A cloud fusion-based multi-carrier water leakage data analysis method and system

Abstract

The application relates to the technical field of water leakage positioning, and discloses a multi-carrier water leakage data analysis method and system based on cloud fusion, which comprises the following steps: performing format standardization processing on multi-source operation data to obtain multi-carrier data; performing multi-scale decomposition on the multi-carrier data to obtain multi-scale feature components; performing hierarchical aggregation on the coupling features of the multi-scale feature components to obtain multi-scale cross-carrier coupling features; constructing a benchmark coupling metric sequence according to historical normal working condition data, comparing the multi-scale cross-carrier coupling features with the benchmark coupling metric sequence, obtaining a decomposition scale and a corresponding abnormal carrier that have a characteristic deviation, and decoupling an abnormal coupling component; matching a cooperative matching rule of the multi-scale cross-carrier coupling features according to the decomposition scale and the corresponding abnormal carrier, and performing reverse tracing analysis on the abnormal coupling component to obtain a water leakage position judgment result; and the application can improve the efficiency of the multi-carrier water leakage data analysis based on cloud fusion.

Description

Technical Field

[0001] This invention relates to the field of leakage location technology, and in particular to a cloud-based multi-carrier leakage data analysis method and system. Background Technology

[0002] In the current process of analyzing and processing water supply network leaks, there is a lack of unified standardization and time-series alignment methods for multi-source operational data from different carriers such as pressure, flow, and acoustics. Data from different sampling frequencies are difficult to directly fuse and analyze, and it is impossible to fully explore the cross-scale coupling and correlation characteristics between multi-source data. This can easily lead to confusion between the overall operational trend of the network and local leakage anomalies, resulting in insufficient extraction of effective anomaly information and leakage feature identification being easily affected by irrelevant operating conditions.

[0003] Existing leak detection methods struggle to establish stable baseline coupling evolution relationships, failing to effectively decouple and separate true abnormal coupling components. Furthermore, they lack the capability for anomaly backpropagation analysis based on pipeline topology, resulting in significant deviations in leak location determination. The overall processing performance of leak identification and localization falls short of the practical needs of pipeline network operation and maintenance. Therefore, improving the ability to analyze and locate leak data across multiple carriers in water supply networks has become an urgent problem to be solved. Summary of the Invention

[0004] This invention provides a cloud-based, multi-carrier leakage data analysis method and system to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, this invention provides a cloud-based, multi-carrier leakage data analysis method, comprising: D1. Standardize the format of the multi-source operation data of the water supply network on different carriers to obtain the multi-carrier data of the water supply network. D2. Based on the coupling strength of the multi-carrier data, perform multi-scale decomposition on the multi-carrier data to obtain the multi-scale feature components of the multi-carrier data; D3. Perform hierarchical aggregation of the coupling features of the multi-scale feature components to obtain the multi-scale cross-carrier coupling features of the water supply network. D4. Based on the historical normal operating data of the water supply network, construct the benchmark coupling measurement sequence of the water supply network, compare the deviation of the multi-scale cross-carrier coupling characteristics with the benchmark coupling measurement sequence to identify the decomposition scale and corresponding abnormal carriers of the water supply network that have deviated from the characteristics, and decouple the abnormal coupling components from the abnormal carriers. D5. Based on the decomposition scale and the corresponding abnormal carrier, match the collaborative matching rules of the multi-scale cross-carrier coupling features, and perform reverse tracing analysis on the abnormal coupling components according to the topological connection relationship of the water supply network and the collaborative matching rules to obtain the leakage location determination result of the water supply network.

[0006] In a preferred embodiment, the step of standardizing the format of the multi-source operational data of the water supply network to obtain multi-carrier data of the water supply network includes: Acquire multi-source operational data of the water supply network, including network pressure data, flow data, and acoustic data; Data cleaning is performed on the multi-source operational data to obtain standardized multi-carrier data of the water supply network; The standardized multi-carrier data is time-aligned, and carrier data with different sampling frequencies are mapped to a unified time reference to obtain the multi-carrier data of the water supply network.

[0007] In a preferred embodiment, the step of performing multi-scale decomposition on the multi-carrier data based on the coupling strength of the multi-carrier data to obtain the multi-scale feature components of the multi-carrier data includes: Cross-correlation analysis is performed on the multi-carrier data to obtain the coupling correlation strength between the multi-carrier data; Based on the coupling strength, the optimal decomposition scale parameter of the multi-carrier data is determined. When the coupling strength is higher than a preset strength threshold, a fine decomposition scale is selected, and when the coupling strength is lower than the preset strength threshold, a coarse decomposition scale is selected. Based on the optimal decomposition scale parameter, the multi-carrier data is decomposed into multi-scale components to obtain the multi-scale feature components of the water supply network. The multi-scale feature components include coarse-scale components that characterize the global operating trend of the network and fine-scale components that characterize local abnormal disturbances.

[0008] In a preferred embodiment, the hierarchical aggregation of the coupling features of the multi-scale feature components to obtain the multi-scale cross-carrier coupling features of the water supply network includes: The coupling correlation features between pressure data, flow data and acoustic data in the coarse-scale component and the fine-scale component are extracted to obtain the coarse-scale carrier coupling features and fine-scale carrier coupling features of the water supply network. Based on the scale hierarchy relationship between the coarse-scale carrier coupling characteristics and the fine-scale carrier coupling characteristics, the aggregate hierarchical structure of the water supply network is obtained; Based on the aforementioned aggregation hierarchy, the fine-scale carrier coupling features are first aggregated, and then the aggregation result is combined and fused with the coarse-scale carrier coupling features across scales to obtain the aggregated coupling features of the water supply network. The aggregation coupling features are fused across levels to obtain the multi-scale cross-carrier coupling features of the water supply network.

[0009] In a preferred embodiment, the step of constructing a baseline coupling metric sequence for the water supply network based on historical normal operating data of the water supply network, comparing the deviation of the multi-scale cross-carrier coupling characteristics with the baseline coupling metric sequence to identify the decomposition scale and corresponding abnormal carriers where the water supply network's characteristics deviate, and decoupling the abnormal coupling components from the abnormal carriers includes: The historical normal operating condition data of the water supply network is obtained, and the historical normal operating condition data is divided into time-series slices to obtain the normal operating condition time-series segments of the historical normal operating condition data. Extract the normal coupling metric values ​​of the normal operating condition time sequence segment at different decomposition scales to obtain the benchmark coupling metric sequence of the water supply network at different decomposition scales; The coupling metric value in the multi-scale cross-carrier coupling feature is compared with the normal coupling metric value in the benchmark coupling metric sequence to obtain the coupling deviation degree of the water supply network. Based on the coupling deviation, the target decomposition scale of the water supply network is determined, and the carrier corresponding to the target decomposition scale is identified as an abnormal carrier. The abnormal coupling component is separated from the abnormal carrier.

[0010] In a preferred embodiment, separating the anomalous coupling component from the anomalous carrier includes: Obtain the abnormal coupling metric values ​​of the abnormal carrier at different decomposition scales; The abnormal coupling metric value is compared with the normal coupling metric value at the corresponding scale in the benchmark coupling metric sequence to obtain the coupling deviation ratio of the abnormal carrier. Arrange the coupling deviation ratios to obtain the cross-scale deviation ratio matrix of the abnormal carrier; The off-diagonal elements of the cross-scale deviation ratio matrix are extracted, and the extracted elements are summed to obtain the total scale interference under different decomposition scales. The total scale interference is normalized to obtain the inter-scale interference weight coefficients at different decomposition scales. The abnormal coupling component of the water supply network is calculated using the coupling deviation ratio, the inter-scale interference weight coefficient, the abnormal coupling metric, and the normal coupling metric.

[0011] In a preferred embodiment, the formula for calculating the abnormal coupling component is as follows: ; in, Indicates the target decomposition scale index. This represents the summation index for traversing all decomposition scales. This represents the total number of decomposition scales. Indicates that the abnormal carrier is in the first Anomaly coupling metrics at each decomposition scale Indicating the first in the reference coupling metric sequence Normal coupling metric at each decomposition scale Indicates that the abnormal carrier is in the first Coupling deviation ratio at each decomposition scale Indicates the first Inter-scale interference weighting coefficients for each decomposition scale Indicates that the abnormal carrier is in the first Anomalous coupling components at each decomposition scale.

[0012] In a preferred embodiment, the step of matching the multi-scale cross-carrier coupling features according to the decomposition scale and corresponding anomalous carriers, and performing reverse tracing analysis on the anomalous coupling components according to the topological connection relationship of the water supply network and the collaborative matching rules to obtain the leakage location determination result of the water supply network includes: The decomposition scale and the corresponding anomaly carrier are encoded and mapped to obtain the anomaly identifier sequence of the anomaly carrier under different decomposition scales; Based on the anomaly identifier sequence, a collaborative matching rule for the multi-scale cross-carrier coupling feature is matched from a preset rule base; Adjacency analysis is performed on the topological connection relationship of the water supply network to obtain the topological adjacency matrix between multiple carriers in the water supply network; Based on the topological adjacency matrix and the cooperative matching rule, path backtracking analysis is performed on the abnormal coupling components to obtain the propagation contribution sequence of the abnormal coupling components on the topological path; The propagation contribution sequence is sorted and compared, and the topological position corresponding to the largest propagation contribution in the sorting result is determined as the leakage location determination result of the water supply network.

[0013] In a preferred embodiment, the step of performing path backtracking analysis on the anomalous coupling components based on the topological adjacency matrix and the cooperative matching rules to obtain the propagation contribution sequence of the anomalous coupling components on the topological path includes: Starting from the pipeline monitoring node corresponding to the abnormal carrier, the abnormal propagation path of the abnormal carrier is obtained by traversing in reverse along the adjacency relationship in the topological adjacency matrix. The abnormal coupling components of the intermediate nodes on the abnormal propagation path are accumulated to obtain the path cumulative abnormality of the abnormal propagation path. According to the collaborative matching rules, the cumulative anomaly amount of the path is weighted and assigned to obtain the weighted anomaly amount of the anomaly propagation path. The weighted outliers are normalized and compared to obtain the propagation contribution of the outlier propagation path. The propagation contribution values ​​are arranged in path order to obtain the propagation contribution value sequence of the abnormal coupling component on the topological path.

[0014] To address the aforementioned problems, the present invention also provides a cloud-based multi-carrier leakage data analysis system, the system comprising: The data standardization preprocessing module is used to standardize the format of multi-source operational data of the water supply network on different carriers to obtain multi-carrier data of the water supply network. The multi-scale feature decomposition module is used to perform multi-scale decomposition on the multi-carrier data according to the coupling strength of the multi-carrier data to obtain the multi-scale feature components of the multi-carrier data. The coupling feature hierarchical aggregation module is used to hierarchically aggregate the coupling features of the multi-scale feature components to obtain the multi-scale cross-carrier coupling features of the water supply network. The abnormal coupling component decoupling identification module is used to construct a benchmark coupling metric sequence of the water supply network based on historical normal operating data of the water supply network, compare the deviation of the multi-scale cross-carrier coupling features with the benchmark coupling metric sequence, identify the decomposition scale and corresponding abnormal carrier of the water supply network where the features deviate, and decouple the abnormal coupling components from the abnormal carrier. The reverse propagation leakage location module is used to match the multi-scale cross-carrier coupling features with the decomposition scale and corresponding abnormal carrier according to the collaborative matching rules, and to perform reverse tracing analysis on the abnormal coupling components according to the topological connection relationship of the water supply network and the collaborative matching rules, so as to obtain the leakage location determination result of the water supply network.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention performs format standardization and time-series alignment processing on multi-carrier operation data, and completes multi-scale decomposition by combining coupling strength. It can effectively separate the global operation trend of the pipeline network from local abnormal disturbances. Then, through hierarchical aggregation, multi-scale cross-carrier coupling characteristics are obtained, which fully explores the cross-scale and cross-carrier intrinsic correlation between pressure, flow, and acoustic multi-source data, and improves the utilization depth of multi-source data information in water supply network leakage scenarios.

[0016] 2. This invention constructs a baseline coupling evolution relationship based on historical normal operating conditions to complete deviation comparison and decouple abnormal coupling components. Then, it combines the pipeline network topology connection relationship to implement reverse propagation and backtracking analysis of abnormal components, accurately locating the topological location corresponding to the leak, reducing interference caused by irrelevant operating conditions, and improving the reliability of leak identification and the accuracy of leak location determination. Attached Figure Description

[0017] Figure 1 A flowchart illustrating a cloud-based multi-carrier leakage data analysis method according to an embodiment of the present invention; Figure 2 A functional module diagram of a cloud-based multi-carrier leakage data analysis system provided in an embodiment of the present invention; The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0018] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0019] This application provides a cloud-integrated multi-carrier water leakage data analysis method. The executing entity of this cloud-integrated multi-carrier water leakage data analysis method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application embodiment: a server, a terminal, etc. In other words, the cloud-integrated multi-carrier water leakage data analysis method can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster. The server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0020] Reference Figure 1The diagram shown is a flowchart illustrating a cloud-based multi-carrier leakage data analysis method according to an embodiment of the present invention. In this embodiment, the cloud-based multi-carrier leakage data analysis method includes: D1. Standardize the format of the multi-source operation data of the water supply network on different carriers to obtain the multi-carrier data of the water supply network. In this embodiment of the invention, the step of standardizing the format of multi-source operational data of the water supply network to obtain multi-carrier data of the water supply network includes: Acquire multi-source operational data of the water supply network, including network pressure data, flow data, and acoustic data; Data cleaning is performed on the multi-source operational data to obtain standardized multi-carrier data of the water supply network; The standardized multi-carrier data is time-aligned, and carrier data with different sampling frequencies are mapped to a unified time reference to obtain the multi-carrier data of the water supply network.

[0021] All data collection records are read from the pressure, flow, and acoustic acquisition devices deployed on-site in the water supply network to obtain complete multi-source operational data of the water supply network. The multi-source operational data of the water supply network completely includes network pressure data, flow data, and acoustic data. All original records reported by the acquisition devices are completely retained for each type of data. All original records completely carry the actual network operation status perceived by the corresponding monitoring points at the time of acquisition by the devices, without discarding the valid original record content reported by the devices.

[0022] Data cleaning is performed on each of the acquired multi-source operational data of the water supply network. Every record within the network's pressure, flow, and acoustic data is traversed, and completely duplicate records are identified and deleted. Invalid records that are empty or exceed the normal operating range of the network are directly removed. For monitoring points with missing records, the missing content is filled and corrected based on the valid operational records of adjacent times before and after the point. All network pressure, flow, and acoustic data are converted to a unified storage format, and the layout of fields within each type of data is uniformly adjusted. After all processing is completed, standardized multi-carrier data of the water supply network is directly generated.

[0023] The system reads the time stamp information of each record in the standardized multi-carrier data of the water supply network, distinguishes and identifies the inherent actual sampling frequencies of the network pressure data, flow data, and acoustic data, selects a fixed time base as the common reference standard for all carrier data, extracts the original time stamp corresponding to each record of the network pressure data, flow data, and acoustic data, and maps each record to the selected unified time base according to the original time stamp. The arrangement of records of carrier data with different sampling frequencies is adjusted so that the operation records of the same network time under different carriers can be matched with each other in the time dimension. When the sampling frequencies of different carriers are significantly different or there are missing timestamps in some areas, linear resampling interpolation is used to fill in the time series points. After all records have completed the time mapping process, the multi-carrier data of the water supply network is obtained directly.

[0024] The beneficial effects are that this processing flow fully realizes the acquisition, cleaning, and time-series alignment of multi-source operational data of the water supply network, outputting multi-carrier data of the water supply network with a regular format and a completely unified time base, eliminating data differences caused by different acquisition devices in terms of storage format and sampling period, ensuring that the three types of data—pressure, flow, and acoustic—can achieve mutual correspondence at the time-series level, providing a reliable and complete input foundation for subsequent leakage analysis processes such as multi-carrier data coupling strength analysis and multi-scale decomposition, and avoiding the distortion of subsequent feature extraction results caused by the disordered original data format and time misalignment.

[0025] D2. Based on the coupling strength of the multi-carrier data, perform multi-scale decomposition on the multi-carrier data to obtain the multi-scale feature components of the multi-carrier data; In this embodiment of the invention, the step of performing multi-scale decomposition on the multi-carrier data based on the coupling strength of the multi-carrier data to obtain the multi-scale feature components of the multi-carrier data includes: Cross-correlation analysis is performed on the multi-carrier data to obtain the coupling correlation strength between the multi-carrier data; Based on the coupling strength, the optimal decomposition scale parameter of the multi-carrier data is determined. When the coupling strength is higher than a preset strength threshold, a fine decomposition scale is selected, and when the coupling strength is lower than the preset strength threshold, a coarse decomposition scale is selected. Based on the optimal decomposition scale parameter, the multi-carrier data is decomposed into multi-scale components to obtain the multi-scale feature components of the water supply network. The multi-scale feature components include coarse-scale components that characterize the global operating trend of the network and fine-scale components that characterize local abnormal disturbances.

[0026] The process iterates through all time-series records in the multi-carrier data of the water supply network, and performs cross-correlation analysis on each pair of complete time-series data sequences corresponding to different carriers. It compares the synchronous changes of different carrier data at various time positions, and statistically analyzes the actual correlation degree when the data of one type of carrier changes, and the linkage influence level between different carriers is fully statistically analyzed to obtain the coupling correlation strength between the multi-carrier data of the water supply network. The coupling correlation strength output by the cross-correlation analysis takes values ​​in the range of [-1, 1]. The closer the value is to 1, the stronger the linkage between carriers. The closer it is to -1, the negative linkage is represented. The closer it is to 0, the almost no linkage relationship is represented.

[0027] The preset strength threshold is determined by collecting cross-correlation analysis results of multi-carrier data under a large number of historical normal operation and leakage fault conditions of the water supply network, statistically analyzing the distribution range of coupling correlation strength under different conditions, and selecting a value that can distinguish between strong and weak linkage states between carriers. In this embodiment, 0.6 can be selected as the preset strength threshold. In actual engineering, the distribution of coupling correlation strength under historical conditions can be statistically analyzed, and the quantile of the cross-correlation coefficient under normal conditions can be taken as the threshold. The calculated coupling correlation strength is compared with the preset strength threshold item by item. If the coupling correlation strength is higher than the preset strength threshold, a fine decomposition scale is directly selected. If the coupling correlation strength is lower than the preset strength threshold, a coarse decomposition scale is directly selected. This completes the determination of the optimal decomposition scale parameter for multi-carrier data.

[0028] The optimal decomposition scale parameters that have been determined are invoked to perform a complete multi-scale decomposition process on the multi-carrier data of the water supply network. The continuous and complete multi-carrier data time series information is split into two types of non-overlapping time series components. One type of component retains the slowly changing operation information over a large area of ​​the network, which is used to reflect the overall operation trend of the network over time and generate a coarse-scale component that represents the global operation trend of the network. The other type of component retains the short-term rapid fluctuation information generated at local points in the network, which is used to reflect the short-term fluctuation changes generated at local locations in the network and generate a fine-scale component that represents local abnormal disturbances. The coarse-scale component and the fine-scale component are merged and summarized to directly obtain the multi-scale feature components of the water supply network.

[0029] The beneficial effects are as follows: by fully exploring the linkage relationship between the temporal changes of data from different carriers through cross-correlation analysis to obtain the coupling correlation strength, by relying on the preset strength threshold determined by pre-statistical data of various operating conditions of the pipeline network to complete the adaptive selection of scale, and then by multi-scale decomposition to separate the global operating trend and local abnormal disturbances, it gets rid of the limitation that the fixed decomposition scale cannot adapt to the different linkage states between carriers, separates the normal large-scale operating changes of the pipeline network from the local disturbance information caused by leakage, reduces the masking effect of the global operating trend of the pipeline network on the local leakage anomaly signal, provides a basic time-series component for the subsequent hierarchical aggregation of coupling features, and improves the identification effect of leakage-related anomaly information.

[0030] D3. Perform hierarchical aggregation of the coupling features of the multi-scale feature components to obtain the multi-scale cross-carrier coupling features of the water supply network. In this embodiment of the invention, the hierarchical aggregation of the coupling features of the multi-scale feature components to obtain the multi-scale cross-carrier coupling features of the water supply network includes: The coupling correlation features between pressure data, flow data and acoustic data in the coarse-scale component and the fine-scale component are extracted to obtain the coarse-scale carrier coupling features and fine-scale carrier coupling features of the water supply network. Based on the scale hierarchy relationship between the coarse-scale carrier coupling characteristics and the fine-scale carrier coupling characteristics, the aggregate hierarchical structure of the water supply network is obtained; Based on the aforementioned aggregation hierarchy, the fine-scale carrier coupling features are first aggregated, and then the aggregation result is combined and fused with the coarse-scale carrier coupling features across scales to obtain the aggregated coupling features of the water supply network. The aggregation coupling features are fused across levels to obtain the multi-scale cross-carrier coupling features of the water supply network.

[0031] The entire time-series content of pressure, flow, and acoustic data contained in the already generated coarse-scale component and fine-scale component is read separately. The linkage change patterns of pressure, flow, and acoustic data over time are analyzed in the coarse-scale component, and the linkage change patterns of pressure, flow, and acoustic data over time are analyzed in the fine-scale component. The linkage and correlation information between the three types of data under the coarse-scale component is completely extracted to generate the coarse-scale carrier coupling characteristics of the water supply network. The linkage and correlation information between the three types of data under the fine-scale component is completely extracted to generate the fine-scale carrier coupling characteristics of the water supply network.

[0032] By analyzing the hierarchical attributes of the information carried by coarse-scale carrier coupling features and fine-scale carrier coupling features, we can identify that coarse-scale carrier coupling features carry overall operational information of the pipeline network that changes slowly over a large area, while fine-scale carrier coupling features carry disturbance information of rapid fluctuations at local points in the pipeline network. Fine-scale carrier coupling features belong to lower-level detailed information, while coarse-scale carrier coupling features belong to upper-level general information. Based on this hierarchical correspondence between the two, which progresses from lower-level details to upper-level general information, we can construct an organizational structure that divides the network into upper-level information nodes and lower-level information nodes, thus obtaining the aggregated hierarchical structure of the water supply pipeline network.

[0033] Strictly following the information processing sequence defined within the established aggregation hierarchy, the aggregation operation is prioritized for fine-scale carrier coupling features located at lower-level information nodes. These fine-scale carrier coupling features, scattered across different monitoring points, are merged and integrated according to their node affiliation within the aggregation hierarchy. After merging and integration, a fine-scale coupling aggregation result is obtained. This result is then combined with the coarse-scale carrier coupling features located at upper-level information nodes, performing cross-scale splicing and fusion. This ensures that the detailed information from the lower-level aggregation is fully integrated into the overall upper-level information, with information matched one-to-one according to the time dimension. No valid correlation information originally carried within the two types of features is discarded. After processing, the aggregated coupling features of the water supply network are obtained.

[0034] Cross-level fusion processing is performed on the generated aggregated coupling features to open up the information transmission path between the upper and lower nodes inside the aggregated coupling features, eliminate the information isolation between information at different scale levels, and allow the overall coupling information at the coarse scale level and the local coupling information at the fine scale level to refer to and complement each other. The coupling information belonging to different scales and carriers is integrated into a complete and unified feature set. After processing, the multi-scale cross-carrier coupling features of the water supply network are obtained.

[0035] The beneficial effects are as follows: the coupling correlation features between different monitoring carriers are extracted from the coarse-scale component and the fine-scale component, respectively; an aggregation hierarchical structure is built based on the inherent scale hierarchy relationship of the two types of features; the fine-scale carrier coupling feature aggregation is completed first in a bottom-up processing order, and then cross-scale splicing and fusion are carried out; finally, cross-level fusion is performed to generate multi-scale cross-carrier coupling features. This fully preserves the correlation information between multiple carriers at both the global operation level and the local disturbance level of the pipeline network, avoids the fragmentation of feature information at different scales, and provides a complete feature input for subsequent deviation comparison with the baseline coupling evolution relationship. This ensures that subsequent leakage anomaly identification can simultaneously take into account the overall operating conditions of the pipeline network and local anomaly disturbances.

[0036] D4. Based on the historical normal operating data of the water supply network, construct the benchmark coupling measurement sequence of the water supply network, compare the deviation of the multi-scale cross-carrier coupling characteristics with the benchmark coupling measurement sequence to identify the decomposition scale and corresponding abnormal carriers of the water supply network that have deviated from the characteristics, and decouple the abnormal coupling components from the abnormal carriers. In this embodiment of the invention, the step of constructing a benchmark coupling metric sequence for the water supply network based on historical normal operating data of the water supply network, comparing the deviation between the multi-scale cross-carrier coupling characteristics and the benchmark coupling metric sequence to identify the decomposition scale and corresponding abnormal carriers where the water supply network's characteristics deviate, and decoupling the abnormal coupling components from the abnormal carriers includes: The historical normal operating condition data of the water supply network is obtained, and the historical normal operating condition data is divided into time-series slices to obtain the normal operating condition time-series segments of the historical normal operating condition data. Extract the normal coupling metric values ​​of the normal operating condition time sequence segment at different decomposition scales to obtain the benchmark coupling metric sequence of the water supply network at different decomposition scales; The coupling metric value in the multi-scale cross-carrier coupling feature is compared with the normal coupling metric value in the benchmark coupling metric sequence to obtain the coupling deviation degree of the water supply network. Based on the coupling deviation, the target decomposition scale of the water supply network is determined, and the carrier corresponding to the target decomposition scale is identified as an abnormal carrier. The abnormal coupling component is separated from the abnormal carrier.

[0037] The step of separating the anomalous coupling component from the anomalous carrier includes: Obtain the abnormal coupling metric values ​​of the abnormal carrier at different decomposition scales; The abnormal coupling metric value is compared with the normal coupling metric value at the corresponding scale in the benchmark coupling metric sequence to obtain the coupling deviation ratio of the abnormal carrier. Arrange the coupling deviation ratios to obtain the cross-scale deviation ratio matrix of the abnormal carrier; The off-diagonal elements of the cross-scale deviation ratio matrix are extracted, and the extracted elements are summed to obtain the total scale interference under different decomposition scales. The total scale interference is normalized to obtain the inter-scale interference weight coefficients at different decomposition scales. The abnormal coupling component of the water supply network is calculated using the coupling deviation ratio, the inter-scale interference weight coefficient, the abnormal coupling metric, and the normal coupling metric.

[0038] The calculation formula for the abnormal coupling component is as follows: ; in, Indicates the target decomposition scale index. This represents the summation index for traversing all decomposition scales. This represents the total number of decomposition scales. Indicates that the abnormal carrier is in the first Anomaly coupling metrics at each decomposition scale Indicating the first in the reference coupling metric sequence Normal coupling metric at each decomposition scale Indicates that the abnormal carrier is in the first Coupling deviation ratio at each decomposition scale Indicates the first Inter-scale interference weighting coefficients for each decomposition scale Indicates that the abnormal carrier is in the first Anomalous coupling components at each decomposition scale.

[0039] The system reads all operational records from periods when there were no leaks in the water supply network, obtaining complete historical normal operating data. It then performs time-series slicing on the entire set of historical normal operating data according to continuous time intervals. Each slice extracts a continuous and uninterrupted range of network operation time. The slicing process completely preserves the chronological order of the original data within each time segment and does not modify the network operation records stored within the slice. After the slicing is completed, the normal operating time-series fragments of the historical normal operating data are obtained.

[0040] For each segment of normal operating condition time series that has been divided, the multi-scale decomposition processing logic of the previous multi-carrier data is used to perform scale splitting processing on the data within the segment. Within each decomposition scale that has been divided, normal coupling metric values ​​between pressure data, flow data, and acoustic data are extracted. The coupling metric values ​​use the cross-covariance between different carrier time series as a physical index to characterize the strength of coupling and linkage between multiple carriers. The normal coupling metric values ​​corresponding to all normal operating condition time series segments at each decomposition scale are collected. All the extracted normal coupling metric values ​​are arranged in an orderly manner according to the chronological order. After the arrangement is completed, the benchmark coupling metric sequence of the water supply network at different decomposition scales is obtained.

[0041] Retrieve the multi-scale cross-carrier coupling features obtained from the current pipeline network measurement, extract the coupling metric value corresponding to each decomposition scale within the multi-scale cross-carrier coupling features, for the same decomposition scale, extract the normal coupling metric value stored at the same decomposition scale position in the benchmark coupling metric sequence, perform a difference comparison operation on the two sets of metric information under the same decomposition scale, and statistically analyze the difference level between the current measured coupling state and the historical normal coupling state. After the comparison processing is completed, the coupling deviation degree of the water supply network is obtained.

[0042] The coupling deviation degree is calculated for each decomposition scale. The coupling deviation degree corresponding to each decomposition scale is compared with the pre-determined deviation discrimination conditions one by one. The decomposition scale that meets the deviation discrimination conditions is defined as the target decomposition scale of the water supply network. The data acquisition carrier that outputs the collected information to the target decomposition scale and generates characteristic offset is directly judged as an abnormal carrier.

[0043] For the identified abnormal carriers, signal separation processing is carried out to separate the coupling information caused by the normal operation of the pipeline network and the coupling information caused by water leakage disturbance. After complete component decoupling processing, the abnormal coupling component is separated from the abnormal carrier.

[0044] Retrieve the full time-series records corresponding to the identified anomalous carriers, reuse the existing multi-scale decomposition processing flow to perform scale splitting on the time-series records of the anomalous carriers, and fully collect the measurement information corresponding to the coupling state of the carriers at each decomposition scale to obtain the anomalous coupling measurement values ​​of the anomalous carriers at different decomposition scales.

[0045] For each decomposition scale, the abnormal coupling metric value corresponding to that decomposition scale is extracted. At the same time, the normal coupling metric value stored at the same decomposition scale position in the baseline coupling metric sequence is extracted. The ratio of the two sets of metrics at the same decomposition scale position is calculated. The abnormal coupling metric value is compared with the normal coupling metric value of the corresponding scale. A set of calculation results is generated for each decomposition scale. The set of calculation results is the coupling deviation ratio of the abnormal carrier.

[0046] The coupling deviation ratios calculated from all decomposition scales are arranged in a regular manner according to the scale dimension and the carrier dimension. The rows of the matrix correspond to different decomposition scales, and the columns of the matrix correspond to the various data carriers involved in the calculation. Each set of coupling deviation ratios is filled into the corresponding position inside the matrix. After the arrangement process is completed, the cross-scale deviation ratio matrix of the abnormal carriers is obtained.

[0047] Read all elements stored in the cross-scale deviation ratio matrix line by line, skipping the elements stored in the main diagonal position of the matrix, and extracting only all the non-diagonal elements of each row except the main diagonal. Perform cumulative processing on all the extracted non-diagonal elements of the same row. The result of each row's cumulative processing represents the overall situation of the cross-interference of the corresponding decomposition scale by other scales. After all rows are processed, the total scale interference under different decomposition scales is obtained.

[0048] Read the total scale interference corresponding to each decomposition scale, calculate the sum of all scale interference, divide the total scale interference corresponding to a single decomposition scale by the sum of all scale interference, complete the normalization process of the total scale interference, eliminate the numerical influence of the overall total, and obtain the scale interference weight coefficients under different decomposition scales.

[0049] The generated coupling deviation ratio, inter-scale interference weight coefficient, abnormal coupling metric, and normal coupling metric are retrieved for comprehensive calculation. During the calculation, the inter-scale interference weight coefficient and coupling deviation ratio are matched according to the scale correspondence. The cross-interference contribution formed by all decomposition scales is statistically analyzed. The signal part caused by the cross-interference between each decomposition scale is deducted from the abnormal coupling metric. Then, the coupling contribution of the normal operating condition of the pipeline represented by the normal coupling metric is removed. The signal content generated purely by leakage disturbance is extracted. After processing, the abnormal coupling component of the water supply pipeline is calculated.

[0050] The target decomposition scale index is taken from the target decomposition scale obtained after the coupling deviation determination. The summation index of all decomposition scales is used to completely cover all the scales generated after multi-scale decomposition. The total number of decomposition scales is the total number of scales actually obtained after multi-scale decomposition of multi-carrier data. The abnormal coupling metric value of the abnormal carrier at the corresponding decomposition scale comes from the output result of the step of retrieving the abnormal coupling metric value of the abnormal carrier at different decomposition scales. The normal coupling metric value of the corresponding decomposition scale in the benchmark coupling metric sequence is taken from the benchmark coupling metric sequence of the water supply network at different decomposition scales. The coupling deviation ratio of the abnormal carrier at the corresponding decomposition scale comes from the coupling deviation ratio of the abnormal carrier. The inter-scale interference weight coefficient is set by statistically analyzing the strength of interference between different decomposition scales under multiple historical network conditions. The abnormal coupling component of the abnormal carrier at the corresponding decomposition scale is the final output of this operation process.

[0051] The calculation process first iterates and sums all decomposition scales. In the summation operation, the inter-scale interference weight coefficient and coupling deviation ratio corresponding to each decomposition scale are multiplied and then accumulated. The overall result obtained by summation is used to convert the scale cross-interference ratio. The signal part caused by mutual interference between scales is deducted from the abnormal coupling metric value. Then, the coupling contribution corresponding to the normal working condition of the pipeline represented by the benchmark coupling metric sequence is removed. Finally, the abnormal coupling component of the abnormal carrier generated only by leakage disturbance is obtained at the corresponding decomposition scale. The dimensions of all inputs and outputs in the entire calculation process are kept completely consistent. The abnormal coupling component output by the calculation can be directly sent to the subsequent reverse tracing analysis process to complete the leakage point location processing.

[0052] With other input conditions remaining unchanged, when the overall value of the inter-scale interference weight coefficient increases, the proportion of deduction caused by scale cross-interference will increase accordingly, and the value of the abnormal coupling component of the final output anomaly carrier at the corresponding decomposition scale will decrease accordingly. When the overall value of the inter-scale interference weight coefficient decreases, the proportion of deduction caused by scale cross-interference will decrease accordingly, and the value of the abnormal coupling component of the final output anomaly carrier at the corresponding decomposition scale will increase accordingly. When the normal coupling metric value of the corresponding decomposition scale in the baseline coupling metric sequence increases, the value of the abnormal coupling component of the final output anomaly carrier at the corresponding decomposition scale will decrease accordingly. When the abnormal coupling metric value of the anomaly carrier at the corresponding decomposition scale increases, the value of the abnormal coupling component of the final output anomaly carrier at the corresponding decomposition scale will increase accordingly. The output results fully and accurately reflect the true coupling anomaly level caused by leakage disturbance at the current decomposition scale.

[0053] The beneficial effects are as follows: relying on real, fault-free historical normal operating data of the water supply network, a benchmark coupling metric sequence is constructed through time-series slicing. This serves as a reference benchmark for the normal operation of the network. The coupling deviation degree is obtained by comparing the real-time multi-scale cross-carrier coupling features with the benchmark coupling metric sequence. This can accurately locate the target decomposition scale where feature deviation occurs and the corresponding abnormal carrier. After obtaining the coupling deviation ratio through ratio calculation, a cross-scale deviation ratio matrix is ​​constructed. The total scale interference is obtained by extracting the off-diagonal elements of the cumulative matrix rows. Then, the inter-scale interference weight coefficient is obtained through normalization conversion. Combined with multiple sets of metric information, comprehensive calculation and decoupling are performed to obtain the abnormal coupling components. This effectively eliminates the signal influence caused by the cross-interference between different decomposition scales, filters the coupling signals generated by the normal operating conditions of the network, and outputs pure leakage disturbance-related features. This provides reliable input data for subsequent reverse tracing analysis of network leakage points and reduces the adverse effects of daily operating condition fluctuations of the network on leakage fault location.

[0054] D5. Based on the decomposition scale and the corresponding abnormal carrier, match the collaborative matching rules of the multi-scale cross-carrier coupling features, and perform reverse tracing analysis on the abnormal coupling components according to the topological connection relationship of the water supply network and the collaborative matching rules to obtain the leakage location determination result of the water supply network.

[0055] In this embodiment of the invention, the step of matching the multi-scale cross-carrier coupling features according to the decomposition scale and the corresponding abnormal carrier, and performing reverse tracing analysis on the abnormal coupling components according to the topological connection relationship of the water supply network and the collaborative matching rules to obtain the leakage location determination result of the water supply network includes: The decomposition scale and the corresponding anomaly carrier are encoded and mapped to obtain the anomaly identifier sequence of the anomaly carrier under different decomposition scales; Based on the anomaly identifier sequence, a collaborative matching rule for the multi-scale cross-carrier coupling feature is matched from a preset rule base; Adjacency analysis is performed on the topological connection relationship of the water supply network to obtain the topological adjacency matrix between multiple carriers in the water supply network; Based on the topological adjacency matrix and the cooperative matching rule, path backtracking analysis is performed on the abnormal coupling components to obtain the propagation contribution sequence of the abnormal coupling components on the topological path; The propagation contribution sequence is sorted and compared, and the topological position corresponding to the largest propagation contribution in the sorting result is determined as the leakage location determination result of the water supply network.

[0056] The step of performing path backtracking analysis on the anomalous coupling components based on the topological adjacency matrix and the cooperative matching rules to obtain the propagation contribution sequence of the anomalous coupling components on the topological path includes: Starting from the pipeline monitoring node corresponding to the abnormal carrier, the abnormal propagation path of the abnormal carrier is obtained by traversing in reverse along the adjacency relationship in the topological adjacency matrix. The abnormal coupling components of the intermediate nodes on the abnormal propagation path are accumulated to obtain the path cumulative abnormality of the abnormal propagation path. According to the collaborative matching rules, the cumulative anomaly amount of the path is weighted and assigned to obtain the weighted anomaly amount of the anomaly propagation path. The weighted outliers are normalized and compared to obtain the propagation contribution of the outlier propagation path. The propagation contribution values ​​are arranged in path order to obtain the propagation contribution value sequence of the abnormal coupling component on the topological path.

[0057] Extract the category information of the identified decomposition scales, and simultaneously extract the corresponding equipment category and monitoring point information of the anomalous carriers. Perform coding mapping processing on the combination relationship between decomposition scales and anomalous carriers. Each combination of decomposition scales and anomalous carriers is assigned a unique corresponding identifier code. The code contains the scale attributes and carrier point attributes of the current combination, without omitting any identified anomalous combination. Arrange all the generated identifier codes in the order of scale hierarchy to form an ordered sequence set, and obtain the anomalous identifier sequence of anomalous carriers under different decomposition scales.

[0058] The pre-defined rule base is constructed by collecting a large amount of real-world field measurement data on pipeline leakage conditions, statistically analyzing the objective laws of leakage disturbance propagation and diffusion within the pipeline under different anomaly identification sequence conditions, and organizing the corresponding constraint relationships between scale, carrier, and signal propagation mode. The base stores the binding index relationship between each type of anomaly identification sequence and collaborative matching rules. The generated anomaly identification sequence is retrieved, and the anomaly identification sequence is compared with each index entry stored in the pre-defined rule base. The record in the index entry that completely matches the current anomaly identification sequence is located, and the corresponding constraint content is read from the record to obtain the collaborative matching rules of multi-scale cross-carrier coupling features.

[0059] The complete pipeline layout file and physical connectivity data of all monitoring nodes of the water supply network were retrieved. Adjacency analysis of the topological connection relationship of the water supply network was carried out. All monitoring nodes with multi-carrier acquisition devices were traversed one by one within the network. The other monitoring nodes directly connected to each monitoring node through the pipeline were confirmed one by one. The different relationships of direct pipeline connection and indirect cross-node connection were distinguished. Only the adjacency status of direct pipeline connection between nodes was recorded. Each row and column of the matrix corresponds to each monitoring node equipped with multi-carrier acquisition device within the network. The internal cells of the matrix are used to mark whether there is a direct pipeline connection between row nodes and column nodes. The adjacency status information between all pairs of nodes was completely recorded. After parsing the connectivity information of all nodes, the topological adjacency matrix between multi-carrier devices in the water supply network was obtained.

[0060] The topological adjacency matrix obtained from the parsing is read, and the cooperative matching rules that have been searched and matched are also read. A complete path backtracking analysis is performed on the abnormal coupling components output in the previous decoupling. The reverse process of the leakage disturbance signal propagating along the pipeline is simulated. The contribution level of each topological path to the current abnormal coupling component is deduced one by one. The contribution value calculated for each topological path is recorded completely. All contribution values ​​are sorted and summarized according to the correspondence of the topological paths to obtain the propagation contribution sequence of the abnormal coupling component on the topological path.

[0061] Extract all propagation contribution values ​​stored within the propagation contribution sequence, perform full sorting and comparison processing, sort and distinguish according to the actual size of the values, filter out the propagation contribution value with the largest value among all values, locate the pipeline topology location bound to the entry with the largest propagation contribution value, and directly determine the pipeline topology location as the result of the water supply pipeline leakage location determination.

[0062] Locate the actual pipeline monitoring node corresponding to the abnormal carrier, set this monitoring node as the starting node for the traversal operation, and continuously perform a reverse traversal operation along the reverse transmission direction of the leakage disturbance signal based on the direct adjacency connection status of the nodes recorded in the topological adjacency matrix. Each traversal jumps from the current node to the upstream node directly connected to it, and the traversal continues iteratively until there are no new unvisited upstream nodes. Record all pipeline nodes passed in sequence during the traversal process, connect the passed nodes in the order of traversal to form a complete path, and summarize all paths that can transmit abnormal disturbances to the starting node to obtain the abnormal propagation path of the abnormal carrier.

[0063] For each acquired anomaly propagation path, all intermediate nodes contained in the path are fully read, and the original values ​​of the anomaly coupling components corresponding to each intermediate node are retrieved. Along the single anomaly propagation path, the values ​​of the anomaly coupling components corresponding to all intermediate nodes on the path are continuously accumulated item by item. The anomaly disturbance information of all intermediate nodes on the same path is merged and summarized into a single total value. The total value obtained by this summary is the path cumulative anomaly amount of the anomaly propagation path.

[0064] The collaborative matching rules are weight allocation criteria derived from historical leakage cases in the preset rule base. The rules define different weight ratios for abnormal signals along different paths for different decomposition scales and carrier types. The rules read the already matched collaborative matching rules and, according to the ratio requirements recorded in the rules, assign corresponding weight coefficients to the accumulated abnormal amount of each abnormal propagation path. The weight coefficients are used to correct the attenuation effect of differences in pipeline length and diameter on abnormal signals. The accumulated abnormal amount of the path and the assigned weight coefficients are calculated and corrected accordingly to eliminate numerical deviations caused by differences in pipeline hardware structure. After processing, the weighted abnormal amount of the abnormal propagation path is obtained.

[0065] The weighted anomaly values ​​of all anomaly propagation paths are collected after weight correction. The sum of all weighted anomalies is calculated. The weighted anomaly value of a single anomaly propagation path is divided by the sum of all weighted anomalies of all paths to complete the normalization comparison process, eliminate the numerical interference caused by the fluctuation of the total anomaly value of the pipeline network, and calculate the propagation contribution of each anomaly propagation path.

[0066] Referring to the physical arrangement of the actual pipeline network topology, without disrupting the original topological logical relationship between each path, the propagation contribution calculated for each abnormal propagation path is arranged sequentially to form an ordered set of values, thus obtaining the propagation contribution sequence of the abnormal coupling component on the topology path.

[0067] The beneficial effects are as follows: by generating anomaly identifier sequences through encoding mapping to complete the retrieval and invocation of collaborative matching rules, and parsing the topological adjacency matrix by combining the actual pipeline connection information of the pipeline network, the reverse traversal of the anomaly propagation path is completed by using the monitoring node corresponding to the anomaly carrier as the traversal starting point. The cumulative anomaly quantity statistics of the path are completed in sequence, the weight allocation based on collaborative matching rules is completed, and the weighted anomaly quantity normalization conversion is completed to generate a complete propagation contribution sequence. The topological position corresponding to the maximum propagation contribution is selected as the leakage judgment point. The reverse tracing analysis is carried out by fully combining the physical topology of the pipeline network and the propagation characteristics of leakage disturbance, reducing the computational interference caused by non-faulty paths, improving the degree of fit between the leakage location judgment result and the actual fault point of the pipeline network, and realizing the accurate location of water supply network leakage faults based on multi-carrier coupling characteristics.

[0068] like Figure 2 The diagram shown is a functional block diagram of a cloud-based multi-carrier leakage data analysis system provided in an embodiment of the present invention.

[0069] The cloud-based fusion-based multi-carrier leak data analysis system 100 described in this invention can be installed in electronic devices. Depending on the functions implemented, the cloud-based fusion-based multi-carrier leak data analysis system 100 may include a data standardization preprocessing module 101, a multi-scale feature decomposition module 102, a coupled feature hierarchical aggregation module 103, an abnormal coupled component decoupling and identification module 104, and a reverse propagation leak location module 105. The modules described in this invention can also be referred to as units, which are a series of computer program segments that can be executed by the processor of an electronic device and perform a fixed function, stored in the memory of the electronic device.

[0070] In this embodiment, the functions of each module / unit are as follows: The data standardization preprocessing module 101 is used to perform format standardization processing on the multi-source operation data of the water supply network on different carriers to obtain the multi-carrier data of the water supply network. The multi-scale feature decomposition module 102 is used to perform multi-scale decomposition on the multi-carrier data according to the coupling strength of the multi-carrier data to obtain the multi-scale feature components of the multi-carrier data. The hierarchical aggregation module 103 is used to hierarchically aggregate the coupling features of the multi-scale feature components to obtain the multi-scale cross-carrier coupling features of the water supply network. The abnormal coupling component decoupling identification module 104 is used to construct a benchmark coupling measurement sequence of the water supply network based on the historical normal operating data of the water supply network, compare the deviation of the multi-scale cross-carrier coupling features with the benchmark coupling measurement sequence, so as to identify the decomposition scale and corresponding abnormal carrier of the water supply network where the features deviate, and decouple the abnormal coupling components from the abnormal carrier. The reverse propagation leakage location module 105 is used to match the multi-scale cross-carrier coupling feature cooperative matching rule according to the decomposition scale and the corresponding abnormal carrier, and to perform reverse tracing analysis on the abnormal coupling component according to the topological connection relationship of the water supply network and the cooperative matching rule to obtain the leakage location determination result of the water supply network.

[0071] In the several embodiments provided by this invention, it should be understood that the disclosed methods and systems can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0072] The modules described as separate components may or may not be physically separate. The components shown as modules 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 modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0073] Furthermore, the functional modules in the various embodiments of the present invention 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 in the form of hardware plus software functional modules.

[0074] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0075] This application embodiment can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0076] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A cloud-based, multi-carrier leakage data analysis method, characterized in that, The method includes: D1. Standardize the format of the multi-source operation data of the water supply network on different carriers to obtain the multi-carrier data of the water supply network. D2. Based on the coupling strength of the multi-carrier data, perform multi-scale decomposition on the multi-carrier data to obtain the multi-scale feature components of the multi-carrier data; D3. Perform hierarchical aggregation of the coupling features of the multi-scale feature components to obtain the multi-scale cross-carrier coupling features of the water supply network. D4. Based on the historical normal operating data of the water supply network, construct the benchmark coupling measurement sequence of the water supply network, compare the deviation of the multi-scale cross-carrier coupling characteristics with the benchmark coupling measurement sequence to identify the decomposition scale and corresponding abnormal carriers of the water supply network that have deviated from the characteristics, and decouple the abnormal coupling components from the abnormal carriers. D5. Based on the decomposition scale and the corresponding abnormal carrier, match the collaborative matching rules of the multi-scale cross-carrier coupling features, and perform reverse tracing analysis on the abnormal coupling components according to the topological connection relationship of the water supply network and the collaborative matching rules to obtain the leakage location determination result of the water supply network.

2. The multi-carrier leakage data analysis method based on cloud fusion as described in claim 1, characterized in that, The process of standardizing the format of multi-source operational data of the water supply network yields multi-carrier data of the water supply network, including: Acquire multi-source operational data of the water supply network, including network pressure data, flow data, and acoustic data; Data cleaning is performed on the multi-source operational data to obtain standardized multi-carrier data of the water supply network; The standardized multi-carrier data is time-aligned, and carrier data with different sampling frequencies are mapped to a unified time reference to obtain the multi-carrier data of the water supply network.

3. The multi-carrier leakage data analysis method based on cloud fusion as described in claim 1, characterized in that, The step of performing multi-scale decomposition on the multi-carrier data based on the coupling strength of the multi-carrier data to obtain the multi-scale feature components of the multi-carrier data includes: Cross-correlation analysis is performed on the multi-carrier data to obtain the coupling correlation strength between the multi-carrier data; Based on the coupling strength, the optimal decomposition scale parameter of the multi-carrier data is determined. When the coupling strength is higher than a preset strength threshold, a fine decomposition scale is selected, and when the coupling strength is lower than the preset strength threshold, a coarse decomposition scale is selected. Based on the optimal decomposition scale parameter, the multi-carrier data is decomposed into multi-scale components to obtain the multi-scale feature components of the water supply network. The multi-scale feature components include coarse-scale components that characterize the global operating trend of the network and fine-scale components that characterize local abnormal disturbances.

4. The multi-carrier leakage data analysis method based on cloud fusion as described in claim 3, characterized in that, The hierarchical aggregation of the coupling features of the multi-scale feature components to obtain the multi-scale cross-carrier coupling features of the water supply network includes: The coupling correlation features between pressure data, flow data and acoustic data in the coarse-scale component and the fine-scale component are extracted to obtain the coarse-scale carrier coupling features and fine-scale carrier coupling features of the water supply network. Based on the scale hierarchy relationship between the coarse-scale carrier coupling characteristics and the fine-scale carrier coupling characteristics, the aggregate hierarchical structure of the water supply network is obtained; Based on the aforementioned aggregation hierarchy, the fine-scale carrier coupling features are first aggregated, and then the aggregation result is combined and fused with the coarse-scale carrier coupling features across scales to obtain the aggregated coupling features of the water supply network. The aggregation coupling features are fused across levels to obtain the multi-scale cross-carrier coupling features of the water supply network.

5. The multi-carrier leakage data analysis method based on cloud fusion as described in claim 1, characterized in that, The process involves constructing a baseline coupling metric sequence for the water supply network based on historical normal operating data, comparing the multi-scale cross-carrier coupling characteristics with the baseline coupling metric sequence to identify the decomposition scale and corresponding anomalous carriers where the water supply network exhibits characteristic deviations, and decoupling anomalous coupling components from the anomalous carriers. This includes: The historical normal operating condition data of the water supply network is obtained, and the historical normal operating condition data is divided into time-series slices to obtain the normal operating condition time-series segments of the historical normal operating condition data. Extract the normal coupling metric values ​​of the normal operating condition time sequence segment at different decomposition scales to obtain the benchmark coupling metric sequence of the water supply network at different decomposition scales; The coupling metric value in the multi-scale cross-carrier coupling feature is compared with the normal coupling metric value in the benchmark coupling metric sequence to obtain the coupling deviation degree of the water supply network. Based on the coupling deviation, the target decomposition scale of the water supply network is determined, and the carrier corresponding to the target decomposition scale is identified as an abnormal carrier. The abnormal coupling component is separated from the abnormal carrier.

6. The multi-carrier leakage data analysis method based on cloud fusion as described in claim 5, characterized in that, The step of separating the anomalous coupling component from the anomalous carrier includes: Obtain the abnormal coupling metric values ​​of the abnormal carrier at different decomposition scales; The abnormal coupling metric value is compared with the normal coupling metric value at the corresponding scale in the benchmark coupling metric sequence to obtain the coupling deviation ratio of the abnormal carrier. Arrange the coupling deviation ratios to obtain the cross-scale deviation ratio matrix of the abnormal carrier; The off-diagonal elements of the cross-scale deviation ratio matrix are extracted, and the extracted elements are summed to obtain the total scale interference under different decomposition scales. The total scale interference is normalized to obtain the inter-scale interference weight coefficients at different decomposition scales. The abnormal coupling component of the water supply network is calculated using the coupling deviation ratio, the inter-scale interference weight coefficient, the abnormal coupling metric, and the normal coupling metric.

7. The multi-carrier leakage data analysis method based on cloud fusion as described in claim 6, characterized in that, The calculation formula for the abnormal coupling component is as follows: ； in, Indicates the target decomposition scale index. This represents the summation index for traversing all decomposition scales. This represents the total number of decomposition scales. Indicates that the abnormal carrier is in the first Anomaly coupling metrics at each decomposition scale Indicating the first in the reference coupling metric sequence Normal coupling metric at each decomposition scale Indicates that the abnormal carrier is in the first Coupling deviation ratio at each decomposition scale Indicates the first Inter-scale interference weighting coefficients for each decomposition scale Indicates that the abnormal carrier is in the first Anomalous coupling components at each decomposition scale.

8. The multi-carrier leakage data analysis method based on cloud fusion as described in claim 1, characterized in that, The step of matching the multi-scale cross-carrier coupling features according to the decomposition scale and corresponding abnormal carrier, and performing reverse tracing analysis on the abnormal coupling components according to the topological connection relationship of the water supply network and the collaborative matching rules to obtain the leakage location determination result of the water supply network includes: The decomposition scale and the corresponding anomaly carrier are encoded and mapped to obtain the anomaly identifier sequence of the anomaly carrier under different decomposition scales; Based on the anomaly identifier sequence, a collaborative matching rule for the multi-scale cross-carrier coupling feature is matched from a preset rule base; Adjacency analysis is performed on the topological connection relationship of the water supply network to obtain the topological adjacency matrix between multiple carriers in the water supply network; Based on the topological adjacency matrix and the cooperative matching rule, path backtracking analysis is performed on the abnormal coupling components to obtain the propagation contribution sequence of the abnormal coupling components on the topological path; The propagation contribution sequence is sorted and compared, and the topological position corresponding to the largest propagation contribution in the sorting result is determined as the leakage location determination result of the water supply network.

9. The multi-carrier leakage data analysis method based on cloud fusion as described in claim 1, characterized in that, The step of performing path backtracking analysis on the anomalous coupling components based on the topological adjacency matrix and the cooperative matching rules to obtain the propagation contribution sequence of the anomalous coupling components on the topological path includes: Starting from the pipeline monitoring node corresponding to the abnormal carrier, the abnormal propagation path of the abnormal carrier is obtained by traversing in reverse along the adjacency relationship in the topological adjacency matrix. The abnormal coupling components of the intermediate nodes on the abnormal propagation path are accumulated to obtain the path cumulative abnormality of the abnormal propagation path. According to the collaborative matching rules, the cumulative anomaly amount of the path is weighted and assigned to obtain the weighted anomaly amount of the anomaly propagation path. The weighted outliers are normalized and compared to obtain the propagation contribution of the outlier propagation path. The propagation contribution values ​​are arranged in path order to obtain the propagation contribution value sequence of the abnormal coupling component on the topological path.

10. A cloud-based, multi-carrier leakage data analysis system, characterized in that, The system for implementing the cloud-based multi-carrier leakage data analysis method as described in claim 1 includes: The data standardization preprocessing module is used to standardize the format of multi-source operational data of the water supply network on different carriers to obtain multi-carrier data of the water supply network. The multi-scale feature decomposition module is used to perform multi-scale decomposition on the multi-carrier data according to the coupling strength of the multi-carrier data to obtain the multi-scale feature components of the multi-carrier data. The coupling feature hierarchical aggregation module is used to hierarchically aggregate the coupling features of the multi-scale feature components to obtain the multi-scale cross-carrier coupling features of the water supply network. The abnormal coupling component decoupling identification module is used to construct a benchmark coupling metric sequence of the water supply network based on historical normal operating data of the water supply network, compare the deviation of the multi-scale cross-carrier coupling features with the benchmark coupling metric sequence, identify the decomposition scale and corresponding abnormal carrier of the water supply network where the features deviate, and decouple the abnormal coupling components from the abnormal carrier. The reverse propagation leakage location module is used to match the multi-scale cross-carrier coupling features with the decomposition scale and corresponding abnormal carrier according to the collaborative matching rules, and to perform reverse tracing analysis on the abnormal coupling components according to the topological connection relationship of the water supply network and the collaborative matching rules, so as to obtain the leakage location determination result of the water supply network.

Images