A method and system for tracing and accurately locating site pollution sources

Abstract

The present invention relates to the technical field of contaminated site investigation and monitoring, and discloses a method and system for tracing and accurately locating site pollution sources. The method comprises laying out initial detection points in the area to be detected and obtaining soil detection information, generating soil impact values, optimizing the layout of the detection points, and then obtaining water quality detection data and water level elevation data of the target detection points. Combining these data, a target flow field map of groundwater is constructed, and the flow field map is input into a pre-constructed pollution tracing model to ultimately determine the location of the pollution source. By combining the physical and chemical properties of the soil, groundwater detection data, and water level elevation data, the migration path of pollutants in the soil and groundwater is comprehensively simulated. This multi-level data integration and optimization analysis method significantly improves the accuracy of pollution source tracing and effectively solves the problem of inaccurate positioning caused by relying solely on groundwater detection data in the prior art.

Description

Technical Field

[0001] The present invention relates to the technical field of contaminated site investigation and monitoring, and more specifically, to a method and system for tracing and accurately locating site pollution sources. Background Art

[0002] The source tracing analysis of surface soil and groundwater pollution in public areas is particularly important. Although relevant source tracing technologies have been involved in existing technologies, there are still some shortcomings.

[0003] For example, the patent application with publication number CN114236075A provides a groundwater pollution visualization monitoring and early warning system and method. The technical solution includes an acquisition device, a data transmission device, a cloud server and a display terminal. The system collects data such as groundwater organic matter and heavy metals. The cloud server stores, analyzes, visualizes and warns, and the display terminal displays the results, thereby improving the timeliness and quality of the early warning. The patent application with publication number CN114527206A provides a sulfonamide antibiotic groundwater pollution tracing method and system. The technical solution collects traceability area information, collects groundwater samples and pre-treats them, determines the content of sulfonamide antibiotics in the water samples, and determines that the water sample is abnormal if the analysis data exceeds the preset threshold, initiates emergency response and monitoring, and calculates pollution emission parameters to obtain pollution source information.

[0004] Although existing technologies provide solutions for groundwater testing and using testing data to trace the source of pollution, both soil and groundwater play an important role in the spread of pollution. For example, the permeability of soil directly affects the flow rate and direction of groundwater, which in turn affects the speed and direction of pollutant propagation. Therefore, it is difficult to achieve accurate pollution tracing by relying solely on groundwater testing data.

[0005] In view of this, the present invention proposes a method and system for tracing and accurately locating site pollution sources to solve the above problems. Summary of the Invention

[0006] In order to overcome the above-mentioned defects of the prior art, the present invention provides a method and system for tracing and accurately locating site pollution sources.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] In the first aspect, a method for tracing and accurately locating the source of pollution on a site includes:

[0009] H initial detection points are arranged in the detection area, soil detection information corresponding to the H initial detection points is obtained respectively, soil impact values ​​are generated according to the soil detection information, the H initial detection points are optimized according to the soil impact values, and W target detection points are determined, where H>W;

[0010] Obtain water quality test data and water level elevation data corresponding to W target detection points respectively, and construct a target flow field map based on the soil impact value, water quality test data, and water level elevation data. The target flow field map is a groundwater flow field map of the area to be detected;

[0011] The target flow field map is input into the pre-built pollution source tracing model to obtain the location of the pollution source.

[0012] Furthermore, the method of arranging H initial detection points in the area to be detected includes:

[0013] Determine the area of ​​the area to be detected, establish a two-dimensional coordinate system with any boundary point of the area to be detected as the origin, determine the boundary range based on the area and the two-dimensional coordinate system, generate the X-axis coordinate of the initial detection point within the X-axis range, generate the Y-axis coordinate of the initial detection point within the Y-axis range, and use a preset random number generation algorithm to generate H initial detection points. The boundary range includes the X-axis range and the Y-axis range.

[0014] Furthermore, the method for determining the boundary range based on the area of ​​the region and the two-dimensional coordinate system includes:

[0015] Arrange the boundary vertices of the area to be detected in order, calculate the sum of the cross products of the coordinates of adjacent vertices in sequence, and take half of the absolute value of the sum of the cross products as the area of ​​the area;

[0016] According to the extreme values ​​of the boundary vertices of the area to be detected, the X-axis range and the Y-axis range are determined to form a boundary range.

[0017] Furthermore, the soil detection information includes a soil surface image and soil detection data. The method for generating a soil impact value based on the soil detection information includes:

[0018] A first influence value is generated based on a soil surface image, a second influence value is generated based on soil detection data, and a soil influence value is determined based on the first influence value, the second influence value, and a preset weight.

[0019] Furthermore, the method for determining W target detection points includes:

[0020] S101: taking the soil impact value corresponding to each initial detection point as a data point, and constructing H data points;

[0021] S102: Calculate the Euclidean distance between all data points and generate a distance matrix;

[0022] S103: Treat each data point as a separate cluster;

[0023] S104: Determine the two clusters with the smallest Euclidean distance in the distance matrix and merge them into a new cluster;

[0024] S105: Update the distance matrix and calculate the Euclidean distance between the new cluster and other clusters;

[0025] S106: Repeat the above S104-S105 until the number of clusters is reduced to W, select the initial detection point corresponding to any data point from each cluster as the target detection point, and output W target detection points.

[0026] Furthermore, the method of generating the distance matrix includes:

[0027] DM= ;

[0028] Among them, DM is the H×H distance matrix, For the Data points and The Euclidean distance between data points, ≤H, ≤H.

[0029] Furthermore, the method for constructing the target flow field map based on the soil impact value, water quality detection data and water level elevation data includes:

[0030] Obtain geographic information data corresponding to the area to be tested, input the geographic information data into the hydrological simulation software, establish a watershed model grid, input water level elevation data into the hydrological simulation software as the starting condition for watershed model grid calculation, input soil impact value into the hydrological simulation software to simulate the resistance of groundwater flow, input water quality detection data into the hydrological simulation software to simulate the migration behavior of pollutants, and export the target flow field map after the simulation is completed.

[0031] Furthermore, the method for constructing the pollution source tracing model includes:

[0032] Acquire a sample data set, wherein the sample data set includes a historical target flow field map and a historical pollution source location;

[0033] Divide the sample data set into a sample training set and a sample test set, and build a regression network;

[0034] The historical target flow field map in the sample training set is used as the input data of the regression network, and the historical pollution source locations in the sample training set are used as the output data of the regression network. The regression network is trained to obtain an initial regression network for predicting the real-time pollution source locations.

[0035] The initial regression network is tested using a sample test set, and the output of the initial regression network that satisfies a preset error value is used as the pollution source tracing model.

[0036] In a second aspect, a system for tracing and accurately locating pollution sources on a site is provided, which is used to implement the aforementioned method for tracing and accurately locating pollution sources on a site, including:

[0037] Data processing module: used to arrange H initial detection points in the detection area, obtain soil detection information corresponding to the H initial detection points, generate soil impact values ​​based on the soil detection information, optimize the H initial detection points based on the soil impact values, and determine W target detection points, where H>W;

[0038] Flow field construction module: used to obtain water quality detection data and water level elevation data corresponding to W target detection points respectively, and construct a target flow field map based on the soil impact value, water quality detection data and water level elevation data. The target flow field map is the groundwater flow field map of the area to be detected;

[0039] Source tracing module: used to input the target flow field map into the pre-built pollution source tracing model to obtain the location of the pollution source.

[0040] Compared with the prior art, the present invention has the following beneficial effects:

[0041] The present invention first arranges H initial detection points in the area to be detected, obtains soil detection information corresponding to the H initial detection points respectively, generates soil impact values ​​based on the soil detection information, optimizes the H initial detection points according to the soil impact values, and then obtains water quality detection data and water level elevation data corresponding to W target detection points respectively, constructs a target flow field map based on the soil impact values, water quality detection data and water level elevation data, and finally inputs the target flow field map into a pre-built pollution tracing model to obtain the location of the pollution source. The present invention then combines soil detection information (such as soil physical and chemical properties), groundwater detection data, water level elevation data and a method of optimizing the layout of detection points to comprehensively simulate the migration path of pollutants in soil and groundwater. This multi-level data integration and analysis method improves the accuracy of pollution source tracing, thereby solving the problem of inaccurate positioning caused by relying solely on groundwater detection data in the existing technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 This is a flow chart of a method for tracing and accurately locating site pollution sources in the present invention;

[0043] Figure 2 This is a schematic diagram of the structure of a system for tracing and accurately locating site pollution sources in the present invention;

[0044] Figure 3 Schematic diagram for determining the boundary range in the present invention. DETAILED DESCRIPTION

[0045] The following will provide a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0046] Example 1

[0047] See also Figure 1 As shown, this embodiment discloses a method for tracing and accurately locating the source of pollution on a site, including:

[0048] S10: H initial detection points are arranged in the detection area, soil detection information corresponding to each of the H initial detection points is obtained, soil impact values ​​are generated based on the soil detection information, the H initial detection points are optimized based on the soil impact values, and W target detection points are determined, where H>W;

[0049] In this embodiment, the area to be detected may be a transition area connecting point source pollution from industrial enterprises and surrounding areas with dense human activity.

[0050] The method of arranging H initial detection points in the area to be detected includes:

[0051] Determine the area of ​​the area to be detected, establish a two-dimensional coordinate system with any boundary point of the area to be detected as the origin, determine the boundary range based on the area and the two-dimensional coordinate system, generate the X-axis coordinate of the initial detection point within the X-axis range, generate the Y-axis coordinate of the initial detection point within the Y-axis range, and use a preset random number generation algorithm to generate H initial detection points. The boundary range includes the X-axis range and the Y-axis range.

[0052] Methods for determining the boundary range based on the area and two-dimensional coordinate system include:

[0053] Arrange the boundary vertices of the area to be detected in order and calculate the sum of the cross products of the coordinates of adjacent vertices in sequence. And add the last closed edge , take half of the absolute value of the sum of the cross products as the area of ​​the region. The details are as follows:

[0054] ;

[0055] in, is the area of ​​the region, is the total number of boundary vertices of the area to be detected, For the The X-axis coordinates of the boundary vertices, For the The Y-axis coordinates of the boundary vertices, For the The X-axis coordinates of the boundary vertices, For the The Y-axis coordinates of the boundary vertices, and Respectively The X-axis and Y-axis coordinates of the boundary vertices, is the minimum X-axis coordinate of the boundary vertex, is the maximum X-axis coordinate of the boundary vertex, is the minimum Y-axis coordinate value among the boundary vertices, is the maximum Y-axis coordinate of the boundary vertex, ≤ Directly take the extreme value of the vertex coordinate to determine the X-axis range and Y-axis range , forming a minimum enclosing rectangle (i.e., the bounding range). Linear operations on vertex coordinates can accurately calculate the area of ​​any simple polygon, without relying on complex geometric partitioning. This method is highly efficient and applicable to both convex and concave polygons. Determining the boundary by traversing the extreme values ​​of vertex coordinates is simple and efficient, allowing for rapid localization of the spatial extent of an area and providing basic parameters for subsequent detection or visualization.

[0056] It should be noted that if Figure 3 As shown, is the X-axis range, The Y-axis range is calculated in this embodiment by using the coordinates of the boundary vertices to calculate the area of ​​the area to be detected. The area of ​​the area is obtained in advance. Therefore, according to the above method, the area and projection range of the area to be detected can be obtained simultaneously through the coordinates of the boundary points. For irregular areas, the projection range is determined by the minimum and maximum values ​​of the X and Y coordinates of all boundary points. In this way, the boundary range of the area is clearly defined for the layout of the initial detection points.

[0057] It should be added that the random number generation algorithm can be a Markov random walk method. The Markov random walk method generally refers to a particle or object moving in a series of random steps in a discrete space (such as a one-dimensional, two-dimensional or higher-dimensional grid). In each step, the object moves from the current position to the next adjacent position, and the probability of selecting the next position depends only on the current position and does not depend on the previous trajectory. In this embodiment, the initial detection points are generated by the Markov random walk method, which utilizes the randomness and memorylessness of the algorithm to randomly distribute the detection points in the area to be detected. Specifically, in each step, the Markov random walk method generates the position of the next detection point based on the position of the current detection point, and the probability of selecting the position depends only on the current state (position) and does not depend on the previous trajectory. This method ensures that the distribution of the detection points is random and diverse, which helps to fully cover the area to be detected, thereby more effectively collecting soil detection data.

[0058] It should be noted that soil detection information includes soil surface images and soil detection data. Soil detection data includes but is not limited to soil pH value, organic matter content value and water content value. Taking fluorine pollutants as an example, under normal circumstances, a higher pH value will reduce the adsorption capacity of fluorine pollutants, thereby increasing their mobility in the environment. The moisture content of the soil directly affects the solubility and mobility of pollutants. The organic matter content value refers to the proportion of organic matter in the soil (such as animal and plant residues, microorganisms and their decomposition products) to the total mass of the soil. The organic matter in the soil can usually significantly affect the adsorption characteristics of fluorine pollutants, enhance their retention effect in the soil, and reduce their possibility of migration to groundwater. Under saturated conditions, soil is more likely to allow fluorine pollutants to migrate through the water phase, while under dry conditions, the diffusion of fluorine pollutants mainly depends on gas phase migration or solid phase adsorption. The above-mentioned soil pH value, organic matter content value and water content value can all be obtained through corresponding sensors. For example, the organic matter content value can be obtained through an electrochemical sensor or a near-infrared spectroscopy sensor.

[0059] The method for generating soil impact value based on soil detection information includes:

[0060] A first influence value is generated based on a soil surface image, a second influence value is generated based on soil detection data, and a soil influence value is determined based on the first influence value, the second influence value, and a preset weight.

[0061] The method for generating a first influence value based on a soil surface image includes:

[0062] The soil surface image is analyzed to obtain a soil porosity value, a soil density value, and a soil particle ratio, and a first impact value is calculated based on the soil porosity value, the soil density value, and the soil particle ratio.

[0063] It can be understood that the soil porosity value refers to the size of the pores in the soil (i.e., the gaps between particles), the soil density value refers to the ratio of the area of ​​the soil pores to the total area of ​​the soil, and the soil particle ratio refers to the ratio of large-particle soil to the total particle soil. Among them, large-particle soil means that during the soil surface image recognition process, the soil particle diameter must meet the preset diameter threshold.

[0064] It should be added that the analysis of soil surface images can be carried out through existing image processing and computer vision technologies. For example, morphological operations (such as dilation, erosion, opening and closing operations) are used to process noise and details in the image, extract the structural characteristics of particles (such as size and shape) and the distribution of pore space, quantify the soil density value and particle ratio, and use edge detection algorithms (such as Canny edge detection and Sobel operator) to identify the edges of soil particles and pores. Edge detection helps to quantify the size and shape of soil particles and calculate the soil porosity value.

[0065] Methods for calculating the first impact value based on the dimensionless soil porosity value, soil density value, and soil particle ratio include:

[0066] FIV= ;

[0067] Where FIV is the first impact value, which reflects the migration speed and penetration capacity of pollutants in the soil. The larger the value, the faster the migration and the stronger the penetration. is the soil density value (normalized value), is the soil particle ratio, is the soil porosity value (normalized value); 、 、 is the weight coefficient (dimensionless), which is used to adjust the relative contribution of each variable and can be calibrated by experimental data or domain knowledge; is a proportional constant (dimensionless) used for overall scaling to ensure that the FIV value is within a reasonable range and can be determined by data fitting. Indicates positive influencing factors (soil pore and particle ratio). Increasing the pore or particle ratio will increase the first impact value; the denominator Indicates a negative influencing factor (soil density). An increase in soil density will reduce the first influence value. In practical applications, if the soil density is close to zero (rare in theory, because soil density is usually greater than 0), a very small constant can be added, such as ,in To prevent numerical errors.

[0068] In this embodiment, soil density value and soil porosity value are used as examples. A larger soil porosity value indicates more or larger pores, which allows water and pollutants to penetrate and flow more easily. Therefore, pollutants migrate faster and have stronger permeability in soil with high porosity value. A larger soil density value means fewer gaps between particles and a tighter soil structure, which significantly reduces the permeability of water and pollutants because pollutants need to overcome greater resistance during the flow process. Therefore, the larger the first influence value, the faster the pollutants migrate in the soil and the stronger the permeability.

[0069] The method for generating a second influence value according to soil detection data includes:

[0070] The calculation of the Second Impact Value (SIV) quantifies the organic matter content by nonlinear combination ( ), pH value (PH) and water content value ( ), the comprehensive impact characterization of this embodiment is as follows:

[0071] SIV= ;

[0072] Where SIV is the second impact value. The larger the value, the faster the pollutant migrates. is the organic matter content value, is the soil pH value, is the water content value (normalized to [0,1]).

[0073] for Sensitivity weight, greater than 0; for Offset constant, greater than or equal to 1; for Action intensity adjustment factor; is the organic matter-water content coupling coefficient, which is greater than 0; is the global proportional constant; is a very small constant, ensuring that the denominator is non-zero. To capture extreme Migration-promoting effect: Enhanced sensitivity in acidic / alkaline regions; constant avoid The effect is reset to zero over time; the denominator is used for Control the intensity of the action. Simulating the antagonistic effect of the two: molecules Directly promotes migration (positive correlation); denominator Characterizes organic matter retention (negative correlation), low hour Effect enhancement (significant adsorption in arid areas), is a very small constant.

[0074] In this embodiment, taking the organic matter content value as an example, when the organic matter content value is larger, the organic matter in the soil can usually significantly affect the adsorption characteristics of fluorine pollutants and enhance their retention effect in the soil. Therefore, from the above content, it can be seen that when the second influence degree value is larger, it indicates that the migration speed of the pollutants in the soil is faster and the penetration ability is stronger.

[0075] Methods for determining soil impact values ​​include:

[0076] The soil influence value is obtained by linearly weighting the first influence value and the second influence value. A specific example of this embodiment is as follows:

[0077] SOV= FIV+ SIV;

[0078] Among them, SOV is the soil impact value, 、 All are preset weights.

[0079] It should be added that the above 、 They are all pre-set through expert experience and stored in the database; the method of pre-setting weights through expert experience includes: selecting multidisciplinary experts, integrating their experience judgment using the Delphi method or the analytic hierarchy process (AHP), verifying the rationality of the weights by combining historical data calibration and scenario testing, and finally storing the normalized weights in the database and supporting dynamic updates to ensure that the model retains the depth of domain knowledge and is scientific and maintainable (this is existing technology and will not be elaborated here).

[0080] In this embodiment, the first impact value is based on the soil surface image and reflects the impact of the physical structural properties of the soil (such as porosity, density, and particle ratio) on the migration of pollutants. The physical properties of the soil will directly affect the infiltration rate and diffusion path of the pollutants. The second impact value is based on soil detection data and mainly reflects the impact of the chemical properties of the soil (such as pH value, organic matter content, water content, etc.) on the behavior of pollutants. The chemical properties determine the solubility, adsorption, and chemical reactivity of pollutants in the soil. By considering the physical and chemical properties separately, the migration and retention behavior of pollutants in the soil can be more comprehensively evaluated, because these two aspects of characteristics will affect the diffusion and infiltration of pollutants in the actual environment.

[0081] The method for determining W target detection points includes:

[0082] S101: taking the soil impact value corresponding to each initial detection point as a data point, and constructing H data points;

[0083] S102: Calculate the Euclidean distance between all data points and generate a distance matrix;

[0084] S103: Treat each data point as a separate cluster;

[0085] S104: Determine the two clusters with the smallest Euclidean distance in the distance matrix and merge them into a new cluster;

[0086] S105: Update the distance matrix and calculate the Euclidean distance between the new cluster and other clusters;

[0087] S106: Repeat the above S104-S105 until the number of clusters is reduced to W, select the initial detection point corresponding to any data point from each cluster as the target detection point, and output W target detection points.

[0088] Methods for generating distance matrices include:

[0089] DM= ;

[0090] Among them, DM is the H×H distance matrix, For the Data points and The Euclidean distance between data points, ≤H, ≤H.

[0091] In this embodiment, the Euclidean distance between all initial detection points is calculated, and clusters with the smallest distances are continuously merged until the number of clusters is reduced to W. Each cluster represents an area with similar soil influence values. This method ensures that the selected W target detection points can represent the different characteristics of the entire area to be detected, making the detection results more representative. The soil influence values ​​are calculated in combination with the soil physical characteristics (first influence value) and chemical characteristics (second influence value), ensuring that the comprehensive characteristics of each detection point are fully considered. Subsequently, the initial detection points are optimized using a hierarchical clustering method to ensure that the selected W target detection points can represent areas with different soil characteristics. This global optimization method can more comprehensively reflect the diversity of the entire detection area, avoid bias, and improve the representativeness and accuracy of the detection.

[0092] S20: Acquire water quality detection data and water level elevation data corresponding to W target detection points respectively, and construct a target flow field map based on the soil impact value, the water quality detection data, and the water level elevation data. The target flow field map is a groundwater flow field map of the area to be detected;

[0093] In this embodiment, the water quality test data includes but is not limited to the average flow rate value and the diffusion coefficient value. The diffusion coefficient value refers to the diffusion rate of the pollutant in the groundwater due to the molecular motion. Specifically, it quantifies the rate at which the pollutant diffuses to the low concentration area under the drive of the concentration gradient in the water body. The diffusion coefficient value can usually be obtained by querying the corresponding data, or can also be obtained by experiments conducted by people in this field. For example, the pollutant can be a fluorine pollutant, and the diffusion coefficient value of the fluorine pollutant in the groundwater is approximately .

[0094] The water level elevation data mentioned above refers to the elevation of groundwater at the target detection point. Specifically, a corresponding monitoring well is opened at each target detection point, and a water level meter or data recorder is used to measure the elevation of the groundwater level (water surface height) in different monitoring wells to generate the water level elevation corresponding to each target detection point.

[0095] Methods for constructing target flow field maps based on soil impact values, water quality test data, and water level elevation data include:

[0096] Obtain geographic information data corresponding to the area to be tested, input the geographic information data into the hydrological simulation software, establish a watershed model grid, input water level elevation data into the hydrological simulation software as the starting condition for watershed model grid calculation, input soil impact value into the hydrological simulation software to simulate the resistance of groundwater flow, input water quality detection data into the hydrological simulation software to simulate the migration behavior of pollutants, and export the target flow field map after the simulation is completed.

[0097] In this embodiment, geographic information data includes but is not limited to terrain elevation, surface features, underground geological structure and other data. Terrain elevation refers to the height of each surface point relative to a reference plane. Terrain elevation affects the flow direction and speed of surface and groundwater. The elevation difference forms a hydraulic gradient, and water will naturally flow from high to low. Therefore, when constructing the model, terrain elevation is used to determine the driving force of groundwater flow. Surface features describe various physical and biological properties of the surface. Surface features affect the rate and concentration of pollutants infiltrating into the underground through the surface. Underground geological structure refers to the distribution, composition, thickness and other characteristics of rock and soil layers below the surface. The underground geological structure determines the hydraulic conductivity and permeability of the underground medium. The hydrological simulation software can be MODFLOW, FEFLOW and other software.

[0098] It should be added that the water level elevation data of each target detection point is input into the hydrological simulation software as the initial calculation condition of the watershed model grid, because the water level elevation data provides the initial state of the groundwater head, which is used to simulate the flow direction and speed of groundwater. In addition, the soil influence value of each target detection point is input into the hydrological simulation software. These data include the physical properties (such as porosity and permeability) and chemical properties (such as organic matter content and pH value) of the soil, which are used to calculate the resistance to groundwater flow.

[0099] This embodiment combines soil testing information to generate soil impact values, water quality testing data, and water level elevation data, fully accounting for various factors affecting groundwater flow and pollutant diffusion (such as soil physical and chemical properties, water flow velocity and direction, and pollutant diffusion rate). This helps establish a more accurate groundwater flow field model that comprehensively reflects the hydrological conditions and pollutant migration characteristics within the area to be tested, thereby improving the accuracy and reliability of model simulation results. Geographic information data, water level elevation data, soil impact values, and water quality testing data are combined and input into hydrological simulation software to form a multi-level comprehensive model. This model can better reflect the dynamic changes of groundwater and pollutants under different conditions, adapt to complex environmental conditions, improve the ability to predict future groundwater flow trends and pollutant diffusion paths, and provide strong support for environmental monitoring and governance decision-making.

[0100] S30: Input the target flow field map into the pre-built pollution source tracing model to obtain the location of the pollution source.

[0101] The construction methods of pollution source tracing model include:

[0102] Acquire a sample data set, wherein the sample data set includes a historical target flow field map and a historical pollution source location;

[0103] Divide the sample data set into a sample training set and a sample test set, and build a regression network;

[0104] The historical target flow field map in the sample training set is used as the input data of the regression network, and the historical pollution source locations in the sample training set are used as the output data of the regression network. The regression network is trained to obtain an initial regression network for predicting the real-time pollution source locations.

[0105] The initial regression network is tested using a sample test set, and the output of the initial regression network that satisfies a preset error value is used as the pollution source tracing model. The initial regression network is preferably a deep neural network model.

[0106] It can be understood that the target flow field map is a visual representation of the flow pattern of groundwater in a specific area. It shows information such as the flow direction, velocity, and flow velocity changes of groundwater. Since the flow of groundwater is the main carrier for the diffusion and migration of pollutants in the underground environment, the target flow field map can clearly show the potential migration path of pollutants in groundwater. In a hydrological environment, pollutants usually diffuse along the flow direction of groundwater. By analyzing the flow direction and velocity in the target flow field map, the migration route of pollutants can be determined, the direction of their source can be traced, and the location of the pollution source can be inferred.

[0107] The method for building a pollution source tracing model uses historical data to train a regression network (such as a deep neural network), which learns and captures the complex relationship between the target flow field map and the location of the pollution source. The historical target flow field maps and corresponding pollution source locations in the training dataset provide examples. The regression network uses these examples to learn how different flow field patterns affect the diffusion path of pollutants and the ultimate location of the pollution source. The deep neural network model, due to its powerful nonlinear mapping capabilities, can identify and learn the underlying complex relationship between the target flow field map (such as water flow direction, velocity distribution, etc.) and the location of the pollution source. The trained network can predict the origin of pollutants, that is, the location of the pollution source, from the real-time flow field map.

[0108] This embodiment first arranges H initial detection points in the detection area, obtains soil detection information corresponding to the H initial detection points respectively, generates soil impact values ​​based on the soil detection information, optimizes the H initial detection points based on the soil impact values, and then obtains water quality detection data and water level elevation data corresponding to W target detection points respectively. A target flow field map is constructed based on the soil impact values, water quality detection data, and water level elevation data. Finally, the target flow field map is input into a pre-built pollution tracing model to obtain the location of the pollution source. This embodiment combines soil detection information (such as soil physical and chemical properties), groundwater detection data, water level elevation data, and a method of optimizing the layout of detection points to comprehensively simulate the migration path of pollutants in soil and groundwater. This multi-level data integration and analysis method improves the accuracy of pollution source tracing, thereby solving the problem of inaccurate positioning caused by relying solely on groundwater detection data in the existing technology.

[0109] Example 2

[0110] See also Figure 2 As shown, based on the same inventive concept, this embodiment discloses a system for tracing and accurately locating site pollution sources. For details not provided in this embodiment, please refer to the description of the relevant parts in Example 1. The system includes:

[0111] Data processing module: used to arrange H initial detection points in the detection area, obtain soil detection information corresponding to the H initial detection points, generate soil impact values ​​based on the soil detection information, optimize the H initial detection points based on the soil impact values, and determine W target detection points, where H>W;

[0112] In this embodiment, the area to be detected may be a transition area connecting point source pollution from industrial enterprises and surrounding areas with dense human activity.

[0113] The method of arranging H initial detection points in the area to be detected includes:

[0114] Determine the area of ​​the area to be detected, establish a two-dimensional coordinate system with any boundary point of the area to be detected as the origin, determine the boundary range based on the area and the two-dimensional coordinate system, generate the X-axis coordinate of the initial detection point within the X-axis range, generate the Y-axis coordinate of the initial detection point within the Y-axis range, and use a preset random number generation algorithm to generate H initial detection points. The boundary range includes the X-axis range and the Y-axis range.

[0115] Methods for determining the boundary range based on the area and two-dimensional coordinate system include:

[0116] ;

[0117] in, is the area of ​​the region, is the total number of boundary vertices of the area to be detected, For the The X-axis coordinates of the boundary vertices, For the The Y-axis coordinates of the boundary vertices, For the The X-axis coordinates of the boundary vertices, For the The Y-axis coordinates of the boundary vertices, is the minimum X-axis coordinate of the boundary vertex, is the maximum X-axis coordinate of the boundary vertex, is the minimum Y-axis coordinate value among the boundary vertices, The maximum Y-axis coordinate of the boundary vertex.

[0118] The method for generating soil impact value based on soil detection information includes:

[0119] A first influence value is generated based on a soil surface image, a second influence value is generated based on soil detection data, and a soil influence value is determined based on the first influence value, the second influence value, and a preset weight.

[0120] The method for generating a first influence value based on a soil surface image includes:

[0121] The soil surface image is analyzed to obtain a soil porosity value, a soil density value, and a soil particle ratio, and a first impact value is calculated based on the soil porosity value, the soil density value, and the soil particle ratio.

[0122] Methods for calculating the first impact value based on soil porosity, soil density, and soil particle ratio include:

[0123] FIV= ;

[0124] Where FIV is the first impact value, which reflects the migration speed and penetration capacity of pollutants in the soil. The larger the value, the faster the migration and the stronger the penetration. is the soil density value, is the soil particle ratio, is the soil porosity value.

[0125] In this embodiment, soil density value and soil porosity value are used as examples. A larger soil porosity value indicates more or larger pores, which allows water and pollutants to penetrate and flow more easily. Therefore, pollutants migrate faster and have stronger permeability in soil with high porosity value. A larger soil density value means fewer gaps between particles and a tighter soil structure, which significantly reduces the permeability of water and pollutants because pollutants need to overcome greater resistance during the flow process. Therefore, the larger the first influence value, the faster the pollutants migrate in the soil and the stronger the permeability.

[0126] The method for generating a second influence value according to soil detection data includes:

[0127] SIV= ;

[0128] Where SIV is the second impact value. The larger the value, the faster the pollutant migrates. is the organic matter content value, is the soil pH value, is the water content value (normalized to [0,1]).

[0129] Methods for determining soil impact values ​​include:

[0130] SOV= FIV+ SIV;

[0131] Among them, SOV is the soil impact value, 、 All are preset weights.

[0132] It should be added that the above 、 They are all pre-set through expert experience and stored in the database.

[0133] The data processing module includes an initialization unit, a matrix construction unit, a first processing unit, a merging processing unit, an updating unit and a second processing unit.

[0134] Initialization unit: used to take the soil impact value corresponding to each initial detection point as a data point, and then construct H data points;

[0135] Matrix construction unit: used to calculate the Euclidean distance between all data points to generate a distance matrix;

[0136] The first processing unit: used to treat each data point as a separate cluster;

[0137] Merge processing unit: used to determine the two clusters with the smallest Euclidean distance in the distance matrix and merge them into a new cluster;

[0138] Update unit: used to update the distance matrix and calculate the Euclidean distance between the new cluster and other clusters;

[0139] The second processing unit is used to repeatedly merge new clusters and update the distance matrix until the number of clusters is reduced to W, select the initial detection point corresponding to any data point from each cluster as the target detection point, and output W target detection points.

[0140] Methods for generating distance matrices include:

[0141] DM= ;

[0142] Among them, DM is the H×H distance matrix, For the Data points and The Euclidean distance between data points, ≤H, ≤H.

[0143] Flow field construction module: used to obtain water quality detection data and water level elevation data corresponding to W target detection points respectively, and construct a target flow field map based on the soil impact value, water quality detection data and water level elevation data. The target flow field map is the groundwater flow field map of the area to be detected;

[0144] In this embodiment, the water quality test data includes but is not limited to the average flow rate value and the diffusion coefficient value. The diffusion coefficient value refers to the diffusion rate of the pollutant in the groundwater due to the molecular motion. Specifically, it quantifies the rate at which the pollutant diffuses to the low concentration area under the drive of the concentration gradient in the water body. The diffusion coefficient value can usually be obtained by querying the corresponding data, or can also be obtained by experiments conducted by people in this field. For example, the pollutant can be a fluorine pollutant, and the diffusion coefficient value of the fluorine pollutant in the groundwater is approximately .

[0145] The water level elevation data mentioned above refers to the elevation of groundwater at the target detection point. Specifically, a corresponding monitoring well is opened at each target detection point, and a water level meter or data recorder is used to measure the elevation of the groundwater level (water surface height) in different monitoring wells to generate the water level elevation corresponding to each target detection point.

[0146] Methods for constructing target flow field maps based on soil impact values, water quality test data, and water level elevation data include:

[0147] Obtain geographic information data corresponding to the area to be tested, input the geographic information data into the hydrological simulation software, establish a watershed model grid, input water level elevation data into the hydrological simulation software as the starting condition for watershed model grid calculation, input soil impact value into the hydrological simulation software to simulate the resistance of groundwater flow, input water quality detection data into the hydrological simulation software to simulate the migration behavior of pollutants, and export the target flow field map after the simulation is completed.

[0148] Source tracing module: used to input the target flow field map into the pre-built pollution source tracing model to obtain the location of the pollution source.

[0149] The construction methods of pollution source tracing model include:

[0150] Acquire a sample data set, wherein the sample data set includes a historical target flow field map and a historical pollution source location;

[0151] Divide the sample data set into a sample training set and a sample test set, and build a regression network;

[0152] The historical target flow field map in the sample training set is used as the input data of the regression network, and the historical pollution source locations in the sample training set are used as the output data of the regression network. The regression network is trained to obtain an initial regression network for predicting the real-time pollution source locations.

[0153] The initial regression network is tested using a sample test set, and the output of the initial regression network that satisfies a preset error value is used as the pollution source tracing model. The initial regression network is preferably a deep neural network model.

[0154] It can be understood that the target flow field map is a visual representation of the flow pattern of groundwater in a specific area. It shows information such as the flow direction, velocity, and flow velocity changes of groundwater. Since the flow of groundwater is the main carrier for the diffusion and migration of pollutants in the underground environment, the target flow field map can clearly show the potential migration path of pollutants in groundwater. In a hydrological environment, pollutants usually diffuse along the flow direction of groundwater. By analyzing the flow direction and velocity in the target flow field map, the migration route of pollutants can be determined, the direction of their source can be traced, and the location of the pollution source can be inferred.

[0155] The method for building a pollution source tracing model uses historical data to train a regression network (such as a deep neural network), which learns and captures the complex relationship between the target flow field map and the location of the pollution source. The historical target flow field maps and corresponding pollution source locations in the training dataset provide examples. The regression network uses these examples to learn how different flow field patterns affect the diffusion path of pollutants and the ultimate location of the pollution source. The deep neural network model, due to its powerful nonlinear mapping capabilities, can identify and learn the underlying complex relationship between the target flow field map (such as water flow direction, velocity distribution, etc.) and the location of the pollution source. The trained network can predict the origin of pollutants, that is, the location of the pollution source, from the real-time flow field map.

[0156] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters, weights and thresholds in the formulas are set by technicians in this field according to actual conditions.

[0157] The above embodiments can be implemented in whole or in part via software, hardware, firmware, or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product comprises one or more computer instructions or computer programs. When loaded or executed on a computer, the processes or functions described in accordance with the embodiments of the present invention are fully or partially performed. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another. For example, the computer instructions can be transferred from one website, computer, server, or data center to another website, computer, server, or data center via a wired or wireless network. The computer-readable storage medium can be any available medium accessible by a computer, or a data storage device such as a server or data center that contains a collection of one or more available media. The available medium can be magnetic media (e.g., floppy disks, hard disks, or magnetic tapes), optical media (e.g., DVDs), or semiconductor media. The semiconductor media can be a solid-state drive.

[0158] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed in this invention can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present invention.

[0159] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0160] In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only one type. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0161] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0162] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0163] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

[0164] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for tracing and accurately locating pollution sources on a site, characterized in that: include: H initial detection points are arranged in the detection area, soil detection information corresponding to the H initial detection points is obtained respectively, soil impact values ​​are generated according to the soil detection information, the H initial detection points are optimized according to the soil impact values, and W target detection points are determined, where H>W; The soil detection information includes a soil surface image and soil detection data. The method for generating a soil impact value based on the soil detection information includes: generating a first influence value based on a soil surface image, generating a second influence value based on soil detection data, and determining a soil influence value based on the first influence value, the second influence value, and a preset weight; The method for generating a first influence value based on a soil surface image includes: Analyze the soil surface image to obtain a soil porosity value, a soil density value, and a soil particle ratio, and calculate a first impact value based on the soil porosity value, the soil density value, and the soil particle ratio; The method for generating a second influence value according to soil detection data includes: The calculation of the second impact value is to quantify the comprehensive impact of organic matter content, pH value and water content through nonlinear combination; The method for determining W target detection points includes: S101: taking the soil impact value corresponding to each initial detection point as a data point, and constructing H data points; S102: Calculate the Euclidean distance between all data points and generate a distance matrix; S103: Treat each data point as a separate cluster; S104: Determine the two clusters with the smallest Euclidean distance in the distance matrix and merge them into a new cluster; S105: Update the distance matrix and calculate the Euclidean distance between the new cluster and other clusters; S106: Repeat S104-S105 until the number of clusters is reduced to W, select the initial detection point corresponding to any data point from each cluster as the target detection point, and output W target detection points; Obtain water quality test data and water level elevation data corresponding to W target detection points respectively, and construct a target flow field map based on the soil impact value, water quality test data, and water level elevation data. The target flow field map is a groundwater flow field map of the area to be detected; The target flow field map is input into the pre-built pollution source tracing model to obtain the location of the pollution source.

2. The method for tracing and accurately locating the pollution source of a site according to claim 1, characterized in that: The method for arranging H initial detection points in the area to be detected includes: Determine the area of ​​the area to be detected, establish a two-dimensional coordinate system with any boundary point of the area to be detected as the origin, determine the boundary range based on the area and the two-dimensional coordinate system, generate the X-axis coordinate of the initial detection point within the X-axis range, generate the Y-axis coordinate of the initial detection point within the Y-axis range, and use a preset random number generation algorithm to generate H initial detection points. The boundary range includes the X-axis range and the Y-axis range.

3. The method for tracing and accurately locating the pollution source of a site according to claim 2, characterized in that: The method for determining the boundary range based on the area of ​​the region and the two-dimensional coordinate system includes: Arrange the boundary vertices of the area to be detected in order, calculate the sum of the cross products of the coordinates of adjacent vertices in sequence, and take half of the absolute value of the sum of the cross products as the area of ​​the area; According to the extreme values ​​of the boundary vertices of the area to be detected, the X-axis range and the Y-axis range are determined to form a boundary range.

4. The method for tracing and accurately locating the pollution source of a site according to claim 1, characterized in that: The method for generating a distance matrix comprises: DM= ; Among them, DM is the H×H distance matrix, For the Data points and The Euclidean distance between data points, ≤H, ≤H.

5. The method for tracing and accurately locating the pollution source of a site according to claim 1, characterized in that: The method for constructing a target flow field map based on soil impact value, water quality detection data and water level elevation data includes: Obtain geographic information data corresponding to the area to be tested, input the geographic information data into the hydrological simulation software, establish a watershed model grid, input water level elevation data into the hydrological simulation software as the starting condition for watershed model grid calculation, input soil impact value into the hydrological simulation software to simulate the resistance of groundwater flow, input water quality detection data into the hydrological simulation software to simulate the migration behavior of pollutants, and export the target flow field map after the simulation is completed.

6. The method for tracing and accurately locating the pollution source of a site according to claim 1, characterized in that: The method for constructing the pollution source tracing model includes: Acquire a sample data set, wherein the sample data set includes a historical target flow field map and a historical pollution source location; Divide the sample data set into a sample training set and a sample test set, and build a regression network; The historical target flow field map in the sample training set is used as the input data of the regression network, and the historical pollution source locations in the sample training set are used as the output data of the regression network. The regression network is trained to obtain an initial regression network for predicting the real-time pollution source locations. The initial regression network is tested using a sample test set, and the output of the initial regression network that satisfies a preset error value is used as the pollution source tracing model.

7. A system for tracing and accurately locating site pollution sources, which is used to implement the method for tracing and accurately locating site pollution sources according to any one of claims 1 to 6, characterized in that: include: Data processing module: used to arrange H initial detection points in the detection area, obtain soil detection information corresponding to the H initial detection points, generate soil impact values ​​based on the soil detection information, optimize the H initial detection points based on the soil impact values, and determine W target detection points, where H>W; Flow field construction module: used to obtain water quality detection data and water level elevation data corresponding to W target detection points respectively, and construct a target flow field map based on the soil impact value, water quality detection data and water level elevation data. The target flow field map is the groundwater flow field map of the area to be detected; Source tracing module: used to input the target flow field map into the pre-built pollution source tracing model to obtain the location of the pollution source.

Images