A rural water supply intelligent management method and system based on digital twinning

Abstract

This invention discloses a smart management method and system for rural water supply based on digital twins, relating to the field of rural water supply technology. By deploying data acquisition devices at key nodes, it acquires flow and pressure indicators in real time, forming an initial operational dataset, and uses data analysis algorithms to identify information gaps. If a gap exceeds a threshold, an edge computing module is activated to aggregate data, generating a dynamic control indicator set. A machine learning classification algorithm is then used to optimize the response speed sequence, extracting high-priority events and quickly locating problem nodes using graph theory algorithms. Subsequently, historical and current data are integrated to update the resource allocation scheme, dynamically adjusting the sampling frequency to form a closed-loop feedback mechanism and generate alarm signals. This invention, through the combination of data-driven approaches and intelligent algorithms, significantly improves the efficiency of water supply network problem detection and resource allocation, ensuring the stable operation of the water supply system.

Description

Technical Field

[0001] This invention relates to the field of rural water supply technology, and in particular discloses a smart management method and system for rural water supply based on digital twins. Background Technology

[0002] Rural water supply management is an important area for ensuring people's livelihood and promoting rural revitalization.

[0003] From a technical perspective, the biggest challenge in rural water supply management lies in achieving dynamic control over complex water supply networks and improving management accuracy with limited resources. The primary challenge is the difficulty in obtaining real-time status information for each link in the water supply network. Key indicators such as water volume and pressure are often missing due to insufficient monitoring methods, directly leading to a lack of comprehensive understanding of the system's actual operation by managers. Furthermore, due to the inability to obtain this dynamic information in a timely manner, managers struggle to quickly identify the root cause of problems and take targeted measures when facing emergencies. For example, when an anomaly occurs in a certain area of ​​the pipeline network, it is difficult to quickly locate the specific location, requiring extensive investigation and consuming significant time and manpower.

[0004] Therefore, under the condition of limited resources, how to build a management mechanism that can reflect the real-time operation status of the water supply network and support rapid problem location and optimized resource allocation has become a key issue in the intelligent transformation of rural water supply. Summary of the Invention

[0005] This invention provides a method and system for intelligent management of rural water supply based on digital twins, aiming to achieve intelligent management of the entire process of rural water supply network from data collection to fault handling and optimization iteration.

[0006] One aspect of this invention relates to a smart management method for rural water supply based on digital twins, comprising the following steps: S100. Real-time data is collected by data acquisition devices deployed at key nodes of the water supply network, covering the water supply network from source to end, to obtain the initial operating dataset of the water supply network, including real-time data such as flow rate and pressure. S200. Based on the initial running dataset, apply data analysis algorithms to process the continuous data stream, identify abnormal fluctuation patterns, and determine potential information gap areas in the water supply network, where information gap areas refer to transmission segments with missing data. S300. If the information gap area determined by the data analysis algorithm exceeds the preset threshold, the edge computing module is activated to perform local aggregation of the data in the information gap area to obtain a refined dynamic control index set, which includes the aggregated flow velocity value and pressure change value. S400. Input the refined dynamic control index set into the machine learning classification algorithm for classification, determine the priority level of problem discovery, and obtain the classified response speed optimization sequence, wherein the response speed optimization sequence is sorted according to the urgency of the event. S500. Extract high-priority events from the classified response speed optimization sequence, apply graph theory algorithm to construct a connectivity graph of the water supply network for the high-priority events, and determine the specific node coordinates for rapid location, where the specific node coordinates correspond to the key connection points of the transmission pipeline. S600: After obtaining the specific node coordinates for rapid positioning, the entire tracking record is updated through a distributed database. By integrating historical data and current indicators, an updated resource-limited allocation scheme is obtained, which optimizes the distribution of maintenance resources. S700 If the deviation of the flow index is greater than the deviation of the pressure index in the updated resource-limited allocation scheme, the sampling frequency of the data acquisition device is adjusted to obtain an optimized dynamic control feedback loop. An alarm signal is generated based on the optimized dynamic control feedback loop and distributed to the terminal device. It is then determined whether the response speed meets the threshold requirement, and the final problem discovery closed-loop mechanism is determined.

[0007] Further, step S100 includes: S110. Obtain flow and pressure indicators through data acquisition devices deployed at key nodes of the water supply network; S120. Spatial mapping of flow and pressure indicators between node coordinates and topology structure to obtain mapping results containing geographic location information; S130. Based on the mapping results, pipe diameter and pipe material properties, determine the pipe resistance, and trigger the terminal flow verification logic based on the abnormal fluctuations after the pipe resistance is correlated with the water supply period to obtain the terminal flow. S140. The terminal flow is summarized and the data is cleaned and denoised to obtain the initial operating dataset of the water supply network.

[0008] Further, step S200 includes: S210. Obtain the initial running dataset and construct a continuous data flow matrix, which contains flow and pressure values ​​aligned to a timestamp sequence. S220. Traverse the continuous data stream matrix to identify null gaps caused by numerical interruptions, and retrieve the synchronous operation data of adjacent nodes in the pipeline network topology based on the time location information of the null gaps. S230. Calculate the interpolation error of the synchronous operation data. If the interpolation error exceeds the preset value, mark the null gap as an abnormal fluctuation mode caused by the transmission link failure. S240. Based on the spatial distribution trajectory of the abnormal fluctuation pattern in the pipeline topology, delineate the range of pipe sections where data loss occurred, and determine the information gap area in the water supply network that is the data loss transmission segment.

[0009] Further, step S300 includes: S310. Statistically identify the cumulative link length of the information fault area in the pipeline topology, and use the cumulative link length as the basis for determining the fault scale. S320. If the fault scale determination criteria exceed the preset threshold, an edge collaboration trigger signal is sent to the edge computing module at the physical boundary of the information fault area. After receiving the edge collaboration trigger signal, the S330 edge computing module collects high-frequency discrete sampling data and performs local aggregation operation to obtain the aggregated flow velocity value and pressure change value. S340. Based on the aggregated flow rate and pressure change values, a structured encapsulation is performed to generate a refined set of dynamic control indicators.

[0010] Further, step S400 includes: S410. Extract the aggregated flow velocity and pressure change values ​​from the refined dynamic control index set, and convert the flow velocity and pressure change values ​​into a vectorized index feature set through a preset feature extraction model. S420. Input the vectorized indicator feature set into the machine learning classification algorithm to obtain the priority-labeled feature subset; S430. Based on the priority-labeled feature subset, an initial response speed sequence is generated using a sequence generation model, and the initial response speed sequence is optimized and adjusted using an association rule mining algorithm to obtain an optimized response speed sequence. S440. Obtain the classified response speed optimization sequence based on the optimized response speed sequence, wherein the response speed optimization sequence is sorted according to the urgency of the event.

[0011] Further, step S500 includes: S510. Extract urgent events whose urgency exceeds a preset threshold from the classified response speed optimization sequence; S520. Retrieve topology data of the water supply network for emergency events. The topology data is used to construct a connectivity matrix that includes pipe attributes and flow loads. S530. Calculate the node weights and edge weights distribution through the connectivity matrix, and use the Dijkstra algorithm to determine the key hubs in the water supply network connectivity graph generated by the connectivity matrix. S540. Match the physical location of key hubs with spatial coordinates in the geographic information system. The spatial coordinates are used to determine the specific node coordinates for rapid positioning.

[0012] Further, step S600 includes: S610. Obtain the specific node coordinates for rapid positioning, and use the specific node coordinates to retrieve the full tracking record from the distributed database. The full tracking record integrates historical maintenance data and current traffic indicators. S620. Calculate resource allocation weights based on the entire tracking record. Resource allocation weights are generated by weighting historical maintenance data with current traffic indicators. S630. Perform multi-objective programming calculations on the available resource set according to the resource allocation weights to obtain a resource-limited allocation scheme; S640: Analyze the limited resource allocation scheme to perform spatial scheduling and optimize the distribution of maintenance resources.

[0013] Further, step S700 includes: S710: Analyze the limited resource allocation scheme and compare the flow index deviation with the pressure index deviation. If the flow index deviation is greater than the pressure index deviation, send a frequency modulation command. S720: Response frequency modulation command generates high-frequency sampled data stream, and the high-frequency sampled data stream is used to obtain an optimized dynamic control feedback loop; S730: Distribute alarm signals to terminal devices in a loop according to optimized dynamic control feedback, and receive feedback timestamps returned by terminal devices to calculate response speed; S740. Verify whether the response speed meets the preset threshold requirements and determine the final problem discovery closed-loop mechanism.

[0014] Another aspect of the present invention relates to a digital twin-based intelligent management system for rural water supply, and a digital twin-based intelligent management method for rural water supply, comprising: The initial operational dataset acquisition module is used to collect real-time data through data acquisition devices deployed at key nodes of the water supply network, covering the water supply network from source to end, and obtain the initial operational dataset of the water supply network, where real-time data includes flow and pressure indicators. The information gap region determination module is used to process continuous data streams using data analysis algorithms based on the initial running dataset, identify abnormal fluctuation patterns, and determine potential information gap regions in the water supply network, where an information gap region refers to a transmission segment with missing data. The dynamic control indicator set acquisition module is used to activate the edge computing module to locally aggregate the data in the information gap area if the information gap area determined by the data analysis algorithm exceeds a preset threshold, so as to obtain a refined dynamic control indicator set. The dynamic control indicator set includes aggregated flow velocity values ​​and pressure change values. The response speed optimization sequence acquisition module is used to input the refined dynamic control index set into the machine learning classification algorithm for classification, determine the priority level of problem discovery, and obtain the classified response speed optimization sequence, wherein the response speed optimization sequence is sorted according to the urgency of the event. The specific node coordinate determination module is used to extract high-priority events from the classified response speed optimization sequence, apply graph theory algorithms to construct a connectivity graph of the water supply network for the high-priority events, and determine the specific node coordinates for rapid location, where the specific node coordinates correspond to the key connection points of the transmission pipeline; The resource-limited allocation scheme acquisition module is used to obtain the specific node coordinates for rapid positioning, update the entire tracking record through a distributed database, integrate historical data and current indicators, and obtain an updated resource-limited allocation scheme, which optimizes the distribution of maintenance resources. The problem discovery closed-loop mechanism determination module is used to adjust the sampling frequency of the data acquisition device if the deviation of the flow index is greater than the deviation of the pressure index in the updated limited resource allocation scheme, to obtain an optimized dynamic control feedback loop, generate an alarm signal based on the optimized dynamic control feedback loop and distribute it to the terminal device, determine whether the response speed meets the threshold requirement, and determine the final problem discovery closed-loop mechanism.

[0015] The beneficial effects achieved by this invention are as follows: 1. The present invention provides a digital twin-based intelligent management method and system for rural water supply, which addresses issues such as information gaps caused by data loss in water supply networks, event response priority judgment, and uneven resource allocation in business scenarios. It proposes an integrated and directly implementable solution. By clarifying the specific implementation details of the core algorithm, the basis and adaptation methods for setting key parameters, the deployment and interaction rules of hardware devices, and the complete execution logic of digital twin technology and closed-loop feedback mechanisms, it achieves intelligent management of the entire process of rural water supply networks, from data acquisition to fault handling and optimization iteration.

[0016] 2. This invention combines data-driven approaches with intelligent algorithms to achieve an accuracy of ≥95% in identifying abnormal fluctuations in the water supply network, an information fault location error of ≤50m, and a response time of ≤30 minutes for high-priority events. This significantly improves the efficiency of water supply network problem detection and resource allocation. At the same time, it is suitable for the actual scenario of rural water supply networks characterized by "wide coverage, scattered nodes, limited resources, and poor communication conditions," reducing implementation costs and maintenance difficulties, and ensuring the stable operation of the water supply system. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating an embodiment of the intelligent rural water supply management method based on digital twins according to the present invention. Figure 2This is a functional block diagram of an embodiment of the intelligent rural water supply management system based on digital twins of the present invention.

[0018] Explanation of icon numbers: 10. Initial running dataset acquisition module; 20. Information gap region determination module; 30. Dynamic control indicator set acquisition module; 40. Response speed optimization sequence acquisition module; 50. Specific node coordinate determination module; 60. Resource limited allocation scheme acquisition module; 70. Problem discovery closed-loop mechanism determination module. Detailed Implementation

[0019] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0020] like Figure 1 As shown, the first embodiment of the present invention proposes a smart management method for rural water supply based on digital twins, including the following steps: Step S100: Collect real-time data by data acquisition devices deployed on key nodes of the water supply network, covering the water supply network from source to end, and obtain the initial operating dataset of the water supply network, wherein the real-time data includes flow indicators and pressure indicators.

[0021] This step involves real-time data acquisition and initial dataset construction for the water supply network. Data acquisition devices such as flow sensors, pressure sensors, and water quality sensors are deployed at key nodes in the rural water supply network, including water sources, booster pump stations, main pipelines, branch pipelines, and end-user interfaces, to achieve full-link coverage from the water source (such as reservoirs and wells) to the end user.

[0022] The core of the collected real-time data includes: 1. Flow rate index: The value range is 0.1~50m³ / h (adjusted according to pipe specifications, 0.5~50m³ / h for main pipes and 0.1~5m³ / h for branch pipes), and the sampling frequency is 1 time / 5~15 minutes by default; 2. Pressure index: The value range is 0.1~1.2MPa (normal water supply pressure 0.2~0.6MPa, booster pump station outlet pressure 0.6~1.2MPa), and the sampling frequency is the same as the flow index.

[0023] The collected real-time data undergoes preliminary noise reduction (removing outliers exceeding ±30% of the normal range) and timestamp alignment, and is integrated to form an initial operational dataset that covers the entire water supply chain and is time-series continuous. The dataset has a sampling period of 5 to 15 minutes, and the daily data volume of a single node is 96 to 288 records, providing basic data support for subsequent data analysis and digital twin modeling.

[0024] Digital twins refer to the construction of virtual models that are completely corresponding to and mapped in real time to the physical entities (pipelines, pumping stations, sensors, etc.) of rural water supply networks. Through real-time data-driven operation, they enable visualized monitoring, simulation analysis, and optimized scheduling of the water supply network's operational status. The core is "virtual mapping + real-time linkage".

[0025] The initial running dataset is a dataset of real-time flow, pressure, and other data collected by acquisition devices at key nodes of the water supply network, covering the entire water supply chain (from source to end). After preliminary processing, it forms a time-series continuous and uniformly formatted dataset, which serves as the basis for subsequent data analysis.

[0026] In this embodiment, digital twin technology refers to a technical system that uses the network topology, geographic information, equipment parameters, and operational data of the rural water supply network as a basis to construct a 1:1 virtual model. Through real-time data acquisition and transmission, millisecond-level data synchronization between the physical entity and the virtual model is achieved. Based on the virtual model, the system enables visualized monitoring of the water supply network's operational status, hydraulic simulation analysis, fault simulation and deduction, and resource allocation optimization. Its core implementation process includes three stages: virtual model construction, real-time synchronization of physical and virtual data, and virtual model simulation and optimization application. Specific implementation details are integrated into each step of the method and are disclosed in detail at the corresponding stages to ensure that those skilled in the art can build a digital twin system and implement it in practice based on the specification.

[0027] 1. Digital Twin Virtual Model Construction Stage Based on the physical topology data, geographic information data, equipment attribute data, and hydraulic characteristic data of the rural water supply network, a three-dimensional virtual model is constructed using BIM+GIS integrated modeling technology. The modeling steps are as follows: (1) Basic data collection: Collect the latitude and longitude coordinates, pipe diameter / material / length / burial depth, pump power / head, sensor deployment location / collection parameters, and other basic data of the physical entities of the water supply network, such as water source, pumping station, pipeline, valve, and sensor, to form a standardized modeling dataset; (2) BIM refined modeling: Use Revit software to construct a refined three-dimensional model of each physical entity of the water supply network, and define the hydraulic characteristics of the pipeline (friction resistance coefficient, local... (2) Resistance coefficient), equipment operating parameters (rated flow rate, rated pressure) and other attributes; (3) GIS geographic mapping: import the BIM model into ArcGIS software to complete the accurate mapping between the three-dimensional model and the actual geographic space, so that the geographic coordinates of the model are completely consistent with the physical entities; (4) Model lightweighting: use FBX format to perform lightweighting of the BIM+GIS fusion model, adapt to the needs of edge computing and cloud deployment, and reduce the computing power consumption of the model operation; (5) Model parameter configuration: set up dynamically adjustable parameter interfaces in the virtual model, including flow rate, pressure, sampling frequency, resource allocation weight, etc., to realize the parameter linkage between the model and the physical entities.

[0028] 2. Real-time synchronization stage of physical and virtual data The physical water supply network and the virtual model are synchronized in real time at the millisecond level by using edge computing + 5G / IoT communication. The synchronization mechanism is as follows: (1) Data acquisition layer: Sensors deployed at key nodes collect real-time data such as flow and pressure. After local cleaning and aggregation by the edge computing module, the data is transmitted to the cloud digital twin platform through the 5G / NB-IoT Internet of Things protocol; (2) Data parsing layer: The cloud platform sets up a data parsing interface to convert and verify the format of the transmitted real-time data to ensure that the data matches the parameter dimensions of the virtual model; (3) Data synchronization layer: A precise synchronization mechanism with timestamp alignment is established. The error between the real-time data acquisition time of the physical entity and the parameter update time of the virtual model is controlled within 100ms, so as to realize the real-time mapping of the virtual model to the physical entity's operating status; (4) Abnormal data synchronization fault tolerance: If data interruption / loss occurs on the physical side, the virtual model automatically calls the supplementary data of the edge computing module to update the parameters. At the same time, it marks the abnormal data area in the model to ensure the continuity of the synchronization process.

[0029] 3. Virtual Model Simulation and Optimization Application Stage The core simulation and optimization application scenarios of the digital twin virtual model include operation status visualization monitoring, hydraulic characteristic simulation, fault simulation and deduction, and resource allocation optimization. The implementation details of each scenario are as follows: (1) Operation status visualization monitoring: The operation parameters such as flow rate, pressure, and sampling frequency of each node are displayed in real time in the virtual model. The abnormal parameters are marked by color gradient (e.g., red indicates a sudden drop in pressure and yellow indicates a deviation in flow rate) to realize the global visualization of the operation status of the water supply network; (2) Hydraulic characteristic simulation: Based on the pipeline hydraulic parameters and real-time operation data of the virtual model, the Hardy-Cross method is used to perform hydraulic adjustment calculations to simulate the flow and pressure distribution of the water supply network under different operating conditions and predict the trend of changes in the hydraulic characteristics of the pipeline network; (3) Fault simulation and deduction: The fault scenarios such as pipeline leakage, sensor failure, and information gap are simulated in the virtual model to deduce the impact range and propagation speed of the fault on the surrounding nodes and provide a simulation basis for fault handling; (4) Resource allocation optimization: The limited resource allocation scheme is input into the virtual model to simulate the fault handling efficiency and resource consumption under different resource allocation strategies. Through multi-scenario simulation comparison, the resource allocation scheme is optimized to maximize the resource utilization efficiency.

[0030] Step S200: Apply data analysis algorithms to process the continuous data stream based on the initial running dataset, identify abnormal fluctuation patterns, and determine potential information gap areas in the water supply network, where information gap areas refer to transmission segments with missing data.

[0031] This step involves identifying anomalies in water supply data and locating information gaps. The initial operational dataset is input into a preset data analysis algorithm (such as a time series analysis algorithm or an isolated forest algorithm) to perform trend analysis and fluctuation detection on the continuous data stream, identifying abnormal fluctuation patterns in flow and pressure indicators (such as a sudden drop in flow rate ≥20% / h or a sudden change in pressure ≥0.1MPa / 10 minutes).

[0032] Simultaneously, through data continuity verification, information gaps in the water supply network (i.e., transmission segments with missing data, such as a section of pipeline lacking flow / pressure data or data transmission interruptions exceeding 30 minutes) are located. The accuracy of abnormal fluctuation identification is ≥90%, and the information gap location error is ≤50m, ensuring timely detection of data anomalies and potential transmission risks in the water supply network operation.

[0033] Information gaps refer to data loss segments in a water supply network caused by sensor malfunctions, data transmission interruptions, or other reasons. These information gaps prevent the acquisition of real-time operational data such as flow and pressure, making them a weak link in water supply management.

[0034] Step S300: If the information gap area determined by the data analysis algorithm exceeds the preset threshold, the edge computing module is activated to perform local aggregation of the data in the information gap area to obtain a refined dynamic control index set, wherein the dynamic control index set includes the aggregated flow velocity value and pressure change value.

[0035] This step involves information gap processing and dynamic indicator aggregation. A preset threshold for information gap regions is established, with a specific range of 2-8 regions (adjustable based on the size of the water supply network: 2-3 for small networks, 3-5 for medium networks, and 5-8 for large networks). When the number of information gap regions determined by the data analysis algorithm exceeds this preset threshold, the edge computing module is automatically activated.

[0036] Edge computing modules are deployed near key nodes in the water supply network (such as pumping stations and pipeline junctions) to perform local aggregation and interpolation on real-time data from surrounding nodes in areas with information gaps. Redundant data is removed, and core indicators are retained, ultimately resulting in a refined set of dynamic control indicators. This indicator set includes: 1. Flow velocity after aggregation: range of 0.2~3.0m / s (normal operating flow velocity 0.5~2.0m / s); 2. Pressure change value: The value ranges from -0.3 to 0.3 MPa (a negative sign indicates a decrease in pressure, and a positive sign indicates an increase in pressure). The dynamic control indicator set is updated once every 3 to 5 minutes to ensure dynamic control over information gap areas.

[0037] The dynamic control indicator set is a set of core indicators obtained by locally aggregating and refining the data of surrounding nodes in the information fault area. It includes aggregated flow velocity values ​​and pressure change values, and is used to realize dynamic monitoring and control of the information fault area.

[0038] Step S400: Input the refined dynamic control index set into the machine learning classification algorithm for classification, determine the priority level of problem discovery, and obtain the classified response speed optimization sequence, wherein the response speed optimization sequence is sorted according to the urgency of the event.

[0039] This step involves problem prioritization and response sequence optimization. The refined set of dynamic control indicators is input into a machine learning classification algorithm (such as support vector machine or random forest). The algorithm classifies water supply network operation problems into three priority levels based on the dimensions of "event impact scope, urgency, and handling difficulty": 1. High priority: Affecting ≥50 users, pressure / flow exceeding normal range by more than 50%, or posing a risk of pipe rupture; 2. Medium priority: Affecting 20-49 users, with abnormal load / traffic exceeding normal range by 20%-50%; 3. Low priority: Number of affected users < 20, and pressure / traffic abnormalities exceeding the normal range < 20%.

[0040] 4. Based on priority levels, generate an optimized response speed sequence sorted from highest to lowest urgency of events, where the response time for high-priority events is ≤30 minutes, medium-priority events are ≤1 hour, and low-priority events are ≤2 hours, ensuring that urgent issues are handled quickly.

[0041] The response speed optimization sequence is a sequence formed by sorting water supply network operation problems according to their urgency (priority) using a machine learning classification algorithm. This sequence clarifies the response time limit for each type of problem and optimizes the handling efficiency.

[0042] Step S500: Extract high-priority events from the classified response speed optimization sequence, apply graph theory algorithm to construct a connectivity graph of the water supply network for the high-priority events, and determine the coordinates of specific nodes for rapid location, where the coordinates of specific nodes correspond to the key connection points of the transmission pipeline.

[0043] This step involves locating high-priority events and determining node coordinates. High-priority events (such as pipeline leaks and sudden pressure drops) are extracted from the response speed optimization sequence. For these events, graph theory algorithms (such as shortest path algorithms and connectivity analysis algorithms) are applied, combined with the pipeline layout and node distribution of the rural water supply network, to construct a connectivity graph of the water supply network (nodes are pipeline connection points and pumping stations, and edges are pipeline segments).

[0044] A connected graph is a water supply network topology structure built based on graph theory algorithms. It uses key pipeline connection points and pump stations as nodes and pipeline segments as edges to quickly locate the source nodes of abnormal events.

[0045] By analyzing the connectivity graph, the source node of the abnormal event can be located, and the specific node coordinates for rapid positioning (accuracy ≤10m) can be determined. These specific node coordinates correspond to the key connection points of the transmission pipeline (such as pipeline interfaces, valve nodes, and pump station outlet nodes), providing accurate positioning support for maintenance personnel to quickly arrive at the scene and handle the problem.

[0046] Step S600: After obtaining the specific node coordinates for rapid positioning, the entire tracking record is updated through a distributed database. By integrating historical data and current indicators, an updated resource-limited allocation scheme is obtained, which optimizes the distribution of maintenance resources.

[0047] This step involves data updating and optimized allocation of maintenance resources. After obtaining the specific node coordinates for rapid location, data such as node anomaly information, location results, and handling progress are uploaded to the distributed database in real time, updating the entire tracking record of the water supply network operation (record retention period ≥ 1 year).

[0048] The distributed database integrates the node's historical operational data (flow and pressure change data for the past 3 to 12 months) with current dynamic control indicators, and uses resource allocation algorithms (such as genetic algorithms and greedy algorithms) to generate an updated resource-limited allocation scheme by combining the number of maintenance personnel, the distribution of maintenance equipment, and the distance to the faulty node in the rural water supply network.

[0049] The limited resource allocation scheme is a plan that combines historical data with current operating indicators to optimize the allocation of limited resources such as maintenance personnel and equipment. Its core is to maximize the efficiency of fault handling under resource constraints.

[0050] The updated resource allocation scheme optimizes the distribution of maintenance resources (personnel and equipment), clearly defining the responsible person, arrival time, and required equipment for each fault node. For high-priority events, the arrival time of maintenance personnel is ≤30 minutes, and the dispatch time of maintenance equipment is ≤15 minutes, ensuring efficient utilization of maintenance resources.

[0051] Step S700: If the deviation of the flow index is greater than the deviation of the pressure index in the updated resource limited allocation scheme, the sampling frequency of the data acquisition device is adjusted to obtain an optimized dynamic control feedback loop. An alarm signal is generated based on the optimized dynamic control feedback loop and distributed to the terminal device. It is determined whether the response speed meets the threshold requirement, and the final problem discovery closed-loop mechanism is determined.

[0052] This step involves feedback loop optimization and determining the closed-loop mechanism. The absolute values ​​of the flow rate deviation and pressure rate deviation are calculated in the updated resource-limited allocation scheme. 1. Flow rate deviation: Difference between current flow rate and normal flow rate ÷ Normal flow rate × 100%, with a range of -50% to 50%; 2. Pressure index deviation: The difference between the current pressure and the normal pressure ÷ the normal pressure × 100%, with a value range of -50% to 50%.

[0053] 3. If the absolute value of the flow rate deviation is greater than the absolute value of the pressure deviation (i.e., |flow rate deviation| > |pressure deviation|), then adjust the sampling frequency of the data acquisition device in the corresponding area: change the sampling frequency from the default 1 time / 5~15 minutes to 1 time / 1~3 minutes to enhance the sensitivity of flow rate anomaly monitoring and obtain an optimized dynamic control feedback loop.

[0054] Based on the optimized dynamic control feedback loop, tiered alarm signals are generated (high priority: audible and visual alarm, medium priority: SMS alarm, low priority: system notification) and distributed to terminal devices such as maintenance terminals and management personnel's mobile phones. Simultaneously, it is determined whether the response speed meets the threshold requirements (high priority ≤ 30 minutes, medium priority ≤ 1 hour, low priority ≤ 2 hours). If it does, the loop is confirmed as closed; if not, the process reverts to step S400 to re-optimize the response sequence, ultimately establishing a closed-loop mechanism for problem discovery: "data acquisition → anomaly identification → location and handling → feedback optimization," ensuring that any abnormalities in the water supply network are under full control.

[0055] The dynamic control feedback loop is a mechanism that adjusts the sampling frequency of the data acquisition device to enhance the sensitivity of monitoring abnormal indicators, forming a loop of "data acquisition → anomaly identification → optimized sampling → precise monitoring" to ensure dynamic adaptation to the operating status of the water supply network.

[0056] The problem discovery closed-loop mechanism is a complete closed loop covering the entire process of "data collection → anomaly identification → information gap handling → priority classification → location and handling → resource allocation → feedback optimization", ensuring that water supply network operation anomalies are detected in a timely manner, handled quickly, and continuously optimized.

[0057] Furthermore, the intelligent rural water supply management method based on digital twins provided in this embodiment includes step S100 as follows: Step S110: Obtain flow and pressure indicators through data acquisition devices deployed at key nodes of the water supply network.

[0058] The flow rate and pressure index are derived using the following formulas: (1) In formula (1), For the first Key Nodes Flow rate metrics at any given time, in units of ; For the data acquisition device at the basic sampling interval, Internally collected flow volume, unit is ; The basic sampling interval (unit: h, default is 0.01h, i.e. 36s); For the first Key Nodes Pressure reading at any given time (gauge pressure, unit: MPa); For the first Actual measured value of node pressure sensor (absolute pressure, unit: MPa). Atmospheric pressure (unit: MPa, default value is 0.1013 MPa). Number the key nodes ( , (where is the total number of nodes). The control logic of formula (1) is the neighborhood average interpolation logic of the topology network node traffic. Its core value is: 1. Topology-driven interpolation: Based on the physical topology adjacency relationship, it ensures the spatial rationality of the interpolation result; 2. Data completion capability: It effectively solves the problem of state estimation when node traffic data is missing or abnormal; 3. Computational efficiency: It only requires neighborhood node traffic data, has low computational complexity, and is suitable for real-time data completion scenarios.

[0059] Flow and pressure metrics are obtained through data acquisition devices deployed at key nodes in the water supply network. For example, in a smart rural water supply management system, sensors are installed at the junctions of main pipelines and pump station inlets. These devices monitor water flow velocity and pipeline pressure in real time, providing fundamental data for subsequent analysis.

[0060] Step S120: Spatial mapping of flow and pressure indicators between node coordinates and topology structure to obtain mapping results containing geographical location information.

[0061] The mapping result containing geographic location information is obtained using the following formula: (2) In formula (2), For the first The nodes contain mapping results of geographic location information. For the first Node latitude coordinates For the first Longitude coordinates of the node For the first The topological coding of nodes (dimensionless, representing the hierarchy / connection relationship of nodes in the pipeline network topology). The control logic of formula (2) is the spatial mapping logic of multi-dimensional information of pipeline network nodes. Its core value is: 1. Multi-dimensional integration: integrating geographic coordinates, operation indicators and topological relationships into a unified structure, realizing the spatial expression of pipeline network data; 2. Visualization support: the five-tuple structure can be directly connected to the GIS system to realize the geographic visualization and status visualization of the pipeline network; 3. Topological consistency: the topological coding ensures that the spatial mapping results correspond one-to-one with the physical topology of the pipeline network, avoiding information fragmentation.

[0062] The flow rate and pressure indices are spatially mapped between node coordinates and the topology to obtain a mapping result containing geographic location information. Specifically, the topology of the water supply network is first obtained, including pipe connections and node location coordinates, for example, by using a GIS system to label the latitude and longitude coordinates of each node. Then, the collected flow rate indices, such as cubic meters per hour, and pressure indices, such as kPa, are associated with these coordinates. Using spatial interpolation methods, such as Kriging interpolation, the data points are extended to the entire network, forming a mapping result map containing geographic locations, where color gradients represent pressure distribution, thus visually displaying the operating status of the water supply network. This mapping not only reveals local anomalies, such as a sudden drop in pressure at a node, but also provides spatial basis data for calculating pipe resistance. In practical applications, such as in the water supply network of a rural intelligent water supply management system, the mapping result shows that the peak flow rate at nodes in the main urban area reaches 500 cubic meters per hour, while the pressure at suburban nodes remains stable at 300 kPa, which helps identify potential leak points.

[0063] Step S130: Determine the pipe resistance based on the mapping result, pipe diameter, and pipe material properties, and trigger the terminal flow verification logic based on the abnormal fluctuations after the pipe resistance is correlated with the water supply period to obtain the terminal flow.

[0064] Pipeline resistance is calculated using the following formula: (3) In formula (3), For the first Friction resistance along the pipe at the node connection (unit: Pa); The friction factor is dimensionless; 0.02 for cast iron pipe and 0.015 for PE pipe. For the first Length of the pipe connecting the node (unit: m); For the first Pipe diameter of the node connection pipe (unit: m); Density of water (unit: (Default value is 1000). For the first Average flow velocity in the node pipe (unit: m / s). The control logic of formula (3) is the quantitative calculation logic of friction resistance along the water supply network. Its core value is: 1. Physical drive: Based on the classical fluid mechanics formula, it ensures the accuracy and physical rationality of the resistance calculation; 2. Parameter correlation: It incorporates key attributes such as pipe material, pipe diameter, pipe length, and flow velocity into the calculation, and fully reflects the influencing factors of pipe resistance; 3. Engineering adaptation: For the water supply network scenario, it clarifies the resistance coefficient values ​​of different pipe materials, which is convenient for actual engineering applications.

[0065] Terminal traffic is calculated using the following formula: (4) In formula (4), For the first Terminal traffic corresponding to the node (unit: ); for Weight of the water supply period to which the time belongs (dimensionless, 1.2 for peak hours, 1.0 for off-peak hours, and 0.8 for low-peak hours); For the first Rated pressure loss of node pipeline (unit: Pa). The control logic of formula (4) is the resistance-time period joint verification logic of the water supply network terminal flow. Its core value is: 1. Multi-factor coupling correction: Simultaneously introduce the pipeline friction resistance and water supply time period weight to fully reflect the impact of pipeline pressure loss on terminal flow; 2. Anomaly detection trigger: By comparing with the rated pressure loss, it automatically identifies abnormal flow fluctuations and provides quantitative basis for fault diagnosis; 3. Strong scenario adaptability: Customize the time period weight value according to the time period water use characteristics of the water supply network to improve the verification accuracy.

[0066] Based on the mapping results, pipe diameter, and pipe material properties, the pipe resistance is determined. Abnormal fluctuations in the pipe resistance correlated with water supply periods trigger terminal flow verification logic to obtain the terminal flow rate. Specifically, the pipe resistance is estimated using the Darcy-Weisbach formula, where the resistance coefficient depends on the pipe diameter (e.g., 50 mm to 200 mm) and pipe material properties (e.g., the roughness of PVC or cast iron). For example, the roughness of a PVC pipe is 0.0015 mm. Combining the flow rate and pressure data from the mapping results, the resistance value for each pipe segment is calculated, e.g., resistance = 8fLQ. 2 / (π 2 gD 5Here, f is the friction factor, L is the length, Q is the flow rate, g is the gravitational acceleration, and D is the pipe diameter, thus obtaining the overall resistance distribution of the network. Next, the resistance value is correlated with water supply periods, such as the morning peak hours of 7-9 am, to monitor for abnormal fluctuations. For example, a sudden increase in resistance of 20% indicates a possible blockage. This triggers the terminal flow verification logic, which verifies the actual terminal flow by comparing the difference between upstream and downstream flow rates. For example, if the upstream flow rate is 100 cubic meters per hour and the downstream flow rate is 80, the terminal flow rate is verified to be 80, and the model parameters are adjusted. In practice, this method can effectively detect abnormal resistance during off-peak hours at night, thereby preventing pipe burst accidents and obtaining accurate terminal flow data to optimize water supply scheduling.

[0067] Step S140: Summarize the terminal flow and perform data cleaning and noise reduction to obtain the initial operating dataset of the water supply network.

[0068] The initial operational dataset for the water supply network is derived using the following formula: (5) In formula (5), This is the initial operational dataset for the water supply network. Data cleaning and denoising operators (dimensionless, including outlier removal and null filling); This represents the total number of terminal nodes. The set merging operator represents the aggregation of terminal flow across all nodes. The control logic of formula (5) is the aggregation and cleaning logic of terminal flow data in the water supply network. Its core value is: 1. Global aggregation: Merging the scattered terminal verification flow data into a unified set to achieve a centralized presentation of the flow status of the entire network; 2. Quality assurance: Improving data reliability through standardized cleaning and denoising operators to provide high-quality input for subsequent analysis; 3. Process closed loop: Forming a complete initial dataset construction link from "terminal verification flow → global aggregation → data cleaning", receiving the terminal flow verification results, and providing data support for subsequent pipeline network analysis.

[0069] The terminal flow data is aggregated and cleaned and denoised to obtain the initial operational dataset of the water supply network. For example, after aggregating all terminal flow data, median filtering is used to remove noise and outliers are cleaned, thereby improving data quality. Through the above process, efficient monitoring and optimization of the water supply network can be achieved.

[0070] Preferably, the intelligent rural water supply management method based on digital twins provided in this embodiment includes step S200: Step S210: Obtain the initial running dataset and construct a continuous data flow matrix, which contains flow and pressure values ​​aligned to a timestamp sequence.

[0071] Construct a continuous data stream matrix using the following formula: (6) In formula (6), A continuous data stream matrix (three-dimensional matrix); For the first node Flow rate at any given time (unit: ); For the first node Flow rate at any given time (unit: ); For the first node Pressure value at any given time (unit: MPa); For the first node Pressure value at any given time (unit: MPa); For the first timestamp ( , (where is the length of the time series). The control logic of formula (6) is the time-series structured encapsulation logic of water supply network operation data. Its core value is: 1. Three-dimensional structuring: transforming the original data into a three-dimensional matrix to adapt to modern analysis methods such as digital twins and deep learning; 2. Time series alignment: ensuring the time consistency of the data through the timestamp column to avoid time series chaos; 3. Feature integrity: retaining the three core information types of flow, pressure and timestamp at the same time to provide comprehensive input for subsequent analysis.

[0072] First, an initial operational dataset of the water supply network is obtained, for example, by collecting real-time flow and pressure data through sensors deployed on pipeline nodes, thereby constructing a continuous data flow matrix, wherein the continuous data flow matrix aligns these data according to a timestamp sequence to form a structured numerical sequence.

[0073] Step S220: Traverse the continuous data stream matrix to identify null gaps caused by numerical interruptions, and retrieve the synchronous operation data of adjacent nodes in the pipeline network topology based on the time location information of the null gaps.

[0074] The null value marker is derived using the following formula: (7) In formula (7), For the first node Null value flag at any given time (Boolean value, 1 = null, 0 = valid value); The null value is identified. The control logic of formula (7) is the automatic identification logic of null values ​​in water supply network data. Its core value is: 1. Precise positioning: The data interruption point is accurately identified with "node-time" as the granularity; 2. Simple and efficient rules: The judgment rule based on zero flow / pressure value has low computational complexity and is suitable for real-time traversal; 3. Laying the groundwork for subsequent completion: The null value marking result is directly used for the retrieval of synchronous running data of adjacent nodes in the topology, providing a preliminary basis for data completion.

[0075] The synchronization data of topologically adjacent nodes is obtained through the following formula: (8) In formula (8), For the first node Synchronized running data of adjacent nodes in the time topology; For the first The set of topologically adjacent nodes of a node; Adjacent nodes exist Flow rate at any given time (unit: ); Adjacent nodes exist The pressure value at any given moment; Identify adjacent nodes. The control logic of formula (8) is the neighborhood synchronization aggregation logic of the topological network node state. Its core values ​​are: 1. Topology-driven data completion: Based on the physical topological adjacency relationship, it ensures the spatial rationality of the completed data; 2. Synchronization timing guarantee: Strictly aligns to the same time. 1. Avoid timing misalignment interference; 2. Collective storage: Encapsulate the states of neighboring nodes into a unified set to facilitate subsequent interpolation calculations and state analysis.

[0076] The continuous data stream matrix is ​​traversed to identify null gaps caused by numerical interruptions. For example, the time series is scanned in the continuous data stream matrix, and if the flow rate value is missing at a certain timestamp, it is marked as a null gap. Then, based on the time location information of the null gap, the synchronous operation data of the topologically adjacent nodes in the pipeline topology are retrieved. Specifically, the pipeline topology is a graph structure model, including nodes such as pump stations and valves and their connection relationships. Reference values ​​are obtained by querying the same timestamp data of adjacent nodes such as upstream pipeline nodes.

[0077] Step S230: Calculate the interpolation error of the synchronous running data. If the interpolation error exceeds the preset value, mark the null gap as an abnormal fluctuation mode caused by the transmission link failure. Interpolation error is calculated using the following formula: (9) In formula (9), For the first Festival Interpolation error at time (unit: %); For the first node Interpolated flow rate at time (unit: ); For the first node The effective flow value at any given time. The control logic of formula (9) is the logic for verifying the rationality of the interpolated flow result. Its core value is: 1. Time series stability verification: By utilizing the time series stability of the pipeline flow, the interpolation result is verified by comparing adjacent time points; 2. Abnormal pattern recognition: Large errors are directly associated with abnormal scenarios such as transmission link failures, providing a basis for fault location; 3. Data quality assurance: Only the completed results with errors within the threshold are retained to ensure the reliability of the data after repair.

[0078] No. node Interpolated flow rate at time This can be derived from the following formula: (10) In formula (10), For the first The number of neighboring nodes of a node. The control logic of formula (10) is the neighborhood average interpolation logic of the topology network node traffic. Its core value is: 1. Topology-driven: Based on the physical connection relationship, it ensures the spatial rationality of the interpolation result; 2. Data completion: It effectively solves the state estimation problem when node traffic data is missing / abnormal; 3. Computational efficiency: It only requires neighboring node traffic data, has low complexity, and is suitable for real-time completion scenarios.

[0079] Abnormal fluctuation markers are derived using the following formula: (11) In formula (11), Mark abnormal fluctuations (Boolean value, 1 = abnormal, 0 = normal); A threshold is preset for the interpolation error (unit: %, default value is 5). The control logic of formula (11) is the threshold-based identification logic for abnormal fluctuations in water supply network data. Its core value is: 1. Threshold-driven identification: using the preset error threshold as the boundary, it quickly distinguishes between normal fluctuations and abnormal faults; 2. Completion-fault linkage: while completing the data, it completes the anomaly detection, realizing the closed loop of "data repair + fault perception"; 3. Intuitive operation and maintenance support: Boolean flags are directly mapped to fault status, which is convenient for subsequent alarms and location.

[0080] Then, the interpolation error of the synchronous running data is calculated. The interpolation error is estimated by using a linear interpolation method to determine the deviation between the missing value and the adjacent data. For example, the average value of the adjacent nodes is used to fill the gap and the difference is compared. If the deviation exceeds a preset value, such as 5%, the gap is marked as an abnormal fluctuation pattern caused by a transmission link failure. This abnormal fluctuation pattern reflects the interruption of data transmission, which is caused by a problem with the sensor communication link, thus providing a basis for subsequent analysis.

[0081] Step S240: Based on the spatial distribution trajectory of the abnormal fluctuation pattern in the pipeline topology, delineate the range of pipe sections where data loss occurred, and determine the information gap area in the water supply network that is the data loss transmission segment.

[0082] The set of information fault regions is obtained through the following formula: (12) In formula (12), A collection of information fault areas (segment number + node number); Assign a node number to represent the target node in the pipeline network; The pipe section number (dimensionless) represents the node. A connected pipe section; For the first A set of pipe segments connected by nodes; For pipe section number (dimensionless); This represents the total number of moments within the statistical time window, i.e., the total number of time sampling points included in the statistics. For nodes The total number of anomalies within the time window represents the cumulative duration of the information gap. The control logic of formula (12) is the topology-time joint identification logic for the information gap area of ​​the water supply network. Its core value is: 1. Time dimension filtering: only retains the continuous faults with an anomaly rate of more than 50%, and excludes occasional interference; 2. Topology dimension mapping: associates node anomalies with connected pipe segments, and accurately locates the range of pipe segments where data is lost; 3. Aggregated output: outputs in the form of (node, pipe segment) pairs, which facilitates subsequent fault location and operation and maintenance scheduling.

[0083] Based on this, the spatial distribution trajectory of the abnormal fluctuation pattern in the pipeline topology is used to delineate the range of pipe segments where data loss has occurred. For example, the spatial path of multiple consecutive gaps in data loss is traced, and an area is delineated from the starting node to the ending node. This identifies information gaps in the water supply network that represent data loss transmission segments. For instance, in a rural intelligent water supply management system, if a main pipeline segment shows continuous anomalies, that segment is identified as an information gap. In one embodiment, the above process can effectively locate potential fault areas in the water supply network.

[0084] Furthermore, the intelligent rural water supply management method based on digital twins provided in this embodiment includes step S300 as follows: Step S310: Calculate the cumulative link length of the identified information fault regions in the pipeline topology, and use the cumulative link length as the basis for determining the fault scale.

[0085] The cumulative link length in the information gap region is calculated using the following formula: (13) In formula (13), The cumulative link length (in meters) in the information gap region. For the first The length of the root canal segment (unit: m). The control logic of formula (13) is the quantitative evaluation logic of the scale of information faults in the water supply network. Its core value is: 1. Scale quantification: converting discrete fault segments into cumulative lengths to intuitively measure the scope of fault impact; 2. Decision support: providing quantitative basis for fault priority determination and operation and maintenance resource allocation; 3. Topological association: directly calculating based on the set of information fault areas to ensure consistency with the fault location results.

[0086] The water supply network topology is modeled as a graph structure, where nodes represent pump stations or valves and edges represent pipeline links, facilitating the statistical calculation of the cumulative link length of identified information fault areas. In one embodiment, all information fault areas in the network topology are first traversed, for example, for the main pipeline segment of a rural water supply intelligent management system. The affected link length within each fault area is measured one by one, such as the physical distance from the starting node to the ending node. These lengths are accumulated to obtain a total cumulative value; for example, if one fault area covers 3 kilometers of pipeline and another covers 2 kilometers, the total is 5 kilometers. Then, this cumulative link length is used as the basis for determining the fault scale. The determination is made by comparing the cumulative value with a preset threshold to assess the degree of impact. The preset threshold can be set according to the network scale, such as 10 kilometers, to determine whether further intervention is needed.

[0087] Step S320: If the fault scale determination criteria exceed the preset threshold, then send an edge collaboration trigger signal to the edge computing module at the physical boundary of the information fault area.

[0088] The edge-coordinated trigger signal is derived using the following formula: (14) In formula (14), This is the edge collaboration trigger signal (Boolean value, 1 = trigger, 0 = do not trigger); A threshold is preset for the fault size (unit: m, default value is 500). The control logic of formula (14) is the edge collaboration triggering logic driven by the fault size of the water supply network. Its core value is: 1. Triggering on demand: Edge collaboration is only started when the fault size exceeds the threshold, balancing computational efficiency and resource consumption; 2. Fault perception linkage: Directly associates the cumulative length of the fault with information, realizing the closed loop of fault perception and edge scheduling; 3. Real-time response: Boolean trigger signals can be quickly transmitted to the edge module, improving the timeliness of fault handling.

[0089] If the cumulative link length exceeds a preset threshold, the rural water supply intelligent management system automatically sends an edge collaboration trigger signal to the edge computing module at the physical boundary of the information fault area. This signal is transmitted through a wireless communication protocol. The edge computing module is deployed on the boundary node to activate the local response mechanism, thereby ensuring timely handling of large-scale faults.

[0090] Step S330: After receiving the edge collaboration trigger signal, the edge computing module collects high-frequency discrete sampling data and performs local aggregation operation to obtain the aggregated flow velocity value and pressure change value.

[0091] The combined flow rate is obtained using the following formula: (15) In formula (15), For the first root canal segment The combined flow velocity at any given time (unit: m / s); The time window for edge computing aggregation (unit: seconds, default value is 60). For the first root canal segment Interpolation flow of neighboring nodes at time (unit: ); For the first Diameter of the root canal segment (unit: m); Pi is the circumference of a circle. The control logic of formula (15) is the edge aggregation logic of high-frequency flow data of water supply network. Its core value is: 1. Physical driving conversion: the flow velocity is derived from the flow rate and pipe diameter, which conforms to the basic laws of fluid mechanics; 2. Noise smoothing: the time window averaging operation effectively suppresses high-frequency sampling noise and improves data stability; 3. Edge computing adaptation: the local aggregation operation is lightweight and suitable for the real-time computing needs of edge nodes.

[0092] The pressure change value is obtained using the following formula: (16) In formula (16), For the first root canal segment Pressure change over time (unit: MPa / s); For the first root canal segment Interpolated pressure of adjacent nodes at time (unit: MPa); For the first root canal segment Interpolated pressure of adjacent nodes at time (unit: MPa). The control logic of formula (16) is: 1. Noise suppression and trend extraction: By averaging over a time window, high-frequency sensor noise and instantaneous pressure fluctuations are filtered out to extract the true trend of pressure change in the pipe section. 2. Fault perception: Pressure change rate It is a sensitive indicator of pipeline network failures—leakage, pipe bursts, valve adjustments, etc. can all cause abnormal pressure fluctuations, which can be used for early fault warning; 3. Edge computing adaptation: It only relies on local time windows and interpolation pressure of pipe segments, with low computational complexity, and is suitable for real-time execution at edge nodes (such as gateways, RTUs), realizing lightweight deployment of "edge aggregation + cloud analysis" and reducing data transmission pressure.

[0093] After receiving the edge collaboration trigger signal, the edge computing module immediately starts high-frequency discrete sampling data acquisition, such as collecting flow and pressure data 10 times per second, covering the fault boundary area; then it performs local aggregation operations, such as averaging the collected data to obtain aggregated flow velocity and pressure change values. The aggregation operation involves summarizing data within a time window, such as using a sliding window method to calculate the average of the data in the last 5 minutes, thereby smoothing noise and extracting key indicators.

[0094] Step S340: Based on the aggregated flow rate and pressure change values, perform structured encapsulation to generate a refined set of dynamic control indicators.

[0095] The refined set of dynamic control indicators is derived using the following formula: (17) In formula (17), The result is a refined set of dynamic control indicators (structured dataset). The control logic of formula (17) is the structured refinement logic of water supply network fault status data. Its core value is: 1. Data dimensionality reduction: only the core indicators of the faulty pipe section are retained, redundant data is removed, and the efficiency of subsequent processing is improved; 2. Structured encapsulation: the data is organized in the form of (flow rate, pressure change), which is convenient for machine learning, digital twin and other modules to call directly; 3. Accurate fault association: the indicator set is bound to the information gap area to realize accurate perception of fault status.

[0096] The aggregated flow velocity and pressure change values ​​are structured and encapsulated, for example, by organizing these values ​​into a JSON-formatted indicator set, including timestamps, average flow velocity, and pressure difference, thereby generating a refined set of dynamic control indicators. In one embodiment, this process enables precise monitoring of water supply network faults, thereby improving the response efficiency of the rural water supply intelligent management system.

[0097] Preferably, the intelligent management method for rural water supply based on digital twins provided in this embodiment includes step S400: Step S410: Extract the aggregated flow velocity and pressure change values ​​from the refined dynamic control index set, and convert the flow velocity and pressure change values ​​into a vectorized index feature set through a preset feature extraction model.

[0098] The vectorized feature set of indicators is derived using the following formula: (18) In formula (18), For the first Vectorized feature set of root canal segments (4-dimensional vector). It is a normalization operator (maps eigenvalues ​​to the interval [0, 1]); For the first Standard deviation of flow velocity in root canal segment (unit: m / s); For the first Standard deviation of pressure variation in root canal segment (unit: MPa / s); It is a 4-dimensional real space (representing 4 feature dimensions). The control logic of formula (18) is the vectorization and normalization logic of the fault state characteristics of water supply network. Its core value is: 1. Multi-dimensional feature fusion: Combining instantaneous state and historical fluctuations, it comprehensively describes the health of the pipe section; 2. Unified dimensions: Normalization eliminates the differences in dimensions and values ​​between different indicators and improves model compatibility; 3. Machine learning adaptation: The vectorized form can be directly connected to machine learning algorithms such as classification and clustering to realize automated fault diagnosis.

[0099] For a water supply network monitoring system, the first step is to extract aggregated flow velocity and pressure change values ​​from the refined set of dynamic control indicators. For example, these values ​​include the difference between the average flow velocity of the main pipeline section (e.g., 2 meters per second) and the pressure change (e.g., 5 Pa). These values ​​are then transformed using a preset feature extraction model. This feature extraction model can be based on the framework of principal component analysis, mapping the values ​​into multidimensional vectors. For example, flow velocity and pressure can be combined into a vector set with a dimension of 10, thus forming a vectorized set of indicator features.

[0100] Step S420: Input the vectorized indicator feature set into the machine learning classification algorithm to obtain the priority-labeled feature subset.

[0101] Priority labels are derived using the following formula: (19) In formula (19), For the first Priority labels for root canal segments (dimensionless, 1 = high priority, 2 = medium priority, 3 = low priority). This is a random forest classification algorithm (a type of machine learning classification algorithm). The control logic of formula (19) is the machine learning classification logic for the priority of water supply network faults. Its core value is: 1. Intelligent classification: Based on a data-driven machine learning model, it automatically identifies the severity of faults and avoids human judgment bias; 2. Operation and maintenance orientation: The output priority labels can be directly used for emergency repair task sorting to improve operation and maintenance efficiency; 3. Model adaptation: The random forest algorithm is highly adaptable to nonlinear and high-noise data and is suitable for complex pipeline network scenarios.

[0102] When processing water supply network data, the vectorized indicator feature set is input into a machine learning classification algorithm, such as the support vector machine algorithm. This machine learning classification algorithm learns feature patterns through the training dataset. For example, it classifies data on historical water supply events. The machine learning classification algorithm evaluates the weight of each vector and labels its priority. For example, the high-priority feature subset focuses on the part with a sharp drop in pressure, thus obtaining a priority-labeled feature subset. This feature subset is usually reduced to 30% of the original set to highlight key abnormal indicators.

[0103] Step S430: Generate an initial response speed sequence using a sequence generation model based on the priority-labeled feature subset, and then optimize and adjust the initial response speed sequence using an association rule mining algorithm to obtain an optimized response speed sequence.

[0104] The initial response rate sequence is derived using the following formula: (20) In formula (20), The initial response velocity sequence (sorted by node / segment number); For sequence generation model operators (such as LSTM sequence generation). The control logic of formula (20) is the sequence generation logic of the emergency repair response time of water supply network faults. Its core value is: 1. Priority-driven sorting: strictly follow the fault priority to ensure that high-priority faults receive the fastest response; 2. Intelligent time allocation: the sequence generation model can optimize the response order by integrating multiple factors to avoid local optima; 3. Strong executability: the output ordered sequence can be directly connected to the emergency repair scheduling system to generate specific task assignments.

[0105] The optimized response speed sequence is obtained using the following formula: (twenty one) In formula (21), The optimized initial response rate sequence; For association rule mining algorithm operators; For pipe section and The correlation coefficient (dimensionless, range [-1, 1]). The control logic of formula (21) is the association rule optimization logic of the response speed sequence. Its core value is: 1. Correlation-driven cleaning: noise is removed by using the physical correlation of pipe segments, and the real response mode is retained; 2. Regularized sequence reconstruction: missing data is filled in by association mining to improve the integrity of the sequence; 3. Provide high-quality input for closed loop: the optimized sequence can significantly improve the decision efficiency and accuracy of subsequent dynamic control feedback loop.

[0106] An initial response rate sequence is generated using a sequence generation model based on a priority-labeled feature subset. For example, a recurrent neural network (RNN) model can be used as the input feature subset. The RNN model generates an initial response rate sequence through layer-by-layer processing, which may include response steps from slow to fast. Subsequently, the initial response rate sequence is optimized and adjusted using an association rule mining algorithm. This algorithm analyzes the associations between features, such as using the Apriori algorithm to scan the sequence and find frequent itemsets. For example, if the pressure change value exceeds a threshold, a fast response rule is associated, and redundant parts in the initial response rate sequence are adjusted to obtain an optimized response rate sequence, where the sequence elements are refined into a more efficient order.

[0107] Step S440: Obtain the classified response speed optimization sequence based on the optimized response speed sequence, wherein the response speed optimization sequence is sorted according to the urgency of the event.

[0108] The optimized response speed sequence after classification is obtained using the following formula: (twenty two) In formula (22), Optimize the sequence for the response speed after classification; For sorting operators (by priority) Arranged in descending order, with higher priority first). The control logic of formula (22) is the association rule optimization logic of the emergency repair response sequence of water supply network faults. Its core value is: 1. Global optimization: breaking through the local priority limitation of the initial sequence and realizing the global scheduling optimization of multiple faulty pipe segments; 2. Association perception: using the correlation of pipe segments to mine the collaborative / conflict rules and improve resource utilization; 3. Practical adaptation: the optimized sequence is closer to the on-site operation and maintenance constraints and can be directly used to generate emergency repair orders.

[0109] The optimized response speed sequence is obtained by classifying the response speed sequence, wherein the optimized response speed sequence is sorted according to the urgency of the event, such as placing the high-pressure leakage event at the top of the sequence.

[0110] Furthermore, the intelligent rural water supply management method based on digital twins provided in this embodiment includes step S500: Step S510: Extract emergency events whose urgency exceeds a preset threshold from the classified response speed optimization sequence.

[0111] The set of pipe segments corresponding to high-priority events is obtained by the following formula: (twenty three) In formula (23), The set of pipe segments corresponding to high-priority events. The control logic of formula (23) is the screening and focusing logic of emergency events in the water supply network. Its core value is: 1. Emergency event focusing: quickly locate the highest priority event from multiple fault scenarios to avoid resource dispersion; 2. Alarm classification support: provide accurate pipe segment range for first-level alarms and emergency dispatch; 3. Timing consistency: based on the optimized response sequence screening, ensure that the repair timing is consistent with the priority.

[0112] For the monitoring system of water supply network, firstly, emergency events with an urgency level exceeding a preset threshold are extracted from the classified response speed optimization sequence. For example, if the threshold is set to 80%, the response speed optimization sequence contains multiple events such as pipe rupture or abnormal flow. When the urgency level of an event reaches 85%, it is extracted as an emergency event, and these high-risk parts are dealt with first.

[0113] Step S520: Retrieve the topology data of the water supply network for emergency events. The topology data is used to construct a connectivity matrix that includes pipe attributes and flow loads.

[0114] The water supply network connectivity matrix can be constructed using the following formula: (twenty four) In formula (24), The water supply network connectivity matrix (M×M order, where M is the total number of nodes in the network). For nodes To the node Boundary weights (dimensionless) Indicates no connection. Indicates connection strength. For nodes Rated flow rate of the pipe section (The maximum flow rate of the pipeline network). The control logic of formula (24) is the matrix modeling logic of the water supply pipeline network topology and flow attributes. Its core value is: 1. Topology-load fusion: Unifying the encoding of physical connections and flow capacity to provide an accurate pipeline network model for digital twins; 2. Normalized weights: Normalizing edge weights to eliminate the difference in flow rate of different pipe segments and improve model compatibility; 3. Digital twin adaptation: The matrix form can be directly input into algorithms such as graph neural networks and hydraulic simulation to support real-time simulation and decision-making.

[0115] The retrieval process involves searching for topological data of the water supply network for the aforementioned emergency event. This topological data is typically stored in a database and includes pipe lengths, materials, and connection relationships. For example, for a city's water supply network, the retrieval process involves querying pipe segments related to the event, such as connection data from main road A to branch B. This connection data is used to construct a connectivity matrix that includes pipe attributes and flow loads. Specifically, the connectivity matrix is ​​an M×M matrix, where M is the number of nodes. Each row represents the connection between a node and other nodes. Matrix elements record pipe attributes such as a diameter of 0.5 meters and flow loads such as a water volume of 100 cubic meters per hour. Through this matrix representation, the connectivity of the entire network is quantified. For example, during construction, the matrix is ​​first initialized to zero values, and then non-zero elements are filled in according to the topological data. For example, a pipe attribute value of 1 between node 1 and node 2 indicates a connection, and a flow load value is added to form a complete matrix structure. This matrix not only captures static attributes but also integrates dynamic loads to reflect the real-time status.

[0116] Step S530: Calculate the node weights and edge weight distributions through the connectivity matrix, and use Dijkstra's algorithm to determine the key hubs in the water supply network connectivity graph generated by the connectivity matrix.

[0117] The critical hubs of a water supply network are determined using the following formula: (25) In formula (25), It is a key hub node in the water supply network; For the set of pipeline nodes; For nodes To the node The shortest path length (in meters, solved using Dijkstra's algorithm); For nodes Edge weights (with connected matrix) ); The node operator corresponding to the minimum value is used. The control logic of formula (25) is the weighted identification logic of key hub nodes in the water supply network. Its core value is: 1. Topology-weight fusion: It considers both the physical distance between nodes and the connection flow weight, which is more in line with the characteristics of the water supply system than the traditional degree centrality; 2. Global connectivity orientation: The identified hub node is the optimal dispatch center of the whole network, which can maximize the emergency response efficiency and fault impact control; 3. Algorithm interpretability: Based on Dijkstra's shortest path, the result is intuitive and interpretable, which is easy for operation and maintenance personnel to understand and trust.

[0118] The node weights and edge weights are calculated using the connectivity matrix. For example, node weights can be calculated based on degree centrality, i.e., the sum of the number of connections a node has. A node connected to 5 pipes would have a weight of 5. Edge weights, on the other hand, consider pipe length and flow rate. For instance, an edge weight is the ratio of length to flow rate, such as 10 meters / 100 cubic meters = 0.1. The Dijkstra algorithm, a shortest path algorithm, is used to determine key nodes in the water supply network connectivity graph generated from the connectivity matrix. This algorithm calculates the minimum distance to each node by gradually expanding from the starting node using a greedy strategy. For example, in a connected graph, the matrix is ​​converted into a graph structure, with nodes representing water stations and edges representing pipes. Dijkstra's algorithm initializes the distance array to infinity, and then starts the relaxation operation from the source node, updating the distance values ​​of adjacent nodes. For example, if the edge weight from the source node to node A is 2, then the distance of A is updated to 2. A priority queue is used to select the node with the smallest distance to continue expanding until the distances of all nodes are determined, thereby finding the key hub with the shortest path. For example, the central pump station is identified as a hub because it has the smallest cumulative edge weight. This method ensures efficient location of core nodes in complex networks.

[0119] Key hub nodes are hub nodes in water supply networks that simultaneously integrate topological connectivity and pipeline physical weights. After global traversal using the shortest path algorithm, they are identified as having the shortest cumulative weighted path and play a core control role in network connectivity and scheduling.

[0120] Step S540: Match the physical location of the key hub with the spatial coordinates in the geographic information system. The spatial coordinates are used to determine the specific node coordinates for rapid positioning. The specific node coordinates for rapid localization are obtained using the following formula: (26) In formula (26), The specific node coordinates (latitude and longitude) for rapid location; Latitude (in degrees) of key hub nodes; The longitude (unit: °) of the key hub node. The control logic of formula (26) is the spatial mapping logic between the digital twin network and the physical world. Its core value is: 1. Topology-spatial alignment: transforming the abstract key hub node into navigable latitude and longitude coordinates, and opening up the connection between the digital model and the physical site; 2. Rapid positioning support: the coordinates are directly connected to the GIS and navigation system to realize the accurate navigation of emergency repair personnel / equipment; 3. Operation and maintenance closed loop implementation: from fault identification, hub positioning to spatial coordinates, the closed loop of "digital perception-physical disposal" is completed.

[0121] The physical location of the key hub is matched with spatial coordinates in a geographic information system (GIS), such as latitude and longitude coordinates (120.5, 30.2). This spatial coordinates are then matched using a GIS database to determine the specific node coordinates for rapid location, thereby enabling on-site response. For example, this process can improve the emergency response efficiency of water supply networks.

[0122] Preferably, the intelligent management method for rural water supply based on digital twins provided in this embodiment includes step S600: Step S610: Obtain the specific node coordinates for rapid positioning, and use the specific node coordinates to retrieve the full tracking record from the distributed database. The full tracking record integrates historical maintenance data and current traffic indicators.

[0123] The complete tracking record of key hub nodes is derived using the following formula: (27) In formula (27), For full-process tracking and recording; For distributed database retrieval operators; It is a distributed database (storing historical maintenance data and historical flow / pressure indicators). The control logic of formula (27) is the association retrieval logic between the spatial coordinates of pipeline nodes and operation and maintenance data. Its core value is: 1. Accurate data association: Directly locate the full file of nodes through latitude and longitude coordinates to avoid ambiguity in node number / name; 2. Operation and maintenance decision support: Integrate historical maintenance and operation indicators to provide a basis for root cause analysis of faults and preventive maintenance; 3. Distributed architecture adaptation: Adapt to the characteristics of rural pipeline networks that are widely dispersed and have a large amount of data to ensure query efficiency.

[0124] For water supply network monitoring systems, the first step is to obtain the coordinates of specific nodes for rapid location. These coordinates are typically derived from a geographic information system (GIS), such as the latitude and longitude of a key connection point (120.3, 30.1) used to identify pipeline junctions. In one embodiment, the specific node coordinates are used to retrieve a distributed database to obtain a complete tracking record. This record integrates historical maintenance data with current flow indicators. For example, a distributed database, such as a Hadoop-based system, stores maintenance logs associated with node coordinates and real-time sensor data. The retrieval process involves key-value queries, using node coordinates as the key to extract records. For instance, a query might return that the node underwent 5 maintenance visits in the past year and that the current flow rate is 80 cubic meters per hour, thus forming a fused record. The historical maintenance data in the fused record includes fault types such as leaks and maintenance dates, while the current flow indicator is collected in real-time by sensors. The fusion method integrates historical data and current indicators through time series analysis to form a complete timeline view. This timeline view helps analyze the evolution of node status.

[0125] Step S620: Calculate the resource allocation weight based on the full tracking record. The resource allocation weight is generated by weighting historical maintenance data with the current flow index.

[0126] The resource allocation weights for key hub nodes are derived using the following formula: (28) In formula (28), Assign weights to resources (dimensionless, range [0, 1]); This is the weighting coefficient for historical data (dimensionless, default value is 0.6). This is the current indicator weighting coefficient (dimensionless, default value is 0.4). ); Weighting of historical maintenance data ( , This represents the number of historical failures. (Total number of maintenance operations) The current indicator weight ( , For the current flow deviation, (Rated flow). The control logic of formula (28) is the weighted evaluation logic of pipeline node resource demand. Its core value is: 1. Historical-current dual-dimensional integration: It considers both the long-term health status of nodes (historical faults) and the short-term operational anomalies (current flow), making the evaluation more comprehensive; 2. Adjustable weight: Through , Strategies can be flexibly adjusted, such as improving efficiency in emergency scenarios. Focus on current anomalies; 3. Dimensionless normalization: The result is normalized to [0, 1], which facilitates priority comparison and resource scheduling among multiple nodes.

[0127] Resource allocation weights are calculated based on the entire tracking record. These weights are generated by weighting the historical maintenance data with the current flow rate indicator. For example, a weighted average formula is used to calculate the resource allocation weights. The weight factor for historical maintenance data is 0.6, and the weight is calculated based on maintenance frequency; higher frequencies result in higher weights. The weight factor for the current flow rate indicator is 0.4. The weight is increased if the flow rate exceeds a threshold, considering the degree of flow anomalies. Specifically, historical data is first quantified; for example, the number of maintenance visits is converted into scores (e.g., 5 visits correspond to a score of 50). Then, this is combined with a flow rate indicator (e.g., 80 cubic meters, converted to a score of 40), and a weighted sum is performed to obtain the total weight (e.g., 50). 0.6+40 0.4 = 46, thus quantifying resource priority.

[0128] Step S630: Perform multi-objective programming calculations on the available resource set according to the resource allocation weights to obtain a resource-limited allocation scheme.

[0129] The resource-limited allocation scheme is derived using the following formula: (29) In formula (29), For resource-limited allocation schemes; This refers to the collection of available maintenance resources (personnel, equipment, and materials). For resources The utility value (dimensionless, representing resource maintenance efficiency). For resources The scheduling cost (dimensionless, representing the cost of resource transportation / usage); This is the optimal solution operator for multi-objective programming. The control logic of formula (29) is the multi-objective optimization scheduling logic for emergency repair resources. Its core values ​​are: 1. Multi-objective collaborative optimization: It takes into account node demand, resource efficiency and scheduling cost at the same time, avoiding resource waste caused by a single objective; 2. Demand-oriented allocation: Resource allocation weights. 3. Optimal under cost constraints: Maximize repair efficiency and cost-effectiveness under the premise of limited total resources.

[0130] Based on the resource allocation weights, perform multi-objective programming calculations on the available resource set to obtain a limited resource allocation scheme. For example, the multi-objective programming uses a linear programming method, with objectives including minimizing response time and maximizing coverage. The available resource set includes maintenance teams and equipment inventory. Constraints are set in the calculation process, such as the total resources not exceeding 10 teams. The optimization variable is the allocation amount for each node. The solution is iteratively solved using a solver such as CPLEX, and a scheme is generated, such as allocating 3 teams to high-weight nodes.

[0131] Step S640: Analyze the limited resource allocation scheme and perform spatial scheduling to optimize the distribution of maintenance resources.

[0132] The target location for resource scheduling is determined using the following formula: (30) In formula (30), For resources The location of the scheduling target; For space scheduling operators; For resources The transportation distance threshold (unit: km). The control logic of formula (30) is the maintenance resource space scheduling optimization logic, the core value of which is: 1. Distance constraint adaptation: through the transportation distance threshold 1. Adapts to the complex terrain and dispersed resources of rural pipeline networks, avoiding ineffective scheduling; 2. Core node orientation: Always focuses on key hub nodes. 3. The project is feasible: the scheduling results are navigable latitude and longitude coordinates, which directly support emergency repair dispatch and on-site operations.

[0133] The resource-limited allocation scheme is analyzed to perform spatial scheduling and optimize the distribution of maintenance resources. For example, spatial scheduling is achieved through path planning algorithms such as A... The algorithm maps the limited resource allocation scheme to geospatial space and adjusts team routes to cover multiple nodes, thereby achieving efficient distribution.

[0134] Furthermore, the intelligent rural water supply management method based on digital twins provided in this embodiment includes step S700 as follows: Step S710: Analyze the limited resource allocation scheme and compare the flow index deviation with the pressure index deviation. If the flow index deviation is greater than the pressure index deviation, send a frequency modulation command.

[0135] The deviation of flow indicators in the resource allocation scheme is derived using the following formula: (31) In formula (31), Deviation of flow indicators in resource allocation scheme (unit: ); Actual traffic flow at key hub nodes; The rated flow rate for key hub nodes. The control logic of formula (31) is the quantitative assessment logic for flow anomalies at key hub nodes. Its core value is: 1. Accurate anomaly perception: Directly compare the actual flow rate with the rated flow rate. The absolute value form ensures that the deviation is unambiguous and intuitively reflects the degree of anomaly; 2. Control decision support: Provide a quantitative judgment standard for "prioritizing control of flow / pressure deviation" and avoid blind control; 3. Digital twin adaptation: It can directly access real-time monitoring data or digital twin hydraulic simulation results to achieve dynamic perception and closed-loop control.

[0136] The deviation of the stress index in the resource allocation plan is calculated using the following formula: (32) In formula (32), Pressure index deviation (unit: MPa); The actual pressure on key hub nodes; Rated pressure for key hub nodes. The control logic of formula (32) is the quantitative assessment logic for pressure anomalies in key hub nodes. Its core value is: 1. Intuitive anomaly perception: The absolute value form eliminates the directional difference between overpressure and underpressure, directly reflecting the severity of pressure deviation from the design; 2. Precise control decision: It provides a quantitative judgment standard for "prioritizing control of flow / pressure deviation", avoiding blind control; 3. Native digital twin: It can directly access real-time monitoring data or digital twin hydraulic simulation results to achieve dynamic perception and closed-loop control.

[0137] For a water supply network monitoring system, the first step is to analyze a limited resource allocation scheme. This scheme is generated based on previously calculated weights. For example, in the limited resource allocation scheme, maintenance resources allocated to a specific pipeline node might be two teams and a backup pump. The system then compares the deviations in flow rate and pressure. The flow rate deviation is calculated as the difference between the current measured value and the historical average. For instance, if the current flow rate is 90 cubic meters per hour and the historical average is 70 cubic meters per hour, the deviation is 20 cubic meters. The pressure deviation is similarly calculated as the difference between the current pressure and the standard pressure, such as a deviation of 0.5 MPa. If the flow rate deviation is greater than the pressure deviation, a frequency modulation command is sent. For example, if the flow rate deviation is 20 cubic meters per hour and the pressure deviation is 0.5 MPa, the flow rate deviation is considered large, and the sensor is instructed to increase its sampling frequency from once per minute to once per second. This modulation command is transmitted to the field equipment via a network protocol such as MQTT.

[0138] Step S720: Response frequency modulation command generates high-frequency sampling data stream, and the high-frequency sampling data stream is used to obtain an optimized dynamic control feedback loop.

[0139] The optimized dynamic control feedback loop is derived from the following formula: (33) In formula (33), For optimized dynamic control feedback loop (including sampling frequency, indicator set, response sequence); A set of dynamic control indicators under high-frequency sampling; The new sampling frequency. The control logic of formula (33) is the closed-loop construction logic of high-frequency sampling dynamic regulation of the water supply system. Its core value is: 1. Sampling frequency driven optimization: by improving the sampling frequency... Upgrading from "minute-level perception" to "second-level / sub-second-level perception" significantly improves anomaly detection sensitivity; 2. Refined indicator set: high-frequency sampling It includes more transient features, which can support more accurate deviation calculation and fault diagnosis; 3. Control response closed loop: integrate sampling, sensing and control into a feedback loop to achieve continuous optimization of "sensing-decision-execution-re-sensing".

[0140] In response to the frequency modulation command, a high-frequency sampling data stream is generated. This high-frequency sampling data stream involves real-time acquisition of flow velocity and volume data within the pipeline, forming a continuous data sequence. An optimized dynamic control feedback loop is obtained using this high-frequency sampling data stream. For example, the data stream is input into a control algorithm through a feedback loop mechanism. This control algorithm employs the proportional-integral-derivative (PID) control principle. First, it analyzes abnormal patterns in the data stream, such as sudden flow peaks, and then adjusts the valve opening to stabilize the system. The optimization process includes iteratively updating control parameters, such as reducing the integral time constant based on data stream deviations, thereby forming a closed-loop response. This reduces the delay in the rural water supply intelligent management system from detection to adjustment to the second level. This dynamic control through the feedback loop lies in the real-time fusion of new data and model predictions. For example, in a pipeline leakage scenario, the high-frequency data stream shows a rapid increase in flow; the loop mechanism immediately calculates the deviation and generates a correction signal, thereby optimizing the overall network stability.

[0141] Step S730: Distribute alarm signals to terminal devices in a loop according to the optimized dynamic control feedback, and receive the feedback timestamp returned by the terminal devices to calculate the response speed; The average response speed is calculated using the following formula: (34) In formula (34), Average response time (unit: s); This represents the total number of terminal devices. For the first The timestamp (in seconds) of each terminal device receiving the alarm. For the first The sending timestamp of each alarm (unit: s). The control logic of formula (34) is the quantitative evaluation logic of emergency alarm response efficiency. Its core value is: 1. Average delay measurement: By eliminating abnormal fluctuations of a single terminal through arithmetic averaging, it more robustly reflects the overall response capability of the system; 2. Closed-loop optimization support: It provides a quantifiable evaluation standard for the performance iteration of dynamic control feedback loop (such as comparing high-frequency sampling). Low frequency sampling ); 3. Operation and maintenance experience orientation: directly related to the latency of operation and maintenance personnel receiving alarms, it is a key indicator for improving the speed of emergency response.

[0142] Based on the optimized dynamic control feedback loop, alarm signals are distributed to terminal devices. For example, the alarm signal contains node location and deviation details and is sent to the maintenance personnel's mobile device via a wireless network. The feedback timestamp returned by the terminal device is received to calculate the response speed. For example, the timestamp shows that the interval from the alarm being issued to the confirmation is 30 seconds.

[0143] Step S740: Verify whether the response speed meets the preset threshold requirements and determine the final problem discovery closed-loop mechanism.

[0144] The state of the problem discovery closed-loop mechanism is obtained through the following formula: (35) In formula (35), Status of the closed-loop mechanism for problem discovery ( =Closed-loop effective =Needs optimization); The response speed threshold is 30 (unit: s, default value is 30). The control logic of formula (35) is the validity verification logic of the problem discovery closed loop. Its core value is: 1. Threshold judgment: transforming the fuzzy "whether the response meets the standard" into a clear binary classification state, which facilitates automatic decision-making by the system; 2. Closed loop driven optimization: through state judgment, a continuous improvement closed loop of "run-evaluation-optimization-re-run" is formed to ensure the long-term high efficiency of the system; 3. Operation and maintenance operability: the state result is directly connected to the operation and maintenance platform, and the Optimize state can automatically trigger alarms to prompt engineers to intervene in optimization.

[0145] Verify whether the response speed meets the preset response speed threshold requirement, for example, the preset response speed threshold is 30s. If it meets the requirement, determine the final problem discovery closed-loop mechanism. This problem discovery closed-loop mechanism ensures the complete process from deviation detection to problem resolution, thereby improving the emergency efficiency of the water supply network.

[0146] Please see Figure 2This embodiment provides a digital twin-based intelligent management system for rural water supply, which implements a digital twin-based intelligent management method for rural water supply. The method includes an initial operational dataset acquisition module 10, an information gap area determination module 20, a dynamic control indicator set acquisition module 30, a response speed optimization sequence acquisition module 40, a specific node coordinate determination module 50, a resource-limited allocation scheme acquisition module 60, and a problem discovery closed-loop mechanism determination module 70. The initial operational dataset acquisition module 10 collects real-time data through data acquisition devices deployed at key nodes of the water supply network, covering the water supply network from source to end. The system obtains an initial operational dataset of the water supply network, including real-time data such as flow and pressure indicators. An information gap region determination module 20 processes the continuous data stream using data analysis algorithms based on the initial operational dataset, identifies abnormal fluctuation patterns, and determines potential information gap regions in the water supply network, where an information gap region refers to a transmission segment with missing data. A dynamic control indicator set acquisition module 30, if the information gap region determined by the data analysis algorithm exceeds a preset threshold, activates the edge computing module to locally aggregate the data in the information gap region, obtaining a refined dynamic control indicator set, which includes aggregated data from the previous stage. The system includes: a flow rate value and pressure change value; a response speed optimization sequence acquisition module 40, which inputs the refined dynamic control index set into a machine learning classification algorithm for classification, determines the priority level of problem discovery, and obtains a classified response speed optimization sequence, where the response speed optimization sequence is sorted according to the urgency of the event; a specific node coordinate determination module 50, which extracts high-priority events from the classified response speed optimization sequence, applies graph theory algorithms to construct a connectivity graph of the water supply network for high-priority events, and determines the specific node coordinates for rapid location, where the specific node coordinates correspond to the key connection points of the transmission pipeline; a resource-limited allocation scheme acquisition module 60, which, after obtaining the specific node coordinates for rapid location, updates the entire tracking record through a distributed database, integrates historical data and current indicators, and obtains an updated resource-limited allocation scheme, where the resource-limited allocation scheme optimizes the distribution of maintenance resources; and a problem discovery closed-loop mechanism determination module 70, which, if the flow rate index deviation in the updated resource-limited allocation scheme is greater than the pressure index deviation, adjusts the sampling frequency of the data acquisition device to obtain an optimized dynamic control feedback loop, generates an alarm signal based on the optimized dynamic control feedback loop and distributes it to the terminal equipment, determines whether the response speed meets the threshold requirement, and determines the final problem discovery closed-loop mechanism.

[0147] The intelligent rural water supply management method and system based on digital twins provided in this embodiment have the following beneficial effects compared with the prior art: 1. This embodiment achieves intelligent and refined management of rural water supply networks by relying on digital twin technology through "full-link data acquisition → anomaly and fault identification → edge computing completion → priority classification → precise positioning → resource optimization → feedback closed loop"; 2. The key technical highlights of this embodiment are edge computing completion of information gaps (solving the problem of weak data transmission in rural water supply networks), precise node positioning based on graph theory (improving fault handling efficiency), and dynamic feedback loop optimization (adapting to real-time changes in the water supply network). At the same time, the method's engineering feasibility is ensured by clearly defined thresholds and numerical ranges (such as information gap thresholds of 2 to 8 and response time limits of ≤30 minutes). 3. Focusing on the pain points of rural water supply networks, such as "wide coverage, scattered nodes, and easy interruption of data transmission", the entire process uses digital and intelligent means to achieve real-time control of water supply operation status, rapid handling of faults, and efficient utilization of resources, thereby ensuring the safety and stability of rural water supply.

[0148] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention. Clearly, those skilled in the art can make various alterations and modifications to the invention without departing from its spirit and scope. Thus, if these modifications and modifications of the invention fall within the scope of the claims and their equivalents, the invention is also intended to include these modifications and modifications.

Claims

1. A method for intelligent management of rural water supply based on digital twins, characterized in that, Includes the following steps: S100. Real-time data is collected by data acquisition devices deployed at key nodes of the water supply network, covering the water supply network from source to terminal, to obtain the initial operating dataset of the water supply network, wherein the real-time data includes flow rate indicators and pressure indicators. S200. Based on the initial running dataset, apply data analysis algorithms to process the continuous data stream, identify abnormal fluctuation patterns, and determine potential information gap areas in the water supply network, wherein the information gap area refers to a transmission segment with missing data. S300. If the information fault area determined by the data analysis algorithm exceeds a preset threshold, the edge computing module is activated to perform local aggregation on the data of the information fault area to obtain a refined dynamic control index set, wherein the dynamic control index set includes aggregated flow velocity values ​​and pressure change values. S400. Input the refined dynamic control index set into the machine learning classification algorithm for classification, determine the priority level of problem discovery, and obtain the classified response speed optimization sequence, wherein the response speed optimization sequence is sorted according to the urgency of the event. S500. Extract high-priority events from the classified response speed optimization sequence, apply graph theory algorithm to the high-priority events to construct a connectivity graph of the water supply network, and determine the specific node coordinates for rapid location, wherein the specific node coordinates correspond to the key connection points of the transmission pipeline. S600: After obtaining the specific node coordinates for rapid positioning, the entire tracking record is updated through a distributed database. Historical data and current indicators are integrated to obtain an updated resource-limited allocation scheme, wherein the resource-limited allocation scheme optimizes the distribution of maintenance resources. S700 If the deviation of the flow index is greater than the deviation of the pressure index in the updated resource-limited allocation scheme, the sampling frequency of the data acquisition device is adjusted to obtain an optimized dynamic control feedback loop. An alarm signal is generated based on the optimized dynamic control feedback loop and distributed to the terminal device. It is then determined whether the response speed meets the threshold requirement, and the final problem discovery closed-loop mechanism is determined.

2. The method for intelligent management of rural water supply based on digital twins according to claim 1, characterized in that, Step S100 includes: S110. Obtain flow and pressure indicators through data acquisition devices deployed at key nodes of the water supply network; S120. Spatial mapping of the flow rate index and the pressure index between node coordinates and topology structure to obtain a mapping result containing geographical location information; S130. Based on the mapping result, the pipe resistance is determined in conjunction with the pipe diameter and pipe material properties. Based on the abnormal fluctuations of the pipe resistance in relation to the water supply period, the terminal flow verification logic is triggered to obtain the terminal flow. S140. The terminal flow is summarized and the data is cleaned and denoised to obtain the initial operating dataset of the water supply network.

3. The method for intelligent management of rural water supply based on digital twins according to claim 2, characterized in that, Step S200 includes: S210. Obtain the initial running dataset and construct a continuous data flow matrix, wherein the continuous data flow matrix contains flow and pressure values ​​aligned by timestamp sequence; S220. Traverse the continuous data stream matrix to identify the null value gaps generated by the numerical interruption points, and retrieve the synchronous operation data of the topologically adjacent nodes in the pipeline network topology based on the time positioning information of the null value gaps. S230. Calculate the interpolation error of the synchronous operation data. If the interpolation error exceeds a preset value, mark the null gap as an abnormal fluctuation mode caused by transmission link failure. S240. Based on the spatial distribution trajectory of the abnormal fluctuation pattern in the pipeline topology, delineate the range of pipe sections where data loss occurred, and determine the information gap region in the water supply network that is the data loss transmission segment.

4. The method for intelligent management of rural water supply based on digital twins according to claim 1, characterized in that, Step S300 includes: S310. Statistically analyze the cumulative link length of the identified information fault regions in the pipeline topology, and use the cumulative link length as the basis for determining the fault scale. S320. If the fault scale determination criteria exceed a preset threshold, an edge collaboration trigger signal is sent to the edge computing module at the physical boundary of the information fault region. S330. After receiving the edge collaboration trigger signal, the edge computing module collects high-frequency discrete sampling data and performs local aggregation operation to obtain the aggregated flow velocity value and pressure change value. S340. Based on the aggregated flow rate and pressure change values, a structured encapsulation is performed to generate a refined set of dynamic control indicators.

5. The method for intelligent management of rural water supply based on digital twins according to claim 4, characterized in that, Step S400 includes: S410. Extract the aggregated flow velocity values ​​and pressure change values ​​from the refined dynamic control index set, and convert the flow velocity values ​​and pressure change values ​​into a vectorized index feature set through a preset feature extraction model. S420. Input the vectorized indicator feature set into the machine learning classification algorithm to obtain the priority-labeled feature subset; S430. An initial response speed sequence is generated using a sequence generation model based on the priority-labeled feature subset, and an optimized response speed sequence is obtained by optimizing and adjusting the initial response speed sequence using an association rule mining algorithm. S440. Obtain the classified response speed optimization sequence based on the optimized response speed sequence, wherein the response speed optimization sequence is sorted according to the urgency of the event.

6. The method for intelligent management of rural water supply based on digital twins according to claim 1, characterized in that, Step S500 includes: S510. Extract urgent events whose urgency exceeds a preset threshold from the classified response speed optimization sequence; S520. Retrieve the topology data of the water supply network for the emergency event, wherein the topology data is used to construct a connectivity matrix containing pipe attributes and flow loads. S530. Calculate the node weights and edge weight distributions through the connectivity matrix, and use the Dijkstra algorithm to determine the key hubs in the water supply network connectivity graph generated by the connectivity matrix. S540. Match the physical location of the key hub with the spatial coordinates in the geographic information system, wherein the spatial coordinates are used to determine the specific node coordinates for rapid positioning.

7. The method for intelligent management of rural water supply based on digital twins according to claim 6, characterized in that, Step S600 includes: S610. Obtain the specific node coordinates for rapid positioning, and use the specific node coordinates to retrieve the full tracking record from the distributed database. The full tracking record integrates historical maintenance data and current traffic indicators. S620. Calculate the resource allocation weight based on the full-process tracking record. The resource allocation weight is generated by weighting the historical maintenance data with the current flow index. S630. Perform multi-objective programming calculations on the available resource set according to the resource allocation weights to obtain a limited resource allocation scheme; S640. Analyze the resource-limited allocation scheme and perform spatial scheduling to optimize the distribution of maintenance resources.

8. The method for intelligent management of rural water supply based on digital twins according to claim 1, characterized in that, Step S700 includes: S710: Analyze the limited resource allocation scheme and compare the flow index deviation with the pressure index deviation. If the flow index deviation is greater than the pressure index deviation, send a frequency modulation command. S720. Responding to the frequency modulation command, a high-frequency sampling data stream is generated, and an optimized dynamic control feedback loop is obtained using the high-frequency sampling data stream; S730: Distribute alarm signals to terminal devices in a loop according to optimized dynamic control feedback, and receive feedback timestamps returned by terminal devices to calculate response speed; S740. Verify whether the response speed meets the preset threshold requirement, and determine the final problem discovery closed-loop mechanism.

9. The method for intelligent management of rural water supply based on digital twins according to claim 8, characterized in that, In step S710, the deviation of the flow index in the resource allocation scheme is obtained by the following formula: ； in, For the deviation of the flow rate indicator, Actual traffic flow at key hub nodes. Rated flow rate for key hub nodes; The deviation of the stress index in the resource allocation plan is calculated using the following formula: ； in, This is due to deviation in the pressure index. The actual pressure on key hub nodes Rated pressure for key hub nodes.

10. A smart rural water supply management system based on digital twins, used to execute the smart rural water supply management method based on digital twins as described in any one of claims 1 to 9, characterized in that, include: The initial operating dataset acquisition module is used to collect real-time data through data acquisition devices deployed on key nodes of the water supply network, covering the water supply network from source to end, and obtain the initial operating dataset of the water supply network, wherein the real-time data includes flow indicators and pressure indicators. The information gap region determination module is used to process the continuous data stream using data analysis algorithms based on the initial running dataset, identify abnormal fluctuation patterns, and determine potential information gap regions in the water supply network, wherein the information gap region refers to a transmission segment with missing data. The dynamic control indicator set acquisition module is used to activate the edge computing module to locally aggregate the data in the information gap area if the information gap area determined by the data analysis algorithm exceeds a preset threshold, so as to obtain a refined dynamic control indicator set, wherein the dynamic control indicator set includes aggregated flow velocity values ​​and pressure change values. The response speed optimization sequence acquisition module is used to input the refined dynamic control index set into the machine learning classification algorithm for classification, determine the priority level of problem discovery, and obtain the classified response speed optimization sequence, wherein the response speed optimization sequence is sorted according to the urgency of the event. The specific node coordinate determination module is used to extract high-priority events from the classified response speed optimization sequence, apply graph theory algorithms to the high-priority events to construct a connectivity graph of the water supply network, and determine the specific node coordinates for rapid positioning, wherein the specific node coordinates correspond to the key connection points of the transmission pipeline. The resource-limited allocation scheme acquisition module is used to obtain the specific node coordinates for rapid positioning, update the entire tracking record through a distributed database, integrate historical data and current indicators, and obtain an updated resource-limited allocation scheme, wherein the resource-limited allocation scheme optimizes the distribution of maintenance resources. The problem discovery closed-loop mechanism determination module is used to adjust the sampling frequency of the data acquisition device if the deviation of the flow index is greater than the deviation of the pressure index in the updated limited resource allocation scheme, to obtain an optimized dynamic control feedback loop, generate an alarm signal based on the optimized dynamic control feedback loop and distribute it to the terminal device, determine whether the response speed meets the threshold requirement, and determine the final problem discovery closed-loop mechanism.

Images