Water pollution tracing method and system based on pollution propagation directed graph

Abstract

The application discloses a water quality pollution tracing method and system based on a pollution propagation directed graph. After identifying a water quality pollution event, the water quality pollution tracing method screens out monitoring points located upstream of a pollution point as candidate points, calculates the correlation between the water quality monitoring data of each candidate point and the water quality monitoring data of an adjacent downstream candidate point, and constructs a pollution propagation directed graph based on the correlation calculation result. The pollution propagation directed graph can accurately reflect the propagation path of pollutants in the river network topological structure. Based on the pollution propagation directed graph, pollution sources are clustered. The clustering result can clearly reveal the propagation path and propagation direction of pollutants in the river network topological structure, automatic division of pollution propagation groups is realized, the pollution source can be quickly and accurately located, and the accuracy and reliability of the water quality pollution tracing result are improved.

Description

Technical Field

[0001] This invention relates to the field of pollution source tracing technology, and in particular, to a water pollution source tracing method and system, electronic equipment, and computer-readable storage medium based on a directed graph of pollution propagation. Background Technology

[0002] Water pollution source tracing typically refers to tracking and locating the source of pollutants when abnormalities occur in river water quality. Only by achieving efficient and accurate source tracing can environmental supervision be fundamentally strengthened and pollution input be promptly blocked. In existing technologies, common water pollution source tracing methods mainly rely on manual, periodic water sampling and analysis at pre-set monitoring points. Water quality is assessed based on monitoring data. These monitoring points are usually located at existing automatic water quality monitoring stations or sampling stations. Once an abnormality is detected, environmental personnel often need to rely on experience to infer possible polluting enterprises and conduct on-site verification, making the source tracing process time-consuming and labor-intensive. In addition, existing technologies have proposed some automatic water pollution source tracing methods, such as water quality fingerprinting, pollutant concentration inference, and machine learning-based source tracing models. However, water quality fingerprinting and pollutant concentration inference rely heavily on prior knowledge and manual features, resulting in low source tracing accuracy. Existing source tracing models treat monitoring stations as independent entities, ignoring the influence of river topology and flow direction, leading to logical errors in the source tracing results and poor accuracy and reliability. Summary of the Invention

[0003] This invention provides a water pollution source tracing method and system based on a directed graph of pollution propagation, an electronic device, and a computer-readable storage medium, which can improve the accuracy and reliability of water pollution source tracing results.

[0004] According to one aspect of the present invention, a method for tracing water pollution sources based on a directed graph of pollution propagation is provided, comprising the following: Water pollution events can be identified and pollution sites can be determined based on water quality monitoring data from various monitoring points. Upstream monitoring points of pollution sites are selected as candidate sites for potential pollution sources. The correlation between water quality monitoring data of each candidate site and water quality monitoring data of adjacent downstream candidate sites is calculated, and a directed graph of pollution propagation is constructed based on the correlation calculation results. Pollution sources were clustered based on the directed graph of pollution propagation, and water pollution source tracing analysis was conducted based on the pollution source clustering results to determine the pollution source of this water pollution incident.

[0005] Furthermore, the process of calculating the correlation between the water quality monitoring data of each candidate point and the water quality monitoring data of adjacent downstream candidate points, and constructing a directed graph of pollution propagation based on the correlation calculation results, includes the following: Calculate the transmission lag time of each candidate point relative to the contaminated point; Time offset is performed based on the transmission lag time corresponding to each candidate point to set a search window and obtain the water quality monitoring data sequence of each candidate point within the search window; Align the water quality monitoring data sequences of each candidate point and the adjacent downstream candidate points within the search window during the pollution occurrence period, and calculate the comprehensive similarity coefficient between the two water quality monitoring data sequences. The comprehensive similarity coefficient is compared with the preset similarity threshold. If the comprehensive similarity coefficient is greater than the similarity threshold, a directed edge is constructed from the current candidate point to the adjacent downstream candidate point. The weight of the directed edge is set to the comprehensive similarity coefficient. After traversing all candidate points, a directed graph of pollution propagation is constructed.

[0006] Furthermore, the comprehensive similarity coefficient is calculated based on the following formula: ; ; ; in, S Represents the overall similarity coefficient. r This represents the Pearson correlation coefficient. Represents the Spearman correlation coefficient. , Indicates the weighting coefficient. This indicates the first water quality monitoring data sequence of the current candidate site. k Each water quality monitoring data value, This represents the first water quality monitoring data sequence of adjacent downstream candidate points. k Each water quality monitoring data value, This represents the average value of the water quality monitoring data sequence for the current candidate location. This represents the average value of water quality monitoring data sequences from adjacent downstream candidate sites. Indicates the length of the water quality monitoring data sequence. express The rank of the water quality monitoring data sequence at the current candidate site. express Rank in the water quality monitoring data sequence of adjacent downstream candidate sites This represents the rank mean of the water quality monitoring data sequence for the current candidate location. This represents the rank mean of the water quality monitoring data sequence of adjacent downstream candidate points.

[0007] Furthermore, the process of calculating the transmission lag time of each candidate point relative to the contaminated point includes the following: The upstream of the pollution point is divided into multiple upstream river segments. For each upstream river segment, the lag time required for pollutants to be transmitted from the upstream candidate node to the adjacent downstream candidate node within that upstream river segment is calculated. Based on the number of upstream river segments corresponding to each candidate node, the lag time is accumulated segment by segment to obtain the transmission lag time of each candidate point relative to the pollution point.

[0008] Furthermore, the process of identifying water pollution events based on water quality monitoring data from various monitoring points includes the following: Based on the current water quality monitoring data of any monitoring point, determine whether there is a decline in the water quality category level, and calculate the relative change rate of the current water quality monitoring data of any monitoring point compared with the middle value of the sliding window formed by the historical water quality monitoring data of the past several days. If the relative change rate is greater than the preset relative change rate threshold and there is a decline in the water quality category level, then it is determined that a water pollution event has occurred at that monitoring point.

[0009] Furthermore, the process of clustering pollution sources based on the directed graph of pollution propagation includes the following: Construct a random walk transition probability matrix based on a directed graph of pollution propagation; Singular value decomposition is performed on the random walk transition probability matrix to obtain the right singular vector matrix, and the first few elements of the right singular vector matrix are selected. d Given a right singular vector, we obtain the low-dimensional embedding row vector for each candidate point. All low-dimensional embedding row vectors are normalized, and pollution source clustering is performed based on the normalized low-dimensional embedding row vectors.

[0010] Furthermore, the process of conducting water pollution source tracing analysis based on pollution source clustering results to determine the pollution source of this water pollution incident includes the following: For pollution groups obtained by pollution source clustering, starting from the downstream candidate nodes with significant pollution characteristics in the pollution group, we trace back upstream step by step until we locate the earliest candidate point with the same pollution characteristics and no upstream pollution input, and then determine it as the pollution source of this pollution event.

[0011] In addition, the present invention also provides a water pollution source tracing system based on a directed graph of pollution propagation, comprising: The water pollution incident identification module is used to identify water pollution incidents and determine the pollution points based on water quality monitoring data from various monitoring points. The pollution propagation directed graph construction module is used to screen upstream monitoring points of pollution points as candidate points of potential pollution sources, calculate the correlation between water quality monitoring data of each candidate point and water quality monitoring data of adjacent downstream candidate points, and construct a pollution propagation directed graph based on the correlation calculation results. The water pollution source tracing and analysis module is used to cluster pollution sources based on the directed graph of pollution propagation, and to perform water pollution source tracing analysis based on the pollution source clustering results to determine the pollution source of this water pollution incident.

[0012] In addition, the present invention also provides an electronic device, including a processor and a memory, wherein the memory stores a computer program, and the processor executes the steps of the method described above by calling the computer program stored in the memory.

[0013] In addition, the present invention provides a computer-readable storage medium for storing a computer program for tracing water pollution sources based on a directed graph of pollution propagation, wherein the computer program executes the steps of the method described above when running on a computer.

[0014] The present invention has the following beneficial effects: The water pollution source tracing method based on a directed pollution propagation graph of the present invention, after identifying a water pollution event, first selects monitoring points upstream of the pollution point as candidate points, calculates the correlation between the water quality monitoring data of each candidate point and the water quality monitoring data of adjacent downstream candidate points, and constructs a directed pollution propagation graph based on the correlation calculation results. This graph can accurately reflect the propagation path of pollutants in the river network topology. Then, pollution source clustering is performed based on the directed pollution propagation graph. The clustering results can clearly reveal the propagation path and direction of pollutants in the river network topology, realize the automatic division of pollution propagation groups, and quickly and accurately locate pollution sources, thereby improving the accuracy and reliability of water pollution source tracing results.

[0015] In addition, the water pollution source tracing system based on the pollution propagation directed graph of the present invention also has the above-mentioned advantages.

[0016] In addition to the objectives, features, and advantages described above, the present invention has other objectives, features, and advantages. The invention will now be described in further detail with reference to the figures. Attached Figure Description

[0017] The accompanying drawings, which form part of this application, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings: Figure 1 This is a schematic flowchart of a water pollution source tracing method based on a directed graph of pollution propagation, according to a preferred embodiment of this application. Figure 2 yes Figure 1 A schematic diagram of the sub-process of step S2; Figure 3 yes Figure 1A schematic diagram of the sub-process of step S3; Figure 4 This is a schematic diagram of the module structure of a water pollution source tracing system based on a directed graph of pollution propagation, according to another embodiment of this application. Detailed Implementation

[0018] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0019] Reference Figure 1 A preferred embodiment of this application provides a water pollution source tracing method based on a directed graph of pollution propagation, including the following: Step S1: Identify water pollution events and determine the pollution locations based on water quality monitoring data from each monitoring point; Step S2: Select upstream monitoring points of the pollution point as candidate points of potential pollution sources, calculate the correlation between the water quality monitoring data of each candidate point and the water quality monitoring data of adjacent downstream candidate points, and construct a directed graph of pollution propagation based on the correlation calculation results. Step S3: Cluster pollution sources based on the directed graph of pollution propagation, and conduct source tracing analysis of water pollution based on the pollution source clustering results to determine the pollution source of this water pollution incident.

[0020] It is understood that the water pollution source tracing method based on the directed pollution propagation graph in this embodiment, after identifying a water pollution event, first selects monitoring points upstream of the pollution point as candidate points, calculates the correlation between the water quality monitoring data of each candidate point and the water quality monitoring data of adjacent downstream candidate points, and constructs a directed pollution propagation graph based on the correlation calculation results. This graph can accurately reflect the propagation path of pollutants in the river network topology. Then, pollution source clustering is performed based on the directed pollution propagation graph. The clustering results can clearly reveal the propagation path and direction of pollutants in the river network topology, realize the automatic division of pollution propagation groups, and quickly and accurately locate pollution sources, thereby improving the accuracy and reliability of water pollution source tracing results.

[0021] In step S1, the water quality monitoring data for each monitoring point can use at least one water quality indicator selected from pH, dissolved oxygen, turbidity, conductivity, ammonia nitrogen, total phosphorus, total nitrogen, and permanganate index. By analyzing the water quality monitoring data from each monitoring point, water pollution events can be automatically identified, and the pollution points, i.e., the monitoring points where the water pollution events occurred, can be determined. Furthermore, monitoring points are typically set at fixed monitoring sections and sewage outlets.

[0022] In addition, the process of identifying water pollution events based on water quality monitoring data from various monitoring points includes the following: Based on the current water quality monitoring data of any monitoring point, it is determined whether a decline in water quality category has occurred. The relative rate of change of the current water quality monitoring data at any monitoring point compared to the median of a sliding window composed of historical water quality monitoring data from the past few days is calculated. If the relative rate of change is greater than a preset relative rate of change threshold, and a decline in water quality category has occurred, then a water pollution event is determined to have occurred at that monitoring point. The condition that the relative rate of change is greater than the preset relative rate of change threshold can be expressed as: ; in, This indicates that a certain monitoring point is in t Real-time water quality monitoring data, This represents the median of a sliding window composed of historical water quality monitoring data from the past few days. Represents the smoothing constant. This represents the relative rate of change threshold, which can be adjusted based on the characteristics of the water quality indicator or historical data from the monitoring station. Historical water quality monitoring data can be selected from the past 3, 5, or 7 days; the specific choice depends on the actual situation and is not limited here. For example, taking historical water quality monitoring data from the past three days (72 hours) can be represented as... In addition, when water quality monitoring data includes at least two water quality indicators, a water pollution event is determined to have occurred if the relative change rate of any one of the water quality indicators is greater than the preset relative change rate threshold and a decline in the water quality category level occurs.

[0023] It is understandable that existing technologies typically compare water quality monitoring data from each sampling (e.g., hourly) with water quality classification standards to determine if the water quality category level has decreased. If it has decreased, a water pollution event is identified. The accuracy of this method for identifying water pollution events needs improvement because even if the water quality category level decreases by one level, the individual water quality monitoring data may not show significant fluctuations. The decrease in the water quality category level may be due to small fluctuations in multiple monitoring data points simultaneously, but this does not necessarily indicate a water pollution event. This invention, in addition to determining whether the water quality category has decreased, also calculates the relative rate of change of the current water quality monitoring data compared to the median of a sliding window composed of historical water quality monitoring data from the past few days. If the relative rate of change is greater than a preset relative rate of change threshold, and a decrease in the water quality category level also occurs, a water pollution event is identified. This method can sensitively detect situations where water quality monitoring parameters deteriorate significantly. When both a decrease in the water quality category level and at least one significantly deteriorated water quality monitoring data point occur simultaneously, it means that a water pollution event has definitely occurred, thus improving the accuracy of water pollution event identification.

[0024] In other embodiments of the present invention, a change quantity index can be used to replace the aforementioned relative change rate index. For example, the change quantity between the current water quality monitoring data of any monitoring point and the median of the sliding window formed by the historical water quality monitoring data of the past several days can be calculated. If the change quantity is greater than a preset change quantity threshold, and at the same time a decrease in the water quality category level occurs, a water pollution event is determined to have occurred. The change quantity being greater than the preset change quantity threshold can be expressed as: , This indicates a preset threshold for the amount of change, which can be adjusted according to the characteristics of the water quality indicator. For example, for dissolved oxygen, A concentration of 1 mg / L can be used for ammonia nitrogen. A concentration of 0.2 mg / L can be used for total nitrogen. A concentration of 0.5 mg / L can be used for total phosphorus. A concentration of 0.03 mg / L is acceptable for turbidity. 50 NTU can be used for conductivity. A value of 50 μS / cm can be used for the permanganate index. A concentration of 1.5 mg / L can be taken, depending on the pH. 0.5 is acceptable.

[0025] In addition, before collecting water quality monitoring data from each monitoring point, the monitoring instruments at the monitoring points must be subject to quality control verification. Only if the monitoring instruments pass the quality control verification can the collected data be considered valid data; otherwise, it will be considered invalid data.

[0026] In addition, in step S2, based on the river main and tributary network topology and combined with GIS water system data, multiple monitoring points located upstream of the pollution point are selected as candidate points for potential pollution sources, thus obtaining a set of candidate points for potential pollution sources. Optionally, to improve efficiency, upstream monitoring points within a preset distance from the pollution point can be used as candidate points. This preset distance can be set as needed, and its value is generally in the range of 50km to 200km. Then, the correlation between the water quality monitoring data of each candidate point and the water quality monitoring data of adjacent downstream candidate points is calculated, that is, the correlation between the water quality monitoring data of two adjacent candidate points is calculated to assess the pollution association strength between two adjacent candidate points, and a directed pollution propagation graph is constructed based on the correlation calculation results to accurately reflect the propagation path of pollutants in the river network topology.

[0027] Among them, such as Figure 2 As shown, the process of calculating the correlation between the water quality monitoring data of each candidate point and the water quality monitoring data of adjacent downstream candidate points, and constructing a directed graph of pollution propagation based on the correlation calculation results, includes the following: Step S21: Calculate the transmission lag time of each candidate point relative to the contaminated point; Step S22: Perform time offset based on the transmission lag time corresponding to each candidate point to set a search window and obtain the water quality monitoring data sequence of each candidate point within the search window; Step S23: Align the water quality monitoring data sequences of each candidate point and the adjacent downstream candidate point within the search window during the pollution occurrence period, and calculate the comprehensive similarity coefficient between the two water quality monitoring data sequences. Step S24: Compare the comprehensive similarity coefficient with the preset similarity threshold. If the comprehensive similarity coefficient is greater than the similarity threshold, construct a directed edge from the current candidate point to the adjacent downstream candidate point. The weight of the directed edge is set to the comprehensive similarity coefficient. After traversing all candidate points, a directed graph of pollution propagation is constructed.

[0028] Specifically, based on the river network topology and the locations of multiple candidate points, the upstream of each pollution point is divided into multiple upstream segments, meaning that each pair of adjacent candidate points forms an upstream segment. For each upstream segment, the average flow velocity is calculated using Manning's formula, which is expressed as: , Indicates the average flow velocity. R Indicates the hydraulic radius. S The gradient is represented by the water level gradient. Then, the lag time required for pollutants to travel from the upstream candidate point to the downstream candidate point in the upstream river segment is calculated using the following formula: , Indicates the lag time. L This represents the length of the upstream river segment. Finally, based on the number of upstream river segments corresponding to each candidate node, the lag time is segmented and accumulated to obtain the position of each candidate point. i Relative to pollution sites j Total transmission latency .

[0029] Then, taking into account the uncertainties such as velocity fluctuations, diffusion, and degradation in actual rivers, the present invention considers the total transport lag time... Based on this, a time offset was applied, and a search window was set with a range of: A represents the time offset, typically 5-12 hours, which allows us to obtain the water quality monitoring data sequence for each candidate location within the search window. ,in, Indicates candidate nodes i exist tThe water quality monitoring data at any given time (e.g., data on 8 water quality indicators) can be effectively covered by a search window to cover uncertainties caused by fluctuations in water flow velocity, diffusion, degradation, and monitoring errors, avoiding bias from data at a single moment. Furthermore, for ease of calculation, the water quality monitoring data has been pre-processed using min-max standardization.

[0030] Next, the water quality monitoring data sequences of each candidate point and its adjacent downstream candidate point within the search window are aligned within the time period of pollution occurrence, ensuring that the two water quality monitoring data sequences are of the same length and time-aligned. Then, the comprehensive similarity coefficient between the two water quality monitoring data sequences is calculated. Optionally, the comprehensive similarity coefficient can be calculated based on the following formula: ; ; ; Among them, among them, S Represents the overall similarity coefficient. r This represents the Pearson correlation coefficient. Represents the Spearman correlation coefficient. , Indicates the weighting coefficient. This indicates the first water quality monitoring data sequence of the current candidate site. k Each water quality monitoring data value, This represents the first water quality monitoring data sequence of adjacent downstream candidate points. k Each water quality monitoring data value, This represents the average value of the water quality monitoring data sequence for the current candidate location. This represents the average value of water quality monitoring data sequences from adjacent downstream candidate sites. Indicates the length of the water quality monitoring data sequence. express The rank of the water quality monitoring data sequence at the current candidate site. express Rank in the water quality monitoring data sequence of adjacent downstream candidate sites This represents the rank mean of the water quality monitoring data sequence for the current candidate location. This represents the rank mean of the water quality monitoring data sequence of adjacent downstream candidate points.

[0031] It is understood that this invention uses the Pearson correlation coefficient to capture the consistency of the overall trend of pollutant concentration, and the Spearman correlation coefficient to capture the consistency of extreme values, abrupt change points and ranking order. It is more sensitive to nonlinear relationships and abnormal peaks. Then, the two are weighted and fused to comprehensively evaluate the similarity between two water quality monitoring data sequences from the perspective of overall trend consistency and local feature consistency. This can accurately assess the pollution association strength between two adjacent candidate nodes, thereby improving the accuracy of pollutant transmission path identification.

[0032] Finally, the overall similarity coefficient will be used. S Similarity threshold with preset By comparison, if the comprehensive similarity coefficient is greater than the similarity threshold, it is considered that the pollution association between the current candidate node and the adjacent downstream candidate node is strong. Then, a directed edge is constructed from the current candidate point to the adjacent downstream candidate point. The weight of the directed edge is set to the corresponding comprehensive similarity coefficient. After traversing all candidate points, that is, after completing the pollution association strength assessment between all adjacent candidate points, a directed graph of pollution propagation can be constructed. This directed graph can accurately reflect the propagation path of pollutants in the river network topology.

[0033] In addition, such as Figure 3 As shown, in step S3, the process of clustering pollution sources based on the directed graph of pollution propagation includes the following: Step S31: Construct a random walk transition probability matrix based on the directed graph of pollution propagation; Step S32: Perform singular value decomposition on the random walk transition probability matrix to obtain the right singular vector matrix, and select the first few elements from the right singular vector matrix. d Given a right singular vector, we obtain the low-dimensional embedding row vector for each candidate point. Step S33: Normalize all low-dimensional embedding row vectors and perform pollution source clustering based on the normalized low-dimensional embedding row vectors.

[0034] Specifically, according to the random walk theory, we first construct the random walk transition probability matrix based on the directed graph of contamination propagation, which can be expressed as: ,in, This represents the random walk transition probability matrix. This represents the weight matrix of directed edges, where the elements are... Indicates the upstream candidate point i Pointing to adjacent downstream candidate points j The weight of the directed edge. D Represent an out-degree diagonal matrix, whose diagonal elements are , indicating the first i The sum of the weights of all outgoing edges of a node, with all non-diagonal elements being 0.

[0035] Then, the random walk transition probability matrix Singular value decomposition can be represented as: , U Let represent the left singular vector matrix, where the left singular vectors represent the probability distribution of a given node reaching various nodes in the network after a long random walk. Represents a singular value matrix. V Let represent the right singular vector matrix, where the right singular vectors represent the probability distribution of data originating from a certain node and ultimately converging to another node. Since this invention focuses on which upstream nodes influence the current contamination level of a node, it selects the right singular vector matrix to capture upstream propagation relationships. This is achieved by selecting the first... d From the right singular vectors, the random walk transition probability matrix can be extracted. The main information is used to obtain the low-dimensional embedding row vector for each candidate point, which can be represented as: , Indicates candidate nodes i The low-dimensional embedded row vectors, Represents a right singular vector matrix V The i Before the journey d Each element.

[0036] Next, to eliminate the magnitude differences between the low-dimensional embedding row vectors of each candidate node and enhance the robustness of subsequent clustering, L2 normalization is performed on each low-dimensional embedding row vector, which can be expressed as: , Indicates candidate nodes i After normalizing the low-dimensional embedding row vectors, the DBSCAN clustering algorithm is used to cluster multiple normalized low-dimensional embedding row vectors to obtain the pollution source clustering results. The clustering process is as follows: (1) Initialization: All node data is marked as "unvisited", and the pollution source group label is initialized to C = 0; (2) Select the starting point: Randomly select an unvisited node p and mark p as "visited"; (3) Neighborhood query: Query all data points contained in the ε-neighborhood of node p (i.e., the spherical region with radius ε centered at p); (4) Determine the core point: If the number of points in the ε-neighborhood of p is greater than or equal to the preset minimum number of points MinPts, then p is marked as a core point and C is assigned to the pollution source group label of p to initialize a new pollution source group and execute step (5); otherwise, p is marked as a noise point (temporarily assigned label -1) and jumps to step (2) to continue processing the next unvisited point; (5) Neighborhood expansion: Add all points in the ε-neighborhood of p to the candidate queue Q; (6) Cluster expansion: Take points q from Q in sequence. If q is "unvisited", mark it as "visited" and query its ε-neighborhood. If the number of points in the ε-neighborhood of q is greater than or equal to MinPts, mark q as a core point and add the points in its ε-neighborhood that have not been added to Q to Q. If q has not been assigned any group label, assign C to q. (7) Complete the current cluster: When Q is empty, the current pollution source group C is completed. Let C = C + 1, and return to step (2). (8) Termination condition: Repeat steps (2) to (7) until all data points have been accessed.

[0037] Furthermore, for pollution groups obtained by clustering pollution sources, starting from the downstream candidate nodes with significant pollution characteristics (i.e., the largest relative rate of change or amount of change) in the pollution group, the process traces upstream step by step until the earliest candidate point showing the same pollution characteristics without upstream pollution input is located, and this point is identified as the pollution source of this pollution event. For example, in a pollution group, if a sewage outlet is located at the upstream end of the pollution group and there are no other effective monitoring sections or emission points upstream, then the sewage outlet can be directly identified as the main responsible source. If a point is simultaneously affected by the combined effects of tributary sewage outlets and main stream pollution, multiple source nodes with zero in-degree (i.e., no upstream pollution input) may appear in the pollution group. These source nodes are the earliest locations to show the same pollution characteristics, and therefore, together they constitute the main responsible source of this complex pollution event.

[0038] In addition, such as Figure 4 As shown, another embodiment of the present invention also provides a water pollution source tracing system based on a directed graph of pollution propagation, preferably employing the water pollution source tracing method based on a directed graph of pollution propagation as described above, including: The water pollution incident identification module is used to identify water pollution incidents and determine the pollution points based on water quality monitoring data from various monitoring points. The pollution propagation directed graph construction module is used to screen upstream monitoring points of pollution points as candidate points of potential pollution sources, calculate the correlation between water quality monitoring data of each candidate point and water quality monitoring data of adjacent downstream candidate points, and construct a pollution propagation directed graph based on the correlation calculation results. The water pollution source tracing and analysis module is used to cluster pollution sources based on the directed graph of pollution propagation, and to perform water pollution source tracing analysis based on the pollution source clustering results to determine the pollution source of this water pollution incident.

[0039] It is understood that the water pollution source tracing system based on the directed pollution propagation graph in this embodiment, after identifying a water pollution event, first selects monitoring points upstream of the pollution point as candidate points, calculates the correlation between the water quality monitoring data of each candidate point and the water quality monitoring data of adjacent downstream candidate points, and constructs a directed pollution propagation graph based on the correlation calculation results. This graph can accurately reflect the propagation path of pollutants in the river network topology. Then, pollution source clustering is performed based on the directed pollution propagation graph. The clustering results can clearly reveal the propagation path and direction of pollutants in the river network topology, realize the automatic division of pollution propagation groups, and quickly and accurately locate pollution sources, thereby improving the accuracy and reliability of water pollution source tracing results.

[0040] In addition, another embodiment of the present invention provides an electronic device including a processor and a memory, wherein the memory stores a computer program, and the processor executes the steps of the method described above by calling the computer program stored in the memory.

[0041] In addition, another embodiment of the present invention provides a computer-readable storage medium for storing a computer program for tracing water pollution sources based on a directed graph of pollution propagation, wherein the computer program executes the steps of the method described above when running on a computer.

[0042] Common computer-readable storage media include: floppy disks, flexible disks, hard disks, magnetic tapes, any other magnetic media, CD-ROMs, any other optical media, punch cards, paper tape, any other physical media with perforated patterns, random access memory (RAM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), flash erasable programmable read-only memory (FLASH-EPROM), any other memory chips or cartridges, or any other media readable by a computer. Instructions may further be transmitted or received by a transmission medium. The term transmission medium can include any tangible or intangible medium used to store, encode, or carry instructions for execution by a machine, and includes digital or analog carrier communication signals or intangible media that facilitate communication of such instructions. Transmission media include coaxial cables, copper wires, and optical fibers, which contain conductors for transmitting a bus of computer data signals.

[0043] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of this application can be implemented in various computer languages, such as the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0044] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0045] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0046] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0047] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0048] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

[0049] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for tracing water pollution sources based on a directed graph of pollution propagation, characterized in that, Includes the following: Water pollution events can be identified and pollution sites can be determined based on water quality monitoring data from various monitoring points. Upstream monitoring points of pollution sites are selected as candidate sites for potential pollution sources. The correlation between water quality monitoring data of each candidate site and water quality monitoring data of adjacent downstream candidate sites is calculated, and a directed graph of pollution propagation is constructed based on the correlation calculation results. Pollution sources were clustered based on the directed graph of pollution propagation, and water pollution source tracing analysis was conducted based on the pollution source clustering results to determine the pollution source of this water pollution incident.

2. The water pollution source tracing method based on a directed graph of pollution propagation as described in claim 1, characterized in that, The process of calculating the correlation between the water quality monitoring data of each candidate point and the water quality monitoring data of adjacent downstream candidate points, and constructing a directed graph of pollution propagation based on the correlation calculation results, includes the following: Calculate the transmission lag time of each candidate point relative to the contaminated point; Time offset is performed based on the transmission lag time corresponding to each candidate point to set a search window and obtain the water quality monitoring data sequence of each candidate point within the search window; Align the water quality monitoring data sequences of each candidate point and the adjacent downstream candidate points within the search window during the pollution occurrence period, and calculate the comprehensive similarity coefficient between the two water quality monitoring data sequences. The comprehensive similarity coefficient is compared with the preset similarity threshold. If the comprehensive similarity coefficient is greater than the similarity threshold, a directed edge is constructed from the current candidate point to the adjacent downstream candidate point. The weight of the directed edge is set to the comprehensive similarity coefficient. After traversing all candidate points, a directed graph of pollution propagation is constructed.

3. The water pollution source tracing method based on a directed graph of pollution propagation as described in claim 2, characterized in that, The overall similarity coefficient is calculated based on the following formula: ； ； ； in, S Represents the overall similarity coefficient. r This represents the Pearson correlation coefficient. Represents the Spearman correlation coefficient. , Indicates the weighting coefficient. This indicates the first water quality monitoring data sequence of the current candidate site. k Each water quality monitoring data value, This represents the first water quality monitoring data sequence of adjacent downstream candidate points. k Each water quality monitoring data value, This represents the average value of the water quality monitoring data sequence for the current candidate location. This represents the average value of water quality monitoring data sequences from adjacent downstream candidate sites. Indicates the length of the water quality monitoring data sequence. express The rank of the water quality monitoring data sequence at the current candidate site. express Rank in the water quality monitoring data sequence of adjacent downstream candidate sites This represents the rank mean of the water quality monitoring data sequence for the current candidate location. This represents the rank mean of the water quality monitoring data sequence of adjacent downstream candidate points.

4. The water pollution source tracing method based on a directed graph of pollution propagation as described in claim 2, characterized in that, The process of calculating the transmission lag time of each candidate point relative to the contaminated point includes the following: The upstream of the pollution point is divided into multiple upstream river segments. For each upstream river segment, the lag time required for pollutants to be transmitted from the upstream candidate node to the adjacent downstream candidate node within that upstream river segment is calculated. Based on the number of upstream river segments corresponding to each candidate node, the lag time is accumulated segment by segment to obtain the transmission lag time of each candidate point relative to the pollution point.

5. The water pollution source tracing method based on a directed graph of pollution propagation as described in claim 1, characterized in that, The process of identifying water pollution events based on water quality monitoring data from various monitoring points includes the following: Based on the current water quality monitoring data of any monitoring point, determine whether there is a decline in the water quality category level, and calculate the relative change rate of the current water quality monitoring data of any monitoring point compared with the middle value of the sliding window formed by the historical water quality monitoring data of the past several days. If the relative change rate is greater than the preset relative change rate threshold and there is a decline in the water quality category level, then it is determined that a water pollution event has occurred at that monitoring point.

6. The water pollution source tracing method based on a directed graph of pollution propagation as described in claim 1, characterized in that, The process of clustering pollution sources based on the directed graph of pollution propagation includes the following: Construct a random walk transition probability matrix based on a directed graph of pollution propagation; Singular value decomposition is performed on the random walk transition probability matrix to obtain the right singular vector matrix, and the first few elements of the right singular vector matrix are selected. d Given a right singular vector, we obtain the low-dimensional embedding row vector for each candidate point. All low-dimensional embedding row vectors are normalized, and pollution source clustering is performed based on the normalized low-dimensional embedding row vectors.

7. The water pollution source tracing method based on a directed graph of pollution propagation as described in claim 1, characterized in that, The process of conducting water pollution source tracing analysis based on pollution source clustering results to determine the pollution source of this water pollution incident includes the following: For pollution groups obtained by pollution source clustering, starting from the downstream candidate nodes with significant pollution characteristics in the pollution group, we trace back upstream step by step until we locate the earliest candidate point with the same pollution characteristics and no upstream pollution input, and then determine it as the pollution source of this pollution event.

8. A water pollution source tracing system based on a directed graph of pollution propagation, characterized in that, include: The water pollution incident identification module is used to identify water pollution incidents and determine the pollution points based on water quality monitoring data from various monitoring points. The pollution propagation directed graph construction module is used to screen upstream monitoring points of pollution points as candidate points of potential pollution sources, calculate the correlation between water quality monitoring data of each candidate point and water quality monitoring data of adjacent downstream candidate points, and construct a pollution propagation directed graph based on the correlation calculation results. The water pollution source tracing and analysis module is used to cluster pollution sources based on the directed graph of pollution propagation, and to perform water pollution source tracing analysis based on the pollution source clustering results to determine the pollution source of this water pollution incident.

9. An electronic device, characterized in that, The method includes a processor and a memory, wherein the memory stores a computer program, and the processor executes the steps of the method as described in any one of claims 1 to 7 by calling the computer program stored in the memory.

10. A computer-readable storage medium for storing a computer program for tracing water pollution sources based on a directed graph of pollution propagation, characterized in that, The computer program, when run on a computer, performs the steps of the method as described in any one of claims 1 to 7.

Images