A sewer network pollutant tracing method, device and system

Abstract

The present application relates to the field of data processing, and particularly relates to a sewer network pollutant tracing method, device and system. The present application fuses multi-source pollution data to construct an uncertainty quantification framework, which is used to determine adaptive pollution source tracing positioning. The experience probability of the relevant historical pollution path is obtained by mining the historical pollution events, and the suspected pollution source point and the suspected pollution path are obtained. Then, the asynchronous strength and the blind area gain of each suspected pollution path are accurately quantified. In the final adaptive tracing positioning process, the experience probability, the asynchronous strength and the blind area gain of the suspected pollution path are fused to obtain an adaptive search radius, and the tracing positioning is performed within the adaptive search radius. The final adaptive search radius obtained by the present application can differentiate the scheduling of search resources, reduce the waste of tracing resources and reduce the tracing cost.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, specifically to a method, equipment, and system for tracing the source of pollutants in drainage pipe networks. Background Technology

[0002] Traditional methods for tracing pollutant sources in urban drainage networks primarily rely on two types of approaches: simulation inversion based on fixed physical models and data-driven algorithms. Physical model-based methods require precise hydraulic parameters and boundary conditions of the network, but actual network parameters are difficult to obtain and highly time-varying, leading to severe model distortion and limited practicality. While data-driven methods reduce reliance on physical models, they often directly input monitoring data into general algorithm frameworks, failing to delve into the unique inherent logic and coupling relationships of drainage network data across three dimensions: topological connectivity, temporal dynamics, and event sparsity. For example, general algorithms cannot effectively express the impact of dynamic changes in water flow direction on the path, or ignore the signal non-independence problem caused by the similar emission patterns of multiple domestic pollution sources. These limitations result in existing methods exhibiting large locational errors, high missed detection rates, and weak robustness when facing source tracing scenarios in real networks where the source point is unknown, release time is unknown, multiple sources are mixed, and background noise is complex.

[0003] In practice, the inventors discovered the following defects in the aforementioned prior art: When tracing pollution sources using existing methods, it is impossible to accurately locate the pollution sources. A large amount of resources need to be invested in investigating possible candidate pollution sources and their search radii. There is a lack of effective methods to evaluate the search radius of pollution sources, and there is a lack of adaptive capabilities. Existing methods for setting the search radius may result in the actual pollution sources being missed or causing excessive waste of resources due to being too large or too small, which is costly. Summary of the Invention

[0004] To address the source tracing problem caused by the lack of effective assessment and search radii for pollution sources in drainage pipe networks, the present invention aims to provide a method, equipment, and system for tracing pollutants in drainage pipe networks. The specific technical solution adopted is as follows: This application provides a method for tracing the source of pollutants in a drainage pipe network, including: A multi-source pollution dataset is constructed, which includes: a pipeline connection topology map, monitoring points, real-time monitoring data of the monitoring points, and their historical pollution source tracing events; the information of the historical pollution source tracing events includes: historical pollution monitoring points, historical pollution paths, and historical pollution source points. The historical pollution monitoring points are the monitoring points where pollution was first detected in the historical pollution source tracing events, the historical pollution source points are the monitoring points closest to the actual pollution source, and the pollution paths are the paths from the historical pollution source points to the historical pollution monitoring points. Using real-time abnormal monitoring points as suspected endpoints, suspected pollution sources and corresponding suspected pollution paths are screened out in the pipeline connection topology map based on the historical pollution paths of the associated monitoring points. The asynchronicity intensity of suspected pollution paths is obtained by analyzing the distribution concentration of pollution propagation time in the history of suspected pollution sources and corresponding suspected pollution paths. Based on the distribution of monitoring points along suspected pollution paths, the blind zone gain of suspected pollution paths is obtained; By fusing the asynchronicity intensity and blind zone gain of the suspected pollution source and the suspected pollution path, an adaptive search radius for the suspected pollution source is obtained, and source tracing and localization are performed within the search radius.

[0005] Furthermore, using real-time abnormal monitoring points as suspected endpoints, suspected pollution sources and corresponding suspected pollution paths are screened in the pipeline connection topology map based on the historical pollution paths of the associated monitoring points, including: For the real-time abnormal monitoring point, by matching the historical pollution monitoring points of historical pollution source tracing events, all relevant historical pollution source tracing events that are the same as the real-time abnormal monitoring point are obtained, the historical pollution source points of the relevant historical pollution source tracing events are obtained, and all paths from the historical pollution source points to the real-time abnormal monitoring point are obtained as a set of historical relevant paths. Based on each historical related path in the historical related path set and the pipeline connection topology, all path segments of the historical related path are obtained, and the timestamps of all historical pollution events corresponding to all endpoints of each path segment are obtained and recorded as the timestamp set of the path segment. The real-time time of the real-time anomaly monitoring point is compared with the timestamp in the timestamp set of the path segment. Suspected pollution source points are screened out from all endpoints of the path segment, and the suspected pollution path is determined.

[0006] Further, the step of comparing the real-time time of the real-time anomaly monitoring point with the timestamps in the timestamp set of the path segment, filtering out suspected pollution source points from all endpoints of the path segment, and determining the suspected pollution path includes: Based on the pipeline connection topology, obtain the set of timestamps of all historical pollution events at the endpoints of the path segment, analyze the time proximity between the real time and the overall time set, and obtain the activity level of the path segment. Filter the activity of all path segments of each historically related path to determine the empirical probability of each historically related path; The endpoints of the paths with the highest empirical probability among all historical related paths are recorded as suspected pollution source points. Based on the pipeline connection topology, the set of reachable paths from all suspected pollution source points to the real-time abnormal monitoring points is denoted as the suspected pollution path corresponding to the suspected pollution source point.

[0007] Furthermore, the step of obtaining the asynchronicity intensity of the suspected pollution path based on the historical distribution concentration of pollution propagation time corresponding to the suspected pollution source point includes: Historical real-time monitoring data of the two endpoints of the suspected contamination path are obtained respectively, and the mutual information of the historical real-time monitoring data of the two endpoints of the suspected contamination path under different time shift multiples is calculated by aligning them. Based on the mutual information of the two endpoints of the suspected contamination path at different time shift multiples, curve fitting is performed to obtain the mutual information curve of the two endpoints of the suspected contamination path. The time shift factor corresponding to the maximum mutual information in the mutual information curves of the two endpoints of the suspected contamination path is taken as the optimal time shift factor; Calculate the time shift difference factor between the optimal time shift factor and each other time shift factor, and determine the peak neighborhood distribution concentration between the optimal time shift factor and each other time shift factor based on the time shift difference factor between the optimal time shift factor and each other time shift factor and the mutual information under each other time shift factor; The asynchronicity intensity of the suspected contamination path is determined based on the peak neighborhood distribution concentration between the optimal time shift factor and other time shift factors.

[0008] Furthermore, obtaining the blind zone gain based on the distribution of monitoring points along the suspected contamination path includes: Based on the pipeline connection topology map and the distribution of the monitoring points, all unmonitored path segments of suspected contamination paths are obtained; the unmonitored path segments are those where the monitoring points do not exist at the path endpoints. Based on all unmonitored path segments of the suspected contamination path, obtain the length of the longest consecutive unmonitored path segment; The blind zone concentration of the suspected pollution path is determined based on the length of the longest consecutive unmonitored path segment and the total length of the suspected pollution path; the blind zone concentration is directly proportional to the length of the longest consecutive unmonitored path segment and inversely proportional to the total length of the suspected pollution path. Based on the blind zone concentration of the suspected contamination path, a preset blind zone gain threshold is used to determine the blind zone gain of the contamination path.

[0009] Furthermore, the specific method for obtaining the blind zone gain coefficient is as follows: Based on the historical pollution source tracing events, some historical pollution events on the suspected pollution path are randomly selected, and the average distance difference between the predicted pollution source point and the actual pollution source of the some historical pollution events is evaluated to obtain the blind zone gain coefficient.

[0010] Furthermore, the step of obtaining the adaptive search radius of the suspected pollution source point by fusing the asynchronicity intensity and blind zone gain of the suspected pollution source point and the suspected pollution path includes: The empirical probability of the historical related paths corresponding to the suspected pollution source point is obtained, the asynchronicity intensity and blind zone gain of the suspected pollution path are obtained, and the search expansion coefficient of the suspected pollution source point is determined by combining the suspected pollution path and the historical related paths corresponding to the suspected pollution source point. The adaptive search radius of the suspected pollution source is obtained by adjusting the preset maximum radius based on the search expansion coefficient of the suspected pollution source.

[0011] Further, the step of adjusting the preset maximum radius to obtain the adaptive search radius of the suspected pollution source point based on the search expansion coefficient of the suspected pollution source point includes: The adaptive search radius of the suspected pollution source is determined by the product of the source tracing probability of the suspected pollution source and the preset maximum search radius.

[0012] This application provides a pollutant tracing device for drainage pipe networks, including a memory, a processor, and a computer program stored in the memory and executable on the processor.

[0013] This application provides a pollutant tracing system for drainage pipe networks, including: Data acquisition module: used to construct a multi-source pollution dataset, which includes: a pipeline connection topology map, monitoring points, real-time monitoring data of the monitoring points and their historical pollution source tracing events; Source tracing analysis module: Used to select suspected pollution sources and corresponding suspected pollution paths in the pipeline connection topology map based on the historical pollution paths of the associated monitoring points, with real-time abnormal monitoring points as suspected endpoints; obtain the asynchronous intensity of suspected pollution paths based on the distribution concentration of the historical pollution propagation time of the suspected pollution sources and corresponding suspected pollution paths; obtain the blind zone gain of suspected pollution paths based on the distribution of monitoring points on suspected pollution paths. Source tracing and localization module: used to obtain the adaptive search radius of the suspected pollution source by fusing the asynchronous intensity and blind zone gain of the suspected pollution source and the suspected pollution path, and to perform source tracing and localization within the search radius.

[0014] The present invention has the following beneficial effects: This application constructs an in-depth analysis system for the activity, asynchronicity, and blind zone gain of suspected pollution paths, achieving dynamic adaptive optimization of the source tracing confidence radius based on the inherent logic of multi-source data from the pipeline network. This, in turn, dynamically schedules search resources to achieve source tracing. Compared to existing technologies, this scheme dynamically adjusts the spatial search range of the source tracing algorithm based on the comprehensive uncertainty of the propagation path from candidate pollution sources to abnormal monitoring points: in low-uncertainty areas with rich path experience, high signal synchronization, and few monitoring blind zones, it automatically shrinks the search radius for precise positioning; in high-uncertainty areas with unclear paths, severe signal dispersion, and long continuous blind zones, it adaptively expands the search range to avoid missing true pollution sources.

[0015] This intelligent scheduling mechanism, tightly coupled with the real-time status and historical experience of the pipeline network, significantly improves the targeting of source tracing searches and the efficiency of computing resource utilization. It enhances the ability to capture pollution sources under complex operating conditions and sparse monitoring environments, effectively addressing the insufficient adaptability of traditional fixed or simple rule-based parameter algorithms when dealing with the dynamic, heterogeneous, and uncertain nature of real pipeline networks. By establishing an intrinsic correlation between data-driven features and core algorithm parameters, this method achieves a substantial improvement in source tracing accuracy and robustness. Attached Figure Description

[0016] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 A flowchart illustrating a method for tracing the source of pollutants in a drainage pipe network, provided in one embodiment of the present invention; Figure 2 This is a schematic diagram of a pollutant tracing device for a drainage pipe network provided in one embodiment of the present invention.

[0018] Figure 3 This is a schematic diagram of a pollutant tracing system for a drainage pipe network, provided as an embodiment of the present invention. Detailed Implementation

[0019] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a method, device, and system for tracing pollutants in a drainage network according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0021] In all division and logarithmic operations covered in this application, a smoothing mechanism is employed to prevent computer program crashes or invalid values ​​from being generated due to a zero denominator or a zero input. Specifically, a positive correction factor is superimposed on the denominator term of the division operation or the argument term of the logarithmic function. For example, the value is This ensures the robustness and feasibility of the algorithm under extreme conditions.

[0022] The following description, in conjunction with the accompanying drawings, details the specific scheme of the method, equipment, and system for tracing pollutants in drainage pipe networks provided by the present invention.

[0023] Example 1: This invention proposes a method, equipment, and system for tracing the source of pollutants in drainage pipe networks. Please refer to [link / reference]. Figure 1 The diagram illustrates a flowchart of a method for tracing the source of pollutants in a drainage pipe network according to an embodiment of the present invention. The method includes the following steps: Step S1: Construct a multi-source pollution dataset, which includes: a pipeline connection topology map, monitoring points, real-time monitoring data of the monitoring points and their historical pollution source tracing events; the information of the historical pollution source tracing events includes: historical pollution monitoring points, historical pollution paths, and historical pollution source points. The historical pollution monitoring points are the monitoring points where pollution was first detected in the historical pollution source tracing events, the historical pollution source points are the monitoring points closest to the actual pollution source, and the pollution paths are the paths from the historical pollution source points to the historical pollution monitoring points.

[0024] It should be noted that the multi-source pollution dataset includes various data types, such as geographic spatiotemporal data, sensor data, and pipeline network relationship data.

[0025] Specifically, in order to implement the method, equipment, and system for tracing pollutants in drainage pipe networks proposed in this embodiment, one specific implementation of this invention involves constructing a multi-source pollution dataset, including: Online monitoring points are deployed to continuously collect high-frequency time-series monitoring data, including water quality parameters (such as specific pollutant concentrations and conductivity) and hydraulic parameters (such as flow rate and velocity) at key nodes. The sampling frequency must be sufficient to capture pollution pulse signals; in this embodiment, the data collection frequency is no less than once per minute. Preferably, the sensor monitoring points should be deployed at pipe intersections and important locations.

[0026] Collect a complete pipeline network topology diagram, including the connection relationship between all monitoring points where sensors are deployed and the pipelines, as well as the spatial geometric length attributes of the pipelines; specifically, the monitoring points are the vertices of the pipeline network topology diagram, and the pipelines are the edges of the pipeline network topology diagram.

[0027] Historical pollution source tracing events are collected, including confirmed pollution event cases. Each record must clearly indicate the location of the pollution source, the location of the monitoring point where the pollution was detected, the event timestamp, and the confirmed historical pollution path. Ultimately, historical pollution monitoring points, historical pollution paths, and historical pollution source points are obtained. Historical pollution monitoring points are the monitoring points that first detected pollution in the historical pollution source tracing events. Historical pollution source points are the monitoring points closest to the actual pollution source. Pollution paths are the paths from historical pollution source points to historical pollution monitoring points.

[0028] In this embodiment, all data must be uniformly connected to the central data platform and undergo time synchronization and quality verification to provide a consistent and reliable data foundation for subsequent multi-level feature mining.

[0029] Step S2: Using real-time abnormal monitoring points as suspected endpoints, filter out suspected pollution sources and corresponding suspected pollution paths in the pipeline connection topology map based on the historical pollution paths of the associated monitoring points.

[0030] To address the fundamental spatial search problem in source tracing—"from which path might the pollutant have traveled?"—it's necessary to mine historical data for preferences in propagation path selection, providing a basic probabilistic framework for subsequent analysis. Simply counting the total number of times pipelines were used can introduce bias due to the uneven temporal distribution of historical events; for example, pipelines that were active in the past but recently idle should be treated differently from those that are currently active. When considering relevant paths ending at the current real-time anomaly monitoring point, the closer the timestamp of a historical pollution event is to the current timestamp, the higher the probability of pollution occurring along that path. The more likely the historical pollution source is to be a current suspected pollution source, thus identifying the suspected pollution path.

[0031] Specifically, in one specific implementation of the present invention, the method for obtaining real-time abnormal monitoring points is as follows: based on the prior experience of experts, the normal range of each hydraulic parameter is preset. If the hydraulic parameter of the monitoring point exceeds the preset range, the monitoring point will experience a real-time abnormality, and the time when the monitoring point experiences an abnormality will be taken as the current time.

[0032] It should be noted that the 'historical pollution source points' and 'actual pollution paths' in this embodiment can be derived from historical manual investigation confirmation records, or from inversion simulation data of historical operating conditions based on high-precision hydraulic models (such as SWMM).

[0033] Step S3: Based on the historical distribution concentration of pollution propagation time corresponding to suspected pollution sources, obtain the asynchronous intensity of suspected pollution paths; Based on historical traces, suspected contamination paths have been identified. Further solutions are needed to address the uncertainty of the time it takes for contamination signals to propagate along the selected paths, injecting a dynamic time dimension into the static path network. Since suspected contamination paths have two monitoring endpoints (upstream and downstream, i.e., the suspected contamination source and the current real-time anomaly monitoring point), when an anomaly occurs at the upstream monitoring point, it will move to the downstream monitoring point after a certain period. If the contamination path remains stable, the time elapsed during the contamination period is relatively predictable. The correlation of contamination changes from the upstream to the downstream monitoring point is concentrated on a specific time delay, with weak asynchronicity. However, the correlation of contamination changes from the upstream to the downstream monitoring point is distributed across multiple time delays, with strong asynchronicity. Therefore, we determine the correlation by evaluating mutual information at arbitrary time delays, and then use the mutual information to assess the asynchronicity of the contamination path.

[0034] Step S4: Based on the distribution of monitoring points along the suspected contamination path, obtain the blind zone gain of the suspected contamination path; After quantifying the asynchronous nature of path experience probability and signal propagation, it is necessary to assess how the limitations of the physical monitoring network amplify the uncertainties identified in the first two layers. This is a key external constraint affecting the confidence level of source tracing inference. Measuring solely by a linear approach, such as the proportion of blind zone pipeline length, fails to distinguish whether blind zones are continuously concentrated or dispersed. Furthermore, the interruption and ambiguity effects of long, continuous blind zones are far greater than the latter. Therefore, before comprehensively assessing source tracing uncertainty, it is essential to analyze the topological concentration of blind zones on candidate paths, i.e., analyze the distribution of monitoring points on suspected contamination paths, and quantify the nonlinear interference gain caused by this "information black hole" on signal source tracing. This allows for the introduction of blind zone gain due to "limited observation capabilities" on top of the "path-time" uncertainty.

[0035] Step S5: By fusing the asynchronicity intensity and blind zone gain of the suspected pollution source and the suspected pollution path, the adaptive search radius of the suspected pollution source is obtained, and source tracing and localization are performed within the search radius; After processing steps S1 to S4, the suspected pollution sources and paths of the real-time anomaly monitoring points can be obtained. Each suspected pollution path has asynchronous intensity and blind zone gain. Therefore, it is necessary to deeply fuse these features that characterize uncertainties in different dimensions to obtain a core quantitative parameter that can directly drive and optimize the source tracing search strategy. Simply linearly weighting and summing these features ignores the differences in their dimensions and coupling relationships, and fails to map the abstract uncertainty measure to a specific spatial search range. Therefore, before constructing the final parameter, a fusion framework needs to be designed to introduce a priori scale of historically successful source tracing, thereby outputting a search radius with clear physical meaning and operational guidance value.

[0036] Preferably, in some implementations of the present invention, using real-time abnormal monitoring points as suspected endpoints, suspected pollution sources and corresponding suspected pollution paths are screened in the pipeline connection topology map based on the historical pollution paths of the associated monitoring points, including: Step S2-1: For the real-time abnormal monitoring point, by matching the historical pollution monitoring points of historical pollution source tracing events, obtain all relevant historical pollution source tracing events that are the same as the real-time abnormal monitoring point, obtain the historical pollution source points of the relevant historical pollution source tracing events, and obtain all paths from the historical pollution source points to the real-time abnormal monitoring point as a set of historical relevant paths.

[0037] Input the historical pollution event database (containing confirmed historical pollution source points S, historical monitoring points M, and timestamps τ of historical pollution source tracing events) and the pipeline topology connection diagram. Use the monitoring points with real-time anomalies as historical pollution monitoring points, take the time of the current real-time anomaly monitoring point as the current time, and filter the historical source tracing events with the current real-time anomaly monitoring point as the historical pollution monitoring point from the historical pollution source tracing events to obtain several historically related pollution paths, forming a set of historically related pollution paths.

[0038] Specifically, each path in the historically related pollution set is a real pollution source tracing event path, and the monitoring points where anomalies were detected in the historical paths are the same as the monitoring points where anomalies are detected in the current real-time paths, that is, they have the same path endpoint; the starting points of each path in the historically related pollution set may be different.

[0039] Step S2-2: Based on each historical related path in the historical related path set and the pipeline connection topology map, obtain all path segments of the historical related path, and obtain the timestamps of all historical pollution events corresponding to all endpoints of each path segment as the timestamp set of the path segment.

[0040] Because each historically related pollution path consists of multiple path segments, and each path segment has two endpoints (two adjacent path segments share one endpoint), and monitoring points are all located at the endpoints of the path segments, historical pollution events for each path segment can be determined by the monitoring points at the endpoints of the path segments. Furthermore, the more pollution events that occur across the various path segments along a historically related pollution path, the higher the probability that the starting point of that historically related pollution path is the pollution source.

[0041] In this embodiment, for all path segments of a historically related pollution path, a set of timestamps of historical pollution source tracing events for each path segment is obtained. Specifically, for each path segment edge e, a set is formed by collecting the timestamps of all its tagged historical pollution source tracing events. And count the total number of marks for each path segment edge. At the same time, the effective observation time for each pipeline was recorded. This refers to the total time elapsed from when the pipeline was first included in the monitoring system until the current moment, which is used for subsequent normalization processing.

[0042] Step S2-3: Compare the real-time time of the real-time anomaly monitoring point with the timestamps in the timestamp set of the path segment, filter out suspected pollution source points from all endpoints of the path segment, and determine the suspected pollution path.

[0043] Based on historical pollution source tracing events, for a current real-time anomaly monitoring point, relevant historical pollution paths generally represent possible related pollution paths, and the starting point of the path (i.e., the historical pollution source point) is likely the pollution source point of the current real-time anomaly monitoring point. By comparing the current timestamp of the current real-time anomaly monitoring point with the set of timestamps in the path segment, the closer the former is to the latter, the more recent historical pollution events occurred in the path segment before the current time. Therefore, for a relevant historical pollution path, the closer the current timestamp is to the set of timestamps of the path segment corresponding to the relevant historical pollution path, the more likely the historical pollution source point of that historical path is the current suspected pollution source point.

[0044] Furthermore, for suspected contamination pathways, since the flowability of the pipeline may change due to blockage, suspected contamination pathways require further analysis.

[0045] Therefore, in some implementations of this invention, the process of obtaining suspected pollution sources and suspected pollution paths includes: Step S2-3-1: Based on the pipeline connection topology map, obtain the set of timestamps of all historical pollution events at the endpoints of the path segment, analyze the time proximity between the real time and the overall time of the timestamp set, and obtain the activity level of the path segment; Because in the set of all historical pollution events in a path segment, the closer the timestamp of the current moment is to the timestamp of a historical pollution event, the more likely there is a historical pollution event in the path segment that is close to the current moment. Therefore, the proximity can be assessed by comparing the timestamp of the pollution event with the timestamp of the current moment, and thus the activity level of the path segment can be obtained. That is, the greater the proximity, the more historical pollution events occurred in the path segment in the period closer to the current moment. Based on the empirical frequency assumption, the activity level of pollution events in the corresponding path segment is greater.

[0046] In one specific implementation of this invention, the activity level of path segment e at the current time t is expressed by the formula: ,in This represents the activity level of path segment e at the current time t. The time decay function assigns a higher weight to historical pollution events on the path segment that are closer to the current time. This indicates the degree of proximity between the timestamp of time t and the timestamp of historical pollution event i. represents the preset decay constant, used to evaluate the time decay rate, so that the inherent frequencies of different pollution events are comparable; t is the timestamp of the current moment; exp() represents an exponential function with the natural constant as the base; This represents the timestamp of historical pollution source tracing event i related to pipeline e; This represents the total monitoring time from pipeline e to the current point; log() represents the logarithmic function with the natural constant as the base, used to prevent the denominator from being too large.

[0047] The attenuation constant is determined by the median interval between adjacent historical pollutant events across all historical path segments; a default value can be set during system initialization or when historical events are insufficient. =30 days, used as the starting parameter for experience.

[0048] It should be noted that in the absence of historical data or during the system cold start phase, the activity level A(e,t) of all path segments can be set to 1 by default, that is, it is assumed that all path segments have the same initial activity level, and dynamic updates will be performed after sufficient event data has been accumulated.

[0049] Step S2-3-2: Filter the activity of all path segments of each historically related path to determine the empirical probability of each historically related path; In some specific embodiments of the present invention, since the historical correlation path consists of multiple path segments, according to the weakest link effect (used to describe the overall capability of a system being limited by the weakest part), the path segment with the lowest activity limits the activity of the entire historical correlation path, and the lowest activity is selected as the empirical probability of the historical correlation path.

[0050] It should be noted that in actual engineering practice, depending on the characteristics of the pipeline network, the mean method can also be adopted, that is, the average activity of all path segments in the historical related paths can be used as the empirical probability of the historical related paths.

[0051] Step S2-3-3: Obtain the path endpoints with the highest empirical probability among all historical related paths and record them as suspected pollution source points; Since a higher historical correlation path probability indicates more recent pollution events, the starting point of the historical correlation path (i.e., the historical pollution source point) is more likely to be a pollution source point based on the frequency of pollution events. Therefore, for monitoring points with current real-time anomalies, the starting point of the historical correlation pollution path (i.e., the historical pollution source point) is usually a possible pollution source point. Thus, multiple starting points with higher historical correlation path pollutant probabilities are selected as suspected pollution source points.

[0052] Preferably, in the specific operation process, the empirical probability of selecting historical pollution-related paths can be obtained by using a threshold method, and the empirical threshold can be set to 0.6; the threshold can also be determined according to the actual situation, and this embodiment does not limit it.

[0053] Step S2-3-4: Based on the pipeline connection topology map, traverse all the reachable paths from the suspected pollution source points to the real-time abnormal monitoring points and record them as the suspected pollution path corresponding to the suspected pollution source point.

[0054] Since the possible contamination paths may change with the pipeline route, it is necessary to traverse all reachable paths from the suspected contamination source to the real-time anomaly monitoring point as suspected contamination paths.

[0055] Preferably, in some implementations of the present invention, the asynchronicity strength of the suspected pollution path is obtained based on the degree of concentration of the historical pollution propagation time of the suspected pollution source point corresponding to the suspected pollution path, including: Step S3-1: Obtain historical real-time monitoring data of the two endpoints of the suspected contamination path respectively, and calculate the mutual information of the historical real-time monitoring data of the two endpoints of the suspected contamination path under different time shift multiples by aligning them.

[0056] Since monitoring data at historical monitoring points can show the possible spread of pollution along pollution paths in the past, the more stable the spread path, the stronger the correlation between the historical monitoring data at the pollution source and the real-time abnormal monitoring points at a specific time shift; when the spread path is unstable, the correlation is lower. Thus, the correlation under different time delays can be represented by mutual information.

[0057] As one embodiment, the steps for acquiring mutual information at different time shift multiples include: acquiring upstream and downstream monitoring points (i.e., pollution source and current real-time abnormal monitoring point) of suspected pollution paths; inputting time-series data X(t) and Y(t) of synchronously sampled hydraulic parameters of paired upstream monitoring point U and downstream monitoring point D over a long period (e.g., several consecutive months); the sampling frequency must be consistent and sufficiently high, recommended to be no less than 1 sample / minute (or 5 minutes / time depending on the hydraulic dynamics of the pipeline network) to capture the propagation dynamics of pollution pulse signals. Before calculating mutual information, the original time-series data needs to be preprocessed, including but not limited to: using moving average filtering to suppress high-frequency noise; performing Z-score standardization (or minimum-maximum normalization) to eliminate the impact of signal baseline drift and amplitude differences between different monitoring points on statistical correlation measurement. For each pair of preprocessed data, within a predetermined reasonable time shift search range... Within this timeframe, the time shift is calculated point by point with a fixed step size (such as an integer multiple of the sampling interval). Mutual information under Mutual information calculation requires binning and discretizing continuous data, and estimation using joint and marginal probability distributions. Its advantage lies in its ability to capture nonlinear statistical dependencies. Among these, It can be estimated based on the maximum theoretical hydraulic residence time of the pipeline network. The above processes are all technical means commonly used by those skilled in the art in the process of mutual information calculation, and the specific principles and logic are not elaborated or limited here.

[0058] Step S3-2: Based on the historical real-time monitoring data of the two endpoints of the suspected contamination path at different time shift multiples, curve fitting is performed to obtain the mutual information curve of the two endpoints of the suspected contamination path.

[0059] Since the data obtained at different time shifts are discrete, mutual information of a continuous distribution can be obtained by fitting a curve.

[0060] Preferably, the curve fitting method can be a least-squares nonlinear fitting algorithm. The specific operation process is a commonly used technique known to those skilled in the art and will not be described in detail here.

[0061] Step S3-3: Take the time shift factor corresponding to the maximum mutual information in the mutual information curves of the two endpoints of the suspected contamination path as the optimal time shift factor.

[0062] The maximum value in the mutual information curve indicates the strongest correlation between upstream and downstream monitoring points (endpoints) of a suspected contamination path under the corresponding time shift factor difference. Therefore, the time shift factor corresponding to the maximum value of the mutual information curve can be taken as the optimal time shift factor.

[0063] Specifically, the global maximum point is identified from the mutual information curve, and the time shift corresponding to the maximum point is... This is the time shift factor that has the strongest correlation from upstream monitoring point U to downstream monitoring point D, and is denoted as the optimal time shift factor.

[0064] Step S3-4: Calculate the time shift difference factor between the optimal time shift factor and each other time shift factor, and determine the peak neighborhood distribution concentration between the optimal time shift factor and each other time shift factor based on the time shift difference factor between the optimal time shift factor and each other time shift factor and the mutual information under each other time shift factor.

[0065] Since the horizontal axis of the mutual information curve is the time shift factor, the smaller the difference between other time shift factors and the optimal time shift factor, the larger the mutual information value corresponding to other time shift factors, which indicates that the distribution concentration of the peak neighborhood near the mutual information curve is greater, that is, the mutual information curve is more concentrated near the peak.

[0066] As one example, the calculation method for the neighborhood distribution concentration of other time shift factors and the optimal time shift factor includes: Time shift factor The difference factor from the optimal time shift factor is: .

[0067] Time shift factor The peak neighborhood distribution concentration is calculated as follows: ,in , They represent time shift factors respectively. and optimal time shift multiple Corresponding mutual information.

[0068] Step S3-5: Determine the asynchronicity intensity of the suspected contamination path based on the peak neighborhood distribution concentration between the optimal time shift multiple and other time shift multiples.

[0069] Since the mutual information is maximized at the optimal time shift factor, it indicates the strongest correlation between the monitoring data of upstream and downstream monitoring points at the corresponding time shift factor differences. When the mutual information curve is closely distributed around the optimal time shift factor, the fewer peaks around the curve, the higher the certainty of pollution signal propagation from upstream and downstream monitoring points along the suspected pollution path, indicating a stable pollution path and weak asynchrony. Therefore, the asynchrony strength of a suspected pollution path is determined by the peak neighborhood concentration of all other time shift factors; the greater the peak neighborhood concentration of other time shift factors, the weaker the asynchrony strength of the suspected pollution path.

[0070] As one example, the asynchronicity intensity of a suspected pollution path is determined by all time shift factors of the mutual information curve formed by upstream monitoring point U and downstream monitoring point D, calculated using the following formula: ,in Indicates the asynchronicity intensity of suspected contamination pathways. Indicates time shift factor The peak neighborhood distribution concentration; norm() represents the normalization function, which takes the arctangent function as the normalization function to normalize the molecule, eliminate the influence of the original signal amplitude or mutual information absolute magnitude on the result, and make Dq a dimensionless quantity that purely describes the distribution shape.

[0071] Preferably, in some implementations of the embodiments of the present invention, obtaining the blind zone gain of a suspected contamination path based on the distribution of monitoring points along the suspected contamination path includes: Continuous monitoring blind zones result in pollutants undergoing complete physicochemical changes (such as degradation and adsorption) without intermediate records, leading to feature loss. Compared to dispersed blind zones, the uncertainty introduced by continuous blind zones increases non-linearly; therefore, blind zone gain is introduced for compensation.

[0072] Step S4-1: Based on the pipeline connection topology map and the distribution of the monitoring points, obtain all unmonitored path segments of the suspected contamination path.

[0073] Input the pipeline network topology diagram, obtain all path segments through suspected contamination paths, and distinguish between monitored and unmonitored path segments.

[0074] Preferably, at least one online monitoring sensor is present at the endpoint of the monitored path segment; no online monitoring sensor is present at the endpoint of the unmonitored path segment.

[0075] Step S4-2: Based on all unmonitored path segments of the suspected contamination path, obtain the length of the longest continuous unmonitored path segment.

[0076] In suspected pollution pathways, unmonitored path segments indicate the existence of continuous, unknown paths. These unknown paths may exhibit further pollution changes. However, the unknown impact of shorter path segments is generally limited by adjacent monitored path segments, while longer, continuous path segments are more difficult to define. Therefore, the more consecutive unmonitored path segments there are, the greater the overall impact of shorter, continuous, or discontinuous monitored path segments on the entire pathway, and the greater the resulting unknown impact. Thus, the longest consecutive unmonitored path segment is selected to represent the overall unknown impact segment of the suspected pollution pathway.

[0077] Step S4-3: Determine the blind zone concentration of the suspected pollution path based on the length of the longest continuous unmonitored path segment and the total length of the suspected pollution path.

[0078] Since the longest unmonitored segment in a pollution path is usually an area where the direction of pollutants cannot be accurately determined, the longer the longest continuous unmonitored segment, the greater the impact on the final source tracing. In contrast, the shorter unmonitored segment has a smaller impact on the final source tracing. Therefore, the blind zone concentration is calculated by combining the length of the longest continuous unmonitored segment of a suspected pollution path with the total length of the suspected pollution path. The longer the longest continuous unmonitored segment, the greater the blind zone concentration.

[0079] Preferably, the calculation method for blind zone concentration can be written as: ,in represents the concentration of blind spots in suspected contamination pathways; exp() is an exponential function with the natural constant as the base. The longest consecutive unmonitored road segment of a suspected contamination route; Indicates the total length of the suspected contamination path; This represents the blind zone gain coefficient.

[0080] Preferably, the blind zone gain coefficient This data is obtained through historical analysis. As one example, confirmed historical pollution events and their corresponding actual propagation paths are collected. The statistical relationship between the blind zone concentration ratio of each path and the final source location error of the event (such as the estimated spatial distance between the source point and the actual source point) is calculated. The α value is determined through regression analysis to ensure that the gain model matches the actual observational uncertainty of a specific pipe network. In cases of system initialization or lack of calibration data, the coefficients... It can be set within an empirical range, such as 0.5 to 2.0, as an initial value for calculation.

[0081] Step S4-4: Based on the blind zone concentration of the suspected contamination path, a preset blind zone gain threshold is used to determine the blind zone gain of the contamination path.

[0082] Since a large blind zone may cause gain failure, a preset blind zone gain threshold is used as the upper limit of gain.

[0083] Preferably, the blind zone gain threshold is selected as 10 based on experience, but it can also be determined according to the specific implementation situation. This embodiment does not limit it.

[0084] Preferably, in some implementations of the present invention, by fusing the asynchronicity intensity and blind zone gain of the suspected pollution source and the suspected pollution path, an adaptive search radius for the suspected pollution source is obtained, and source tracing and localization are performed within the search radius, including: Step S5-1: Obtain the empirical probability of the historical related paths corresponding to the suspected pollution source point, obtain the asynchronicity intensity and blind zone gain of the suspected pollution path, and determine the search expansion coefficient of the suspected pollution source point by combining the suspected pollution path and the historical related paths corresponding to the suspected pollution source point.

[0085] Since each suspected pollution source corresponds to one or more suspected pollution paths, the search expansion coefficient of a suspected pollution source is obtained by integrating the empirical probability, asynchronicity intensity, and blind zone gain of all suspected pollution paths corresponding to a suspected pollution source. The larger the search expansion coefficient, the more uncontrollable the pollution propagation process of the suspected pollution source is. Therefore, the search expansion coefficient is proportional to the asynchronicity intensity and blind zone gain.

[0086] The empirical probability of a suspected contamination path is calculated in the same way as the empirical probability of historically related contamination paths, determined by the minimum activity level of all path segments of the suspected contamination path, which will not be elaborated here.

[0087] As one example, the search expansion coefficient is expressed by the formula: ,in This represents the search expansion coefficient of suspected pollution sources; This indicates the number of suspected pollution pathways corresponding to suspected pollution sources. This represents the empirical probability of a suspected contamination path k. Indicates the asynchronicity strength of the suspected contamination path k; This represents the normalized blind zone gain of the suspected contamination path k, where This is the blind zone gain after normalization (e.g., mapping to the [1, 2] interval via maximum and minimum normalization).

[0088] Step S5-2: Based on the search expansion coefficient of the suspected pollution source point, adjust the preset maximum radius to obtain the adaptive search radius of the suspected pollution source point.

[0089] Because the larger the search expansion coefficient of a suspected pollution source, the more uncontrollable the pollution propagation process of the suspected pollution source is (uncertain time, large blind zone), and the lower the possibility that there is a real pollution source near the suspected pollution source. Therefore, a larger search radius needs to be set to prevent omissions.

[0090] As one embodiment, the adaptive search radius size of the preset maximum radius is adjusted as follows: ,in Indicates the adaptive search radius of suspected pollution sources; This indicates the preset maximum search radius; This represents the search expansion coefficient of suspected pollution sources.

[0091] Preferably, the preset maximum search radius is obtained based on historical data. Based on historical pollution source tracing events, the search expansion coefficient for corresponding historical suspected pollution sources is predicted. The actual search radius is then used, based on the distance between historical actual pollution sources and suspected pollution sources. To obtain the true maximum search radius, where Indicates the true maximum search radius. This indicates the true search radius for historical pollution source tracing points. This represents the search expansion coefficient for historical pollution source tracing points. It is used when the system lacks historical case data. It can be set to an empirical default value based on the scale of a typical urban drainage network, such as 500 meters, and then dynamically updated after accumulating enough successful cases.

[0092] Preferably, after obtaining the adaptive search radius of the suspected pollution source, search resources that match the search radius are mobilized to locate the actual pollution source within the search radius, and the actual pollution path is gradually investigated based on the empirical probability of the suspected pollution path of the suspected pollution source.

[0093] In summary, this invention integrates multi-source pollution data to construct an uncertainty quantification framework for determining adaptive pollution source tracing and localization. By mining the empirical probabilities of relevant historical pollution paths through historical pollution events, suspected pollution sources and paths are obtained. Furthermore, the asynchronicity intensity and blind zone gain of each suspected pollution path are precisely quantified. In the final adaptive source tracing and localization process, the empirical probabilities, asynchronicity intensity, and blind zone gain of the suspected pollution paths are integrated to obtain an adaptive search radius, within which source tracing and localization are performed. This invention's final adaptive search radius allows for differentiated scheduling of search resources, reducing waste of tracing resources and lowering tracing costs.

[0094] It should be noted that the execution algorithm of the drainage network pollutant source tracing method provided herein is not inherently related to any particular computer, virtual system, or other device. Various general-purpose systems can also be used in conjunction with the teachings herein. Based on the above description, the required structure for constructing such a system is obvious. Furthermore, the embodiments of the present invention are not directed to any particular programming language. It should be understood that the content of the invention described herein can be implemented using various programming languages, and the above description of specific languages ​​is for the purpose of disclosing the best mode of implementation of the invention.

[0095] Numerous specific details are set forth in the specification provided herein. However, it will be understood that embodiments of the invention may be practiced without these specific details. In some instances, well-known methods, structures, and techniques have not been shown in detail so as not to obscure the understanding of this specification.

[0096] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0097] Example 2: This invention also proposes a pollutant tracing system for drainage pipe networks; please refer to [link / reference]. Figure 2 The diagram shows a structural diagram of a pollutant tracing system for a drainage network provided by an embodiment of the present invention. The system includes: a data acquisition module 301, a tracing analysis module 302, and a tracing location module 303.

[0098] The data acquisition module 301 is used to construct a multi-source pollution dataset, which includes: a pipeline connection topology map, monitoring points, real-time monitoring data of the monitoring points and their historical pollution source tracing events.

[0099] The source tracing analysis module 302 is used to select suspected pollution sources and corresponding suspected pollution paths in the pipeline connection topology map based on the historical pollution paths of the associated monitoring points, with real-time abnormal monitoring points as suspected endpoints; to obtain the asynchronous intensity of the suspected pollution path based on the distribution concentration of the historical pollution propagation time of the suspected pollution source points and the suspected pollution path based on the distribution of monitoring points on the suspected pollution path; and to obtain the blind zone gain of the suspected pollution path based on the distribution of monitoring points on the suspected pollution path.

[0100] The source tracing and localization module 303 is used to obtain the adaptive search radius of the suspected pollution source by fusing the asynchronous intensity and blind zone gain of the suspected pollution source and the suspected pollution path, and to perform source tracing and localization within the search radius.

[0101] It should be noted that the system provided in the above embodiments is only an example of the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer device can be divided into different functional modules to complete all or part of the functions described above. In addition, the embodiment of the method for tracing pollutants in drainage pipe networks provided in the above embodiments belongs to the same concept, and its specific implementation process is detailed in the method embodiments, which will not be repeated here.

[0102] Example 3: This invention also proposes a pollutant tracing device for drainage pipe networks; please refer to [link / reference]. Figure 3 The computer device includes a memory 401, a processor 402, and a computer program 403 stored in the memory 401 and running on the processor 402. When the processor 402 executes the computer program 403, the computer device can execute any of the aforementioned methods for tracing pollutants in drainage pipe networks.

[0103] Furthermore, embodiments of this application also protect an apparatus that may include a memory and a processor, wherein the memory stores executable program code, and the processor is used to call and execute the executable program code to perform a method for tracing pollutants in a drainage network provided in embodiments of this application.

[0104] This embodiment can divide the device into functional modules based on the above method example. For example, each module can correspond to a separate function, or two or more functions can be integrated into one processing module. The integrated module can be implemented in hardware. It should be noted that the module division in this embodiment is illustrative and only represents one logical functional division. In actual implementation, there may be other division methods.

[0105] When each module is divided according to its function, the device may also include a data acquisition module, a source tracing analysis module, a source tracing location module, etc. It should be noted that all relevant content of each step involved in the above method embodiments can be referenced to the functional description of the corresponding functional module, and will not be repeated here.

[0106] It should be understood that the apparatus provided in this embodiment is used to perform the above-described method for tracing the source of pollutants in a drainage network, and therefore can achieve the same effect as the above-described implementation method.

[0107] When using integrated units, the device may include a processing module and a storage module. When applied to a workpiece, the processing module can be used to control and manage the workpiece's operations. The storage module can be used to support the execution of program code by the workpiece.

[0108] The processing module may be a processor or a controller, which can implement or execute various exemplary logic blocks, modules, and circuits contained in conjunction with the disclosure of this application. The processor may also be a combination of functions that implement computing capabilities, such as a combination of one or more microprocessors, a combination of digital signal processing (DSP) and a microprocessor, etc., and the storage module may be a memory.

Claims

1. A method for tracing the source of pollutants in a drainage pipe network, characterized in that, include: A multi-source pollution dataset is constructed, which includes: a pipeline connection topology map, monitoring points, real-time monitoring data of the monitoring points, and their historical pollution source tracing events; the information of the historical pollution source tracing events includes: historical pollution monitoring points, historical pollution paths, and historical pollution source points. The historical pollution monitoring points are the monitoring points where pollution was first detected in the historical pollution source tracing events, the historical pollution source points are the monitoring points closest to the actual pollution source, and the pollution paths are the paths from the historical pollution source points to the historical pollution monitoring points. Using real-time abnormal monitoring points as suspected endpoints, suspected pollution sources and corresponding suspected pollution paths are screened out in the pipeline connection topology map based on the historical pollution paths of the associated monitoring points. The asynchronicity intensity of suspected pollution paths is obtained by analyzing the distribution concentration of pollution propagation time in the history of suspected pollution sources and corresponding suspected pollution paths. Based on the distribution of monitoring points along suspected pollution pathways, the blind zone gain of suspected pollution pathways is obtained; By fusing the asynchronicity intensity and blind zone gain of the suspected pollution source and the suspected pollution path, an adaptive search radius for the suspected pollution source is obtained, and source tracing and localization are performed within the search radius.

2. The method for tracing the source of pollutants in a drainage pipe network according to claim 1, characterized in that, The process of using real-time abnormal monitoring points as suspected endpoints and filtering suspected pollution sources and corresponding suspected pollution paths based on the historical pollution paths of associated monitoring points in the pipeline connection topology map includes: For the real-time abnormal monitoring point, by matching the historical pollution monitoring points of historical pollution source tracing events, all relevant historical pollution source tracing events that are the same as the real-time abnormal monitoring point are obtained, the historical pollution source points of the relevant historical pollution source tracing events are obtained, and all paths from the historical pollution source points to the real-time abnormal monitoring point are obtained as a set of historical relevant paths. Based on each historical related path in the historical related path set and the pipeline connection topology, all path segments of the historical related path are obtained, and the timestamps of all historical pollution events corresponding to all endpoints of each path segment are obtained and recorded as the timestamp set of the path segment. The real-time time of the real-time anomaly monitoring point is compared with the timestamp in the timestamp set of the path segment. Suspected pollution source points are screened out from all endpoints of the path segment, and the suspected pollution path is determined.

3. The method for tracing the source of pollutants in a drainage pipe network according to claim 2, characterized in that, The step of comparing the real-time time of the real-time anomaly monitoring point with the timestamps in the timestamp set of the path segment, filtering out suspected pollution source points from all endpoints of the path segment, and determining the suspected pollution path includes: Based on the pipeline connection topology, obtain the set of timestamps of all historical pollution events at the endpoints of the path segment, analyze the time proximity between the real time and the overall time set, and obtain the activity level of the path segment. Filter the activity of all path segments of each historically related path to determine the empirical probability of each historically related path; The endpoints of the paths with the highest empirical probability among all historical related paths are recorded as suspected pollution source points. Based on the pipeline connection topology, the set of reachable paths from all suspected pollution source points to the real-time abnormal monitoring points is denoted as the suspected pollution path corresponding to the suspected pollution source point.

4. The method for tracing the source of pollutants in a drainage pipe network according to claim 1, characterized in that, The method of obtaining the asynchronicity intensity of suspected pollution paths based on the historical distribution concentration of pollution propagation time corresponding to suspected pollution sources includes: Historical real-time monitoring data of the two endpoints of the suspected contamination path are obtained respectively, and the mutual information of the historical real-time monitoring data of the two endpoints of the suspected contamination path under different time shift multiples is calculated by aligning them. Based on the mutual information of the two endpoints of the suspected contamination path at different time shift multiples, curve fitting is performed to obtain the mutual information curve of the two endpoints of the suspected contamination path. The time shift factor corresponding to the maximum mutual information in the mutual information curves of the two endpoints of the suspected contamination path is taken as the optimal time shift factor; Calculate the time shift difference factor between the optimal time shift factor and each other time shift factor, and determine the peak neighborhood distribution concentration between the optimal time shift factor and each other time shift factor based on the time shift difference factor between the optimal time shift factor and each other time shift factor and the mutual information under each other time shift factor; The asynchronicity intensity of the suspected contamination path is determined based on the peak neighborhood distribution concentration between the optimal time shift factor and other time shift factors.

5. The method for tracing the source of pollutants in a drainage pipe network according to claim 1, characterized in that, The process of obtaining blind zone gain based on the distribution of monitoring points along suspected contamination pathways includes: Based on the pipeline connection topology map and the distribution of the monitoring points, all unmonitored path segments of suspected contamination paths are obtained; the unmonitored path segments are those where the monitoring points do not exist at the path endpoints. Based on all unmonitored path segments of the suspected contamination path, obtain the length of the longest consecutive unmonitored path segment; The blind zone concentration of the suspected pollution path is determined based on the length of the longest consecutive unmonitored path segment and the total length of the suspected pollution path; the blind zone concentration is directly proportional to the length of the longest consecutive unmonitored path segment and inversely proportional to the total length of the suspected pollution path. Based on the blind zone concentration of the suspected contamination path, a preset blind zone gain threshold is used to determine the blind zone gain of the contamination path.

6. The method for tracing the source of pollutants in a drainage pipe network according to claim 5, characterized in that, The specific method for obtaining the blind zone gain coefficient is as follows: Based on the historical pollution source tracing events, some historical pollution events on the suspected pollution path are randomly selected, and the average distance difference between the predicted pollution source point and the actual pollution source of the some historical pollution events is evaluated to obtain the blind zone gain coefficient.

7. The method for tracing the source of pollutants in a drainage pipe network according to claim 3, characterized in that, The adaptive search radius of the suspected pollution source point is obtained by fusing the asynchronicity intensity and blind zone gain of the suspected pollution source point and the suspected pollution path, including: The empirical probability of the historical related paths corresponding to the suspected pollution source point is obtained, the asynchronicity intensity and blind zone gain of the suspected pollution path are obtained, and the search expansion coefficient of the suspected pollution source point is determined by combining the suspected pollution path and the historical related paths corresponding to the suspected pollution source point. The adaptive search radius of the suspected pollution source is obtained by adjusting the preset maximum radius based on the search expansion coefficient of the suspected pollution source.

8. The method for tracing the source of pollutants in a drainage pipe network according to claim 7, characterized in that, The step of adjusting the preset maximum radius to obtain the adaptive search radius of the suspected pollution source point based on the search expansion coefficient of the suspected pollution source point includes: The adaptive search radius of the suspected pollution source is determined by the product of the source tracing probability of the suspected pollution source and the preset maximum search radius.

9. A source tracing device for pollutants in a drainage pipe network, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of a method for tracing the source of pollutants in a drainage pipe network as described in any one of claims 1 to 8.

10. A pollutant tracing system for drainage pipe networks, characterized in that, The system includes the following modules: Data acquisition module: used to construct a multi-source pollution dataset, which includes: a pipeline connection topology map, monitoring points, real-time monitoring data of the monitoring points, and their historical pollution source tracing events; the information of the historical pollution source tracing events includes: historical pollution monitoring points, historical pollution paths, and historical pollution source points. The historical pollution monitoring points are the monitoring points where pollution was first detected in the historical pollution source tracing events, the historical pollution source points are the monitoring points closest to the actual pollution source, and the pollution paths are the paths from the historical pollution source points to the historical pollution monitoring points. Source tracing analysis module: Used to select suspected pollution sources and corresponding suspected pollution paths in the pipeline connection topology map based on the historical pollution paths of the associated monitoring points, with real-time abnormal monitoring points as suspected endpoints; obtain the asynchronous intensity of suspected pollution paths based on the distribution concentration of the historical pollution propagation time of the suspected pollution sources and corresponding suspected pollution paths; obtain the blind zone gain of suspected pollution paths based on the distribution of monitoring points on suspected pollution paths. Source tracing and localization module: used to obtain the adaptive search radius of the suspected pollution source by fusing the asynchronous intensity and blind zone gain of the suspected pollution source and the suspected pollution path, and to perform source tracing and localization within the search radius.

Images