A method for tracing sources of groundwater pollution in production enterprises

Abstract

The present application relates to the field of groundwater pollution source tracing, and particularly relates to a method for tracing the source of groundwater pollution in a production enterprise. First, the original concentration values of different pollution factors obtained are standardized to obtain the concentration values of different pollution factors after standardization, principal component analysis is performed, the covariance matrix is calculated and characteristic decomposition is performed, the load is obtained and the main control pollution factor is screened; then, based on the concentration values of different pollution factors after standardization, a pollution source candidate area preliminary screening algorithm is used to obtain the pollution source candidate area; finally, based on the pollution source candidate area, a steady-state weighted convection-dispersion point source joint inversion algorithm is used to calculate the position coordinates of the final pollution source. The technical problems of dimension disaster caused by multi-pollutant high-dimensional data, slow calculation speed, high cost and low precision caused by too large search space of preliminary positioning of pollution sources, unreasonable source tracing weight allocation caused by neglecting the importance difference of pollutants and neglecting the convection effect of groundwater are solved.

Description

Technical Field

[0001] This invention relates to the field of groundwater pollution source tracing, and more particularly to a method for tracing groundwater pollution sources in operating enterprises. Background Technology

[0002] With the rapid development of industrialization, urbanization, and agricultural intensification, groundwater pollution has evolved from localized, sporadic events into a global, systemic environmental crisis. Operating enterprises, especially chemical production plants, have complex and interconnected production facilities, equipment, pipelines, and tanks that are often concealed. Leaks or seepage can occur due to raw material storage, production operations, pipeline damage, ground cracking, and waste disposal, leading to the entry of hazardous substances into the groundwater system. This has become a major point source of groundwater pollution. Characteristic pollutants emitted by these plants, such as heavy metals, organochlorides, petroleum hydrocarbons, PFAS, and perfluorinated compounds, are characterized by their high degree of concealment, long migration distances, and significant lag effects. Once they enter the groundwater system, they often take years or even decades to appear in downstream monitoring wells, greatly increasing the difficulty of tracing pollution sources. Traditional groundwater pollution tracing methods suffer from slow speed, high cost, low accuracy, cumbersome procedures, and a tendency to generalize from specific points. Therefore, there is an urgent need to provide a method for tracing groundwater pollution sources in operating enterprises, shortening the tracing time, improving accuracy, and providing a scientific basis for identifying pollution sources in these enterprises. Summary of the Invention

[0003] This invention provides a method for tracing the source of groundwater pollution in operating enterprises, which solves the problem that traditional methods directly analyze all pollutants, making it impossible to effectively shorten the tracing time and improve the accuracy of tracing. It also addresses the problem that traditional methods cannot reduce environmental risks from the source and protect the safety of groundwater resources.

[0004] The present invention provides a method for tracing the source of groundwater pollution in operating enterprises, comprising the following steps:

[0005] S1. Standardize the original concentration values ​​of different pollutants to obtain standardized concentration values ​​of different pollutants; perform principal component analysis based on the standardized concentration values ​​of different pollutants, calculate the covariance matrix and perform eigenvalue decomposition to obtain loadings and screen the main controlling pollutants.

[0006] S2. Based on the standardized concentration values ​​of different pollutants, a preliminary screening algorithm for pollution source candidate regions is adopted to obtain the pollution source candidate regions. Based on the pollution source candidate regions, the location coordinates of the final pollution source are calculated through a steady-state weighted convection-diffusion point source joint inversion algorithm.

[0007] Preferably, S1 specifically includes:

[0008] Perform eigenvalue decomposition on the covariance matrix to obtain eigenvalues ​​and corresponding eigenvectors. Each component in the eigenvector is the loading of the corresponding pollutant factor in the principal components.

[0009] Preferably, S1 specifically includes:

[0010] The obtained eigenvalues ​​are sorted in descending order, and the contribution rate is calculated as the proportion of each eigenvalue to the sum of all eigenvalues.

[0011] Preferably, S1 specifically includes:

[0012] Based on the contribution rate, the number of principal components to be retained is determined by setting a cumulative contribution rate threshold; within the range of retained principal components, pollutants whose absolute load value exceeds the load threshold are included as the main control pollutants in the set of main control pollutants.

[0013] Preferably, S1 specifically includes:

[0014] The comprehensive weight of the main pollution factor is obtained by multiplying the absolute value of the loading of the main pollution factor by the corresponding contribution rate and summing them.

[0015] Preferably, S2 specifically includes:

[0016] In the preliminary screening algorithm for candidate pollution sources, the correlation between the standardized concentration values ​​of different pollution factors and the Euclidean distance from the candidate pollution source to the corresponding sampling point is calculated to obtain the concentration-distance correlation of the main pollution factor.

[0017] Preferably, S2 specifically includes:

[0018] Based on the concentration-distance correlation of the main controlling pollutants, a comprehensive weight of the main controlling pollutants is introduced to obtain a comprehensive score; candidate pollution sources whose comprehensive scores exceed the set correlation threshold are included in the pollution source candidate area.

[0019] Preferably, S2 specifically includes:

[0020] In the steady-state weighted convection-diffusion point source joint inversion algorithm, the weighted sum of squared errors between the original concentration value and the theoretically predicted concentration of the main pollutant at the sampling point is used as the objective function. Based on the candidate pollution source region, through iteration, the location coordinates and source strength of the candidate pollution source that minimizes the objective function value are output, and the final location coordinates of the pollution source are obtained.

[0021] Preferably, S2 specifically includes:

[0022] In the steady-state weighted convection-diffusion point source joint inversion algorithm, based on the source strength of the main controlling pollutant, a normalization factor and a convection-diffusion attenuation term are introduced to calculate the theoretically predicted concentration.

[0023] The beneficial effects of the technical solution of the present invention are:

[0024] 1. By performing principal component analysis, calculating the covariance matrix, and performing eigenvalue decomposition on the standardized concentration values ​​of different pollutants, the loadings of each pollutant and the contribution rates of the principal components were obtained. This allowed for the screening of the main controlling pollutants and the calculation of their comprehensive weights, achieving efficient dimensionality reduction of pollutants and accurate identification of key pollutants. This significantly reduced data complexity and highlighted the contribution of the dominant pollution sources.

[0025] 2. By employing a preliminary screening algorithm for pollution source candidate areas, the concentration-distance correlation of the main controlling pollutant factors is weighted and summed, and a correlation threshold is set, the candidate areas for pollution sources are quickly delineated. This significantly reduces the search space for subsequent accurate inversion, improves the overall efficiency of source tracing calculations, and enhances the robustness of the preliminary screening, providing a scientific basis for pollution source tracing and identification of enterprises in production.

[0026] 3. By using the steady-state weighted convection-diffusion point source joint inversion algorithm, with the weighted sum of squared errors between the measured concentration and the theoretically predicted concentration at the sampling point as the objective function, and combining iterative methods with the Levenberg-Marquardt algorithm or particle swarm optimization, the precise location and quantitative estimation of candidate pollution source positions and source strengths are achieved. This accurately reconstructs the pollution diffusion process and provides visualized source tracing results of pollution leakage intensity and spatial coordinates. Attached Figure Description

[0027] Figure 1 This is a flowchart of a method for tracing the source of groundwater pollution in an operating enterprise, as described in this invention. Detailed Implementation

[0028] To further illustrate the technical means and effects adopted by the present invention to achieve the intended purpose, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments; based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0029] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0030] The following description, in conjunction with the accompanying drawings, details a specific scheme for a groundwater pollution tracing method for operating enterprises provided by the present invention.

[0031] See attached document Figure 1The diagram illustrates a flowchart of a method for tracing the source of groundwater pollution in an operating enterprise, provided by an embodiment of the present invention. The method includes the following steps:

[0032] S1. Standardize the original concentration values ​​of different pollutants to obtain standardized concentration values ​​of different pollutants; perform principal component analysis based on the standardized concentration values ​​of different pollutants, calculate the covariance matrix and perform eigenvalue decomposition to obtain the loading and screen the main controlling pollutants.

[0033] Sensors, such as PID sensors, electrochemical sensors, and UV-Vis photometric sensors, are deployed within the factory premises to form a network of regular or irregular sampling points. These points record the raw concentration values ​​of different pollutants in real time, creating a pollutant concentration matrix. The rows of the concentration matrix represent the location coordinates of different spatial sampling points, and the columns represent different types of pollutants, such as volatile organic compounds, heavy metals, and inorganic ions.

[0034] The original concentration values ​​of different pollutants in the pollutant concentration matrix are standardized using the Z-score standardization algorithm to obtain the standardized concentration values ​​of different pollutants, thereby obtaining the standardized pollutant concentration matrix. The Z-score standardization algorithm is a well-known technique in the art and will not be described in detail here.

[0035] Principal component analysis was performed on the standardized concentration values ​​of different pollutants to calculate the covariance matrix and perform eigenvalue decomposition, thereby obtaining the loadings and screening the main controlling pollutants.

[0036] Specifically, by performing a transpose multiplication operation on the standardized pollutant concentration matrix, a symmetric square matrix, namely the covariance matrix, is obtained. The diagonal elements of the covariance matrix are the variances of the standardized concentration values ​​of each pollutant itself, while the off-diagonal elements are the covariances of the standardized concentration values ​​between two different pollutants. A positive covariance indicates that the standardized concentration values ​​of the two pollutants tend to increase or decrease simultaneously, possibly sharing a pollution source; a negative covariance indicates that the standardized concentration values ​​of the two pollutants have an inverse relationship; and a covariance close to zero indicates that the standardized concentration values ​​of the two pollutants are essentially independent.

[0037] Eigenvalue decomposition is performed on the covariance matrix using numerical methods such as QR decomposition or exponential iteration to obtain a set of eigenvalues ​​and corresponding eigenvectors. The magnitude of the eigenvalues ​​directly reflects the overall variability of the standardized concentration values ​​of different pollutants across all sampling points along the corresponding eigenvector direction. Each eigenvector is a column vector with a length equal to the number of pollutant species. Loadings are obtained based on the eigenvectors from the eigenvalue decomposition process. Each component in the eigenvector is the loading of the corresponding pollutant in the principal component. The larger the absolute value of the loading, the more significant the contribution of the pollutant to the spatial variation pattern represented by the principal component.

[0038] After obtaining all eigenvalues, they are sorted in descending order, and the proportion of each eigenvalue to the sum of all eigenvalues ​​is calculated, i.e., the contribution rate. This quantifies the explanatory power of the corresponding principal component for the variation of the entire pollution concentration field. The pollution concentration field refers to the concentration distribution of different pollutants at various sampling points within the entire plant area. A cumulative contribution rate threshold is set, and the threshold is accumulated one by one starting from the first principal component until it is met. This determines the number of principal components to be retained. The selection of the cumulative contribution rate threshold ensures that the retained principal components can represent the original concentration values, while achieving significant dimensionality compression. The cumulative contribution rate threshold is determined based on the cumulative variance explanation ratio method and combined with the Jolliffe empirical criterion. Specifically, the covariance matrix is ​​eigenvalued and sorted in descending order of eigenvalues. The contribution rate of each principal component and the cumulative contribution rate are calculated. Principal components with eigenvalues ​​greater than 0.7 are prioritized, and within this range, the smallest number of principal components whose cumulative contribution rate first reaches 0.80 to 0.90 is selected as the number of principal components to be retained. This ensures that the retained principal components can explain most of the spatial variation information of the pollution factor concentration field.

[0039] Within the retained principal components, the absolute loading value of each contaminant factor in all eigenvectors is examined one by one. If the absolute loading value of a contaminant factor exceeds a preset loading threshold in at least one principal component, the contaminant factor is included in the set of principal contaminant factors as a controlling contaminant factor. The loading threshold screening mechanism ensures that only contaminants that have a significant impact on at least one principal component are considered controlling contaminants; the loading threshold is determined according to the Kaiser-Varimax loading significance criterion, and the sample size is the number of sampling points.

[0040] For each controlling pollutant, the absolute values ​​of the loadings of the controlling pollutant in all retained principal components are iterated, multiplied by the corresponding contribution rate, and summed to obtain the comprehensive weight of the controlling pollutant. The higher the comprehensive weight, the stronger the dominance of the controlling pollutant in the entire pollution system.

[0041] The formula for calculating the comprehensive weight of the main controlling pollutants is as follows:

[0042] in, Indicates the first The comprehensive weight of each major pollution factor is used to quantify the relative importance of the major pollution factors in pollution source tracing; Index indicating the controlling pollutant; Indicates the preceding The cumulative contribution rate of each principal component is used to indicate that only principal components whose cumulative contribution rate meets the cumulative contribution rate threshold are considered. It is the number of principal components that need to be retained to meet the cumulative contribution rate threshold requirement; Indicates the first Contribution rate of each principal component; Indicates the first The eigenvalues ​​corresponding to each principal component; This represents the sum of the eigenvalues ​​of all principal components, used as a normalization factor to make... =1; It is the number of all principal components obtained through eigenvalue decomposition; Indicates the first The primary controlling pollutant was in the... The absolute value of the load in each principal component is used to eliminate the influence of positive and negative directions, focusing only on the contribution intensity; This represents the set of indexes for the main controlling pollution factors.

[0043] S2. Based on the standardized concentration values ​​of different pollutants, a preliminary screening algorithm for pollution source candidate regions is used to obtain pollution source candidate regions. Based on the pollution source candidate regions, the location coordinates of the final pollution source are calculated through a steady-state weighted convection-diffusion point source joint inversion algorithm.

[0044] A preliminary screening algorithm for pollution source candidate regions is adopted to perform a weighted summation of the correlation between the standardized concentration values ​​of different pollution factors and the Euclidean distance from the candidate pollution source to each sampling point, so as to quickly delineate the candidate pollution source region and thus significantly reduce the search space for subsequent accurate inversion.

[0045] The preliminary screening algorithm for pollution source candidate areas traverses all possible candidate pollution sources within the entire enterprise plant area. For each candidate pollution source, the Euclidean distance between the candidate pollution source and all sampling points is calculated one by one, reflecting the spatial geometric relationship of the pollution factor spreading outward from the potential source.

[0046] For each controlling pollutant, the correlation between the standardized concentration values ​​at all sampling points and the Euclidean distance from the candidate pollution source to the corresponding sampling point is evaluated to obtain the concentration-distance correlation of the controlling pollutant. Furthermore, a comprehensive weight of the controlling pollutant is introduced as an adjustment coefficient to reflect the importance of different pollutants in contributing to the overall pollution event. The higher the comprehensive weight of the controlling pollutant, the greater the influence of its concentration-distance correlation on the determination of the candidate pollution source area, ensuring greater focus on the key substances that truly drive pollution distribution. The weighted correlations of all controlling pollutants are summed to obtain a comprehensive score. A higher comprehensive score indicates a higher degree of agreement between the measured spatial distribution of concentrations and the theoretical diffusion model when candidate pollution sources are assumed to be pollution sources.

[0047] A relevance threshold is set, and only candidate pollution sources with a comprehensive score exceeding the relevance threshold are included in the pollution source candidate region. The choice of the relevance threshold directly affects the size of the pollution source candidate region.

[0048] The formula for determining candidate pollution source areas is expressed as follows:

[0049] in, Indicates candidate areas for pollution sources; Indicates the first The correlation between the standardized concentration values ​​of the first controlling pollutant and the Euclidean distance from the candidate pollution source to the corresponding sampling point, i.e., the first... The concentration-distance correlation of each major controlling pollutant can be calculated using the Pearson correlation coefficient. Indicates the first The first sampling point at the sampling point Standardized concentration values ​​of each major controlling pollutant; Indicates candidate pollution sources up to the first Euclidean distance between sampling points; These are the location coordinates of the candidate pollution sources; The correlation threshold is used to control the strictness of the initial screening. It traverses all candidate pollution source locations to calculate the comprehensive score set, further calculates the mean and standard deviation of the comprehensive score set, and sets the correlation threshold as the sum of the mean and standard deviation of the comprehensive score. This indicates the overall score; This indicates the weighted correlation of the main controlling pollutant.

[0050] By using a preliminary screening algorithm for pollution source candidate regions, candidate pollution source regions can be initially identified from a massive number of possible locations, which can significantly improve computational efficiency and enhance the robustness of source tracing.

[0051] The steady-state weighted convection-diffusion point source joint inversion algorithm uses the weighted sum of squared errors between the measured concentration at the sampling point (i.e., the original concentration of the main controlling pollutant at the sampling point) and the theoretically predicted concentration as the objective function. By iteratively adjusting the location coordinates and source strength of candidate pollution sources, the objective function is minimized, thereby achieving precise location and quantitative source tracing of pollution sources. The source strength is the mass flux of pollutants continuously released into the groundwater system by the candidate pollution source per unit time. The initial source strength can be automatically set by the existing particle swarm optimization algorithm. The initialization method is random initialization, which is carried out within the feasible region that satisfies the physical range constraints. The physical range constraints are used to ensure that the initial source strength value conforms to the basic physical laws of groundwater pollutant release, including a positive source strength, the theoretically predicted concentration and the measured concentration being consistent in magnitude, the pollutants exhibiting diffusion attenuation characteristics along the propagation path, and the pollutant release being a continuous positive mass flux. This avoids generating initial solutions that do not conform to physical meaning and improves the stability and convergence efficiency of the inversion optimization process.

[0052] For each candidate pollution source in the current iteration, the Euclidean distance from each sampling point to the candidate pollution source is calculated to reflect the spatial scale of the spread of pollutants from the source to the sampling point. At the same time, the directional relationship between the groundwater flow direction obtained through hydrogeological survey and the line connecting the sampling point and the candidate pollution source is calculated, and the cosine value is obtained. The cosine value ranges from -1 to 1. A positive cosine value indicates the downstream direction, a negative cosine value indicates the upstream direction, and a zero cosine value indicates the direction perpendicular to the groundwater flow. This ensures that the influence of groundwater flow on the asymmetric distribution of pollutants is captured and avoids the erroneous assumption of simple symmetric diffusion.

[0053] The theoretically predicted concentration is calculated independently for each major pollutant. The calculation process starts from the source strength, multiplies it by a normalization factor that includes pi, dispersion coefficient, and Euclidean distance to obtain the basic diffusion amplitude, and multiplies it by a convection-diffusion attenuation term to simulate convection-diffusion coupling. In the downstream direction, the closer the cosine value is to 1, the slower the attenuation and the higher the theoretically predicted concentration. In the upstream direction, the closer the cosine value is to -1, the more severe the attenuation and the faster the theoretically predicted concentration decreases. In the vertical direction, the attenuation is moderate.

[0054] The measured concentration of each sampling point and each main pollutant is subtracted from the corresponding theoretically predicted concentration to obtain the residual. The residual is squared to eliminate the influence of positive and negative offsets, and then multiplied by the comprehensive weight of the main pollutant. The weighted squared residuals of all sampling points and all main pollutants are summed to obtain the objective function value under the location coordinates and source strength combination of the current candidate pollution source. The smaller the objective function value, the closer the theoretically predicted concentration is to the measured concentration distribution.

[0055] A steady-state inversion optimization model is constructed, and the steady-state weighted convection-diffuse point source joint inversion algorithm is mathematically implemented and formalized. The steady-state inversion optimization model is as follows:

[0056] in, Indicates the location coordinates of the final pollution source; This indicates the final source strength of each major pollutant, used to quantify the leakage intensity; Indicates the first The ultimate source strength of the main controlling pollutant; This represents the combination of parameters that minimizes the objective function value; Indicates search domain restrictions; Indicates the first The source strength of the main controlling pollutants; Indicates the first The source strength of the main controlling pollutant satisfies the constraint that it must be greater than zero; Let denot be the objective function of the steady-state inversion optimization model, and let the outer summation be over all The sampling points are accumulated, and the inner summation is the accumulation of each of the main controlling pollution factors in the set of main controlling pollution factors; Indicates the first At the sampling point, the first The measured concentration values ​​of the main controlling pollutants, i.e., the first At the sampling point, the first The original concentration values ​​of the main controlling pollutants; Indicates the first The dispersion coefficients of the main controlling pollutants are obtained directly from the hydrogeological survey reports. This represents the convection-diffusion attenuation term, used to introduce directionality and diffusion effects; The scalar value representing groundwater flow velocity is calculated using the following formula: , and The velocity component is obtained from hydrogeological surveys, which include borehole exploration, groundwater monitoring well deployment, water level monitoring, pumping tests, and groundwater sampling and analysis. It is combined with data such as those published by geological survey institutions and regional hydrogeological data from national or local hydrogeological databases to form a hydrogeological survey report. Indicates the direction of groundwater flow from the candidate pollution source to the first The cosine of the angle between the line segments at each sampling point is calculated using the following formula: ; Indicates the direction of groundwater flow and the coordinates of the candidate pollution source location. To the Coordinates of each sampling point The angle between the line segments; Indicates the first The location coordinates of each sampling point; Indicates the theoretically predicted concentration; Indicates the normalization factor; Indicates the basic diffusion range; Represents the residual.

[0057] In the steady-state inversion optimization model, the Levenberg-Marquardt algorithm or particle swarm optimization is used for iteration. Iteration constraints include: the source strength is always greater than zero, and the location of candidate pollution sources is limited to the candidate pollution source region. The iteration stopping criterion is that the change in the objective function is less than a preset iteration threshold based on actual conditions, or the maximum number of iterations is reached.

[0058] After the iteration converges, the location coordinates of the candidate pollution source that minimizes the objective function value are used as the location coordinates of the final pollution source, and the source strength of each controlling pollution factor is used as the pollution continuous release intensity.

[0059] In summary, a method for tracing the source of groundwater pollution in operating enterprises has been developed.

[0060] The order of the embodiments is for illustrative purposes only and does not represent the superiority or inferiority of the embodiments; the processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired results; in some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0061] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0062] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for tracing the source of groundwater pollution in an operating enterprise, characterized in that, Includes the following steps: S1. Standardize the original concentration values ​​of different pollutants to obtain standardized concentration values ​​of different pollutants; perform principal component analysis based on the standardized concentration values ​​of different pollutants, calculate the covariance matrix and perform eigenvalue decomposition to obtain loadings and screen the main controlling pollutants, and determine the comprehensive weight of the main controlling pollutants. S2. Based on the standardized concentration values ​​of different pollutants, a preliminary screening algorithm for pollution source candidate regions is used to obtain pollution source candidate regions; based on the pollution source candidate regions, the location coordinates of the final pollution source are calculated through a steady-state weighted convection-diffusion point source joint inversion algorithm. The preliminary screening algorithm for the pollution source candidate area includes: traversing all candidate pollution sources within the enterprise's factory area and calculating the Euclidean distance between each candidate pollution source and all sampling points; for each major control pollution factor, evaluating the correlation between the standardized concentration value at all sampling points and the Euclidean distance from the candidate pollution source to the corresponding sampling point using the Pearson correlation coefficient to obtain the concentration-distance correlation of the major control pollution factor; introducing a comprehensive weight for the major control pollution factor and summing the weighted correlations of all major control pollution factors to obtain a comprehensive score; and including candidate pollution sources whose comprehensive scores exceed the correlation threshold in the pollution source candidate area. The steady-state weighted convection-diffusion point source joint inversion algorithm includes: calculating the Euclidean distance from each sampling point to the candidate pollution source one by one; for each controlling pollution factor, based on the source strength of the controlling pollution factor, introducing a normalization factor and a convection-diffusion attenuation term, and independently calculating the theoretical predicted concentration of the controlling pollution factor; weighting the residuals between the original concentration values ​​of each controlling pollution factor at each sampling point and the corresponding theoretical predicted concentrations with the comprehensive weight of the controlling pollution factor, and summing the weighted squared residuals of all sampling points and all controlling pollution factors as the objective function; based on the candidate pollution source region, through iteration, outputting the location coordinates of the candidate pollution source that minimizes the objective function value and the source strength of each controlling pollution factor, thus obtaining the final location coordinates of the pollution source.

2. The method for tracing the source of groundwater pollution in an operating enterprise according to claim 1, characterized in that, S1 specifically includes: Perform eigenvalue decomposition on the covariance matrix to obtain eigenvalues ​​and corresponding eigenvectors. Each component in the eigenvector is the loading of the corresponding pollutant factor in the principal components.

3. The method for tracing the source of groundwater pollution in an operating enterprise according to claim 2, characterized in that, S1 specifically includes: The obtained eigenvalues ​​are sorted in descending order, and the contribution rate is calculated as the proportion of each eigenvalue to the sum of all eigenvalues.

4. The method for tracing the source of groundwater pollution in an operating enterprise according to claim 3, characterized in that, S1 specifically includes: Based on the contribution rate, the number of principal components to be retained is determined by setting a cumulative contribution rate threshold; within the range of retained principal components, pollutants whose absolute load value exceeds the load threshold are included as the main control pollutants in the set of main control pollutants.

5. The method for tracing the source of groundwater pollution in an operating enterprise according to claim 4, characterized in that, S1 specifically includes: The comprehensive weight of the main pollution factor is obtained by multiplying the absolute value of the loading of the main pollution factor by the corresponding contribution rate and summing them.

6. The method for tracing the source of groundwater pollution in an operating enterprise according to claim 1, characterized in that, S2 specifically includes: the convection-diffusion attenuation term is determined based on the cosine value of the angle between the groundwater flow direction and the line connecting the candidate pollution source to the sampling point; wherein the positive or negative value of the cosine value indicates whether the sampling point is downstream, upstream, or vertical relative to the groundwater flow direction.

Images