Method and system for pre-judging leakage risk of secondary water supply pipe network by integrating digital twinning

Abstract

The invention provides an integrated digital twinborn secondary water supply pipe network leakage risk pre-judgment method and system, and relates to the technical field of water supply pipe networks, and the method comprises the steps: constructing a secondary water supply pipe network total factor digital twinborn model according to the modeling factors of a secondary water supply pipe network; a water supply medium in a secondary water supply pipe network is used as a research object, and a water supply medium digital twinborn layer is constructed; the water supply medium digital twinning layer is fused to a secondary water supply pipe network total factor digital twinning model; according to the fused digital twinborn model, extracting and analyzing a pipe network pressure mean value to obtain pipe network pressure threshold intervals and leakage performance characteristics of different water supply partitions; constructing a deep learning model according to the leakage performance characteristics and the pipe network pressure threshold intervals of the different water supply partitions; and according to a risk pre-judgment result output by the deep learning model, performing visual labeling in the fused digital twinborn model, and displaying the working state of the secondary water supply pipe network in real time.

Description

Technical Field

[0001] This invention relates to the field of water supply network technology, specifically to a method and system for predicting leakage risks in secondary water supply networks that integrates digital twins. Background Technology

[0002] As a crucial component of urban water supply systems, secondary water supply networks directly impact residents' water safety and the quality of water services. Furthermore, the complex environments in which these networks are laid mean that the pipes and auxiliary equipment are prone to aging and damage over time, leading to leaks. Leaks not only waste water resources but can also cause pressure imbalances, water pollution, and even land subsidence. Traditional methods of data analysis and risk prediction operate relatively independently, failing to achieve integrated integration of real-time risk feature extraction, intelligent prediction, and visualization. This results in insufficient timeliness, accuracy, and intuitiveness in leak risk prediction, making it difficult to meet the needs of refined operation and maintenance management of secondary water supply networks.

[0003] Therefore, we now provide a method and system for predicting leakage risks in secondary water supply networks using integrated digital twins. Summary of the Invention

[0004] To address the aforementioned technical problems, the present invention aims to provide a method and system for predicting leakage risks in secondary water supply networks using integrated digital twins.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for predicting leakage risks in secondary water supply networks integrating digital twins, the method comprising: Based on the modeling elements of the secondary water supply network, a full-element digital twin model of the secondary water supply network is constructed. Taking the water supply medium in the secondary water supply network as the research object, a digital twin layer of the water supply medium is constructed; and the digital twin layer of the water supply medium is integrated into the full-element digital twin model of the secondary water supply network. Based on the fused digital twin model, the average pressure of the pipeline network is extracted and analyzed to obtain the pressure threshold range and leakage characteristics of different water supply areas. Based on the leakage characteristics and the pressure threshold range of the pipeline network in different water supply zones, a deep learning model is constructed. Based on the risk prediction results output by the deep learning model, the model is visualized and labeled in the fused digital twin model to display the working status of the secondary water supply network in real time.

[0006] Furthermore, the process of constructing a full-element digital twin model of the secondary water supply network includes: The modeling elements of a secondary water supply network include network body elements, auxiliary equipment elements, spatial geographic elements, and basic attribute elements; data of the corresponding modeling elements are acquired to form spatial data, attribute data, and numerical data of the secondary water supply network; the spatial data, attribute data, and numerical data of the secondary water supply network are preprocessed to obtain a standardized network basic database; Based on professional digital twin modeling software, a three-dimensional model is constructed using multi-dimensional data from the pipeline network database to create a corresponding digital twin model of all elements of the secondary water supply network.

[0007] Furthermore, the process of constructing a digital twin layer for the water supply medium includes: Real-time acquisition of modeling data for the water supply medium, including the medium's physical properties, flow properties, and spatiotemporal properties; Based on computational fluid dynamics technology and real-time modeling data of the water supply medium, a digital twin layer of the water supply medium is constructed; spatial matching rules and data linkage interfaces are established between the digital twin layer of the water supply medium and the full-element digital twin model, and the digital twin layer of the water supply medium is integrated into the full-element digital twin model of the secondary water supply network.

[0008] Furthermore, based on the fused digital twin model, the process of extracting and analyzing the average pressure of the pipeline network includes: The fused digital twin model is divided into several water supply zones, and the pressure monitoring data of several water supply zones within a preset time window is extracted and recorded as the pressure monitoring data sequence of the corresponding water supply zone within the time window. The pressure monitoring data sequence is smoothed to obtain a smoothed pressure monitoring data sequence; based on the smoothed pressure monitoring data sequence, the average pressure of the corresponding water supply zone within the time window is calculated as the pressure benchmark value of the zone. Based on the smoothed pressure monitoring data sequence and average pressure, calculate the corresponding pressure standard deviation and coefficient of variation; based on the historical average pressure and historical pressure standard deviation of the historical collection period of the corresponding water supply zone, confirm the pipeline pressure threshold range of the corresponding water supply zone. Set the leakage rate, and extract the leakage performance characteristics corresponding to the leakage based on the leakage rate and the corresponding pressure monitoring data sequence. The leakage performance characteristics include absolute pressure value characteristics, pressure change rate characteristics, and cross-zone linkage characteristics.

[0009] Furthermore, the process of extracting absolute pressure features, pressure change rate features, and cross-regional linkage features includes: If the real-time smoothed pressure monitoring data sequence of the water supply area is continuously lower than the corresponding water supply area's pipeline pressure threshold range, and the duration exceeds the preset time threshold, then the corresponding absolute pressure value feature is obtained. Based on the real-time smoothed pressure monitoring data sequence of the water supply zone and the corresponding pressure change time difference, the corresponding pressure change rate characteristics are obtained. When the leakage is large, the pressure anomaly in the water supply zone where the leakage point is located will spread to the adjacent zones, causing the average pressure of the adjacent zones to drop. Based on the pressure coordination change coefficient, the corresponding cross-zone linkage characteristics can be obtained.

[0010] Furthermore, based on the leakage characteristics and the pipeline pressure threshold ranges of different water supply zones, the process of constructing a deep learning model includes: The absolute pressure characteristics, pressure change rate characteristics, cross-zone linkage characteristics, and corresponding leakage volume of different water supply zones are quantified into leakage feature vectors. By simulating leakage scenarios in a full-element digital twin model of a secondary water supply network using digital twin layers of different water supply media, and combining this with historical leakage records of the secondary water supply network, a dataset sample is constructed; the dataset sample includes normal operation samples and leakage samples. Multi-class labels are used to define the risk level of the dataset samples; the leakage feature vectors of the corresponding dataset samples are divided into 4 categories, namely level 0 risk, level 1 risk, level 2 risk and level 3 risk, and thus a label set is formed. A deep learning model is constructed using a CNN-LSTM hybrid model, denoted as the leakage risk prediction model. The leakage risk prediction model structure consists of an input layer, a feature extraction layer, a temporal modeling layer, a fully connected layer, and an output layer. The dataset samples were divided into training set, validation set and test set in a ratio of 7:2:1 for training, thereby obtaining the trained leakage risk prediction model. The leakage feature vectors obtained in real time from the integrated digital twin model of the secondary water supply network are input into the trained leakage risk prediction model, and the model outputs the risk level and the corresponding probability.

[0011] Furthermore, based on the visualization interface of the digital twin model, different colors and styles of annotation elements are used according to the risk level to achieve an intuitive display of the risk status.

[0012] A second aspect of the present invention also provides a secondary water supply network leakage risk prediction system integrating digital twins, including: a digital model construction module, a model fusion module, a feature analysis module and an intelligent prediction module; The digital model building module is used to construct a full-element digital twin model of the secondary water supply network based on the modeling elements of the secondary water supply network; and to construct a digital twin layer of the water supply medium as the research object. The model fusion module is used to fuse the digital twin layer of the water supply medium into the full-element digital twin model of the secondary water supply network; The feature analysis module is used to extract and analyze the average pressure of the pipeline network based on the fused digital twin model, and to obtain the pressure threshold range and leakage characteristics of the pipeline network in different water supply areas. The intelligent prediction module is used to construct a deep learning model based on the leakage characteristics and the pipeline pressure threshold range of different water supply zones; based on the risk prediction results output by the deep learning model, the model is visualized and labeled in the fused digital twin model to display the working status of the secondary water supply network in real time.

[0013] Compared with existing technologies, the beneficial effects of this invention are as follows: By integrating digital twin technology to construct a full-element model of the secondary water supply network and fusing the water supply medium layer, precise mapping and dynamic linkage between the physical entity of the network and the digital model are achieved. This breaks through the limitations of traditional methods that separate the analysis of network elements and medium characteristics, providing comprehensive and three-dimensional data support for leakage risk prediction. Through pressure data extraction and multi-dimensional feature analysis of the fused model, the pressure benchmarks and leakage characteristics of different water supply zones can be accurately defined, laying a solid feature foundation for risk prediction. By constructing a targeted deep learning model and combining it with visualization annotation technology, intelligent prediction and real-time intuitive display of leakage risks are achieved, significantly improving the timeliness and accuracy of leakage risk identification and helping maintenance personnel quickly locate risk points and grasp the network's operating status. The overall method forms a complete technical chain from model construction, feature extraction, intelligent prediction to visualization display, effectively improving the refinement and intelligence level of secondary water supply network operation and maintenance management, reducing resource waste and safety hazards caused by leakage, and ensuring the stable operation of the water supply system. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0015] Figure 1 A schematic diagram illustrating the principle of a method for predicting leakage risks in secondary water supply networks using integrated digital twins.

[0016] Figure 2 A schematic diagram of a module for predicting leakage risks in a secondary water supply network that integrates digital twins. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be described in detail below. Obviously, the described embodiments are merely some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other implementation methods obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0018] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0019] Example 1: like Figure 1 As shown, a method for predicting leakage risks in secondary water supply networks integrating digital twins is described, the method comprising the following steps: Step S1: Based on the modeling elements of the secondary water supply network, construct a full-element digital twin model of the secondary water supply network; In practical applications, the modeling elements of secondary water supply networks include, but are not limited to, network body elements, auxiliary equipment elements, spatial geographic elements, and basic attribute elements. It should be noted that network body elements include, but are not limited to, pipe material, pipe diameter, pipe length, pipe route, and connection method; auxiliary equipment elements include, but are not limited to, water pumps, valves, pressure reducing valves, water meters, pressure sensors, and flow sensors; spatial geographic elements include, but are not limited to, network burial depth, building layout, and water supply zone boundaries; and basic attribute elements include, but are not limited to, network commissioning years, maintenance records, design pressure, and design flow rate. In this embodiment, data of corresponding modeling elements are acquired to form spatial data, attribute data, and numerical data of the secondary water supply network. The spatial data refers to the positional relationships between water supply pipes and different devices within the secondary water supply network. The attribute data refers to the corresponding descriptive language of the secondary water supply network, such as pipe burial depth and pipe connection methods. The numerical data refers to the numerical data of the corresponding modeling elements in the secondary water supply network, such as pipe length.

[0020] In this embodiment, the spatial data, attribute data, and numerical data of the secondary water supply network are preprocessed to obtain a standardized network basic database. The preprocessing includes data verification, data format unification, and data unit unification, etc. The specific preprocessing process will not be described in detail. Based on professional digital twin modeling software, three-dimensional modeling is performed according to the multi-dimensional data in the network basic database to construct a corresponding digital twin model of all elements of the secondary water supply network, realizing the visualization of network elements. It should be noted that once the secondary water supply network is laid out, modifying the modeling elements of the secondary water supply network is quite complex. Therefore, in this embodiment, since the real-time data of the water supply medium can be dynamically monitored, a digital twin model of the water supply medium is constructed to analyze whether there is a risk of leakage in the secondary water supply network, achieving low cost and good results.

[0021] Step S2: Taking the water supply medium in the secondary water supply network as the research object, construct a digital twin layer of the water supply medium; and integrate the digital twin layer of the water supply medium into the full-element digital twin model of the secondary water supply network; It should be noted that taking the water supply medium within the secondary water supply network as an independent research object can reflect the dynamic movement state of the water supply medium within the network, which meets the data requirements for real-time monitoring and prediction of leakage risks in the secondary water supply network.

[0022] In this embodiment, modeling data of the water supply medium is acquired in real time. The modeling data includes, but is not limited to, the physical properties of the medium, the flow properties of the medium, and the spatiotemporal properties of the medium. It should be noted that the physical properties of the medium include, but are not limited to, temperature, density, and dynamic viscosity. The flow properties of the medium include, but are not limited to, flow velocity, flow rate, pressure, and Reynolds number. The spatiotemporal properties of the medium include, but are not limited to, the position, flow trajectory, and residence time of the water supply medium within the pipeline.

[0023] In this embodiment, a digital twin layer of the water supply medium is constructed based on computational fluid dynamics technology and real-time modeling data of the water supply medium. Spatial matching rules and data linkage interfaces are established between the digital twin layer of the water supply medium and the full-element digital twin model. The digital twin layer of the water supply medium is integrated into the full-element digital twin model of the secondary water supply network, realizing synchronous rendering and updating of the layer and the model. The integrated full-element digital twin model of the secondary water supply network can simultaneously display the physical structure of the network and the dynamic flow state of the water supply medium inside the network. A unique identifier mapping relationship between the digital twin layer of the water supply medium and the physical network is established, supporting joint query and analysis of the water supply medium and network structural attributes, as well as real-time linkage updates of subsequent data.

[0024] Step S3: Based on the fused digital twin model, extract and analyze the average pressure of the pipeline network to obtain the pressure threshold range and leakage characteristics of different water supply areas. In this embodiment, the steps for extracting and analyzing the average pipeline pressure based on the fused digital twin model are as follows: Step A1: Divide the merged digital twin model into n water supply zones, denoted as follows: , ,……, ; Step A2: Extract pressure monitoring data from n water supply zones within a preset time window T, and record the water supply zones. The pressure monitoring data sequence within the time window is as follows ,in, , The sampling frequency of the pressure monitoring data is consistent with that of the sensor, which marks the start of the time window. It should be noted that the time window T can be adjusted based on the stability of the pipeline network operation.

[0025] Step A3: Process the pressure monitoring data sequence Smoothing preprocessing is performed to obtain smoothed pressure monitoring data sequences. .

[0026] In this embodiment, the moving average method is used to analyze the pressure monitoring data sequence. The pressure monitoring data sequence is smoothed. The calculation formula is as follows: ;in, The half width of a sliding window is generally taken as Determined based on the sampling frequency. The sampling time interval, It needs to fall within the time window Within, boundary points are supplemented using a one-sided moving average.

[0027] Step A4: Based on the smoothed pressure monitoring data sequence Calculate water supply zones Average pressure within time window T The pressure baseline value for this zone is calculated using the following formula: ;in, The number of valid sampling points within the time window T. .

[0028] Step A5: Based on the smoothed pressure monitoring data sequence and average pressure Calculate the corresponding pressure standard deviation. and coefficient of variation ; In this embodiment, the standard deviation Coefficient of variation It should be noted that the coefficient of variation The smaller the value, the more stable the pressure operation of the zone, and the easier it is to identify pressure anomalies when leakage occurs; conversely, the baseline value needs to be further adjusted based on historical data.

[0029] Step A6: According to the corresponding water supply zone Historical average pressure and historical pressure standard deviation over historical data collection periods were used to identify the corresponding water supply zones. The pipeline pressure threshold range; In this embodiment, data is collected from the corresponding water supply zones. Based on nearly six months of leak-free historical operating data, the historical average pressure and historical pressure standard deviation for the corresponding historical data collection period were calculated. The initial boundary of the normal fluctuation range was determined based on the 3σ principle. Then, according to the design pressure of the secondary water supply network, the threshold range was adjusted, with the upper limit not exceeding 90% of the design pressure and the lower limit not lower than 50% of the design pressure, corresponding to the water supply zones. pressure threshold range The calculation formula is as follows: ;in, This is the lower limit empirical correction coefficient; This is the upper limit empirical correction factor, typically set to a value of [value missing]. , It can be adjusted according to the material and service life of different pipelines.

[0030] Step A7: Simulate leakage scenarios with different locations and leakage rates in different water supply zones within the fused digital twin model, and set the leakage rate by combining parameters such as pipe material and diameter in the pipeline network elements. This includes micro-leakage, small to medium-sized leakage, and severe leakage, based on the amount of leakage. and the corresponding pressure monitoring data sequence Extract leakage characteristics corresponding to leakage, including but not limited to absolute pressure characteristics, pressure change rate characteristics, and cross-zone linkage characteristics.

[0031] In step A7, the specific steps for extracting the absolute pressure value feature, pressure change rate feature, and cross-regional linkage feature are as follows: Step A71: Extracting the absolute pressure feature specifically involves: if the water supply zone is located... Real-time smoothed pressure monitoring data sequence Persistently below And the duration exceeds If take Minutes, excluding instantaneous fluctuations, show that the greater the leakage, the more significant the pressure drop below the threshold. With leakage They are positively correlated, and the fitted relationship is as follows: ,in, , The fitting coefficients are determined from the experimental data. This refers to the absolute value characteristic of pressure.

[0032] Step A72: Extracting the pressure change rate feature specifically involves: if the water supply zone is located... Real-time smoothed pressure monitoring data sequence The rate of decrease increases sharply, defining the characteristic of the pressure change rate. ,in, This refers to the time difference of pressure change. In this implementation, if during normal operation... When leaking And for two or more consecutive sampling cycles, microleakage Small to medium leakage In case of severe leakage .

[0033] Step A73: Extracting cross-zone linkage features specifically involves: when the leakage is large, the water supply zone where the leakage point is located... The abnormal pressure can spread to adjacent partitions. Let j be a neighboring partition of i, resulting in neighboring partitions average pressure A decrease of 5%-10%, and and The pressure change trend shows a synchronous decay, which can be determined by the pressure coordination change coefficient, denoted as the cross-regional linkage characteristic, and the formula is: ;in, This represents the covariance.

[0034] It should be noted that, during normal operation of the secondary water supply network, When leakage exists, And it decreases as the leakage increases.

[0035] Step S4: Construct a deep learning model based on the leakage characteristics and the pressure threshold range of the pipeline network in different water supply zones; based on the risk prediction results output by the deep learning model, perform visualization annotation in the fused digital twin model to display the working status of the secondary water supply network in real time.

[0036] In this embodiment, the absolute pressure characteristics, pressure change rate characteristics, cross-zone linkage characteristics, and corresponding leakage amounts of different water supply zones are quantified into leakage feature vectors.

[0037] By simulating leakage scenarios in the full-element digital twin model of the secondary water supply network using digital twin layers of different water supply media, and combining them with historical leakage records of the secondary water supply network, a dataset sample is constructed. The dataset sample includes normal operation samples and leakage samples, and the dataset sample covers the absolute pressure value characteristics, pressure change rate characteristics, and cross-zone linkage characteristics of different water supply zones, as well as the corresponding leakage amount.

[0038] Multi-class labels are used to define the risk level of the dataset samples. The leakage feature vectors are divided into four categories: Level 0 risk (no risk, normal operation), Level 1 risk (minor leakage risk), Level 2 risk (small to medium leakage risk), and Level 3 risk (serious leakage risk), forming a label set. .

[0039] In this embodiment, a deep learning model, denoted as the leakage risk prediction model, is constructed using a CNN-LSTM hybrid model. This model combines the feature extraction capabilities of Convolutional Neural Networks (CNN) with the temporal data processing advantages of Long Short-Term Memory Networks (LSTM) to accurately capture the spatiotemporal correlation features of pressure sequences and the temporal evolution of leakage risk. The leakage risk prediction model structure consists of an input layer, a feature extraction layer, a temporal modeling layer, a fully connected layer, and an output layer, as detailed below: 1. Input layer: The input dimension is ( ), where m is the number of features in the dataset samples, and k is the time step size, which is 10-15, corresponding to 10-15 minutes of stress time-series data, i.e., the input is the time-series feature matrix for each sample. .

[0040] 2. Feature Extraction Layer: Two convolutional layers are used: Conv1D and a pooling layer (MaxPooling1D). The first convolutional layer has 32 kernels, a kernel size of 3, and uses ReLU activation. The first pooling layer has a pooling window size of 2 and a stride of 1. The second convolutional layer has 64 kernels, a kernel size of 3, and uses ReLU activation. The second pooling layer has a pooling window size of 2 and a stride of 1. Through convolution and pooling operations, local key features are extracted from the feature matrix, and the output feature map dimension is [dimension missing]. .

[0041] 3. Temporal Modeling Layer: Set up one LSTM layer with 128 hidden units, use tanh as the activation function, and set the dropout coefficient to 0.2 to prevent overfitting. Perform temporal correlation modeling on the feature map output by the convolutional layer to capture the temporal variation pattern of stress anomalies. The output temporal feature vector has a dimension of 128.

[0042] 4. Fully connected layers: Two fully connected layers are set up. The first layer has 64 neurons and the activation function is ReLU; the second layer has 32 neurons and the activation function is ReLU.

[0043] 5. Output Layer: Employs the Softmax activation function to output the probability distributions for four risk levels; the model output is... ,in, The probability of a sample belonging to a corresponding risk level is used to determine the final prediction result, with the level corresponding to the highest probability being taken.

[0044] In this embodiment, the training and optimization process of the leakage risk prediction model is as follows: The Adam optimizer is used, with an initial learning rate of 0.001. An adaptive learning rate strategy is employed, meaning the learning rate decays every 5 epochs with a decay coefficient of 0.95. The cross-entropy loss function is used, which is suitable for multi-class classification tasks. The formula is as follows: ;in, This represents the number of training samples; For the first The true label for each sample, such as the label for Level 1 risk, is... , The model predicts the first Each sample belongs to The probability of level 1 risk. .

[0045] The dataset samples were divided into training, validation, and test sets in a 7:2:1 ratio. The number of training epochs was set to 50, and the batch size was set to 32. An early stopping strategy was adopted during training, i.e., patience=5. Training was stopped when the validation set loss did not decrease for 5 consecutive epochs to avoid overfitting.

[0046] The leakage feature vectors obtained in real time from the integrated digital twin model of the secondary water supply network are input into the trained leakage risk prediction model. The model outputs the risk level and corresponding probability; if the prediction result is... The risk level is assessed, and the flow trajectory data of the digital twin layer of the water supply medium is combined to locate the approximate location of the leak. That is, leakage will cause an abnormal increase in the flow velocity of the medium around the leak point. By the intersection of the abnormal flow velocity area and the abnormal pressure area, the corresponding leak point location range can be narrowed down.

[0047] In this embodiment, the visualization interface of the digital twin model is used to display the risk status intuitively by using different colors and styles of annotation elements according to the risk level: Level 0 risk (no risk): The pipeline model and medium layer are displayed in the default colors (the pipeline is gray and the medium is light blue); Level 1 risk (minor leak): A yellow dashed box is marked within 5m of the leak point, and a yellow arrow is marked on the medium flow trajectory. A warning message "Minor leak risk, investigation recommended" pops up in the upper right corner of the model interface; Level 2 risk (small to medium leak): An orange solid box is marked within 8m of the leak point, and an orange flashing arrow is marked on the medium flow trajectory. An audible and visual warning is triggered simultaneously, and the estimated leakage amount is displayed; Level 3 risk (serious leak): A red thick solid box is marked within 12m of the leak point. The corresponding area of ​​the pipeline model flashes, and a red turbulence effect is displayed on the medium layer. An emergency warning is immediately triggered and pushed to the operation and maintenance terminal, marking the leak location, leakage level, and recommended handling plan.

[0048] It should be noted that the visual annotation is linked in real time with the physical pipeline network and digital twin model. If the maintenance personnel confirm the leakage after investigation and complete the repair, the maintenance record is updated in the model. The model automatically recalculates the pressure threshold range of the partition, optimizes the feature vector weights, iteratively upgrades the deep learning model, and improves the accuracy of subsequent predictions. If it is a misjudgment, the model records the features of the misjudged sample, incorporates it into the next round of training dataset, and corrects the model parameters.

[0049] Furthermore, the deep integration of deep learning models and digital twin models not only enables real-time prediction of leakage risks, but also reduces the difficulty of troubleshooting for maintenance personnel through visual annotation. Combined with historical data and simulation analysis in the model, it provides data support for leakage prevention and pipeline network optimization and renovation, further improving the operational stability of the secondary water supply network.

[0050] Example 2: like Figure 2 As shown, the integrated digital twin secondary water supply network leakage risk prediction system includes, but is not limited to, a digital model construction module, a model fusion module, a feature analysis module, and an intelligent prediction module. The digital model building module is used to construct a full-element digital twin model of the secondary water supply network based on the modeling elements of the secondary water supply network; and to construct a digital twin layer of the water supply medium as the research object. The model fusion module is used to fuse the digital twin layer of the water supply medium into the full-element digital twin model of the secondary water supply network; The feature analysis module is used to extract and analyze the average pressure of the pipeline network based on the fused digital twin model, and to obtain the pressure threshold range and leakage characteristics of the pipeline network in different water supply areas. The intelligent prediction module is used to construct a deep learning model based on the leakage characteristics and the pipeline pressure threshold range of different water supply zones; based on the risk prediction results output by the deep learning model, the model is visualized and labeled in the fused digital twin model to display the working status of the secondary water supply network in real time.

[0051] Optionally, in this embodiment, those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing the hardware related to the terminal device. The program can be stored in a computer-readable storage medium, which may include: flash drive, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0052] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0053] If the integrated units in the above embodiments are implemented as software functional units and sold or used as independent products, they can be stored in the aforementioned computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause one or more electronic devices to execute all or part of the steps of the methods described in the various embodiments of this application.

[0054] In the above embodiments of this application, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0055] In the several embodiments provided in this application, it should be understood that the disclosed application can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection of units or modules may be electrical or other forms.

[0056] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0057] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0058] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for predicting leakage risks in secondary water supply networks integrating digital twins, characterized in that, The method includes: Based on the modeling elements of the secondary water supply network, a full-element digital twin model of the secondary water supply network is constructed. Taking the water supply medium in the secondary water supply network as the research object, a digital twin layer of the water supply medium is constructed; and the digital twin layer of the water supply medium is integrated into the full-element digital twin model of the secondary water supply network. Based on the fused digital twin model, the average pressure of the pipeline network is extracted and analyzed to obtain the pressure threshold range and leakage characteristics of different water supply areas. Based on the leakage characteristics and the pressure threshold range of the pipeline network in different water supply zones, a deep learning model is constructed. Based on the risk prediction results output by the deep learning model, the model is visualized and labeled in the fused digital twin model to display the working status of the secondary water supply network in real time.

2. The method for predicting leakage risk in secondary water supply networks using integrated digital twins as described in claim 1, characterized in that, The process of constructing a full-element digital twin model of a secondary water supply network includes: The modeling elements of a secondary water supply network include network body elements, auxiliary equipment elements, spatial geographic elements, and basic attribute elements; data of the corresponding modeling elements are acquired to form spatial data, attribute data, and numerical data of the secondary water supply network; the spatial data, attribute data, and numerical data of the secondary water supply network are preprocessed to obtain a standardized network basic database; Based on professional digital twin modeling software, a three-dimensional model is constructed using multi-dimensional data from the pipeline network database to create a corresponding digital twin model of all elements of the secondary water supply network.

3. The method for predicting leakage risk in secondary water supply networks using integrated digital twins as described in claim 2, characterized in that, The process of constructing a digital twin layer for the water supply medium includes: Real-time acquisition of modeling data for the water supply medium, including the medium's physical properties, flow properties, and spatiotemporal properties; Based on computational fluid dynamics technology and real-time modeling data of the water supply medium, a digital twin layer of the water supply medium is constructed; spatial matching rules and data linkage interfaces are established between the digital twin layer of the water supply medium and the full-element digital twin model, and the digital twin layer of the water supply medium is integrated into the full-element digital twin model of the secondary water supply network.

4. The method for predicting leakage risk in secondary water supply networks using integrated digital twins as described in claim 3, characterized in that, The process of extracting and analyzing the average pressure of the pipeline network based on the fused digital twin model includes: The fused digital twin model is divided into several water supply zones, and the pressure monitoring data of several water supply zones within a preset time window is extracted and recorded as the pressure monitoring data sequence of the corresponding water supply zone within the time window. The pressure monitoring data sequence is smoothed to obtain a smoothed pressure monitoring data sequence; based on the smoothed pressure monitoring data sequence, the average pressure of the corresponding water supply zone within the time window is calculated as the pressure benchmark value of the zone. Based on the smoothed pressure monitoring data sequence and average pressure, calculate the corresponding pressure standard deviation and coefficient of variation; based on the historical average pressure and historical pressure standard deviation of the historical collection period of the corresponding water supply zone, confirm the pipeline pressure threshold range of the corresponding water supply zone. Set the leakage rate, and extract the leakage performance characteristics corresponding to the leakage based on the leakage rate and the corresponding pressure monitoring data sequence. The leakage performance characteristics include absolute pressure value characteristics, pressure change rate characteristics, and cross-zone linkage characteristics.

5. The method for predicting leakage risk in secondary water supply networks using integrated digital twins as described in claim 4, characterized in that, The process of extracting absolute pressure features, pressure change rate features, and cross-regional linkage features includes: If the real-time smoothed pressure monitoring data sequence of the water supply area is continuously lower than the corresponding water supply area's pipeline pressure threshold range, and the duration exceeds the preset time threshold, then the corresponding absolute pressure value feature is obtained. Based on the real-time smoothed pressure monitoring data sequence of the water supply zone and the corresponding pressure change time difference, the corresponding pressure change rate characteristics are obtained. When the leakage is large, the pressure anomaly in the water supply zone where the leakage point is located will spread to the adjacent zones, causing the average pressure of the adjacent zones to drop. Based on the pressure coordination change coefficient, the corresponding cross-zone linkage characteristics can be obtained.

6. The method for predicting leakage risk in secondary water supply networks using integrated digital twins as described in claim 5, characterized in that, The process of constructing a deep learning model based on leakage characteristics and the pipeline pressure threshold ranges of different water supply areas includes: The absolute pressure characteristics, pressure change rate characteristics, cross-zone linkage characteristics, and corresponding leakage volume of different water supply zones are quantified into leakage feature vectors. By simulating leakage scenarios in a full-element digital twin model of a secondary water supply network using digital twin layers of different water supply media, and combining this with historical leakage records of the secondary water supply network, a dataset sample is constructed; the dataset sample includes normal operation samples and leakage samples. Multi-class labels are used to define the risk level of the dataset samples; the leakage feature vectors of the corresponding dataset samples are divided into 4 categories, namely level 0 risk, level 1 risk, level 2 risk and level 3 risk, and thus a label set is formed. A deep learning model is constructed using a CNN-LSTM hybrid model, denoted as the leakage risk prediction model. The leakage risk prediction model structure consists of an input layer, a feature extraction layer, a temporal modeling layer, a fully connected layer, and an output layer. Data samples are sorted The proportions are divided into training set, validation set and test set for training, and then the trained leakage risk prediction model is obtained. The leakage feature vectors obtained in real time from the integrated digital twin model of the secondary water supply network are input into the trained leakage risk prediction model, and the model outputs the risk level and the corresponding probability.

7. The method for predicting leakage risk in secondary water supply networks using integrated digital twins as described in claim 6, characterized in that, Based on the visualization interface of the digital twin model, different colors and styles of annotation elements are used according to the risk level to achieve an intuitive display of the risk status.

8. A method for predicting leakage risks in secondary water supply networks using integrated digital twins, implementing the method for predicting leakage risks in secondary water supply networks using integrated digital twins as described in any one of claims 1 to 7, characterized in that... include: The module includes a digital model building module, a model fusion module, a feature analysis module, and an intelligent prediction module. The digital model building module is used to construct a full-element digital twin model of the secondary water supply network based on the modeling elements of the secondary water supply network; and to construct a digital twin layer of the water supply medium as the research object. The model fusion module is used to fuse the digital twin layer of the water supply medium into the full-element digital twin model of the secondary water supply network; The feature analysis module is used to extract and analyze the average pressure of the pipeline network based on the fused digital twin model, and to obtain the pressure threshold range and leakage characteristics of the pipeline network in different water supply areas. The intelligent prediction module is used to construct a deep learning model based on the leakage characteristics and the pipeline pressure threshold range of different water supply zones; based on the risk prediction results output by the deep learning model, the model is visualized and labeled in the fused digital twin model to display the working status of the secondary water supply network in real time.

Images