Method for judging stock leakage of water supply pipe network based on DMA (direct memory access) night water volume

Abstract

The invention discloses a method for judging stock leakage of a water supply pipe network based on DMA (Direct Memory Access) night water volume, which is characterized in that the night water volume is relatively stable, and the night flow is finely corrected by creatively integrating lower-level subareas, night water consumption of large users and temperature data, so that the estimated leakage volume closer to the real situation is calculated. The time sequence distribution of the leakage amount ratio is further analyzed, credibility evaluation is introduced, interference caused by instantaneous fluctuation can be effectively filtered out, and finally early warning is triggered when the stock leakage credibility is high enough. According to the method, the accuracy of leakage judgment and the reliability of early warning are remarkably improved, a water affair department can be helped to actively and accurately recognize and manage stock leakage, and powerful support is provided for scientific decision making.

Description

Technical Field

[0001] This application relates to the field of intelligent judgment, and more specifically, to a method for judging the leakage of water supply network inventory based on nighttime water volume DMA. Background Technology

[0002] As a fundamental infrastructure ensuring social production and residents' lives, the safe and stable operation of urban water supply networks is crucial. However, during long-term service, leakage is unavoidable due to various factors such as pipe aging, joint corrosion, geological subsidence, and external loads. This leads to serious water waste and economic losses, and may even trigger secondary disasters such as road collapses and water pollution. Network leakage is categorized into two types: sudden pipe bursts and existing leakage. Sudden pipe bursts typically involve large flow rates and are easily located through sudden pressure drops or water seepage on the road surface. Existing leakage, on the other hand, manifests as background leakage formed by the accumulation of numerous small, continuous, and hidden leaks. This type of leakage is difficult to detect, but its cumulative amount over time is staggering, a major reason for the persistently high production-to-sales ratio of water supply companies. Therefore, developing a system capable of accurately and efficiently identifying and providing early warning of existing leakage in water supply networks is of significant practical importance for improving the precision of water management and achieving energy conservation and emission reduction.

[0003] To address pipeline leakage issues, various detection technologies have been developed within the industry. Traditional leakage detection methods primarily rely on manual listening, inspections, or on-site investigations using specialized equipment such as correlators and leak detectors. While effective in specific scenarios, these methods generally suffer from high labor intensity, low detection efficiency, high costs, and difficulty in covering the entire pipeline network, especially for dispersed and concealed chronic leaks, which are extremely difficult to detect. With the development of information technology, Domain Independent Metering (DMA) zoning metering technology has been widely adopted due to its ability to provide refined management of pipeline networks. By installing flow meters at the inlet of the DMA, management departments can monitor the water supply to specific areas. However, existing leakage assessment methods based on DMA technology still have significant limitations. Most methods rely on a simple calculation of the "production-sales difference," comparing the total water entering the DMA with the total water consumed by all users in that area to estimate the leakage. While this macroscopic calculation method provides a general concept of leakage, it struggles to accurately identify long-term, relatively stable "existing leakage" or "background leakage." Its main drawback is that the calculation results are highly susceptible to interference from various factors, such as fluctuations in normal user water usage behavior, intermittent water usage by large users, and errors in the metering equipment itself. These interfering factors often mask the true leakage signal, leading to distorted leakage assessment results and failing to provide a reliable basis for maintenance decisions.

[0004] Therefore, an optimized method for determining the leakage of water supply network inventory based on nighttime water volume using DMA is needed. Summary of the Invention

[0005] To address the aforementioned technical problems, this application is proposed. An embodiment of this application provides a method for determining the existing leakage in a water supply network based on nighttime water volume (DMA). This method leverages the relatively stable nature of nighttime water consumption and innovatively integrates data from lower-level zones, large users' nighttime water consumption, and temperature to refine the nighttime flow rate, thereby calculating an estimated leakage amount closer to reality. By further analyzing the temporal distribution of this leakage percentage and introducing a reliability assessment, interference caused by instantaneous fluctuations can be effectively filtered out. Finally, an early warning is triggered only when the reliability of the existing leakage is sufficiently high. This method significantly improves the accuracy of leakage assessment and the reliability of early warning, enabling water utilities to proactively and accurately identify and manage existing leakage, providing strong support for scientific decision-making.

[0006] According to one aspect of this application, a method for determining the leakage of water supply network inventory based on nighttime water volume (DMA) is provided, comprising: Collect nighttime water volume data of DMA partitions, nighttime water volume data of lower-level partitions, and nighttime water volume and temperature data of large users within the partitions; Based on the nighttime water volume data of the DMA partition, the nighttime water volume data of the lower-level partition, and the nighttime water volume data and temperature data of large users within the partition, the estimated leakage is calculated. The reliability of the stock leakage is calculated in response to the time-series distribution of the ratio between the estimated leakage and the total daily water supply of the zone meeting the preset conditions. In response to the existing leakage confidence exceeding a preset threshold, a high-confidence leakage warning is triggered.

[0007] Compared with existing technologies, this application provides a method for judging the existing leakage in water supply networks based on nighttime water volume analysis (DMA). This method leverages the relatively stable nature of nighttime water consumption and innovatively integrates nighttime water consumption data from lower-level zones, large users, and temperature data to refine the nighttime flow rate, thereby calculating an estimated leakage amount closer to reality. By further analyzing the temporal distribution of this leakage proportion and introducing a reliability assessment, interference caused by instantaneous fluctuations can be effectively filtered out. An early warning is only triggered when the reliability of the existing leakage is sufficiently high. This method significantly improves the accuracy of leakage judgment and the reliability of early warning, enabling water utilities to proactively and accurately identify and manage existing leakage, providing strong support for scientific decision-making. Attached Figure Description

[0008] The above and other objects, features, and advantages of this application will become more apparent from the more detailed description of the embodiments of this application in conjunction with the accompanying drawings. The drawings are provided to further illustrate the embodiments of this application and form part of the specification. They are used together with the embodiments of this application to explain this application and do not constitute a limitation thereof. In the drawings, the same reference numerals generally represent the same components or steps.

[0009] Figure 1 This is a flowchart of a water supply network inventory leakage determination method based on DMA nighttime water volume according to an embodiment of this application; Figure 2 This is a data flow diagram illustrating the water supply network inventory leakage determination method based on DMA nighttime water volume according to an embodiment of this application. Detailed Implementation

[0010] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described herein.

[0011] As indicated in this application and claims, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" are not specifically singular and may include plural forms. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of explicitly identified steps and elements, which do not constitute an exclusive list, and the method or apparatus may also include other steps or elements.

[0012] While this application makes various references to certain modules of the systems according to embodiments of this application, any number of different modules can be used and run on user terminals and / or servers. The modules described are merely illustrative, and different aspects of the systems and methods may use different modules.

[0013] Flowcharts are used in this application to illustrate the operations performed by the system according to embodiments of this application. It should be understood that the preceding or following operations are not necessarily performed in exact order. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from them.

[0014] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described herein.

[0015] In the technical solution of this application, a method for judging the leakage of water supply network inventory based on nighttime water volume of DMA is proposed. Figure 1 This is a flowchart of a water supply network inventory leakage determination method based on DMA nighttime water volume according to an embodiment of this application. Figure 2 This is a data flow diagram illustrating the water supply network inventory leakage determination method based on DMA nighttime water volume according to an embodiment of this application. Figure 1 and Figure 2 As shown, the water supply network inventory leakage judgment method based on DMA nighttime water volume according to an embodiment of this application includes the following steps: S1, collecting nighttime water volume data of DMA zones, nighttime water volume data of lower-level zones, and nighttime water volume data and temperature data of large users within the zones; S2, calculating the estimated leakage amount based on the nighttime water volume data of DMA zones, nighttime water volume data of lower-level zones, and nighttime water volume data and temperature data of large users within the zones; S3, calculating the inventory leakage confidence level in response to the time-series distribution of the ratio between the estimated leakage amount and the total daily water supply of the zone satisfying a preset condition; S4, triggering a high-confidence leakage warning in response to the inventory leakage confidence level exceeding a preset threshold.

[0016] Specifically, S1 collects nighttime water volume data for the DMA zone, nighttime water volume data for lower-level zones, and nighttime water volume and temperature data for large users within the zone. The DMA zone nighttime water volume data refers to the total water volume flowing into the target independent metering zone during a specific nighttime period; the lower-level zone nighttime water volume data refers to the sum of nighttime water consumption of all lower-level sub-zones within the target zone; and the large user nighttime water volume data within the zone refers to the sum of water consumption of all known large users within the zone who have continuous nighttime water consumption behavior. By acquiring this data, a computational model reflecting the actual operating status of the pipeline network can be established, thus laying a solid quantitative foundation for subsequent identification and early warning of existing leakage.

[0017] In practical implementation, firstly, regarding water volume data collection, the total water inflow of the target DMA zone during the nighttime low-water period (usually 2:00-4:00 AM) is obtained from the water supply SCADA system or online flow meters at the zone boundaries; this is the DMA zone's nighttime water volume data. Simultaneously, if the DMA zone contains smaller sub-zones, the total nighttime water supply of all sub-zones needs to be collected synchronously to form the sub-zone's nighttime water volume data. Furthermore, all large users within the zone with stable nighttime water usage (such as factories and hospitals operating 24 hours a day) need to be identified, and their nighttime water usage is collected through their individual water meters, summarizing this as the large user nighttime water volume data for the zone. This water volume data is the direct basis for calculating the basic leakage rate.

[0018] Specifically, S2 calculates and estimates leakage based on nighttime water volume data from the DMA partition, nighttime water volume data from lower-level partitions, and nighttime water volume and temperature data from large users within the partition. In other words, it transforms discrete, multi-source data into a unified indicator with clear physical meaning—the estimated leakage. This not only quantifies the potential leakage level in the pipe network on a specific day but also provides crucial, pre-corrected input data for subsequent time-series analysis and assessment of existing leakage characteristics.

[0019] Specifically, in a specific example of this application, the estimated leakage can be calculated through the following steps: First, based on the nighttime water volume data of the DMA partition, the nighttime water volume data of the lower-level partition, and the nighttime water volume data of large users within the partition, the basic leakage is calculated, expressed by the formula:

[0020] in, For DMA partition nighttime water volume data, For nighttime water volume data of lower-level zones and This is the nighttime water consumption data for large users within the partition. Basic leakage; Furthermore, the baseline leakage rate is corrected based on temperature data to obtain the estimated leakage rate, expressed by the formula:

[0021] in, Based on the basic leakage rate, Temperature is a factor that affects the environment. As a temperature drift time-sensitive factor, This is a temperature correction factor. The temperature data for that day. As the reference temperature, To estimate the amount of leakage.

[0022] In this process, the temperature correction factor is designed to compensate for changes in water density and metering errors caused by temperature variations, thereby improving the accuracy of leakage calculations.

[0023] In a specific example, the temperature drift time sensitivity factor is 1, and the temperature influence factor is 0.02, that is, In other words, in this specific example, the relationship between temperature difference and temperature correction coefficient is linear, and the temperature influence factor is a linear influence factor. This physical model setting can simplify calculations.

[0024] It should be understood that in the technical solution of this application, the basic leakage is a theoretical leakage value calculated based on pure water balance without any correction for environmental factors, which constitutes the initial benchmark for leakage analysis; the estimated leakage is a corrected value obtained after considering the effect of temperature on the basic leakage, which is a quantitative indicator that is closer to the actual leakage situation of the pipeline network.

[0025] During the experiment, the applicant discovered that soil, as the covering medium for pipelines, possesses high thermal inertia, causing surface temperature changes to take several days to be transmitted to underground pipelines (similar to the lag effect of strong cold air). Simultaneously, the nonlinear characteristics of the pipeline material's temperature response result in drastically different impact patterns on leakage caused by short-term drastic cooling (instantaneous shock) and sustained low temperatures (fatigue accumulation). Existing models rely solely on the current day's temperature for static or linear corrections, lacking a "memory" mechanism for recent weather conditions and failing to distinguish between the aforementioned physical effects, thus affecting the accuracy of leakage estimation. Therefore, in a preferred example of this application, historical temperature data is introduced into the determination of the temperature drift time-series sensitivity factor to quantify the lag and cumulative effect of temperature transmission. In this case, the temperature drift time-series sensitivity factor is not a static constant but a value dynamically generated by the data-driven model, quantifying the comprehensive influence weight of a specific historical temperature sequence on the daily leakage calculation. It should be understood that in the embodiments of this application, the temperature drift time sensitivity factor is a key correction coefficient that defines the degree of influence of temperature deviation on the calculation of leakage. The setting of its value is directly related to the environmental adaptability and temperature time cumulative effect of the temperature correction model.

[0026] Specifically, the process begins by acquiring temperature data for a predetermined number of days prior to the current day's temperature data, and then arranging this data into a historical temperature data sequence. In practice, the length of the historical time window to be traced is first determined based on a pre-set parameter. Subsequently, the system initiates a query request to a database storing historical environmental data (e.g., a meteorological database or a historical archive of a water supply SCADA system). The goal of this query is to extract temperature data for each day covered by the predetermined number of days, starting from the day before the current day. Finally, this step arranges the retrieved temperature data points in chronological order, forming an ordered time series, i.e., the historical temperature data sequence.

[0027] The number of days is a key, configurable parameter that defines the depth or time span of historical information analyzed by the model. Its value (e.g., 7 days, 14 days, or 30 days) is usually set based on water sector expertise, such as typical timing of soil heat conduction or the periodic characteristics of climate change in a specific region.

[0028] Next, time-series propagation encoding is performed on the historical temperature data sequence to obtain the implicit encoding vector of the historical temperature time-series effect propagation. It should be understood that existing temperature correction models rely only on daily temperature data, while the high thermal inertia of soil means that surface temperature changes take several days to propagate to underground pipelines (such as the lag of strong cold air influence), and the response patterns of pipeline materials to instantaneous cooling (impact stress) and sustained low temperatures (fatigue accumulation) are fundamentally different. Existing technologies lack the ability to model historical temperature sequences and cannot distinguish between these two physical mechanisms, necessitating the establishment of a dynamic characterization system for temperature effects through time-series propagation encoding. Therefore, in the technical solution of this application, an intelligent characterization system capable of deeply capturing the spatiotemporal transmission laws of temperature effects is established, transforming discrete temperature readings into dynamic physical codes that can simulate the actual heat conduction process and material response characteristics of underground pipe networks. In this way, the physical cognitive accuracy of existing leakage assessment is reshaped.

[0029] Specifically, firstly, the historical temperature data sequence is segmented to obtain the sequence distribution of historical temperature time-series feature vectors. It should be understood that the transmission of surface temperature fluctuations to underground pipe networks through the soil medium is not an instantaneous process; temperature patterns at different time scales have drastically different physical paths affecting pipe sealing. Therefore, in the technical solution of this application, the historical temperature data sequence is segmented, breaking through the coarse-grained limitation of traditional temperature correction models that rely solely on daily data. By decomposing continuous temperature data into physically meaningful feature segments, a feature foundation capable of characterizing the true thermal dynamics of underground pipe networks can be constructed for subsequent modeling.

[0030] Next, based on the feature distribution of each historical temperature time-series feature vector in the sequence distribution of historical temperature time-series feature vectors, the window size of the historical temperature local semantic enhancement perception window for each historical temperature time-series feature vector is determined. It should be understood that the traditional fixed-window model cannot adapt to the nonlinear characteristics of heat conduction processes in complex pipe network environments. Therefore, in the technical solution of this application, by determining the window size of the historical temperature local semantic enhancement perception window for each historical temperature time-series feature vector, the limitations of empirical parameter settings are broken, enabling the system to autonomously construct a dynamic analysis scale that conforms to the actual thermodynamic response of underground pipe networks based on the physical patterns implicit in the feature vectors (such as the drasticness of temperature changes, trend persistence, etc.). This window mechanism driven by the physical characteristics of data essentially constructs an intelligent interface connecting the digital space and the physical laws of the physical pipe network, thereby significantly improving the environmental adaptability of the temperature correction model.

[0031] Specifically, in the technical solution of this application, the specific process of determining the window size of the historical temperature local semantic enhancement perception window for each historical temperature time-series feature vector includes: First, based on the feature distribution of each historical temperature time-series feature vector in the sequence distribution of historical temperature time-series feature vectors, the nearest neighbor feature distribution entropy of the historical temperature time-series feature vectors is calculated, expressed by the formula:

[0032] in, Indicates exponentiation. For temperature coefficient, and Respectively The and the first A historical temperature time series feature vector express and The Euclidean distance between them for and Historical temperature time series similarity between them For the preset neighborhood, For the second smoothing term, express Compared to Historical temperature time series normalized weighting factor For the first The nearest neighbor feature distribution entropy of a historical temperature time series feature vector; Next, based on the nearest neighbor feature distribution entropy of the historical temperature time-series feature vectors, the window size of the historical temperature local semantic enhancement perception window for each historical temperature time-series feature vector is calculated, expressed by the formula:

[0033] in, For the first The nearest neighbor feature distribution entropy of a historical temperature time series feature vector. For the first smoothing term, To preset the maximum window size, This represents the logarithmic operation with base 2. To preset the maximum window size, No. The window size of the local semantic enhancement perception window for historical temperature time-series feature vectors.

[0034] Furthermore, based on all historical temperature time-series feature vectors within the historical temperature local semantic enhancement perception window, local semantic embedding enhancement is performed on each historical temperature time-series feature vector to obtain the sequence distribution of the historical temperature time-series feature local semantic enhancement encoding vector. It should be understood that while a single temperature feature vector may capture the statistical characteristics within a specific time window (such as the three-day average), it cannot resolve the complex coupling mechanism between temperature dynamics and pipeline material response during that period (such as the nonlinear surge in stress at the cast iron pipe interface during a sudden temperature drop). Therefore, in the technical solution of this application, by deeply mining the physical semantics of temperature evolution within the perception window, a local physical field encoding that faithfully reflects the thermodynamic behavior of underground pipe networks is constructed, thereby optimizing the physical authenticity of the global temperature effect transmission.

[0035] Specifically, in the technical solution of this application, the specific process of performing local semantic embedding enhancement on each historical temperature time-series feature vector to obtain the sequence distribution of the local semantic enhancement encoding vector of the historical temperature time-series features includes: First, the semantic contribution weighting factor of each historical temperature time-series feature vector relative to other historical temperature time-series feature vectors is calculated within the local semantic enhancement perception window of historical temperature, expressed by the formula:

[0036] in, Represents the normalization function. , and The weight matrix is ​​a learnable matrix. for and Feature dimensions, Indicates in Inside, Compared to Semantic contribution weighting factor; Furthermore, based on the semantic contribution weighting factor, local semantic embedding enhancement is performed on each historical temperature time-series feature vector to obtain the sequence distribution of the local semantic enhancement encoding vector of historical temperature time-series features, which is expressed by the formula:

[0037] in, This represents the activation function. The sequence distribution of the local semantic enhancement encoding vector for historical temperature time series features. Local semantic enhancement encoding vector of historical temperature time series features Sequence distribution of encoding vectors for local semantic enhancement of historical temperature time-series features ， These are the 1st, 2nd, and 3rd elements in the sequence distribution of the local semantic enhancement encoding vectors for historical temperature time series features. A local semantic enhancement encoding vector for historical temperature time-series features.

[0038] Subsequently, the sequence distribution of the local semantic enhancement encoding vector of historical temperature time series features is input into a time series effect global propagation encoding network based on the Transformer architecture to obtain the historical temperature time series effect propagation latent encoding vector. It should be understood that while local semantic enhancement encoding can finely characterize the thermodynamic response of a specific time window (such as the instantaneous contraction effect of a metal pipe within a 72-hour window), it cannot model the propagation law of temperature stress along the time axis (such as the thermal hysteresis conduction chain between shallow soil and deep rock). Therefore, in the technical solution of this application, the complete causal chain between historical temperature data and thermodynamic conduction effects is dynamically captured through the global attention mechanism of the Transformer, enabling the temperature correction model to simulate real-world thermodynamic conduction.

[0039] Specifically, in the technical solution of this application, the process of inputting the sequence distribution of the local semantic enhancement encoding vector of historical temperature time series features into the time series effect global propagation encoding network based on the Transformer architecture is expressed by the following formula:

[0040] in, Indicates global context encoding. The implicit encoding vector for propagating the historical temperature time series effect.

[0041] Furthermore, the historical temperature time-series effect propagation implicit coding vector is temporally feature-decoded to obtain the temperature drift time-series sensitivity factor. It should be understood that although the historical temperature time-series effect propagation implicit coding vector contains rich temporal pattern information, its form is complex and cannot be directly substituted into the leakage calculation formula. Therefore, in the technical solution of this application, temporal feature decoding is performed on the historical temperature time-series effect propagation implicit coding vector to achieve the transformation from an abstract, high-dimensional feature representation to a specific, directly applicable physical quantitative indicator.

[0042] In practice, the temperature drift time-series sensitivity factor can be obtained by propagating the historical temperature time-series effect into a hidden encoded vector and passing it through a decoder. The decoder is typically a neural network consisting of one or more fully connected layers, whose weights and bias parameters are determined through training on a large amount of historical data. During this process, the decoder transforms the complex feature vectors learned within the model, which characterize specific patterns (such as the hysteresis and cumulative effects of temperature), into a target value with clear interpretability and practicality—the temperature drift time-series sensitivity factor.

[0043] Specifically, in step S3, the reliability of existing leakage is calculated in response to the temporal distribution of the ratio between the estimated leakage amount and the total daily water supply of the zone meeting a preset condition. It should be understood that existing DMA leakage assessment methods rely on simple production-sales difference calculations, which are easily affected by fluctuations in normal user water usage behavior, intermittent water usage by large users, and metering equipment errors, leading to an inability to accurately identify long-term structural existing leakage. For example, sudden changes in water usage may mask small but continuous leakage signals, causing distortion in leakage assessment. By requiring the temporal distribution of the ratio between the estimated leakage amount and the total daily water supply of the zone to meet preset conditions (such as continuous stability), temporary interference can be effectively filtered out, ensuring that subsequent reliability calculations only target real background leakage. Therefore, in the technical solution of this application, the reliability of existing leakage is calculated in response to the temporal distribution of the ratio between the estimated leakage amount and the total daily water supply of the zone meeting a preset condition. In other words, by calculating the credibility of existing leakage, we can accurately identify the characteristic patterns that truly belong to existing leakage from continuous leakage estimation data, and conduct multi-dimensional comprehensive verification to avoid misjudging occasional water volume fluctuations or data noise as persistent pipeline problems.

[0044] The preset conditions are: the ratio exceeds a preset threshold for 30 consecutive days, and the fluctuation coefficient of the estimated leakage amount is less than a preset threshold for 30 consecutive days. In a more specific implementation, the preset threshold for the ratio is 12%, and the fluctuation coefficient is specifically the standard deviation of the estimated leakage amount for 30 consecutive days. That is, if the estimated leakage rate of the DMA partition remains above 12% for a continuous month, and the estimated leakage amount for that month is relatively stable without drastic fluctuations, then the trigger condition is met, and subsequent confidence calculations are initiated.

[0045] Specifically, in a specific example of this application, the confidence level of stock leakage is calculated using the following formula;

[0046] in, For the monthly production and sales difference of the region, For the number of work orders in each zone, This represents the average number of work orders in the system. The average age of the pipes in each zone. This is to assess the reliability of existing leakage.

[0047] Among them, the monthly production and sales difference of a specific DMA zone refers to the difference between the total water supply and the total chargeable water volume within a natural month. It is a comprehensive indicator for measuring pipeline leakage and metering error, and it is expressed by the formula:

[0048] in, Indicates monthly water supply. Indicates the monthly water usage. The monthly production and sales difference of a zone serves as the most direct and weighted evidence of existing leakage from the perspective of overall water balance. The number of work orders in a zone represents the average number of work orders in the system. The ratio of the two is used to consider historical operation and maintenance experience. If the number of repair work orders in a zone is significantly higher than the average level, it indicates that there is a greater probability of problems in the pipeline network itself in that area, providing operational support for leakage judgment. The average pipe age of a zone provides a basis from the perspective of the physical health of pipeline assets. Older pipelines are naturally more likely to age and corrode, leading to existing leakage, which provides credibility at the physical level for judgment.

[0049] Specifically, in step S4, a high-confidence leakage warning is triggered in response to the existing leakage confidence exceeding a preset threshold. In the technical solution of this application, the preset threshold is 15%. If the system determines that the calculated result of the existing leakage confidence is greater than this preset threshold, a "high-confidence leakage warning" event is automatically triggered.

[0050] Among them, the preset threshold represents the risk tolerance set by the system administrator. It can effectively control the false alarm rate while ensuring a high detection rate, and is an important parameter for balancing the system's sensitivity and specificity.

[0051] It is worth mentioning that the warning can be triggered in various forms. For example, an alarm window may pop up on the monitoring screen of the water supply dispatch center, or the warning information may be sent directly to the maintenance personnel responsible for the area through the system in the form of message push or work order dispatch, so as to guide the maintenance personnel to carry leak detection equipment to the target area for accurate leak investigation and repair.

[0052] In summary, the water supply network leakage assessment method based on nighttime water volume using DMA, as described in this application, is explained. It leverages the relatively stable nature of nighttime water consumption and innovatively integrates data from lower-level zones, large users' nighttime water consumption, and temperature to refine the nighttime flow rate, thereby calculating an estimated leakage amount closer to reality. By further analyzing the temporal distribution of this leakage percentage and introducing a reliability assessment, interference caused by instantaneous fluctuations can be effectively filtered out, ultimately triggering an early warning only when the reliability of the existing leakage is sufficiently high. This method significantly improves the accuracy of leakage assessment and the reliability of early warning, enabling water utilities to proactively and accurately identify and manage existing leakage, providing strong support for scientific decision-making.

[0053] The various embodiments of this disclosure have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best explain the principles, practical application, or improvement of the technology in the market, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. A method for determining the leakage of water supply network inventory based on nighttime water volume (DMA), characterized in that, include: Collect nighttime water volume data of DMA partitions, nighttime water volume data of lower-level partitions, and nighttime water volume and temperature data of large users within the partitions; Based on the nighttime water volume data of the DMA partition, the nighttime water volume data of the lower-level partition, and the nighttime water volume data and temperature data of large users within the partition, the estimated leakage is calculated. The reliability of the stock leakage is calculated in response to the time-series distribution of the ratio between the estimated leakage and the total daily water supply of the zone meeting the preset conditions. In response to the existing leakage confidence exceeding a preset threshold, a high-confidence leakage warning is triggered.

2. The method for judging the leakage of water supply network based on nighttime water volume of DMA as described in claim 1, which calculates and estimates the leakage based on nighttime water volume data of DMA zones, nighttime water volume data of lower-level zones, and nighttime water volume data and temperature data of large users within the zones, includes: The basic leakage is calculated based on the nighttime water volume data of the DMA partition, the nighttime water volume data of the lower-level partition, and the nighttime water volume data of large users within the partition. The baseline leakage rate is corrected based on temperature data to obtain the estimated leakage rate.

3. The method for determining water supply network leakage based on DMA nighttime water volume according to claim 2, characterized in that, Based on the nighttime water volume data of the DMA partition, the nighttime water volume data of the lower-level partition, and the nighttime water volume data of large users within the partition, the basic leakage is calculated, including: The basic leakage is calculated using the following formula: ， in, For DMA partition nighttime water volume data, For nighttime water volume data of lower-level zones and This is the nighttime water consumption data for large users within the partition. Basic leakage; The process of correcting the baseline leakage rate based on temperature data to obtain the estimated leakage rate includes: correcting the baseline leakage rate based on temperature data using the following formula to obtain the estimated leakage rate, wherein the formula is: ， in, Based on the basic leakage rate, Temperature is a factor that affects the environment. As a temperature drift time-sensitive factor, This is a temperature correction factor. The temperature data for that day. As the reference temperature, To estimate the amount of leakage.

4. The method for determining water supply network leakage based on DMA nighttime water volume according to claim 3, characterized in that, The temperature drift time sensitivity factor is 1.

5. The method for determining water supply network leakage based on DMA nighttime water volume according to claim 1, characterized in that, In calculating the reliability of stock leakage in response to the time-series distribution of the ratio between the estimated leakage and the total daily water supply of the zone meeting preset conditions, the preset conditions are that the ratio exceeds a preset threshold for 30 consecutive days and the fluctuation coefficient of the estimated leakage for 30 consecutive days is less than a preset threshold.

6. The method for determining water supply network leakage based on DMA nighttime water volume according to claim 5, characterized in that, The preset threshold is 12%, and the fluctuation coefficient of the estimated leakage over 30 consecutive days is the standard deviation of the estimated leakage over 30 consecutive days.

7. The method for determining water supply network leakage based on DMA nighttime water volume according to claim 1, characterized in that, In response to the time-series distribution of the ratio between estimated leakage and total daily water supply in a given area meeting preset conditions, the reliability of the stock leakage is calculated, including: The reliability of existing leakage is calculated using the following formula: ， in, For the monthly production and sales difference of the region, For the number of work orders in each zone, This represents the average number of work orders in the system. The average age of the pipes in each zone. This is to assess the reliability of existing leakage.

8. The method for determining water supply network leakage based on DMA nighttime water volume according to claim 7, characterized in that, In response to the existing leakage confidence exceeding a preset threshold, a high-confidence leakage warning is triggered, wherein the preset threshold is 15%.

9. The method for determining water supply network leakage based on DMA nighttime water volume according to claim 3, characterized in that, The process of determining the temperature drift time-sensitive factor includes: Get temperature data for the scheduled number of days prior to the current day's temperature data; Arrange the temperature data for the predetermined number of days into a historical temperature data sequence; Time-series propagation coding is performed on historical temperature data sequences to obtain the implicit coding vector of historical temperature time-series effect propagation. Temporal feature decoding is performed on the implicit encoding vector of historical temperature time-series effects to obtain the temperature drift time-series sensitivity factor.

10. The method for determining water supply network inventory leakage based on DMA nighttime water volume according to claim 9, characterized in that, The historical temperature data sequence is time-series propagation encoded to obtain the historical temperature time-series effect propagation latent encoding vector, including: The historical temperature data sequence is segmented to obtain the sequence distribution of historical temperature time-series feature vectors; Based on the feature distribution of each historical temperature time series feature vector in the sequence distribution, semantic transfer encoding is performed on the sequence distribution of historical temperature time series feature vectors to obtain the implicit encoding vector for the propagation of historical temperature time series effects.

Images