key information such as the anomaly location, type determination, and impact assessment are automatically filled in to generate a standardized monitoring report. Finally, the report distribution recipients are determined according to the anomaly level, and the monitoring report is pushed out through various methods. This integrated display method intuitively shows the spatial distribution characteristics of abnormal events, making it easier for managers to quickly grasp the abnormal situation and make handling decisions.

[0044] For example, regarding the pipeline rupture event in area A located in step S5 (location coordinates X=520m, Y=320m), firstly, the abnormal location is marked on the 3D electronic map, using a flashing red icon to represent the rupture event, and displaying the coordinate information and location time (2024-01-01 10:15:30). The impact range is calculated: centered on the rupture point, extending 200 meters upstream (to pressure monitoring point P2) and 300 meters downstream (to flow monitoring point F1), covering a total of 280 residential users and 5 commercial users. A heat map is generated on the 3D map: within 50 meters of the rupture point, it is displayed in dark red (pressure drop > 0.1MPa); within 50-150 meters, it is displayed in orange (pressure drop 0.05-0.1MPa); and within 150-300 meters, it is displayed in yellow (pressure drop < 0.05MPa). The system automatically calls upon the emergency incident report template to generate a monitoring report, which includes: an event summary (pipeline rupture, time of occurrence, and scope of impact), location information (coordinates, location reliability, and error estimation), monitoring data (pressure and flow rate change curves at each monitoring point), impact assessment (statistics of affected users and estimated repair time), and handling recommendations (valve numbers to be closed and backup water supply plans). The emergency level report is immediately pushed to the maintenance team, pipeline dispatch center, and emergency duty personnel via SMS and mobile APP. At the same time, the complete monitoring report is sent to the management department via email for archiving.

[0045] The method provided in this embodiment, on the one hand, achieves unified processing of multi-source heterogeneous monitoring data and rapid discovery of abnormal events by introducing a protocol adaptive fusion mechanism and an anomaly identification method based on time-varying statistical features, thereby improving data processing efficiency and anomaly detection accuracy. On the other hand, by establishing a knowledge graph of pipeline topology spatial association and a hydraulic numerical simulation model, it enables multi-sensor cross-validation and reverse scenario matching analysis of abnormal events, reducing false alarm rate and narrowing the location range of abnormal events. Furthermore, by combining graph theory pressure wave propagation algorithm and multi-dimensional fusion display processing, it achieves precise location and visualization of abnormal locations, shortening fault location time and reducing maintenance costs.

[0046] In some embodiments, the anomaly identification processing of the standardized multi-source time-series dataset based on time-varying statistical features and dynamic control limits to obtain a candidate anomaly event set includes: Step S2.1: Perform sliding time window partitioning on the standardized multi-source time series dataset to obtain time window data.

[0047] For example, in a city's water supply network, the pressure monitoring node in area A collects data once per second, the flow monitoring node collects data once every 30 seconds, and the water quality monitoring node collects data once per minute. Based on the changing characteristics of different parameters, a 1-hour sliding time window (containing 3600 data points) is set for pressure data, a 30-minute sliding time window (containing 60 data points) is set for flow data, and a 2-hour sliding time window (containing 120 data points) is set for water quality data. Taking pressure monitoring node P1 as an example, within the time window from 10:00:00 to 11:00:00 on January 1, 2024, a total of 3600 pressure values ​​are collected. The window slides forward once every minute, keeping the window size constant, thus forming continuous time-series data segments. This sliding window mechanism can capture the dynamic changing characteristics of the data.

[0048] Step S2.2: Perform feature calculation processing on the time window data based on statistical analysis methods to obtain time-varying statistical feature vectors.

[0049] For example, statistical characteristics are calculated for the 1-hour sliding time window data of pressure monitoring node P1 in step S2.1. First, basic statistics are calculated: the arithmetic mean of the 3600 pressure values ​​within the window is 0.45 MPa, and the standard deviation is 0.02 MPa. Then, distribution characteristics are calculated: the skewness of the data is 0.1 (indicating that the data distribution is basically symmetrical), and the kurtosis is 3.0 (indicating that the data distribution is close to a normal distribution). Next, quantile characteristics are calculated: the 25th quantile is 0.43 MPa, the 50th quantile (median) is 0.45 MPa, and the 75th quantile is 0.47 MPa. These statistical characteristics are combined to form a feature vector describing the data distribution characteristics of this time window, which is used for subsequent anomaly detection. As the window slides forward, the statistical characteristics are recalculated to obtain time-varying characteristics reflecting the dynamic changes in the data distribution.

[0050] Step S2.3: Calculate the control limits for the time-varying statistical feature vector based on the historical data distribution to obtain the dynamic control limit parameters.

[0051] For example, the time-varying statistical feature vector obtained in step S2.2 is combined with historical data for the same period to calculate control limits. Taking pressure monitoring node P1 as an example, historical data for the same period in 2023 (January 1, 10:00-11:00) is extracted. Considering the characteristics of winter water use, data from 7 days before and after are selected as references. The historical average pressure for the same period is 0.46 MPa, and the standard deviation is 0.03 MPa. Based on the 3σ principle and combined with seasonal factors, the dynamic control upper limit is calculated to be 0.52 MPa (mean + 2σ), and the lower limit is 0.38 MPa (mean - 2σ). These control limit parameters will be dynamically updated as the time window slides to adapt to changes in the pipeline network operation status.

[0052] Step S2.4: Perform anomaly detection processing on the dynamic control limit parameters based on a multi-level detection strategy to obtain anomaly detection results.

[0053] For example, multi-level anomaly detection is performed on the dynamic control limit parameters calculated in step S2.3. When the pressure value at monitoring point P1 drops to 0.35 MPa at 10:15:00, which is below the dynamic control lower limit of 0.38 MPa, the first-level single-point anomaly detection is triggered. In the following 15 minutes, the pressure continues to drop to 0.32 MPa, and the trend deviates significantly from the historical pattern, triggering the second-level trend anomaly detection. At the same time, it is found that the flow rate of the adjacent flow monitoring node F1 increased by 20% during the same period, reaching 96 m³ / h, triggering the third-level correlation anomaly detection. The trigger time, anomaly type, and detection parameters of each level of detection are recorded to form a complete anomaly detection record.

[0054] Step S2.5: Perform confidence scoring on the anomaly detection results to obtain the anomaly confidence level.

[0055] For example, the confidence score of the anomaly detected in step S2.4 is calculated. For the anomaly detection record of monitoring point P1, the single-point anomaly detection score is 0.8 (the pressure value deviates significantly from the control limit), the trend anomaly detection score is 0.85 (the pressure continues to decrease significantly), and the correlation anomaly detection score is 0.9 (the pressure decrease is strongly correlated with the flow rate increase). According to the reliability weight of each layer of detection (single-point detection 0.3, trend detection 0.3, correlation detection 0.4), the comprehensive confidence score is calculated as: 0.8×0.3 + 0.85×0.3 + 0.9×0.4 = 0.855.

[0056] Step S2.6: Filter the anomaly confidence based on the confidence threshold to obtain the candidate anomaly event set.

[0057] The anomaly confidence scores obtained in step S2.5 are filtered. A confidence threshold of 0.6 is set. The anomaly detected by monitoring point P1 has a confidence score of 0.85, exceeding the threshold, and is therefore included in the candidate anomaly event set. The event occurrence time (2024-01-01 10:15:00), location information (monitoring point P1), anomaly type (pressure anomaly), and influencing parameters (pressure decrease, flow increase) are recorded. Simultaneously, relevant time window data, statistical characteristics, dynamic control limits, and other supporting information are recorded to form a complete anomaly event description, providing data support for subsequent multi-sensor cross-validation.

[0058] In some embodiments, the multi-sensor cross-validation processing of the candidate anomaly event set based on the pipeline network topology spatial association knowledge graph to obtain a high-confidence anomaly event set includes: Step S3.1: Construct a knowledge graph of pipeline topology spatial association based on the candidate abnormal event set and pipeline topology structure to obtain monitoring node attribute data and edge attribute data.

[0059] For example, after a suspected pipe burst occurs in area A of a city's water supply network, a topological spatial association knowledge graph of the area is first constructed. The four monitoring nodes (P1, P2, P3, and F1) in the area are used as vertices of the graph, and their attribute data is recorded: monitoring node P1 is located at coordinates (500, 300), is a pressure sensor, and is installed on a DN300 ductile iron pipe; monitoring node P2 is located at coordinates (500, 500), is a pressure sensor, and is installed on a DN300 ductile iron pipe; monitoring node P3 is located at coordinates (500, 150), is a pressure sensor, and is installed on a DN300 ductile iron pipe; monitoring node F1 is located at coordinates (680, 300), is a flow meter, and is installed on a DN300 ductile iron pipe. The pipeline connection relationships between monitoring nodes are treated as edges of the graph, and edge attribute data are recorded: P1-P2 pipe segment length 200 meters, pipeline resistance coefficient 0.02, valve V1 opening 100%; P1-P3 pipe segment length 150 meters, pipeline resistance coefficient 0.02, valve V2 opening 100%; P1-F1 pipe segment length 180 meters, pipeline resistance coefficient 0.02, valve V3 opening 100%.

[0060] Step S3.2: Based on the graph theory path search algorithm, perform topological distance calculation on the monitoring node attribute data and edge attribute data to obtain the correlation strength coefficient between monitoring points.

[0061] For example, based on the knowledge graph constructed in step S3.1, the Dijkstra shortest path algorithm is used to calculate the topological distances from node P1 to other monitoring nodes. The calculated shortest path length from P1 to P2 is 200 meters, passing through one valve V1; the shortest path length from P1 to P3 is 150 meters, passing through one valve V2; and the shortest path length from P1 to F1 is 180 meters, passing through one valve V3. The correlation strength coefficients are calculated based on the topological distance, pipe diameter ratio, and valve status: the correlation coefficient between P1 and P2 is 0.85 (moderate distance, same pipe diameter, valve fully open); the correlation coefficient between P1 and P3 is 0.9 (relatively close distance, same pipe diameter, valve fully open); and the correlation coefficient between P1 and F1 is 0.88 (moderate distance, same pipe diameter, valve fully open). These correlation strength coefficients reflect the degree of hydraulic influence between monitoring nodes.

[0062] Step S3.3: Based on the correlation strength coefficient, perform a neighboring point search process on the abnormal monitoring points to obtain neighboring monitoring data.

[0063] For example, the correlation strength coefficient calculated in step S3.2 is analyzed, and the correlation strength threshold is set to 0.8. Monitoring points with a correlation strength exceeding the threshold with the anomaly point P1 are screened. The screening results show that P2 (correlation strength 0.85), P3 (correlation strength 0.9), and F1 (correlation strength 0.88) all meet the requirements and are identified as neighboring monitoring points of P1. Monitoring data of these neighboring points before and after the anomaly occurs are collected: the pressure at point P2 increases from 0.45 MPa to 0.48 MPa; the pressure at point P3 decreases from 0.45 MPa to 0.35 MPa; and the flow rate at point F1 increases from 80 m³ / h to 96 m³ / h. The change time of each monitoring value is recorded: the pressure change time at point P2 is 10:15:30.195, the pressure change time at point P3 is 10:15:30.110, and the flow rate change time at point F1 is 10:15:30.150.

[0064] Step S3.4: Based on the principle of pipeline hydraulics, the adjacent monitoring data is processed to verify abnormal behavior, and the verification results are obtained.

[0065] For example, based on the nearby monitoring data obtained in step S3.3, verification was performed according to the typical hydraulic characteristics of a pipeline rupture event. Hydraulic principles indicate that: the pressure upstream of the rupture point will rise slightly (due to the formation of local high impedance at the break point), the pressure downstream will drop significantly (due to pressure wave propagation), and the downstream flow rate will increase significantly (due to pressure release). Comparing with actual monitoring data: the pressure increase of 0.03 MPa at point P2 (upstream) is consistent with the characteristics; the pressure decrease of 0.10 MPa at point P3 (downstream) is consistent with the characteristics; and the flow rate increase of 20% at point F1 (downstream) is consistent with the characteristics. The abnormal behavior of each monitoring point is consistent with the theoretical characteristics of a pipeline rupture.

[0066] Step S3.5: Perform consistency scoring on the verification results to obtain the verification score.

[0067] For example, based on the verification results of step S3.4, the verification consistency score is calculated from three dimensions. Time synchronization dimension: After P1 detects a pressure decrease, P3, F1, and P2 detect changes sequentially, with time intervals of 110ms, 150ms, and 195ms respectively, consistent with the pressure wave propagation law, score 0.9; Amplitude reasonableness dimension: Upstream pressure increases by 0.03MPa, downstream pressure decreases by 0.10MPa, and flow rate increases by 20%, the magnitude of the change is consistent with the burst scale, score 0.85; Type matching dimension: The combination of upstream pressure increase, downstream pressure decrease, and downstream flow rate increase matches the burst event characteristics, score 0.95. The comprehensive verification score is calculated according to the weighted allocation (time 0.3, amplitude 0.3, type 0.4): 0.9×0.3 + 0.85×0.3 + 0.95×0.4 = 0.905.

[0068] Step S3.6: Filter the verification scores based on the verification score threshold to obtain the high-confidence abnormal event set.

[0069] For example, a verification score threshold of 0.7 is set, and the verification scores obtained in step S3.5 are filtered. The comprehensive verification score of this abnormal event is 0.9, which is significantly higher than the threshold of 0.7, so it is identified as a high-confidence abnormal event. Detailed information about the abnormal event is recorded: occurrence time (2024-01-01 10:15:30), abnormality type (pipeline rupture), core monitoring point (P1), associated monitoring points (P2, P3, F1), verification score (0.905), and abnormal characteristics (sudden pressure drop, upstream and downstream pressure difference, sudden increase in flow rate), etc. This event is added to the high-confidence abnormal event set as an important basis for subsequent precise location analysis.

[0070] In some embodiments, the inverse scenario matching analysis of the high-confidence anomaly event set based on the hydraulic numerical simulation model to obtain the preliminary location area includes: Step S4.1: Determine the anomaly type and boundary conditions based on the characteristics of the high-confidence anomaly event set to obtain an anomaly scenario template.

[0071] For example, considering a high-confidence anomaly detected in area A, the anomaly characteristics are analyzed: the pressure at point P1 drops sharply from 0.45 MPa to 0.32 MPa, the pressure at point P2 rises to 0.48 MPa, the pressure at point P3 drops to 0.35 MPa, and the flow rate at point F1 increases by 20%. This combination of characteristics (sudden pressure drop, increased upstream pressure, decreased downstream pressure, and increased flow rate) matches the typical characteristics of a pipeline rupture event. An anomaly scenario template for pipeline rupture is selected, and boundary condition parameters are set: the node pressure is set to 0.1 MPa (considering atmospheric pressure), the flow rate is set to 80 m³ / h (calculated backwards from the increase in flow rate), and the duration is set to steady state (a new equilibrium state has been reached). These boundary condition parameters will be used for subsequent hydraulic simulation calculations.

[0072] Step S4.2: Perform search range division processing on the abnormal scenario template to obtain a candidate location set.

[0073] For example, based on the pipeline rupture anomaly template determined in step S4.1, an initial circular search range with a radius of 300 meters is set, centered on the monitoring node P1 where the anomaly was first detected. Considering the actual layout of the pipeline network, the search range is divided into a grid with a grid spacing of 20 meters, generating candidate locations along the pipeline. After excluding areas covered by buildings and areas without pipelines, a total of 90 valid candidate locations are generated. Each candidate location records its coordinates, the information of the pipe segment it belongs to, and the pipeline network distance from each monitoring point. These candidate locations will be used as potential rupture points for simulation verification.

[0074] Step S4.3: Perform point-by-point simulation calculations on the candidate location set based on the hydraulic simulation model to obtain theoretical response data.

[0075] For example, hydraulic simulation calculations are sequentially performed on the 90 candidate locations generated in step S4.2. Taking a candidate location (X=600m, Y=300m) as an example, the boundary conditions defined in step S4.1 (pressure 0.1MPa, flow rate 80m³ / h) are set at this point, and steady-state hydraulic calculations are performed. After the calculations converge, the theoretical response values ​​of each monitoring point are recorded: theoretical pressure value at point P1 is 0.33MPa, theoretical pressure value at point P2 is 0.47MPa, theoretical pressure value at point P3 is 0.34MPa, and theoretical flow rate at point F1 increases by 18%. The same simulation process is repeated for other candidate locations, and the calculation results for each location are used as a set of theoretical response data for subsequent matching analysis.

[0076] Step S4.4: The theoretical response data is compared and analyzed based on a multi-dimensional similarity evaluation method to obtain a matching score.

[0077] For example, the theoretical response data obtained in step S4.3 is compared with the actual monitoring data in multiple dimensions. Taking the candidate location (X=600m, Y=300m) as an example, the numerical difference is calculated as follows: the pressure deviation at point P1 is 0.01MPa (theoretical 0.33MPa, measured 0.32MPa), the pressure deviation at point P2 is 0.01MPa (theoretical 0.47MPa, measured 0.48MPa), the pressure deviation at point P3 is 0.01MPa (theoretical 0.34MPa, measured 0.35MPa), and the flow deviation at point F1 is 2% (theoretical 18% vs. 20%), with a numerical difference score of 0.92. Trend consistency is calculated: the direction of pressure and flow changes at the four monitoring points is completely consistent with the actual measurements, with a trend consistency score of 0.88. Peak value correspondence is calculated: the temporal sequence of pressure and flow changes conforms to the propagation law, with a peak value correspondence score of 0.85.

[0078] Step S4.5: Perform comprehensive sorting on the matching scores to obtain the optimal matching position.

[0079] For example, the matching score calculated in step S4.4 is weighted and comprehensively calculated. Weighting coefficients are set as follows: numerical difference weight 0.4, trend consistency weight 0.3, and peak correspondence weight 0.3. Taking the candidate position (X=600m, Y=300m) as an example, the comprehensive matching score is calculated as: 0.92×0.4 + 0.88×0.3 + 0.85×0.3 = 0.887. The comprehensive matching scores of the 90 candidate positions are sorted, and the score of this position (0.88) is the highest, while the scores of other positions are all below 0.85. Therefore, the coordinates (X=600m, Y=300m) are determined as the optimal matching position.

[0080] Step S4.6: Based on the optimal matching position, perform regional range delineation processing to obtain the preliminary positioning area.

[0081] For example, taking the optimal matching position (X=600m, Y=300m) determined in step S4.5 as the center, and considering the uncertainties of the hydraulic simulation model, including pipeline parameter errors (±5%), boundary condition errors (±3%), and numerical calculation errors (±2%), a confidence interval with a radius of 80 meters is set. Within this range, the matching score between the simulation results and the measured data at any location is above 0.85. Finally, this circular area (center coordinates X=600m, Y=300m, radius 80 meters) is determined as the preliminary positioning area, providing spatial constraints for subsequent precise positioning.

[0082] In some embodiments, the graph-based pressure wave propagation algorithm is used to optimize the initial location area to obtain the abnormal location, including: Step S5.1: Perform mesh subdivision processing on the preliminary positioning area to obtain candidate location meshes.

[0083] For example, a suspected pipe rupture occurred in area A of a water supply network. The initial location of the ruptured area was a circular region with a radius of 80 meters centered at coordinates (X=600m, Y=300m). The system divided this area using a uniform 2m × 2m grid, generating approximately 5000 grid points. Considering the actual network layout, grid points in areas without pipes, such as building-covered areas and green belts, were removed, retaining 1600 grid points along the pipeline route. Each grid point recorded its coordinates (X, Y), the pipe segment number, and the pipe type. These grid points constituted a candidate location grid, which would be used for subsequent pressure wave propagation analysis.

[0084] Step S5.2: Calculate the pressure wave propagation velocity of the candidate location grid based on the pipeline network physical parameters to obtain propagation velocity data.

[0085] For example, the pressure wave propagation velocity is calculated for each of the 1600 candidate locations obtained in step S5.1. Taking the grid point at coordinates (X=580m, Y=300m) as an example, this point is located on a DN300 ductile iron pipe with the following parameters: pipe diameter 300mm, pipe wall thickness 6mm, elastic modulus 170GPa, and fluid density 1000kg / m³ at a water temperature of 10℃. According to the pressure wave propagation velocity calculation formula, considering the pipe constraints and fluid compressibility, the pressure wave propagation velocity of this pipe section is calculated to be 1200m / s. Similarly, the pressure wave propagation velocity of other pipe sections is calculated: DN200 steel pipe section pressure wave propagation velocity 1400m / s, and DN400 ductile iron pipe section pressure wave propagation velocity 1000m / s. These propagation velocity data will be used to calculate the theoretical time for the pressure wave to reach each monitoring point.

[0086] The formula for calculating the propagation speed of the pressure wave is as follows: ,in, For the propagation speed of pressure waves, Let be the bulk modulus of water. The density of water, The elastic modulus of the pipe is given by [reference]. The inner diameter of the pipe. For pipe wall thickness, This is the pipeline constraint coefficient.

[0087] Step S5.3: Calculate the theoretical propagation time for each monitoring node based on the propagation speed data to obtain time series data.

[0088] For example, using the propagation velocity data calculated in step S5.2, the theoretical propagation time from the candidate location point to each monitoring node is calculated. Taking coordinates (X=580m, Y=300m) as an example, the pressure wave propagation path from this point to the three pressure monitoring points P1, P2, and P3 is calculated: to point P1, it travels 100 meters through a DN300 ductile iron pipe, with a theoretical propagation time of 0.083 seconds; to point P2, it first travels 100 meters through a DN300 ductile iron pipe and then 100 meters through a DN200 steel pipe, with a theoretical propagation time of 0.200 seconds; to point P3, it travels 150 meters through a DN400 ductile iron pipe, with a theoretical propagation time of 0.107 seconds. The same calculation is repeated for each candidate location point to form a complete theoretical propagation time series data.

[0089] Step S5.4: Perform consistency evaluation processing on the time series data to obtain a time consistency score.

[0090] For example, the theoretical propagation time series obtained in step S5.3 is compared and scored with the actual monitoring time. Taking coordinates (X=580m, Y=300m) as an example, the actual monitoring sequence is: pressure reduction was detected at point P1 at 10:15:30.000, at point P3 at 10:15:30.110, and at point P2 at 10:15:30.195. The consistency of the calculated time is compared between the theoretical propagation time and the actual time: propagation order consistency (P1->P3->P2 perfectly matches, score 0.95), and time interval consistency (the difference between theoretical and measured time is less than 0.01 seconds, score 0.90). The overall time consistency score for this point is calculated as: 0.95×0.5 + 0.90×0.5 = 0.925.

[0091] Step S5.5: Perform spatial constraint verification on the time consistency score based on the intensity of anomaly impact to obtain the spatial consistency score.

[0092] For example, spatial constraint verification is performed on the candidate locations with higher scores in step S5.4. Taking coordinates (X=580m, Y=300m) as an example, the pressure change amplitude of each monitoring point is analyzed: the pressure drop at point P1 is 0.13MPa (distance 100m), the pressure drop at point P3 is 0.10MPa (distance 150m), and the pressure drop at point P2 is 0.07MPa (distance 200m). According to the pressure wave attenuation law, the influence intensity should be inversely proportional to the propagation distance, and the spatial attenuation consistency score is calculated to be 0.92.

[0093] Step S5.6: Perform multi-objective optimization processing based on the time consistency score and spatial consistency score to obtain the anomaly location.

[0094] For example, the scores of all candidate locations are comprehensively ranked, with a temporal consistency weight of 0.6 and a spatial consistency weight of 0.4. After calculation, the point at coordinates (X=520m, Y=320m) receives the highest comprehensive score of 0.93 (temporal consistency 0.94, spatial consistency 0.915). Considering the confidence level and potential error of the score, the system sets an error range of ±2 meters around this point. The coordinates (X=520m, Y=320m) are ultimately determined to be an anomaly location, with the positioning error controlled within 2 meters. This result will be used for subsequent visualization and report generation.

[0095] In some embodiments, the multi-dimensional fusion display processing of the abnormal location to generate a monitoring report includes: Step S6.1: The abnormal locations are labeled using a three-dimensional geographic information system to obtain location labeling data.

[0096] For example, regarding a pipeline rupture event in Area A, the coordinates of the abnormal location (X=520m, Y=320m) are imported into a 3D geographic information system. The abnormal point is marked on the electronic map, using a red flashing icon (flashing frequency 1 time / second) to represent the pipeline rupture event, with the icon size set to 20×20 pixels. The marking information includes: event number (AB240101-001), anomaly type (pipeline rupture), occurrence time (2024-01-01-10:15:30), and location accuracy (±2 meters). Simultaneously, pipeline network information is overlaid on the marked point: pipeline type (DN300 ductile iron pipe), installation year (2020), and operating pressure (0.45MPa). This location marking data constitutes the basic display information for the abnormal event.

[0097] Step S6.2: Perform influence range calculation processing based on the location annotation data to obtain influence range data.

[0098] For example, based on the location markings in step S6.1, the impact range of the event is calculated along the pipeline topology. The search extends upstream to pressure monitoring point P2 (200 meters), passing through 6 residential community service pipes, covering 180 residential users; downstream to flow monitoring point F1 (300 meters), passing through 2 commercial area service pipes and 4 residential community service pipes, covering 5 commercial users and 100 residential users. Key facilities within the impact range are recorded: 4 valves (V1-V4), 2 fire hydrants (H1-H2), 3 pressure monitoring points (P1-P3), and 1 flow monitoring point (F1). The system calculates the total water consumption (150 m³ / h) and the service population (approximately 1000 people) within the impact range. This impact range data is used for subsequent analysis and visualization.

[0099] Step S6.3: Perform heat map drawing processing based on the influence range data to obtain spatial distribution data.

[0100] For example, a pressure impact heatmap is generated for the impact range determined in step S6.2. Centered on the burst point, the impact range is divided into three zones: the core zone (0-50 meters, pressure drop > 0.1 MPa) is represented in dark red with a transparency of 20%; the transition zone (50-150 meters, pressure drop 0.05-0.1 MPa) is represented in orange with a transparency of 40%; and the edge zone (150-300 meters, pressure drop < 0.05 MPa) is represented in yellow with a transparency of 60%. User information is overlaid on the heatmap: 50 residential users in the core zone, 130 residential users and 3 commercial users in the transition zone, and 100 residential users and 2 commercial users in the edge zone. This spatial distribution data visually demonstrates the degree of impact of the event.

[0101] Step S6.4: Select a report template based on the spatial distribution data to obtain report template data.

[0102] For example, based on the spatial distribution characteristics in step S6.3, the event is determined to be at the "major pipeline accident" level (affecting more than 200 users), and the emergency event report template is automatically invoked. The template contains five main parts: event summary (including event type, time, location, and scope of impact), on-site data (including pressure and flow curves, user complaint records), impact assessment (including user statistics, duration of water supply interruption), response plan (including valve operation sequence, backup water supply plan), and early warning information (including key time points and precautions). The template sets uniform format requirements for fonts, colors, chart styles, etc.

[0103] Step S6.5: Fill the information into the report template data to obtain the monitoring report.

[0104] For example, based on the emergency incident report template selected in step S6.4, relevant information is automatically filled in. The event summary section is filled in with: Pipeline rupture event (number AB240101-001) occurred on 2024-01-01 at 10:15:30, location coordinates (520m, 320m), and affected area. The on-site data section inserts: 24-hour pressure curves from 3 pressure monitoring points and flow change curve from 1 flow monitoring point. The impact assessment section is filled in with: 280 residential households and 5 commercial households are affected; the estimated water supply interruption duration is 4 hours. The response plan section is generated: Valves V1-V4 need to be closed, and 3 water trucks need to be dispatched to the temporary water supply points H1-H2 in the community. After the report is generated, it is pushed to relevant personnel via SMS and APP.

[0105] Please see Figure 2 , Figure 2A schematic block diagram of the structure of a water network monitoring system 100 based on smart water supply and drainage provided in this application embodiment, as shown below. Figure 2 As shown in the embodiment of this application, the water network monitoring system 100 based on smart water supply and drainage includes: The acquisition module 110 is used to acquire multi-source heterogeneous monitoring data of the water network in real time, and perform unified format conversion and spatiotemporal synchronization processing on the multi-source heterogeneous monitoring data based on the protocol adaptive fusion mechanism to obtain a standardized multi-source time series dataset.

[0106] The anomaly identification module 120 is used to perform anomaly identification processing on the standardized multi-source time series dataset based on time-varying statistical features and dynamic control limits to obtain a candidate anomaly event set.

[0107] The cross-validation processing module 130 is used to perform multi-sensor cross-validation processing on the candidate abnormal event set based on the pipeline network topology spatial association knowledge graph to obtain a high-confidence abnormal event set.

[0108] The matching analysis module 140 is used to perform reverse scenario matching analysis on the high-confidence abnormal event set based on the hydraulic numerical simulation model to obtain the preliminary location area.

[0109] The positioning module 150 is used to perform positioning optimization processing on the preliminary positioning area based on the graph theory pressure wave propagation algorithm to obtain the abnormal location.

[0110] The generation module 160 is used to perform multi-dimensional fusion display processing on the abnormal location and generate a monitoring report.

[0111] It should be noted that those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the system and its modules described above can be referred to the process in the aforementioned embodiment of the water network monitoring method based on smart water supply and drainage, and will not be repeated here.

[0112] The water network monitoring system 100 based on intelligent water supply and drainage provided in the above embodiments can be implemented in the form of a computer program, which can be used in, for example... Figure 3 The terminal device 200 shown is running on it.

[0113] Please see Figure 3 , Figure 3 The following is a schematic block diagram of the structure of a terminal device 200 provided in an embodiment of this application. The terminal device 200 includes a processor 201 and a memory 202, which are connected through a system bus 203. The memory 202 may include a non-volatile storage medium and internal memory.

[0114] The non-volatile storage medium can store a computer program. The computer program includes program instructions, which, when executed by the processor 201, cause the processor 201 to perform any of the aforementioned water network monitoring methods based on smart water supply and drainage.

[0115] The processor 201 provides computing and control capabilities to support the operation of the entire terminal device 200.

[0116] The internal memory provides an environment for the execution of computer programs in non-volatile storage media. When the computer program is executed by the processor 201, the processor 201 can execute any of the above-mentioned water network monitoring methods based on smart water supply and drainage.

[0117] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the terminal device 200 involved in the present application. The specific terminal device 200 may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0118] It should be understood that processor 201 can be a Central Processing Unit (CPU), but it can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among these, the general-purpose processor can be a microprocessor or any conventional processor.

[0119] In some embodiments, the processor 201 is configured to run a computer program stored in memory to perform the following steps: Real-time acquisition of multi-source heterogeneous monitoring data of water network, and unified format conversion and spatiotemporal synchronization processing of the multi-source heterogeneous monitoring data based on protocol adaptive fusion mechanism to obtain standardized multi-source time series dataset; Anomaly identification processing is performed on the standardized multi-source time-series dataset based on time-varying statistical features and dynamic control limits to obtain a candidate set of abnormal events. The candidate abnormal event set is subjected to multi-sensor cross-validation processing based on the knowledge graph of pipeline topology spatial association to obtain a high-confidence abnormal event set. Based on the hydraulic numerical simulation model, a reverse scenario matching analysis was performed on the set of high-confidence abnormal events to obtain the preliminary location area; The preliminary location area is optimized based on the graph theory pressure wave propagation algorithm to obtain the abnormal location. The abnormal locations are processed through multi-dimensional fusion display to generate a monitoring report.

[0120] It should be noted that those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the terminal device 200 described above can be referred to the corresponding process of the aforementioned water network monitoring method based on smart water supply and drainage, and will not be repeated here.

[0121] This application also provides a computer-readable storage medium storing a computer program that, when executed by one or more processors, causes the one or more processors to implement the water network monitoring method based on smart water supply and drainage provided in this application.

[0122] The computer-readable storage medium can be an internal storage unit of the terminal device 200 in the aforementioned embodiments, such as a hard disk or memory of the terminal device 200. The computer-readable storage medium can also be an external storage device of the terminal device 200, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., provided with the terminal device 200.

[0123] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A water network monitoring method based on smart water supply and drainage, characterized in that, include: Real-time acquisition of multi-source heterogeneous monitoring data of water network, and unified format conversion and spatiotemporal synchronization processing of the multi-source heterogeneous monitoring data based on protocol adaptive fusion mechanism to obtain standardized multi-source time series dataset; Anomaly identification processing is performed on the standardized multi-source time-series dataset based on time-varying statistical features and dynamic control limits to obtain a candidate set of abnormal events. The candidate abnormal event set is subjected to multi-sensor cross-validation processing based on the knowledge graph of pipeline topology spatial association to obtain a high-confidence abnormal event set. Based on the hydraulic numerical simulation model, a reverse scenario matching analysis was performed on the set of high-confidence abnormal events to obtain the preliminary location area; The preliminary location area is optimized based on the graph theory pressure wave propagation algorithm to obtain the abnormal location. The abnormal locations are processed through multi-dimensional fusion display to generate a monitoring report.

2. The water network monitoring method based on intelligent water supply and drainage according to claim 1, characterized in that, The anomaly identification process based on time-varying statistical features and dynamic control limits on the standardized multi-source time-series dataset yields a candidate set of anomalous events, including: The standardized multi-source time series dataset is divided into sliding time windows to obtain time window data; The time window data is processed by statistical analysis methods to obtain a time-varying statistical feature vector. Based on the historical data distribution, the time-varying statistical feature vector is processed to calculate the control limit, thereby obtaining the dynamic control limit parameters; An anomaly detection result is obtained by performing anomaly detection processing on the dynamic control limit parameters based on a multi-level detection strategy. The anomaly detection results are subjected to confidence scoring to obtain the anomaly confidence level; The set of candidate abnormal events is obtained by filtering the anomaly confidence based on the confidence threshold.

3. The water network monitoring method based on intelligent water supply and drainage according to claim 1, characterized in that, The candidate anomaly event set is subjected to multi-sensor cross-validation processing based on the pipeline network topology spatial association knowledge graph to obtain a high-confidence anomaly event set, including: Based on the candidate abnormal event set and the pipeline network topology, a pipeline network topology spatial association knowledge graph is constructed to obtain node attribute data and edge attribute data. Based on the graph theory path search algorithm, the node attribute data and edge attribute data are processed by topological distance calculation to obtain the correlation strength coefficient between monitoring points; Based on the correlation strength coefficient, a neighboring point search process is performed on the abnormal monitoring points to obtain neighboring monitoring data; Based on the principles of pipeline hydraulics, the abnormal behavior verification processing of the adjacent monitoring data was performed to obtain the verification results. The verification results are subjected to consistency scoring to obtain a verification score; The verification scores are filtered based on the verification score threshold to obtain the set of high-confidence abnormal events.

4. The water network monitoring method based on intelligent water supply and drainage according to claim 1, characterized in that, The inverse scenario matching analysis based on the hydraulic numerical simulation model on the high-confidence anomaly event set yields a preliminary location area, including: Based on the characteristics of the high-confidence anomaly event set, the anomaly type and boundary conditions are determined to obtain an anomaly scenario template; The abnormal scenario template is subjected to search range division processing to obtain a candidate location set; The candidate location set is simulated point by point using a hydraulic simulation model to obtain theoretical response data. The theoretical response data is compared and analyzed using a multi-dimensional similarity evaluation method to obtain a matching score; The matching scores are then sorted to obtain the optimal matching position. Based on the optimal matching location, the area is delineated to obtain the preliminary positioning area.

5. The water network monitoring method based on intelligent water supply and drainage according to claim 1, characterized in that, The graph-based pressure wave propagation algorithm is used to optimize the initial location of the region to obtain the abnormal location, including: The preliminary positioning area is divided into grids to obtain candidate location grids; The pressure wave propagation velocity of the candidate location grid is calculated based on the pipeline network physical parameters to obtain propagation velocity data. Based on the propagation speed data, the theoretical propagation time of each monitoring node is calculated to obtain time series data; The time series data is subjected to consistency evaluation processing to obtain a time consistency score; Based on the intensity of the anomaly's impact, the temporal consistency score is subjected to spatial constraint verification to obtain a spatial consistency score. The anomaly location is obtained by performing multi-objective optimization based on the time consistency score and spatial consistency score.

6. The water network monitoring method based on intelligent water supply and drainage according to claim 1, characterized in that, The process of performing multi-dimensional fusion display on the abnormal locations and generating a monitoring report includes: The abnormal locations are labeled using a 3D geographic information system to obtain location labeling data. The influence range data is calculated based on the location annotation data. Based on the data of the area of ​​influence, a heat map is drawn to obtain spatial distribution data; Based on the spatial distribution data, a report template is selected to obtain the report template data; The monitoring report is obtained by populating the data in the report template with information.

7. A water network monitoring system based on smart water supply and drainage, characterized in that, include: The acquisition module is used to acquire multi-source heterogeneous monitoring data of the water network in real time, and perform unified format conversion and spatiotemporal synchronization processing on the multi-source heterogeneous monitoring data based on the protocol adaptive fusion mechanism to obtain a standardized multi-source time series dataset. An anomaly identification module is used to perform anomaly identification processing on the standardized multi-source time series dataset based on time-varying statistical features and dynamic control limits to obtain a candidate anomaly event set; The cross-validation processing module is used to perform multi-sensor cross-validation processing on the candidate abnormal event set based on the pipeline topology spatial association knowledge graph to obtain a high-confidence abnormal event set. The matching analysis module is used to perform reverse scenario matching analysis on the high-confidence abnormal event set based on the hydraulic numerical simulation model to obtain the preliminary location area; The positioning module is used to perform positioning optimization processing on the preliminary positioning area based on the graph theory pressure wave propagation algorithm to obtain the abnormal location; The generation module is used to perform multi-dimensional fusion display processing on the abnormal locations and generate a monitoring report.

Images