Water outfall positioning method and system for pipe network pollution coupling inversion based on conductivity

Abstract

The application relates to the technical field of external water positioning, and discloses an external water positioning method and system based on conductivity for pipe network pollution coupling inversion, which comprises the following steps: firstly, a pipe network hydraulic-conductivity coupling model is established, and a node-monitoring point conductivity response kernel function matrix is calculated; an abnormal conductivity period is identified, an abnormal amplitude is calculated, an observed abnormal conductivity vector is generated, and candidate external water infiltration nodes are screened out; a reduced response kernel function matrix is constructed, a regularization inversion equation is constructed in combination with the observed abnormal conductivity vector, and is solved to obtain an external water intensity inversion vector to determine the external water infiltration nodes; and an external water positioning result report is generated based on the reduced response kernel function matrix and the external water intensity inversion vector; the application can improve the accuracy of external water positioning based on conductivity for pipe network pollution coupling inversion.

Description

Technical Field

[0001] This invention relates to the field of external water location technology, and in particular to an external water location method and system based on conductivity-based pipeline pollution coupling inversion. Background Technology

[0002] Locating external water infiltration in pipe networks is a key technology for pipe network operation and maintenance and pollution control. Existing technologies mostly identify external water infiltration through conductivity monitoring, but a precise hydraulic-conductivity coupling model for pipe networks has not yet been constructed. They simply analyze conductivity data and hydraulic parameters independently without considering the convective dispersion transport law of conductivity in the pipe network, nor do they generate a node-monitoring point conductivity response kernel function matrix. This makes it impossible to quantify the influence relationship between infiltration at different nodes on the conductivity of monitoring points. As a result, the correlation analysis between abnormal conductivity signals and infiltration nodes lacks theoretical support, and the physical correlation and accuracy of the location results are low.

[0003] Existing methods for locating external water infiltration nodes mostly employ a single inversion or screening strategy, lacking a multi-level fusion inversion and location framework. They determine candidate nodes solely through simple screening of abnormal time periods, failing to combine regularized inversion to achieve accurate solutions for infiltration intensity. Furthermore, they do not perform multi-source decoupling and location reliability assessment on the location nodes, easily misclassifying coupled sources with similar responses as multiple independent infiltration sources. Moreover, they cannot quantify the reliability of the location results, exhibiting problems such as an excessively large range of candidate nodes, unsound inversion solutions, and a lack of uncertainty assessment for location results. These methods are insufficient to meet the needs of accurately locating external water infiltration nodes in practical engineering. Therefore, improving the accuracy of external water infiltration node location has become an urgent problem to be solved. Summary of the Invention

[0004] This invention provides a method and system for locating external water sources based on conductivity-coupled inversion of pipeline pollution, in order to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides an external water location method based on conductivity-coupled inversion for pipeline pollution, comprising:

[0006] S1. Based on the pipeline network topology, pipe segment attributes and monitoring point locations, a hydraulic-conductivity coupling model of the pipeline network is established, and the node-monitoring point conductivity response kernel function matrix is ​​calculated.

[0007] S2, based on the conductivity time series data of each monitoring point, identify the abnormal conductivity periods and calculate the abnormal amplitude, and generate the observed abnormal conductivity vector;

[0008] S3. Based on the abnormal conductivity period and the hydraulic parameters of the pipeline network, candidate external water infiltration nodes are selected. Based on the candidate external water infiltration nodes, parameters are extracted from the node-monitoring point conductivity response kernel function matrix to construct a reduced response kernel function matrix.

[0009] S4. Based on the reduced response kernel function matrix and the observed abnormal conductivity vector, a regularized inversion equation is constructed and solved to obtain the final external water intensity inversion vector. The external water infiltration node is determined according to the final external water intensity inversion vector.

[0010] S5. Based on the reduced response kernel function matrix and the final external water intensity inversion vector, the external water infiltration node is decoupled from multiple sources and the location confidence is calculated to generate an external water location result report.

[0011] In a preferred embodiment, the step of establishing a hydraulic-conductivity coupling model of the pipeline network based on the pipeline network topology, pipe segment attributes, and monitoring point locations, and calculating the node-monitoring point conductivity response kernel function matrix, includes:

[0012] Acquire pipeline network topology data, pipe segment attribute data, and monitoring point location data;

[0013] Based on the pipeline network topology data, a directed graph model of the pipeline network is established, and steady-state hydraulic calculations are performed on the directed graph model of the pipeline network based on the pipe segment attribute data to obtain the flow rate and velocity of each pipe segment.

[0014] Based on the flow rate and velocity of each pipe segment, a one-dimensional convection-dispersion equation describing the transport process of conductivity in the pipe network is established.

[0015] By traversing all potential infiltration nodes in the pipeline network, and based on the one-dimensional convection dispersion equation, the steady-state conductivity response values ​​generated at all monitoring points when external water with a unit conductivity increment is injected at each potential node are simulated in a forward manner, so as to generate the node-monitoring point conductivity response kernel function matrix.

[0016] In a preferred embodiment, the mathematical expression of the one-dimensional convection dispersion equation is as follows:

[0017] ;

[0018] In the formula, C represents the increase in conductivity in the pipe network water caused by the infiltration of external water. The coordinates are continuous time coordinates, where x represents the axial distance along the pipe segment, and v represents the distance along the pipe segment. e D is the water flow velocity in the pipe section. σ This represents the effective diffusion coefficient within the pipe section.

[0019] In a preferred embodiment, the step of identifying abnormal conductivity periods and calculating the abnormal amplitude based on the conductivity time-series data of each monitoring point, and generating an observed abnormal conductivity vector, includes:

[0020] Based on the obtained conductivity time series data of each monitoring point, the local statistical characteristics of each monitoring point are calculated to obtain the local mean and local standard deviation of each time point.

[0021] Based on the local mean and the local standard deviation, the abnormal detection index of each monitoring point at each time point is calculated and compared with the preset abnormal threshold to identify abnormal moments and generate a set of abnormal conductivity time periods for each monitoring point.

[0022] Based on the conductivity time series data and local mean within the set of abnormal conductivity periods, the conductivity abnormality amplitude within each abnormal period is calculated, and the abnormal amplitudes of all monitoring points are combined into an observed abnormal conductivity vector.

[0023] In a preferred embodiment, the step of screening candidate external water infiltration nodes based on the abnormal conductivity period and the hydraulic parameters of the pipe network includes:

[0024] Based on the set of abnormal conductivity periods at each monitoring point, calculate the time difference of abnormal signals between any two monitoring points.

[0025] Based on the pipeline network topology and the flow velocity of each pipe segment, calculate the hydraulic propagation time between any two monitoring points;

[0026] Based on the time difference of the abnormal signal and the hydraulic propagation time, the monitoring points are screened to obtain monitoring point pairs that meet the propagation consistency condition;

[0027] Based on the monitoring point pairs that meet the propagation consistency condition and the pipeline topology, upstream nodes are traced along the counter-current direction, and all traced upstream nodes are merged to generate a candidate set of external water infiltration nodes.

[0028] In a preferred embodiment, the step of screening monitoring points based on the time difference of the abnormal signal and the hydraulic propagation time to obtain pairs of monitoring points that meet the propagation consistency condition includes:

[0029] Based on the time difference of the abnormal signal and the hydraulic propagation time, the propagation consistency index of the abnormal signal between the two monitoring points is calculated.

[0030] The propagation consistency index is compared with a preset consistency threshold to select monitoring point pairs that meet the propagation consistency conditions.

[0031] In a preferred embodiment, the step of extracting parameters from the node-monitoring point conductivity response kernel function matrix based on the candidate external water infiltration nodes to construct a reduced response kernel function matrix includes:

[0032] Extract the corresponding row of the candidate external water infiltration node from the node-monitoring point conductivity response kernel function matrix to generate a reduced response kernel function matrix.

[0033] In a preferred embodiment, the step of constructing and solving a regularized inversion equation based on the reduced response kernel function matrix and the observed abnormal conductivity vector to obtain the final external water intensity inversion vector, and determining the external water infiltration node based on the final external water intensity inversion vector, includes:

[0034] Based on the reduced response kernel function matrix and the observed abnormal conductivity vector, a regularization term is introduced to construct a regularized inversion equation describing the linear relationship between the monitored anomaly and the node infiltration intensity.

[0035] The regularized inversion equation is solved to derive the corresponding normal equation, and the external water intensity inversion vector related to the regularization parameter is obtained by solving the normal equation.

[0036] The optimal regularization parameter is selected from a set of candidate regularization parameters to minimize the generalized cross-validation index. The normal equation is then resolved based on the optimal regularization parameter to obtain the final external water intensity inversion vector.

[0037] Based on the final external water intensity inversion vector, the normalized contribution rate of each candidate node is calculated, and the external water infiltration node is determined according to the preset contribution rate threshold.

[0038] In a preferred embodiment, the process of decoupling multiple sources and calculating location confidence for the external water infiltration nodes based on the reduced response kernel function matrix and the final external water intensity inversion vector to generate an external water location result report includes:

[0039] Based on the reduced response kernel function matrix and the final external water intensity inversion vector, the response similarity between any two external water infiltration location nodes is calculated.

[0040] Based on the response similarity and the preset similarity threshold, the location nodes are decoupled from multiple sources, and nodes with highly similar responses are merged into an equivalent external water source.

[0041] Based on the reduced response kernel function matrix, the external water intensity inversion vector, and the results of multi-source decoupling, the location confidence index of each final location node is calculated.

[0042] Based on the location reliability index, the location results are classified into levels, and combined with the spatial location of each node, infiltration intensity, normalized contribution rate and multi-source decoupling information, an external water location result report is generated.

[0043] To address the aforementioned problems, this invention also provides an external water location system based on conductivity-coupled inversion of pipeline pollution, the system comprising:

[0044] Conductivity coupling modeling is used to establish a hydraulic-conductivity coupling model of a pipe network based on the pipe network topology, pipe segment properties and monitoring point locations, and generate a node-monitoring point conductivity response kernel function matrix.

[0045] The anomaly identification module is used to identify abnormal conductivity periods and calculate the abnormal amplitude based on the conductivity time series data of each monitoring point, and generate an observed abnormal conductivity vector.

[0046] The regional preliminary screening module is used to calculate the consistency of abnormal signal propagation based on the abnormal conductivity period and the hydraulic parameters of the pipeline network, screen out candidate external water infiltration nodes, and extract the corresponding row from the node-monitoring point conductivity response kernel function matrix to generate a reduced response kernel function matrix.

[0047] The intensity inversion module is used to construct and solve a regularized inversion equation based on the reduced response kernel function matrix and the observed abnormal conductivity vector to obtain the final external water intensity inversion vector, and to determine the external water infiltration node based on the final external water intensity inversion vector.

[0048] The decoupling evaluation module is used to perform multi-source decoupling on multiple external water infiltration nodes and calculate the location confidence based on the reduced response kernel function matrix and the final external water intensity inversion vector, so as to generate an external water location result report.

[0049] Compared with the prior art, the present invention has the following beneficial effects:

[0050] 1. This invention constructs a hydraulic-conductivity coupling model of a pipe network and generates a kernel function matrix of the conductivity response of nodes and monitoring points. This achieves precise quantitative correlation between external water infiltration nodes and conductivity anomalies at monitoring points from a physical mechanism perspective, significantly improving the theoretical scientific rigor and data support for location analysis. The model combines pipe network topology and pipe segment attributes to complete steady-state hydraulic calculations. It accurately describes the convective dispersion transport law of conductivity in the pipe network through a one-dimensional convection dispersion equation. Then, through forward simulation, it obtains the conductivity response values ​​of each monitoring point under unit infiltration intensity. The resulting kernel function matrix can clearly quantify the degree of influence of different node infiltration on each monitoring point, breaking the limitation of independent analysis of hydraulic parameters and conductivity data in traditional methods. This lays a precise model foundation that fits the actual pipe network transport characteristics for subsequent inversion location analysis, effectively avoiding location deviations caused by a lack of physical correlation.

[0051] 2. The multi-level fusion inversion positioning framework designed in this invention, which combines initial screening, regularized inversion, multi-source decoupling, and location reliability assessment, achieves full-process optimization and precision in locating external water infiltration. It addresses the pain points of traditional positioning methods, such as their simplistic approach and lack of verification and evaluation. Initial screening of candidate nodes is completed through anomaly signal propagation consistency analysis, significantly narrowing the inversion range and improving computational efficiency. The introduction of regularized inversion, combined with generalized cross-validation to select optimal parameters, effectively solves the ill-posed nature of the inversion problem, avoids severe oscillations in the solution, and improves the robustness and accuracy of the external water intensity inversion results. Multi-source decoupling and the identification of highly similar equivalent external water sources prevent misclassification of the same infiltration source as multiple independent nodes, restoring the true situation of the infiltration source. Finally, a location reliability index is constructed based on the proportion of fused signal interpretation and the distinguishability of nodes, classifying the positioning results and quantifying uncertainty. This provides a graded reference for engineering investigation, ensuring that the positioning results are not only accurate but also possess clear reliability assessment and practical engineering guidance value. Attached Figure Description

[0052] Figure 1 This is a schematic flowchart of an external water location method based on conductivity-coupled inversion for pipeline pollution, provided in an embodiment of the present invention.

[0053] Figure 2 This is a functional block diagram of an external water location system based on conductivity-coupled inversion of pipeline pollution, provided in an embodiment of the present invention.

[0054] 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

[0055] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0056] This application provides a method for locating external water sources based on conductivity-based pipeline pollution coupling inversion. The execution entity of this 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: a server, a terminal, etc. In other words, the method for locating external water sources based on conductivity-based pipeline pollution coupling inversion 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 providing 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 (CDN), and big data and artificial intelligence platforms.

[0057] Reference Figure 1 The diagram shown is a flowchart illustrating an embodiment of the external water location method based on conductivity-based pipe network pollution coupling inversion provided by the present invention. In this embodiment, the external water location method based on conductivity-based pipe network pollution coupling inversion includes:

[0058] S1. Based on the pipeline network topology, pipe segment attributes and monitoring point locations, a hydraulic-conductivity coupling model of the pipeline network is established, and the node-monitoring point conductivity response kernel function matrix is ​​calculated.

[0059] In this embodiment of the invention, the step of establishing a hydraulic-conductivity coupling model of the pipeline network based on the pipeline network topology, pipe segment attributes, and monitoring point locations, and calculating the node-monitoring point conductivity response kernel function matrix, includes:

[0060] Acquire pipeline network topology data, pipe segment attribute data, and monitoring point location data;

[0061] Based on the pipeline network topology data, a directed graph model of the pipeline network is established, and steady-state hydraulic calculations are performed on the directed graph model of the pipeline network based on the pipe segment attribute data to obtain the flow rate and velocity of each pipe segment.

[0062] Based on the flow rate and velocity of each pipe segment, a one-dimensional convection-dispersion equation describing the transport process of conductivity in the pipe network is established.

[0063] By traversing all potential infiltration nodes in the pipeline network, and based on the one-dimensional convection dispersion equation, the steady-state conductivity response values ​​generated at all monitoring points when external water with a unit conductivity increment is injected at each potential node are simulated in a forward manner, so as to generate the node-monitoring point conductivity response kernel function matrix.

[0064] It should be noted that establishing a directed graph model of a pipeline network involves abstracting the pipeline network into a directed graph in graph theory. The intersections and connection points in the pipeline network are abstracted as graph nodes, and the pipes are abstracted as directed edges connecting the nodes. This model is used to describe the physical connection relationships and fluid flow direction of the pipeline network.

[0065] It should be noted that steady-state hydraulic calculation is based on the pipe network topology and pipe segment properties, and solves the steady-state equations in the pipe network that satisfy the conservation of mass and energy in order to determine the flow rate and flow direction in each pipe segment, as well as the pressure at each node. The pipe segment property data includes at least pipe length, pipe diameter and pipe roughness coefficient.

[0066] For each node except the water source node, the total flow into that node is equal to the total flow out of that node. For each pipe segment, the relationship between its pressure drop and flow rate follows a specific hydraulic formula. By solving this set of equations using an iterative method, the flow rate of each pipe segment and the flow velocity calculated based on the flow rate and pipe diameter can be obtained.

[0067] Furthermore, the mathematical expressions for the steady-state hydraulic calculation equations are as follows:

[0068] ;

[0069] In the formula, I(i) represents the set of pipe segments ending at node i, O(i) represents the set of pipe segments starting at node i, and Q... e d represents the flow rate in the pipe section. i For the water demand of node i, L e D is the length of the pipe section. e Let A be the diameter of the pipe section. e Let f be the cross-sectional area of ​​the pipe section, g be the acceleration due to gravity, and f be the acceleration due to gravity. e Here, is the friction coefficient of the pipe segment, e is the pipe segment index, i is the node index, and h is the node index. e This is to reduce the pressure at both ends of the pipe section.

[0070] Furthermore, the flow rate in the pipe segment is positive when flowing into the node and negative when flowing out.

[0071] It should be noted that the forward simulation, which involves injecting external water with a unit conductivity increment at each candidate node, refers to coupling the network hydraulic model with the conductivity transport model for numerical simulation. Specifically, candidate node i is considered as a continuous conductivity source term with unit intensity, and its boundary conditions are substituted into the overall network transport model, which is composed of the one-dimensional convection dispersion equations of all pipe segments and the node hybrid model. The solution of this coupled model in steady state is then obtained. Given the solution under the given conditions, we obtain the conductivity response values ​​at all monitoring points j, denoted as K(i,j). The steady-state response value represents the steady-state increase in conductivity at monitoring point j caused by continuous injection of external water of unit intensity at node i.

[0072] Furthermore, the pipeline hydraulic model is a mathematical model based on the laws of conservation of mass and energy in fluid mechanics. It is a standard simulation tool in the fields of water supply and drainage engineering and hydraulics. The essence of the conductivity transport model is a physical equation with one-dimensional convection dispersion equation as its core.

[0073] It should be noted that the term "all potential infiltration nodes" refers to a predefined set of nodes that may be subject to external water infiltration, based on the pipeline network topology. This set typically excludes known constant water source nodes and monitoring points themselves, and includes all other intersections, connection points, and other nodes in the pipeline network.

[0074] It should be noted that the dimension of the generated conductivity response kernel function matrix is ​​N. all ×N m , where N all N represents the total number of potential infiltration nodes. m This represents the total number of monitoring points.

[0075] In this embodiment of the invention, the mathematical expression of the one-dimensional convection dispersion equation is as follows:

[0076] ;

[0077] In the formula, C represents the increase in conductivity in the pipe network water caused by the infiltration of external water. The coordinates are continuous time coordinates, where x represents the axial distance along the pipe segment, and v represents the distance along the pipe segment. e D is the water flow velocity in the pipe section. σ This represents the effective diffusion coefficient within the pipe section.

[0078] It should be noted that the one-dimensional convection dispersion equation is an equation that describes the transport and diffusion process of dissolved substances in flowing water in a pipe. This equation treats the pipe section as a one-dimensional channel and shows that the rate of change of conductivity is jointly determined by the time-varying term, the transport term due to convection with the water flow, and the dispersion term caused by the concentration gradient.

[0079] Furthermore, the effective diffusion coefficient within the pipe section reflects both molecular diffusion and turbulent diffusion, and its expression is usually related to the hydraulic characteristics of the pipe section. The corresponding mathematical expression for calculation is as follows:

[0080] ;

[0081] In the formula, D σ α is the effective diffusion coefficient within the pipe section. m The molecular diffusion coefficient is... Let α be the turbulent diffusion coefficient. t It is an empirical coefficient, v e It is the flow rate, D e 'e' represents the pipe segment diameter, and 'e' represents the pipe segment index.

[0082] S2, based on the conductivity time series data of each monitoring point, identify the abnormal conductivity periods and calculate the abnormal amplitude, and generate the observed abnormal conductivity vector;

[0083] In this embodiment of the invention, the step of identifying abnormal conductivity periods and calculating abnormal amplitudes based on conductivity time-series data from each monitoring point, and generating an observed abnormal conductivity vector, includes:

[0084] Based on the obtained conductivity time series data of each monitoring point, the local statistical characteristics of each monitoring point are calculated to obtain the local mean and local standard deviation of each time point.

[0085] Based on the local mean and the local standard deviation, the abnormal detection index of each monitoring point at each time point is calculated and compared with the preset abnormal threshold to identify abnormal moments and generate a set of abnormal conductivity time periods for each monitoring point.

[0086] Based on the conductivity time series data and local mean within the set of abnormal conductivity periods, the conductivity abnormality amplitude within each abnormal period is calculated, and the abnormal amplitudes of all monitoring points are combined into an observed abnormal conductivity vector.

[0087] It should be noted that obtaining conductivity time-series data refers to the sequence of conductivity measurements collected and stored every minute from online water quality sensors deployed at various monitoring points in the pipeline network.

[0088] It should be noted that calculating the local statistical characteristics of each monitoring point is to dynamically capture the slow changes and normal fluctuation range of the background conductivity value, thereby more accurately identifying anomalies caused by the infiltration of external water.

[0089] Define a time window centered at a given time point and with a length of 6 hours. Calculate the arithmetic mean of all conductivity values ​​within the window to obtain the local mean at that time point. Simultaneously, calculate the standard deviation of these values ​​to obtain the local standard deviation at that time point. Then, slide the window forward by one time step and repeat the calculation to generate the corresponding local mean and local standard deviation for each valid time point in the time series.

[0090] Furthermore, the local mean represents the “normal” or “background” level of conductivity at the monitoring point near time point t, reflecting the normal range of conductivity fluctuations around the background level near time point t.

[0091] It should be noted that the mathematical expression used to calculate the anomaly detection index is as follows:

[0092] ;

[0093] Z j (t) represents the anomaly detection index, C j (t) is the measured conductivity value at time point t, μ j (t) is the local background mean at the corresponding time point, σ j (t) is the local standard deviation of the fluctuation at the corresponding time point, where t represents the time point and j is the index of the monitoring point;

[0094] Absolute value operation The absolute distance between the measured value and the background level was calculated. This absolute distance was then divided by the local standard deviation to achieve standardization, making the anomaly detection index a dimensionless value that reflects the significance of the current deviation relative to historical normal fluctuations.

[0095] It should be noted that identifying abnormal moments involves comparing the abnormal detection index with a preset non-negative abnormal threshold. When the abnormal detection index is greater than the preset non-negative abnormal threshold, the moment is determined to be an abnormal moment. All consecutive abnormal moments at monitoring point j are merged to form multiple consecutive abnormal time periods. Each time period is represented by its start and end times. The set of these time periods is the conductivity abnormal time period set for that monitoring point. This conductivity abnormal time period set marks the time range during which the water quality at that monitoring point is significantly disturbed. The preset non-negative abnormal threshold is set to 3.

[0096] It should be noted that the calculation of conductivity anomaly amplitude is performed for each anomaly period in the set of conductivity anomaly periods. For a specific anomaly period, the average difference between the measured conductivity value and its corresponding local background mean value at all times within that period is calculated, and this average value is defined as the anomaly amplitude of that period. This amplitude quantifies the degree to which the average conductivity is higher (positive value) or lower (negative value, but external water infiltration usually leads to an increase) than the normal background level within that anomaly period, reflecting the intensity of the anomaly signal.

[0097] It should be noted that the observed anomalous conductivity vector is a key data structure for integrating and organizing the processing results. The construction process is as follows:

[0098] From all monitoring points, select the time period with the largest amplitude, extract the abnormal amplitude calculated for each monitoring point within that time period, and arrange these amplitude values ​​in order of monitoring point number to form an N-valued array. m The ×1 column vector is the observed anomalous conductivity vector. This vector uses a single number to represent the distribution pattern of the "net anomalous signal intensity" caused by a single external water infiltration event at all monitoring points, after background stripping. Here, N... m This represents the total number of monitoring points.

[0099] S3. Based on the abnormal conductivity period and the hydraulic parameters of the pipeline network, candidate external water infiltration nodes are selected. Based on the candidate external water infiltration nodes, parameters are extracted from the node-monitoring point conductivity response kernel function matrix to construct a reduced response kernel function matrix.

[0100] In this embodiment of the invention, the step of screening candidate external water infiltration nodes based on the abnormal conductivity period and the hydraulic parameters of the pipe network includes:

[0101] Based on the set of abnormal conductivity periods at each monitoring point, calculate the time difference of abnormal signals between any two monitoring points.

[0102] Based on the pipeline network topology and the flow velocity of each pipe segment, calculate the hydraulic propagation time between any two monitoring points;

[0103] Based on the time difference of the abnormal signal and the hydraulic propagation time, the monitoring points are screened to obtain monitoring point pairs that meet the propagation consistency condition;

[0104] Based on the monitoring point pairs that meet the propagation consistency condition and the pipeline topology, upstream nodes are traced along the counter-current direction, and all traced upstream nodes are merged to generate a candidate set of external water infiltration nodes.

[0105] It should be noted that the calculation of the abnormal signal time difference is performed for any two monitoring points. The mathematical expression for the formula to calculate the abnormal signal time difference is as follows: (The formula is extracted from the sets of abnormal conductivity periods for each of the two monitoring points.)

[0106] ;

[0107] Where, Δt j1，j2 The time difference of the abnormal signal is represented by t, where t represents the time point, the median function represents taking the median, j1 and j2 represent any two monitoring points, and T... j1 T j2 This represents the set of periods of abnormal conductivity at each of the two monitoring points.

[0108] Furthermore, the abnormal signal time difference represents the time offset between the observation of the abnormal signal from monitoring point j2 to monitoring point j1. If the abnormal signal time difference is less than zero, it means that the median time of the abnormal signal at monitoring point j1 is later than that at monitoring point j2; otherwise, it is earlier.

[0109] It should be noted that the calculation of hydraulic propagation time is based on the directed graph model of the pipe network and the flow velocity of each pipe segment. For a monitoring point pair (j1, j2), the shortest hydraulic path from j2 to j1 is found along the flow direction in the pipe network topology. The hydraulic propagation time is obtained by summing the propagation times of all pipe segments along this path. The calculation formula is as follows:

[0110] ;

[0111] In the formula, Indicates the hydraulic propagation time, e is the pipe segment index, l e v is the length of the pipe section. e It's speed. This indicates the time required for water to flow through this pipe section. This represents the shortest hydraulic connectivity path from node j2 to node j1 that follows the direction of water flow in the pipe segment.

[0112] It should be noted that the process of tracing upstream nodes to generate a candidate set is performed on each selected pair of monitoring points with consistent propagation. Assuming the time difference of the abnormal signal is greater than zero, starting from the monitoring point j1 with the later signal, all upstream connected nodes are traced along the reverse flow direction of the directed graph of the pipeline network until a node that is also located in the upstream path of monitoring point j2 is encountered. This node is called the "common bifurcation node". All nodes along the reverse flow path from j1 to this "common bifurcation node", including the bifurcation node itself and the nodes connected to it that continue to be traced upstream from the bifurcation node, are included in a candidate region. This process is performed on all selected pairs of monitoring points, and the nodes in all the candidate regions are merged and deduplicated to finally form a candidate set of external water infiltration nodes.

[0113] In this embodiment of the invention, the step of screening monitoring points based on the time difference of the abnormal signal and the hydraulic propagation time to obtain monitoring point pairs that meet the propagation consistency condition includes:

[0114] Based on the time difference of the abnormal signal and the hydraulic propagation time, the propagation consistency index of the abnormal signal between the two monitoring points is calculated.

[0115] The propagation consistency index is compared with a preset consistency threshold to select monitoring point pairs that meet the propagation consistency conditions.

[0116] It should be noted that the propagation consistency index is calculated to quantify the degree of matching between the observed time difference of the anomalous signal and the theoretical hydraulic propagation time. The mathematical expression for calculating the propagation consistency index is as follows:

[0117] ;

[0118] In the formula, γ j1，j2 As a propagation consistency indicator, Δt j1，j2 For abnormal signal time difference, This is the theoretical hydraulic propagation time.

[0119] Furthermore, the closer the propagation consistency index is to 1, the more it matches the time difference of the observed signal with the hydraulic propagation time, and the higher the probability that the signal comes from the upstream of the same hydraulic path; conversely, the closer it is to 0, the less it matches.

[0120] It should be noted that the process of screening monitoring point pairs is as follows: a preset consistency threshold is set, with a value of 0.7; when the propagation consistency index of monitoring point pair (j1, j2) is greater than the preset consistency threshold, the point pair is determined to be a propagation consistent point pair; the set of screened point pairs represents the abnormal signals observed by these monitoring points, which are highly consistent with the hydraulic transport theory of the pipeline network in terms of time sequence, so they are likely to be affected by the sequence of the same upstream infiltration source.

[0121] In this embodiment of the invention, the step of extracting parameters from the node-monitoring point conductivity response kernel function matrix based on the candidate external water infiltration nodes and constructing a reduced response kernel function matrix includes:

[0122] Extract the corresponding row of the candidate external water infiltration node from the node-monitoring point conductivity response kernel function matrix to generate a reduced response kernel function matrix.

[0123] S4. Based on the reduced response kernel function matrix and the observed abnormal conductivity vector, a regularized inversion equation is constructed and solved to obtain the final external water intensity inversion vector. The external water infiltration node is determined according to the final external water intensity inversion vector.

[0124] In this embodiment of the invention, the step of constructing and solving a regularized inversion equation based on the reduced response kernel function matrix and the observed abnormal conductivity vector to obtain the final external water intensity inversion vector, and determining the external water infiltration node based on the final external water intensity inversion vector, includes:

[0125] Based on the reduced response kernel function matrix and the observed abnormal conductivity vector, a regularization term is introduced to construct a regularized inversion equation describing the linear relationship between the monitored anomaly and the node infiltration intensity.

[0126] The regularized inversion equation is solved to derive the corresponding normal equation, and the external water intensity inversion vector related to the regularization parameter is obtained by solving the normal equation.

[0127] The optimal regularization parameter is selected from a set of candidate regularization parameters to minimize the generalized cross-validation index. The normal equation is then resolved based on the optimal regularization parameter to obtain the final external water intensity inversion vector.

[0128] Based on the final external water intensity inversion vector, the normalized contribution rate of each candidate node is calculated, and the external water infiltration node is determined according to the preset contribution rate threshold.

[0129] It should be noted that the mathematical expression of the regularized inversion equation is as follows:

[0130] ;

[0131] Where, ΔC obs It is of dimension N m A vector of observed abnormal conductivity, N × 1, where each element represents the anomalous amplitude at the corresponding monitoring point. m K represents the total number of monitoring points. cand It is a dimension of The reduced response kernel function matrix, whose elements K cand (i, j) represents the steady-state conductivity increment caused by unit intensity infiltration at candidate node i at monitoring point j, and S is the external water intensity vector to be determined. Let J(S) be the model error vector, and J(S) be the objective function to be minimized. The first term... The term called the residual is the squared L2 norm of the difference between the model and the observed data. This is called the Tikhonov regularization term, where λ is the regularization parameter and L is the regularization matrix, using a first-order difference matrix. This indicates the imposition of a spatial smoothness constraint, which requires that the infiltration intensity obtained from the inversion be spatially gradual, avoiding physically unreasonable and violent oscillations.

[0132] It should be noted that the optimization solution for the regularization objective function is obtained by taking its gradient with respect to a vector and setting the gradient to zero, thereby deriving the corresponding normal equation. The mathematical expression of this regularization equation is as follows:

[0133] ;

[0134] In the formula, S*(λ) is the external water intensity inversion vector, K cand It is a dimension of The reduced response kernel function matrix, N m Let λ be the total number of monitoring points, λ be the regularization parameter, L be the regularization matrix, and ΔC be the regularization value. obs It is of dimension N m The observed anomalous conductivity vector is ×1.

[0135] It should be noted that the selection of the optimal regularization parameter based on generalized cross-validation is to avoid the subjectivity of manual selection. The core idea is that a good regularization parameter should minimize the error of the model when predicting "data not used for fitting." The mathematical expression used for calculation is as follows:

[0136] ;

[0137] In the formula, GCV(λ) is the GCV index, S*(λ) is the external water intensity inversion vector, and K cand It is a dimension of The reduced response kernel function matrix, ΔC obs It is of dimension N mThe observed anomalous conductivity vector is N×1, H(λ) is the hat matrix, and trace(H(λ)) is the trace of the hat matrix. m [N] represents the total number of monitoring points. m -trace(H(λ))] 2 Let λ be the square of the model's degrees of freedom, λ be the regularization parameter, and L be the regularization matrix.

[0138] Iterate through a pre-defined set of candidate regularization parameters, calculate GCV(λ) for each parameter, and select the λ that minimizes GCV(λ) as the optimal regularization parameter λ*.

[0139] It should be noted that calculating the final external water intensity inversion vector involves substituting the optimal parameter λ* obtained in the previous step into the normal equation and solving it again. The mathematical expression of the corresponding normal equation is as follows:

[0140] ;

[0141] In the formula, S* is the final external water intensity inversion vector, and K cand It is a dimension of The reduced response kernel function matrix, N m Let λ* be the total number of monitoring points, λ* be the optimal regularization parameter, L be the regularization matrix, and ΔC be the value of the monitoring points. obs It is of dimension N m The observed anomalous conductivity vector is ×1.

[0142] It should be noted that the identification of external water infiltration nodes first involves calculating the normalized contribution rate of each candidate node based on the final external water intensity inversion vector. For each candidate external water infiltration node in the pipe network, the estimated value of external water infiltration intensity at that node, calculated through the coupled inversion model, is first obtained. Then, the estimated values ​​of external water infiltration intensity of all nodes in the candidate external water infiltration node set are summed to obtain the total intensity value. Finally, the intensity estimate of that node is divided by the total intensity value, and the resulting ratio is the normalized contribution rate of that node. This contribution rate quantifies the relative influence of that node on the overall conductivity anomaly signal observed in this study. The higher the contribution rate, the greater the likelihood that the node is identified as an external water infiltration source by the inversion model, and the higher the proportion of its assumed infiltration amount in the total infiltration amount.

[0143] Sort all candidate nodes in descending order of their normalized contribution rate, and set a preset contribution rate threshold of 0.05. Nodes with a normalized contribution rate greater than the preset contribution rate threshold are identified as external water infiltration location nodes.

[0144] S5. Based on the reduced response kernel function matrix and the final external water intensity inversion vector, the external water infiltration node is decoupled from multiple sources and the location confidence is calculated to generate an external water location result report.

[0145] In this embodiment of the invention, the process of performing multi-source decoupling and calculating location confidence for the external water infiltration node based on the reduced response kernel function matrix and the final external water intensity inversion vector to generate an external water location result report includes:

[0146] Based on the reduced response kernel function matrix and the final external water intensity inversion vector, the response similarity between any two external water infiltration location nodes is calculated.

[0147] Based on the response similarity and the preset similarity threshold, the location nodes are decoupled from multiple sources, and nodes with highly similar responses are merged into an equivalent external water source.

[0148] Based on the reduced response kernel function matrix, the external water intensity inversion vector, and the results of multi-source decoupling, the location confidence index of each final location node is calculated.

[0149] Based on the location reliability index, the location results are classified into levels, and combined with the spatial location of each node, infiltration intensity, normalized contribution rate and multi-source decoupling information, an external water location result report is generated.

[0150] It should be noted that the calculation of response similarity is to assess the distinguishability of different localization nodes in the monitoring space, and its mathematical expression is:

[0151] ;

[0152] In the formula, cosθ(i,k) represents the response similarity between the external water infiltration positioning node i and node k, and K... cand (i,·) and K cand (k, ·) are the row vectors corresponding to the localized nodes i and k in the reduced response kernel function matrix, respectively. The L2 norm of a vector is used to represent the vector's L2 norm.

[0153] Furthermore, the cosθ(i,k) response similarity ranges from [-1,1], but is usually [0,1] in this application context; the closer the value is to 1, the more similar the theoretical influence patterns of the two nodes are, and the more difficult it is to distinguish them with limited monitoring point data.

[0154] It should be noted that multi-source decoupling is based on response similarity to determine whether multiple positioning nodes represent the same physical source. A preset similarity threshold of 0.9 is set. For any two positioning nodes i and k, if the response similarity is greater than the preset similarity threshold, they are determined to be coupled sources, meaning that their monitoring response patterns are highly similar, possibly originating from the same infiltration event in spatially close locations or due to the limited resolution of the monitoring network. Otherwise, they are determined to be distinguishable sources. For the node group determined to be coupled sources, they are merged into an equivalent external water source. The equivalent intensity of the equivalent source is the sum of the inversion intensities of each node in the group, and its equivalent position is the centroid position obtained by weighting the spatial coordinates of the nodes with the intensities of each node as weights.

[0155] It should be noted that the location reliability index is calculated to quantify the reliability of each final location node (or equivalent source) i. Its mathematical expression is:

[0156] ;

[0157] In the formula, CI i For location reliability indicators, It is the norm of the response vector of node i. It is the inversion intensity of node i. It is the observed anomaly vector ΔC obs The norm of cosθ max (i) is the maximum response similarity between node i and all other ultimately located nodes;

[0158] The formula consists of the multiplication of two parts: the first part This reflects the proportion of the observed signal explained by that node. An excessively high proportion may indicate that the model is overly reliant on that node and therefore less robust; the second part is 1-cosθ. max (i) reflects the distinguishability of the node from other nodes. The easier it is to distinguish, the higher the uniqueness and the higher the confidence level.

[0159] It should be noted that generating the external water location result report involves systematically integrating and formatting all analysis, calculation, and judgment results. The report content is structured to include the following information: the spatial coordinates of each finally determined external water infiltration point, the estimated infiltration intensity, the normalized contribution rate, the location reliability index and its corresponding confidence level, and multi-source decoupling information. The report is presented in a clear and readable text, table, or graphical format, providing users with a comprehensive, quantitative conclusion on external water location that includes uncertainty assessment, directly supporting subsequent engineering investigation, decision-making, and disposal actions.

[0160] Furthermore, based on the location reliability index value, it is divided into high (CI) i≥0.7), medium (0.4≤CI) i <0.7), low (CI) i <0.4) Level 3.

[0161] like Figure 2 The diagram shown is a functional block diagram of an external water positioning system based on conductivity-based pipeline pollution coupling inversion, provided by an embodiment of the present invention.

[0162] The external water location system 100 based on conductivity-based pipeline pollution coupling inversion described in this invention can be installed in an electronic device. Depending on the functions implemented, the external water location system 100 based on conductivity-based pipeline pollution coupling inversion may include a conductivity coupling modeling module 101, an anomaly identification module 102, a regional preliminary screening module 103, an intensity inversion module 104, and a decoupling evaluation module 105. The module described in this invention can also be called a unit, referring to a series of computer program segments that can be executed by the processor of an electronic device and can perform a fixed function, stored in the memory of the electronic device.

[0163] In this embodiment, the functions of each module are as follows:

[0164] The conductivity coupling modeling is used to establish a hydraulic-conductivity coupling model of the pipeline network based on the pipeline network topology, pipe segment attributes and monitoring point locations, and to generate a node-monitoring point conductivity response kernel function matrix.

[0165] The anomaly identification module is used to identify abnormal conductivity periods and calculate the abnormal amplitude based on the conductivity time series data of each monitoring point, and generate an observed abnormal conductivity vector.

[0166] The regional preliminary screening module is used to calculate the consistency of abnormal signal propagation based on the abnormal conductivity period and the hydraulic parameters of the pipeline network, screen out candidate external water infiltration nodes, and extract the corresponding row from the node-monitoring point conductivity response kernel function matrix to generate a reduced response kernel function matrix.

[0167] The intensity inversion module is used to construct and solve a regularized inversion equation based on the reduced response kernel function matrix and the observed abnormal conductivity vector to obtain the final external water intensity inversion vector, and to determine the external water infiltration node based on the final external water intensity inversion vector.

[0168] The decoupling evaluation module is used to perform multi-source decoupling on multiple external water infiltration nodes and calculate the location confidence based on the reduced response kernel function matrix and the final external water intensity inversion vector, so as to generate an external water location result report.

[0169] 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.

[0170] 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.

[0171] 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.

[0172] 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.

[0173] The embodiments of this application 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.

[0174] 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 method for locating external water sources based on conductivity-coupled inversion of pipe network pollution, characterized in that, The method includes: S1. Based on the pipeline network topology, pipe segment attributes and monitoring point locations, a hydraulic-conductivity coupling model of the pipeline network is established, and the node-monitoring point conductivity response kernel function matrix is ​​calculated. S2, based on the conductivity time series data of each monitoring point, identify the abnormal conductivity periods and calculate the abnormal amplitude, and generate the observed abnormal conductivity vector; S3. Based on the abnormal conductivity period and the hydraulic parameters of the pipeline network, candidate external water infiltration nodes are selected. Based on the candidate external water infiltration nodes, parameters are extracted from the node-monitoring point conductivity response kernel function matrix to construct a reduced response kernel function matrix. S4. Based on the reduced response kernel function matrix and the observed abnormal conductivity vector, a regularized inversion equation is constructed and solved to obtain the final external water intensity inversion vector. The external water infiltration node is determined according to the final external water intensity inversion vector. S5. Based on the reduced response kernel function matrix and the final external water intensity inversion vector, the external water infiltration node is decoupled from multiple sources and the location confidence is calculated to generate an external water location result report.

2. The external water location method based on conductivity-coupled inversion for pipe network pollution as described in claim 1, characterized in that, Based on the pipeline network topology, pipe segment attributes, and monitoring point locations, a hydraulic-conductivity coupling model of the pipeline network is established to calculate the node-monitoring point conductivity response kernel function matrix, including: Acquire pipeline network topology data, pipe segment attribute data, and monitoring point location data; Based on the pipeline network topology data, a directed graph model of the pipeline network is established, and steady-state hydraulic calculations are performed on the directed graph model of the pipeline network based on the pipe segment attribute data to obtain the flow rate and velocity of each pipe segment. Based on the flow rate and velocity of each pipe segment, a one-dimensional convection-dispersion equation describing the transport process of conductivity in the pipe network is established. By traversing all potential infiltration nodes in the pipeline network, and based on the one-dimensional convection dispersion equation, the steady-state conductivity response values ​​generated at all monitoring points when external water with a unit conductivity increment is injected at each potential node are simulated in a forward manner, so as to generate the node-monitoring point conductivity response kernel function matrix.

3. The external water location method based on conductivity-coupled inversion for pipe network pollution as described in claim 2, characterized in that, The mathematical expression of the one-dimensional convection dispersion equation is as follows: ； In the formula, C represents the increase in conductivity in the pipe network water caused by the infiltration of external water. The coordinates are continuous time coordinates, where x represents the axial distance along the pipe segment, and v represents the distance along the pipe segment. e D is the water flow velocity in the pipe section. σ This represents the effective diffusion coefficient within the pipe section.

4. The external water location method based on conductivity-coupled inversion for pipe network pollution as described in claim 1, characterized in that, The process of identifying abnormal conductivity periods and calculating the magnitude of the anomalies based on conductivity time-series data from each monitoring point, and generating an observed abnormal conductivity vector, includes: Based on the obtained conductivity time series data of each monitoring point, the local statistical characteristics of each monitoring point are calculated to obtain the local mean and local standard deviation of each time point. Based on the local mean and the local standard deviation, the abnormal detection index of each monitoring point at each time point is calculated and compared with the preset abnormal threshold to identify abnormal moments and generate a set of abnormal conductivity time periods for each monitoring point. Based on the conductivity time series data and local mean within the set of abnormal conductivity periods, the conductivity abnormality amplitude within each abnormal period is calculated, and the abnormal amplitudes of all monitoring points are combined into an observed abnormal conductivity vector.

5. The external water location method based on conductivity-coupled inversion for pipe network pollution as described in claim 2, characterized in that, The process of selecting candidate external water infiltration nodes based on the abnormal conductivity periods and the hydraulic parameters of the pipe network includes: Based on the set of abnormal conductivity periods at each monitoring point, calculate the time difference of abnormal signals between any two monitoring points. Based on the pipeline network topology and the flow velocity of each pipe segment, calculate the hydraulic propagation time between any two monitoring points; Based on the time difference of the abnormal signal and the hydraulic propagation time, the monitoring points are screened to obtain monitoring point pairs that meet the propagation consistency condition; Based on the monitoring point pairs that meet the propagation consistency condition and the pipeline topology, upstream nodes are traced along the counter-current direction, and all traced upstream nodes are merged to generate a candidate set of external water infiltration nodes.

6. The external water location method based on conductivity-coupled inversion for pipe network pollution as described in claim 5, characterized in that, The step of filtering monitoring points based on the time difference of the abnormal signal and the hydraulic propagation time to obtain pairs of monitoring points that meet the propagation consistency condition includes: Based on the time difference of the abnormal signal and the hydraulic propagation time, the propagation consistency index of the abnormal signal between the two monitoring points is calculated. The propagation consistency index is compared with a preset consistency threshold to select monitoring point pairs that meet the propagation consistency conditions.

7. The external water location method based on conductivity-coupled inversion for pipe network pollution as described in claim 1, characterized in that, The process of extracting parameters from the node-monitoring point conductivity response kernel function matrix based on the candidate external water infiltration nodes to construct a reduced response kernel function matrix includes: Extract the corresponding row of the candidate external water infiltration node from the node-monitoring point conductivity response kernel function matrix to generate a reduced response kernel function matrix.

8. The external water location method based on conductivity-coupled inversion for pipeline pollution as described in claim 1, characterized in that, Based on the reduced response kernel function matrix and the observed abnormal conductivity vector, a regularized inversion equation is constructed and solved to obtain the final external water intensity inversion vector. The external water infiltration nodes are then determined according to this final external water intensity inversion vector, including: Based on the reduced response kernel function matrix and the observed abnormal conductivity vector, a regularization term is introduced to construct a regularized inversion equation describing the linear relationship between the monitored anomaly and the node infiltration intensity. The regularized inversion equation is solved to derive the corresponding normal equation, and the external water intensity inversion vector related to the regularization parameter is obtained by solving the normal equation. The optimal regularization parameter is selected from a set of candidate regularization parameters to minimize the generalized cross-validation index. The normal equation is then resolved based on the optimal regularization parameter to obtain the final external water intensity inversion vector. Based on the final external water intensity inversion vector, the normalized contribution rate of each candidate node is calculated, and the external water infiltration node is determined according to the preset contribution rate threshold.

9. The external water location method based on conductivity-coupled inversion for pipe network pollution as described in claim 1, characterized in that, Based on the reduced response kernel function matrix and the final external water intensity inversion vector, the external water infiltration nodes are decoupled from multiple sources and the location confidence is calculated to generate an external water location result report, including: Based on the reduced response kernel function matrix and the final external water intensity inversion vector, the response similarity between any two external water infiltration location nodes is calculated. Based on the response similarity and the preset similarity threshold, the location nodes are decoupled from multiple sources, and nodes with highly similar responses are merged into an equivalent external water source. Based on the reduced response kernel function matrix, the external water intensity inversion vector, and the results of multi-source decoupling, the location confidence index of each final location node is calculated. Based on the location reliability index, the location results are classified into levels, and combined with the spatial location of each node, infiltration intensity, normalized contribution rate and multi-source decoupling information, an external water location result report is generated.

10. An external water location system based on conductivity-based pipe network pollution coupling inversion, used to implement the external water location method based on conductivity-based pipe network pollution coupling inversion as described in any one of claims 1-9, characterized in that, The system includes: Conductivity coupling modeling is used to establish a hydraulic-conductivity coupling model of a pipe network based on the pipe network topology, pipe segment properties and monitoring point locations, and generate a node-monitoring point conductivity response kernel function matrix. The anomaly identification module is used to identify abnormal conductivity periods and calculate the abnormal amplitude based on the conductivity time series data of each monitoring point, and generate an observed abnormal conductivity vector. The regional preliminary screening module is used to calculate the consistency of abnormal signal propagation based on the abnormal conductivity period and the hydraulic parameters of the pipeline network, screen out candidate external water infiltration nodes, and extract the corresponding row from the node-monitoring point conductivity response kernel function matrix to generate a reduced response kernel function matrix. The intensity inversion module is used to construct and solve a regularized inversion equation based on the reduced response kernel function matrix and the observed abnormal conductivity vector to obtain the final external water intensity inversion vector, and to determine the external water infiltration node based on the final external water intensity inversion vector. The decoupling evaluation module is used to perform multi-source decoupling on multiple external water infiltration nodes and calculate the location confidence based on the reduced response kernel function matrix and the final external water intensity inversion vector, so as to generate an external water location result report.

Images