Underground water pollution intelligent response type permeable reactive barrier design system

Abstract

The invention relates to the technical field of environmental governance, in particular to an intelligent response type permeable reactive barrier design system for underground water pollution. According to the groundwater pollution intelligent response type permeable reactive barrier design system, a plurality of information bases are constructed, and hydrogeology, pollutants, reaction filler and other related information of a target site are comprehensively collected and sorted. A machine learning model and a three-dimensional finite difference model are constructed by using the information, a PRB wall design scheme can be predicted according to pollutant characteristics, and actual site effect evaluation can be performed on the scheme. By defining target conditions and constraint conditions and adopting genetic algorithm optimization, a PRB wall body design scheme which not only meets treatment effect requirements but also conforms to actual conditions is finally obtained, so that a scientific, efficient and comprehensive solution is provided for permeable reactive barrier design in groundwater pollution treatment.

Description

Technical Field

[0001] The invention relates to the technical field of environmental governance, in particular to a groundwater pollution intelligent responsive permeable reaction wall design system. Background Art

[0002] With the increase of industrialization and urbanization activities, the concentration of pollutants in soil and groundwater has continued to rise, posing a major threat to the environment and human health. In particular, the pollution of nitrate nitrogen and heavy metals in drinking water has had a serious impact on the groundwater environment and human health, and effective treatment technologies are urgently needed. Groundwater remediation technology is mainly divided into two categories: ex situ remediation and in situ remediation. Among them, in situ remediation is more popular because of its convenient operation and low cost. Permeable reactive wall (PRB), as an effective method in in situ remediation technology, is widely used to intercept and purify pollutants in groundwater.

[0003] Key factors in PRB design include the width of the reaction wall and the choice of filling materials, which directly affect the reaction effect and service life of the PRB. Accurately determining the thickness of the PRB is the key to ensuring that it can effectively treat pollutants in the long term. Traditional PRB width design methods, such as continuous column experiments based on breakthrough curves and empirical formula calculations, have limitations such as high time cost and insufficient explanatory power for mathematical regression models. These methods often fail to accurately reflect the complexity of the actual groundwater environment and pollutant migration, resulting in deviations between the design results and actual needs.

[0004] Machine learning algorithms are capable of processing complex data relationships and pattern recognition, and have been widely used in environmental assessment and prediction in recent years. Three-dimensional numerical simulation is also an important means to evaluate the effect of PRB. The present invention applies a machine learning model to the prediction of PRB width to improve the efficiency and accuracy of the design, and forms a visualization module through a three-dimensional finite difference model. In summary, the research background of the intelligent responsive permeable reactive wall design system for groundwater pollution mainly revolves around the severity of groundwater pollution problems, the effectiveness of in-situ remediation technology, the importance of PRB design, the limitations of traditional design methods, and the introduction and potential application of machine learning technology. Summary of the invention

[0005] The present invention overcomes the shortcomings of the prior art and provides a groundwater pollution intelligent responsive permeable reaction wall design system.

[0006] To achieve the above-mentioned purpose, the technical solution adopted by the present invention is:

[0007] The present invention discloses a groundwater pollution intelligent responsive permeable reaction wall design system, the design system comprising:

[0008] Constructing a site information database, collecting three-dimensional hydrogeological parameters and drilling data of the target site, and using the drilling data to generate a three-dimensional geological spatial distribution map and flow field distribution of the target site according to spatial interpolation information, and importing the three-dimensional hydrogeological parameters, three-dimensional geological spatial distribution map and flow field distribution of the target site into the site information database;

[0009] Constructing a pollutant information library, obtaining characteristic information of various pollutants, and importing the characteristic information of various pollutants into the pollutant information library; constructing a filler information library, obtaining characteristic information of various reactive fillers, and importing the characteristic information of various reactive fillers into the filler information library;

[0010] Constructing a pairing information library, collecting historical monitoring data, pollution source information, and pollutant migration patterns of the target site, determining available reaction fillers for various pollutants based on the historical monitoring data, pollution source information, and pollutant migration patterns of the target site, and importing the available reaction fillers for various pollutants into the pairing information library;

[0011] Constructing a machine learning model, and importing characteristic information of various pollutants and available reaction fillers into the machine learning model for training, thereby obtaining a trained machine learning model;

[0012] Acquire the three-dimensional hydrogeological parameters, three-dimensional geological spatial distribution map and flow field distribution of the target site in the site information database, and construct a three-dimensional finite difference model of the target site based on the three-dimensional hydrogeological parameters, three-dimensional geological spatial distribution map and flow field distribution of the target site in combination with the finite difference method;

[0013] Obtain pollutant information of the target site, import the pollutant information of the target site into the trained machine learning model, and obtain the PRB wall design plan of the target site; import the PRB wall design plan of the target site into the three-dimensional finite difference model to obtain the final PRB wall design plan of the target site.

[0014] More specifically, the three-dimensional geological spatial distribution map of the target site is generated based on the spatial interpolation information using the borehole data, specifically:

[0015] Acquire drilling data of each borehole in the target site, the drilling data including coordinate information of the drilling location, geological information, hydraulic conductivity, water level data, porosity data, rock layer thickness data and stratum lithology;

[0016] Determine a three-dimensional coordinate system according to the scope of the target site, and locate the coordinates of each borehole in the coordinate system;

[0017] Taking the stratum lithology and thickness data of the drilling location as the basic data point, starting from the bottom layer, for each stratum, the stratum lithology and thickness of each three-dimensional coordinate point in the target site are calculated according to the Kriging interpolation method;

[0018] In the horizontal direction, according to the plane position relationship and thickness change of the boreholes, the interpolation range is gradually expanded to determine the distribution boundary of the same stratum on the plane;

[0019] In the vertical direction, according to the up-down relationship and thickness changes of the strata in different boreholes, the continuous distribution of the strata in the vertical direction is constructed by interpolation;

[0020] For geological structure data, including faults and folds, they are incorporated into the interpolation process as constraints. If there is dislocation of the strata on both sides of the fault, the dislocation is calibrated in three-dimensional space according to the coordinates of the fault position; the bending shape of the fold is constructed in three-dimensional space according to the corresponding shape data in the drilling data;

[0021] By continuously integrating the interpolation results of various strata and geological structures, a three-dimensional geological spatial distribution map of the target site is eventually formed.

[0022] More specifically, the flow field distribution of the target site is generated based on the spatial interpolation information using the borehole data, specifically:

[0023] Using the three-dimensional coordinate system of the target site as a framework, each borehole is positioned according to its coordinates, and the water level data in the borehole is used as a feature basis. Based on the inverse distance weighted interpolation method, the water level value of each three-dimensional coordinate point in the target site is obtained, thereby obtaining the spatial distribution of the water level;

[0024] Based on the hydraulic conductivity data and Darcy's law, the hydraulic gradient of each three-dimensional coordinate point is calculated based on the water level distribution. In the calculation process, the influence of the anisotropy of the formation on the hydraulic conductivity is combined to calculate the hydraulic gradient according to the characteristics of different formations.

[0025] Using the hydraulic conductivity and the calculated hydraulic gradient, the groundwater velocity vector is calculated again according to Darcy's law to determine the direction and velocity of groundwater flow in the corresponding stratum;

[0026] The flow field distribution of the target site is constructed based on the direction and speed of groundwater flow in each stratum and the spatial distribution of water level.

[0027] More specifically, available reactive fillers for various pollutants are determined based on historical monitoring data of the target site, pollution source information, and pollutant migration patterns, specifically:

[0028] Obtain historical monitoring data and pollution source information of the target site;

[0029] Determine all pollutants that have appeared in the target site based on the historical monitoring data, and determine the emission intensity and emission mode of various pollutants that have appeared in the target site from the pollution source information;

[0030] Obtaining the migration rules of various pollutants that have appeared in the target site, and determining the diffusion characteristics of various pollutants that have appeared in the target site according to the migration rules; wherein the diffusion characteristics include diffusion coefficients and diffusion directions;

[0031] and obtaining characteristic information of various pollutants that have appeared in the target site from the pollutant information database;

[0032] Generate characteristic text information of various pollutants that have appeared in the target site based on the characteristic information, emission intensity, emission mode and dispersion characteristics of various pollutants that have appeared in the target site;

[0033] Acquire characteristic information of each reaction filler in the filler information library, introduce a local sensitive attention mechanism, and analyze the attention score between the characteristic information of each reaction filler and the characteristic text information of various pollutants that have appeared in the target site according to the local sensitive attention mechanism;

[0034] The attention scores between the characteristic information of each reactive filler and the characteristic text information of various pollutants that have appeared in the target site are compared and analyzed with the preset score threshold;

[0035] If the attention score between the characteristic information of a certain reaction filler and the characteristic text information of a certain pollutant that has appeared in the target site is greater than a preset score threshold, the reaction filler is marked as an available reaction filler for the corresponding pollutant;

[0036] After the attention scores between the characteristic information of each reaction filler and the characteristic text information of various pollutants that have appeared in the target site are compared and analyzed with the preset score threshold, the available reaction fillers of various pollutants that have appeared in the target site are obtained.

[0037] More specifically, a three-dimensional finite difference model of the target site is constructed based on the three-dimensional hydrogeological parameters, three-dimensional geological spatial distribution map and flow field distribution of the target site and combined with the finite difference method, specifically:

[0038] According to the three-dimensional geological spatial distribution map of the target site, the spatial layout, thickness variation, interlayer contact relationship and geological structural characteristics of different strata are obtained to determine the spatial framework and stratum boundary conditions of the model;

[0039] And according to the flow field distribution of the target site, the spatial distribution of groundwater level, the magnitude and direction of hydraulic gradient and the dynamic characteristics of water flow are obtained;

[0040] According to the geometric shape, size, spatial structure, stratum boundary conditions, groundwater level spatial distribution, hydraulic gradient size and direction, and water flow dynamic characteristics of the target site, the target site is divided into three-dimensional spatial grids to obtain a number of grid nodes;

[0041] For each grid node, the partial differential equation describing groundwater movement is discretized using the finite difference method to obtain the equivalent value of the source-sink intensity of each grid node. At the same time, the zero flux boundary conditions of each grid node are determined based on the three-dimensional geological spatial distribution map and flow field distribution.

[0042] Substitute the equivalent value of the source-sink intensity of each grid node into the corresponding position in the discrete equation, and set the boundary constraints of the discrete equation according to the zero flux boundary conditions determined by the three-dimensional geological spatial distribution map and the flow field distribution;

[0043] Then, the source and sink terms in the discrete equations are adjusted according to the source and sink information in the flow field distribution, and the discrete equations of all grid nodes are combined according to the three-dimensional spatial coordinate relationship to construct a three-dimensional finite difference model of the target site.

[0044] More specifically, a machine learning model is constructed, and characteristic information of various pollutants and available reactive fillers are introduced into the machine learning model for training to obtain a trained machine learning model, specifically:

[0045] Acquiring characteristic information of various pollutants in the pollutant information library, and acquiring available reactive fillers of various pollutants in the pairing information library;

[0046] Search the data network based on the characteristic information of various pollutants and available reactive fillers to obtain PRB wall design solutions for treating various pollutants;

[0047] The characteristic information of various pollutants and available reactive fillers are used as input feature nodes, and the PRB wall design scheme for treating various pollutants is used as output feature nodes; a mapping relationship topology diagram is constructed according to the input feature nodes and the output feature nodes;

[0048] Building a machine learning model based on a graph neural network, converting the connection relationship between the input feature nodes and the output feature nodes in the mapping relationship topology graph into a data structure recognizable by the machine learning model, encoding the information of the nodes and edges in the mapping relationship topology graph according to the graph general relationship and embedding it into the parameter space of the machine learning model for learning and training;

[0049] After each training cycle is completed, the feature contribution of the machine learning model is obtained. If the feature contribution of the machine learning model gradually increases and is greater than a preset threshold, the training is completed and the trained machine learning model is output.

[0050] More specifically, the pollutant information of the target site is obtained, and the pollutant information of the target site is imported into the trained machine learning model to obtain the PRB wall design scheme of the target site; the PRB wall design scheme of the target site is imported into the three-dimensional finite difference model to obtain the final PRB wall design scheme of the target site, specifically:

[0051] Obtaining pollutant information of the target site, importing the pollutant information of the target site into the trained machine learning model, and obtaining a PRB wall design plan of the target site; the PRB wall design plan includes a PRB wall design thickness, position, and filler;

[0052] Importing the PRB wall design of the target site into the three-dimensional finite difference model for evaluation, and obtaining the numerical simulation results of the permeable reactive wall;

[0053] Define target conditions based on minimizing pollutant concentrations and maximizing PRB life, and set constraints based on geological conditions, cost budget, and construction feasibility;

[0054] Determine whether the numerical simulation results of the permeable reaction wall meet the target conditions and the constraint conditions; if the numerical simulation results of the permeable reaction wall meet the target conditions and the constraint conditions, the PRB wall design scheme is directly output as the final PRB wall design scheme;

[0055] Otherwise, a genetic algorithm is introduced, and an iterative optimization cycle is set, and the PRB wall design scheme is adjusted and optimized according to the genetic algorithm. After each iterative optimization cycle is completed, an optimized PRB wall design scheme is obtained;

[0056] The optimized PRB wall design scheme is input into the three-dimensional finite difference model for evaluation to obtain the numerical simulation results of the permeable reaction wall;

[0057] Until the numerical simulation results of the permeable reaction wall of the optimized PRB wall design scheme meet the target conditions and constraint conditions, the iteration is stopped, and the optimized PRB wall is output as the final PRB wall design scheme.

[0058] More specifically, the three-dimensional hydrogeological parameters include strata, water levels, horizontal / vertical permeability coefficients, groundwater flow rates, porosity, and water storage rates.

[0059] More specifically, the characteristic information of the reaction filler includes material, ratio, size, specific surface area and reaction rate constant; the characteristic information of the pollutant includes the negative logarithm of the acid dissociation constant and the molecular weight.

[0060] The present invention solves the technical defects existing in the background technology, and the present invention has the following beneficial effects: This groundwater pollution intelligent responsive permeable reaction wall design system comprehensively collects and organizes relevant information such as hydrogeology, pollutants, reaction fillers, etc. of the target site by constructing multiple information bases. By using this information to construct a machine learning model and a three-dimensional finite difference model, it is possible to predict the PRB wall design scheme according to the characteristics of the pollutants, and to evaluate the actual site effect of the scheme. By defining the target conditions and constraints, and using genetic algorithm optimization, a PRB wall design scheme that meets the treatment effect requirements and conforms to the actual situation is finally obtained, thereby providing a scientific, efficient and comprehensive solution for the design of permeable reaction walls in groundwater pollution control. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, drawings of other embodiments can be obtained based on these drawings without paying creative work.

[0062] Figure 1 Framework diagram of the system for designing a smart responsive permeable reactive wall for local groundwater pollution;

[0063] Figure 2 Flowchart of the system for designing a smart responsive permeable reactive wall for local groundwater pollution. DETAILED DESCRIPTION

[0064] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0065] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0066] like Figure 1 , 2 As shown, the present invention discloses a groundwater pollution intelligent responsive permeable reaction wall design system, the design system comprising:

[0067] S102: constructing a site information database, collecting three-dimensional hydrogeological parameters and drilling data of the target site, and using the drilling data to generate a three-dimensional geological spatial distribution map and flow field distribution of the target site according to spatial interpolation information, and importing the three-dimensional hydrogeological parameters, three-dimensional geological spatial distribution map and flow field distribution of the target site into the site information database;

[0068] S104: constructing a pollutant information library, obtaining characteristic information of various pollutants, and importing the characteristic information of various pollutants into the pollutant information library; constructing a filler information library, obtaining characteristic information of various reactive fillers, and importing the characteristic information of various reactive fillers into the filler information library;

[0069] S106: constructing a pairing information library, collecting historical monitoring data, pollution source information, and pollutant migration rules of the target site, determining available reaction fillers for various pollutants based on the historical monitoring data, pollution source information, and pollutant migration rules of the target site, and importing the available reaction fillers for various pollutants into the pairing information library;

[0070] S108: constructing a machine learning model, and importing characteristic information of various pollutants and available reaction fillers into the machine learning model for training, thereby obtaining a trained machine learning model;

[0071] S110: Acquire the three-dimensional hydrogeological parameters, three-dimensional geological spatial distribution map and flow field distribution of the target site in the site information database, and construct a three-dimensional finite difference model of the target site according to the three-dimensional hydrogeological parameters, three-dimensional geological spatial distribution map and flow field distribution of the target site in combination with the finite difference method;

[0072] S112: Obtain pollutant information of the target site, import the pollutant information of the target site into the trained machine learning model, and obtain the PRB wall design plan of the target site; import the PRB wall design plan of the target site into the three-dimensional finite difference model to obtain the final PRB wall design plan of the target site.

[0073] Among them, the three-dimensional hydrogeological parameters include strata, water level, horizontal / vertical permeability, groundwater flow rate, porosity and water storage rate; the characteristic information of the reaction filler includes material, ratio, size, specific surface area and reaction rate constant; the characteristic information of the pollutant includes the negative logarithm of the acid dissociation constant and molecular weight of the pollutant.

[0074] It should be noted that this groundwater pollution intelligent responsive permeable reactive wall design system comprehensively collects and organizes relevant information such as hydrogeology, pollutants, reactive fillers, etc. of the target site by building multiple information databases. This information is used to build machine learning models and three-dimensional finite difference models. The former can predict the PRB wall design scheme based on the characteristics of pollutants, and the latter can evaluate the actual site effect of the scheme. By defining target conditions and constraints and using genetic algorithm optimization, a PRB wall design scheme that meets the treatment effect requirements (such as minimizing pollutant concentrations and maximizing PRB life) and conforms to actual conditions (geological conditions, cost budget, construction feasibility) is finally obtained, thus providing a scientific, efficient and comprehensive solution for the design of permeable reactive walls in groundwater pollution control.

[0075] More specifically, the three-dimensional geological spatial distribution map of the target site is generated based on the spatial interpolation information using the borehole data, specifically:

[0076] Acquire drilling data of each borehole in the target site, the drilling data including coordinate information of the drilling location, geological information, hydraulic conductivity, water level data, porosity data, rock layer thickness data and stratum lithology;

[0077] Determine a three-dimensional coordinate system according to the scope of the target site, and locate the coordinates of each borehole in the coordinate system;

[0078] Taking the stratum lithology and thickness data of the drilling location as the basic data point, starting from the bottom layer, for each stratum, the stratum lithology and thickness of each three-dimensional coordinate point in the target site are calculated according to the Kriging interpolation method;

[0079] In the horizontal direction, according to the plane position relationship and thickness change of the boreholes, the interpolation range is gradually expanded to determine the distribution boundary of the same stratum on the plane;

[0080] In the vertical direction, according to the up-down relationship and thickness changes of the strata in different boreholes, the continuous distribution of the strata in the vertical direction is constructed by interpolation;

[0081] For geological structure data, including faults and folds, they are incorporated into the interpolation process as constraints. If there is dislocation of the strata on both sides of the fault, the dislocation is calibrated in three-dimensional space according to the coordinates of the fault position; the bending shape of the fold is constructed in three-dimensional space according to the corresponding shape data in the drilling data;

[0082] By continuously integrating the interpolation results of various strata and geological structures, a three-dimensional geological spatial distribution map of the target site is eventually formed.

[0083] Specifically, firstly, the rich borehole data of each borehole in the target site are obtained. These data cover the coordinate information of the borehole location (which is the basis for determining the location of the borehole in space), geological information (such as stratum lithology, which is the key content for constructing geological spatial distribution), hydraulic conductivity (related to groundwater flow), water level data (reflecting groundwater status), porosity data (affecting the water storage performance of the stratum), rock thickness data (important parameters for determining the stratum structure), etc. The coordinate system of the three-dimensional space is determined according to the scope of the target site, which provides a framework for the spatial positioning of all subsequent data. The coordinates of each borehole are located in this coordinate system so that the borehole data can be accurately mapped to the three-dimensional space of the target site. The Kriging interpolation method is used based on the stratum lithology and thickness data of the borehole location as the basic data point. Starting from the bottom layer, calculations are performed for each stratum. In the horizontal direction, the interpolation range is expanded by considering the plane position relationship and thickness change of the boreholes, so as to determine the distribution boundary of the same stratum on the plane. In the vertical direction, interpolation is performed according to the up-down relationship and thickness change of the strata in different boreholes to construct the continuous distribution of the strata in the vertical direction. This interpolation method can reasonably infer the stratigraphic distribution of the entire target site based on limited borehole data. For geological structure data (faults and folds), they are incorporated into the interpolation process as constraints. For faults, if there is dislocation in the strata on both sides of the fault, the dislocation relationship is accurately calibrated in three-dimensional space according to the coordinates of the fault position. For folds, their curved shapes are constructed in three-dimensional space based on the corresponding morphological data in the borehole data. This enables the geological structure to be accurately reflected in the three-dimensional geological space distribution map, reflecting the geological conditions of the target site more realistically. By continuously integrating the interpolation results of various strata and geological structures, the distribution of all strata and the information of geological structures are combined to finally form a complete three-dimensional geological space distribution map of the target site.

[0084] Through the above steps, the limited drilling data can be used to comprehensively and accurately construct the three-dimensional geological spatial distribution map of the target site. This method fully considers various factors such as the lithology, thickness, and geological structure of the strata, and reasonably infers the geological conditions of the entire site through the Kriging interpolation method, so that the constructed three-dimensional geological spatial distribution map can accurately reflect the stratum distribution, geological structure relationship and other information of the target site.

[0085] More specifically, the flow field distribution of the target site is generated based on the spatial interpolation information using the borehole data, specifically:

[0086] Using the three-dimensional coordinate system of the target site as a framework, each borehole is positioned according to its coordinates, and the water level data in the borehole is used as a feature basis. Based on the inverse distance weighted interpolation method, the water level value of each three-dimensional coordinate point in the target site is obtained, thereby obtaining the spatial distribution of the water level;

[0087] Based on the hydraulic conductivity data and Darcy's law, the hydraulic gradient of each three-dimensional coordinate point is calculated based on the water level distribution. In the calculation process, the influence of the anisotropy of the formation on the hydraulic conductivity is combined to calculate the hydraulic gradient according to the characteristics of different formations.

[0088] Using the hydraulic conductivity and the calculated hydraulic gradient, the groundwater velocity vector is calculated again according to Darcy's law to determine the direction and velocity of groundwater flow in the corresponding stratum;

[0089] The flow field distribution of the target site is constructed based on the direction and speed of groundwater flow in each stratum and the spatial distribution of water level.

[0090] Specifically, the three-dimensional coordinate system established in the target site is used as a framework, and each borehole is positioned in this system according to its coordinates. Then, the water level data in the borehole is used as the key basis, and the water level value of each three-dimensional coordinate point in the target site is calculated by the inverse distance weighted interpolation method. The principle of the inverse distance weighted interpolation method is that the closer the unknown point is to the known point, the greater the influence of the known point. This method can reasonably infer the spatial distribution of the water level of the entire site based on the water level data of the borehole. According to the hydraulic conductivity data, combined with Darcy's law (Darcy's law describes the relationship between the flow rate of fluid in porous media and hydraulic gradient and hydraulic conductivity), the hydraulic gradient of each three-dimensional coordinate point is calculated. In this process, the influence of the anisotropy of the formation on the hydraulic conductivity is fully considered. Due to the different characteristics of different formations (such as rock structure, porosity, etc.), the hydraulic conductivity may be different in different directions. Therefore, it is necessary to accurately calculate the hydraulic gradient according to the characteristics of different formations to reflect the actual driving force of groundwater flow.

[0091] Using the known hydraulic conductivity and the calculated hydraulic gradient, the groundwater velocity vector is calculated again according to Darcy's law. The velocity vector contains the direction and velocity information of groundwater flow, and determines the direction and velocity of groundwater flow in the corresponding stratum. Finally, the flow field distribution of the target site is constructed by combining the direction and velocity of groundwater flow in each stratum and the spatial distribution of water level. The flow field distribution can intuitively display the flow state of groundwater in the target site, including information such as water flow velocity, direction, and water level at different locations.

[0092] Through the above steps, the flow field distribution of the target site can be accurately constructed based on the borehole data. This method fully considers factors such as water level data, hydraulic conductivity, and formation characteristics, and uses scientific principles such as inverse distance weighted interpolation and Darcy's law to accurately calculate key elements such as water level distribution, hydraulic gradient, and velocity vector. The constructed flow field distribution can truly and comprehensively reflect the flow state of groundwater in the target site, providing important basic data for studying the movement of groundwater, the migration and diffusion of pollutants, and the design of permeable reaction walls.

[0093] More specifically, available reactive fillers for various pollutants are determined based on historical monitoring data of the target site, pollution source information, and pollutant migration patterns, specifically:

[0094] Obtain historical monitoring data and pollution source information of the target site;

[0095] Determine all pollutants that have appeared in the target site based on the historical monitoring data, and determine the emission intensity and emission mode of various pollutants that have appeared in the target site from the pollution source information;

[0096] Obtaining the migration rules of various pollutants that have appeared in the target site, and determining the diffusion characteristics of various pollutants that have appeared in the target site according to the migration rules; wherein the diffusion characteristics include diffusion coefficients and diffusion directions;

[0097] and obtaining characteristic information of various pollutants that have appeared in the target site from the pollutant information database;

[0098] Generate characteristic text information of various pollutants that have appeared in the target site based on the characteristic information, emission intensity, emission mode and dispersion characteristics of various pollutants that have appeared in the target site;

[0099] Acquire characteristic information of each reaction filler in the filler information library, introduce a local sensitive attention mechanism, and analyze the attention score between the characteristic information of each reaction filler and the characteristic text information of various pollutants that have appeared in the target site according to the local sensitive attention mechanism;

[0100] The attention scores between the characteristic information of each reactive filler and the characteristic text information of various pollutants that have appeared in the target site are compared and analyzed with the preset score threshold;

[0101] If the attention score between the characteristic information of a certain reaction filler and the characteristic text information of a certain pollutant that has appeared in the target site is greater than a preset score threshold, the reaction filler is marked as an available reaction filler for the corresponding pollutant;

[0102] After the attention scores between the characteristic information of each reaction filler and the characteristic text information of various pollutants that have appeared in the target site are compared and analyzed with the preset score threshold, the available reaction fillers of various pollutants that have appeared in the target site are obtained.

[0103] Specifically, first obtain the historical monitoring data and pollution source information of the target site, determine all types of pollutants that have appeared in the site through historical monitoring data, and clarify the emission intensity and emission mode of various pollutants from the pollution source information. At the same time, obtain the migration law of pollutants to determine their diffusion characteristics (diffusion coefficient and diffusion direction), and obtain the characteristic information of pollutants from the pollutant information database. This information comprehensively covers the existence, source and migration of pollutants in the target site.

[0104] Generate characteristic text information of various pollutants in the target site based on the characteristic information, emission intensity, emission mode and dispersion characteristics of pollutants. This process integrates scattered pollutant-related information into a text form that is convenient for subsequent analysis and comparison, and can comprehensively represent the unique properties and behavior patterns of each pollutant.

[0105] After obtaining the characteristic information of each reactive filler in the filler information library, a local sensitive attention mechanism is introduced. This mechanism can analyze the attention score between the reactive filler characteristic information and the pollutant characteristic text information. The attention score reflects the degree of association between the reactive filler and the pollutant. The local sensitive attention mechanism can more accurately capture the key parts of this association, thereby more accurately measuring the degree of match between the two. The attention score is compared and analyzed with the preset score threshold. If the attention score between a reactive filler and a pollutant is greater than the preset score threshold, the reactive filler is marked as an available reactive filler for the corresponding pollutant. By performing such a comparative analysis of the attention scores between all reactive fillers and pollutants, the available reactive fillers for various pollutants in the target site are finally determined.

[0106] Through the above steps, the available reactive fillers for various pollutants can be accurately determined based on the historical monitoring data, pollution source information and pollutant migration rules of the target site. This method comprehensively considers the multi-faceted characteristics of pollutants and integrates them into feature text information. At the same time, it uses the local sensitive attention mechanism to accurately analyze the degree of correlation between reactive fillers and pollutants, and selects available reactive fillers by comparing with the threshold. This provides a scientific basis for the design of permeable reactive walls in groundwater pollution control, ensuring the selection of appropriate reactive fillers to effectively treat pollutants in the target site.

[0107] More specifically, a three-dimensional finite difference model of the target site is constructed based on the three-dimensional hydrogeological parameters, three-dimensional geological spatial distribution map and flow field distribution of the target site and combined with the finite difference method, specifically:

[0108] According to the three-dimensional geological spatial distribution map of the target site, the spatial layout, thickness variation, interlayer contact relationship and geological structural characteristics of different strata are obtained to determine the spatial framework and stratum boundary conditions of the model;

[0109] And according to the flow field distribution of the target site, the spatial distribution of groundwater level, the magnitude and direction of hydraulic gradient and the dynamic characteristics of water flow are obtained;

[0110] According to the geometric shape, size, spatial structure, stratum boundary conditions, groundwater level spatial distribution, hydraulic gradient size and direction, and water flow dynamic characteristics of the target site, the target site is divided into three-dimensional spatial grids to obtain a number of grid nodes;

[0111] For each grid node, the partial differential equation describing groundwater movement is discretized using the finite difference method to obtain the equivalent value of the source-sink intensity of each grid node. At the same time, the zero flux boundary conditions of each grid node are determined based on the three-dimensional geological spatial distribution map and flow field distribution.

[0112] Substitute the equivalent value of the source-sink intensity of each grid node into the corresponding position in the discrete equation, and set the boundary constraints of the discrete equation according to the zero flux boundary conditions determined by the three-dimensional geological spatial distribution map and the flow field distribution;

[0113] Then, the source and sink terms in the discrete equations are adjusted according to the source and sink information in the flow field distribution, and the discrete equations of all grid nodes are combined according to the three-dimensional spatial coordinate relationship to construct a three-dimensional finite difference model of the target site.

[0114] Specifically, the spatial layout, thickness variation, interlayer contact relationship and geological structural characteristics of different strata are obtained from the three-dimensional geological spatial distribution map, which provides the spatial framework and stratigraphic boundary conditions for constructing the three-dimensional finite difference model. For example, the spatial layout of the strata determines the basic framework of the model, and the interlayer contact relationship and geological structural characteristics (such as faults, folds, etc.) will affect the flow path and boundary conditions of groundwater. At the same time, the spatial distribution of groundwater level, the magnitude and direction of hydraulic gradient and the dynamic characteristics of water flow are obtained according to the flow field distribution. These are key information to describe the state of groundwater movement and provide a basis for water flow related calculations in subsequent model construction. The three-dimensional spatial grid is divided according to the geometric shape and size of the target site and the spatial framework, stratigraphic boundary conditions, spatial distribution of groundwater level, magnitude and direction of hydraulic gradient and dynamic characteristics of water flow determined previously, and several grid nodes are obtained. Grid division needs to comprehensively consider multiple factors. For example, in areas with complex geological structures or drastic changes in water flow, finer grids may be required to accurately capture the characteristics of these areas; while in relatively uniform areas, the grid size can be appropriately increased to reduce the amount of calculation.

[0115] For each grid node, the finite difference method is used to discretize the partial differential equation describing groundwater movement, and the equivalent value of the source-sink intensity of each grid node is obtained. The equivalent value of the source-sink intensity reflects the inflow or outflow of groundwater at each node, and is an important parameter in the model to describe the material and energy exchange of the groundwater system. At the same time, the zero flux boundary conditions of each grid node are determined according to the three-dimensional geological spatial distribution map and the flow field distribution. For example, the water-blocking boundary is determined according to the geological structure, and the zero flux condition is set on these boundaries to ensure that the model conforms to the actual geological and water flow conditions. The equivalent value of the source-sink intensity of each grid node is substituted into the corresponding position in the discrete equation, and the boundary constraints of the discrete equation are set according to the determined zero flux boundary conditions, which enables the discrete equation to accurately reflect the actual boundary conditions. Then, the source-sink term in the discrete equation is adjusted according to the source-sink information in the flow field distribution to ensure that the setting of the source-sink term is consistent with the actual groundwater recharge and discharge. Finally, the discrete equations of all grid nodes are combined according to the three-dimensional spatial coordinate relationship to construct a three-dimensional finite difference model of the target site. This model can comprehensively reflect the three-dimensional hydrogeological characteristics and groundwater movement laws of the target site.

[0116] The three-dimensional finite difference model constructed through the above steps can fully and accurately reflect the three-dimensional hydrogeological conditions and groundwater movement laws of the target site. It comprehensively considers multiple factors such as stratigraphic structure, geological structure, groundwater level, hydraulic gradient, etc., and transforms complex actual conditions into a computable mathematical model through reasonable grid division, discretization operation and boundary condition setting. This model provides a reliable foundation for subsequent research on groundwater flow, migration and diffusion of pollutants, etc. It can be used to simulate and predict various phenomena related to groundwater in the target site, and provide effective decision-making support for groundwater pollution control and water resources management.

[0117] More specifically, a machine learning model is constructed, and characteristic information of various pollutants and available reactive fillers are introduced into the machine learning model for training to obtain a trained machine learning model, specifically:

[0118] Acquiring characteristic information of various pollutants in the pollutant information library, and acquiring available reactive fillers of various pollutants in the pairing information library;

[0119] Search the data network based on the characteristic information of various pollutants and available reactive fillers to obtain PRB wall design solutions for treating various pollutants;

[0120] The characteristic information of various pollutants and available reactive fillers are used as input feature nodes, and the PRB wall design scheme for treating various pollutants is used as output feature nodes; a mapping relationship topology diagram is constructed according to the input feature nodes and the output feature nodes;

[0121] Building a machine learning model based on a graph neural network, converting the connection relationship between the input feature nodes and the output feature nodes in the mapping relationship topology graph into a data structure recognizable by the machine learning model, encoding the information of the nodes and edges in the mapping relationship topology graph according to the graph general relationship and embedding it into the parameter space of the machine learning model for learning and training;

[0122] After each training cycle is completed, the feature contribution of the machine learning model is obtained. If the feature contribution of the machine learning model gradually increases and is greater than a preset threshold, the training is completed and the trained machine learning model is output.

[0123] Specifically, first, the characteristic information of various pollutants is obtained from the pollutant information library, and the available reactive fillers of various pollutants are obtained from the pairing information library. Then, the data network is searched based on this information to obtain the PRB (permeable reactive wall) wall design scheme for various pollutant treatment. This step establishes the preliminary connection between the pollutant characteristics and the available reactive fillers and the PRB wall design scheme. The characteristic information of the pollutants and the available reactive fillers are used as input feature nodes, and the PRB wall design scheme is used as the output feature node to construct a mapping relationship topology map. This topology map intuitively shows the relationship structure between the input and output, provides a relationship framework for the construction of subsequent machine learning models, and helps to understand the influence of different factors on the PRB wall design. The machine learning model is constructed based on the graph neural network. In this process, the connection relationship between the input feature nodes and the output feature nodes in the mapping relationship topology map is converted into a data structure that can be recognized by the machine learning model. The information of the nodes and edges in the topology map is encoded according to the graph general relationship and embedded in the parameter space of the machine learning model for learning and training. This method enables the model to learn the complex relationship between the input features (pollutant characteristics and reactive fillers) and the output features (PRB wall design scheme). After each training cycle, the feature contribution of the machine learning model is obtained. The feature contribution reflects the importance of each feature (input feature node) in the model to the output result (PRB wall design plan). When the feature contribution gradually increases and is greater than the preset threshold, it indicates that the model has fully learned the relationship between input and output. At this time, the training is completed and the trained machine learning model is output. This judgment mechanism ensures that the model has sufficient accuracy and reliability.

[0124] The machine learning model constructed and trained through the above steps can effectively learn the complex mapping relationship between pollutant characteristic information, available reactive fillers and PRB wall design schemes. In an innovative way based on graph neural networks, these relationships are integrated into the model structure, and model training is optimized by continuously evaluating feature contributions. The final trained machine learning model can accurately predict the appropriate PRB wall design scheme based on pollutant characteristics and available reactive fillers, providing an efficient and accurate decision support tool for the design of permeable reactive walls in groundwater pollution control.

[0125] More specifically, the pollutant information of the target site is obtained, and the pollutant information of the target site is imported into the trained machine learning model to obtain the PRB wall design scheme of the target site; the PRB wall design scheme of the target site is imported into the three-dimensional finite difference model to obtain the final PRB wall design scheme of the target site, specifically:

[0126] Obtaining pollutant information of the target site, importing the pollutant information of the target site into the trained machine learning model, and obtaining a PRB wall design plan of the target site; the PRB wall design plan includes a PRB wall design thickness, position, and filler;

[0127] Importing the PRB wall design of the target site into the three-dimensional finite difference model for evaluation, and obtaining the numerical simulation results of the permeable reactive wall;

[0128] Define target conditions based on minimizing pollutant concentrations and maximizing PRB life, and set constraints based on geological conditions, cost budget, and construction feasibility;

[0129] Determine whether the numerical simulation results of the permeable reaction wall meet the target conditions and the constraint conditions; if the numerical simulation results of the permeable reaction wall meet the target conditions and the constraint conditions, the PRB wall design scheme is directly output as the final PRB wall design scheme;

[0130] Otherwise, a genetic algorithm is introduced, and an iterative optimization cycle is set, and the PRB wall design scheme is adjusted and optimized according to the genetic algorithm. After each iterative optimization cycle is completed, an optimized PRB wall design scheme is obtained;

[0131] The optimized PRB wall design scheme is input into the three-dimensional finite difference model for evaluation to obtain the numerical simulation results of the permeable reaction wall;

[0132] Until the numerical simulation results of the permeable reaction wall of the optimized PRB wall design scheme meet the target conditions and constraint conditions, the iteration is stopped, and the optimized PRB wall is output as the final PRB wall design scheme.

[0133] Specifically, the pollutant information of the target site is first obtained and input into the trained machine learning model to obtain the PRB wall design scheme of the target site, which includes key information such as the design thickness, location and filler of the PRB wall. The machine learning model is trained based on the mapping relationship between the previously constructed pollutant characteristic information, available reactive fillers and the PRB wall design scheme, so it can generate a preliminary PRB wall design scheme based on the input target site pollutant information. The PRB wall design scheme obtained from the machine learning model is imported into the three-dimensional finite difference model for evaluation to obtain the numerical simulation results of the permeable reactive wall. The three-dimensional finite difference model can simulate the three-dimensional hydrogeological conditions and groundwater movement laws of the target site. Through this model, the effect of the PRB wall design scheme in the actual site can be analyzed, such as the migration of pollutants in the presence of PRB walls. The target conditions are defined based on minimizing the pollutant concentration and maximizing the PRB life, which is considered from the perspective of governance effect and long-term effectiveness of PRB. At the same time, constraints are set according to geological conditions (such as stratum structure, groundwater level, etc.), cost budget (including material cost, construction cost, etc.) and construction feasibility (such as construction difficulty, site accessibility, etc.). These target conditions and constraints provide standards for evaluating whether the PRB wall design scheme is feasible. It is judged whether the numerical simulation results of the permeable reaction wall meet both the target conditions and the constraints. If not, a genetic algorithm is introduced for optimization. Genetic algorithm is an optimization algorithm based on the principle of biological evolution. By setting an iterative optimization cycle, the PRB wall design scheme is adjusted and optimized. After each iterative cycle, the optimized PRB wall design scheme is obtained, and then it is input into the three-dimensional finite difference model for re-evaluation until the numerical simulation results of the optimized scheme meet the target conditions and constraints. At this time, the optimized scheme is output as the final PRB wall design scheme. The transformation from the pollutant information of the target site to a scientific, reasonable, economically feasible and efficient PRB wall design scheme is realized, providing reliable decision support for the design of permeable reaction walls in groundwater pollution control.

[0134] In addition, this system also carries out user interface design:

[0135] Input interface design: Design a user-friendly input interface that allows users to input 3D geological information, contaminant information, and other relevant parameters.

[0136] Output interface design: Design an intuitive output interface to display the optimized PRB design results, including position, thickness, predicted life, etc.

[0137] Visual design: Provides 3D visualization function to help users intuitively understand simulation results and design solutions;

[0138] Modular integration design: Integrate various modules (data processing, machine learning, numerical simulation, optimization algorithm, user interface) into a unified platform.

[0139] Interface design: Design interfaces between modules to ensure data flow and functional coordination.

[0140] In addition, in practical applications, the design system may also include the following steps:

[0141] After the PRB wall is laid out at the target site based on the final PRB wall design plan, the permeability coefficients of several preset position nodes of the PRB wall are obtained at several preset time nodes during the actual treatment process;

[0142] The Lorentz curve fitting algorithm is introduced, and the permeability coefficient data collected at each preset position node at each preset time node are regarded as discrete data points. The time and position are used as variable dimensions, and the variable correspondence relationship in the Lorentz function form is determined. According to the variable dimension and the variable correspondence relationship in the Lorentz function form, each discrete data point is fitted into a permeability coefficient distribution Lorentz curve, and the permeability coefficient distribution Lorentz curve of each preset position node in the PRB wall is obtained;

[0143] Obtaining the treatment accuracy requirement information of the target site, and determining the preset permeability coefficient range of the PRB wall according to the treatment accuracy requirement information of the target site;

[0144] According to the preset permeability range of the PRB wall, the permeability range domain is determined in the permeability distribution Lorentz curve of each preset position node;

[0145] Analyze the length of the curve segment in the permeability coefficient distribution Lorentz curve of each preset position node that is within the permeability coefficient range and the length of the curve segment in the permeability coefficient distribution Lorentz curve that is outside the permeability coefficient range;

[0146] The length of the curve segment in the permeability coefficient distribution Lorenz curve of each preset position node that is located within the permeability coefficient range domain is divided by the length of the curve segment in the permeability coefficient distribution Lorenz curve that is located outside the permeability coefficient range domain, to obtain the compliance segment length ratio coefficient of the PRB wall permeability Lorenz curve of each preset position node;

[0147] Compare the length ratio coefficient of the compliance line segment of the Lorenz curve of the PRB wall permeability coefficient at each preset position node with the preset coefficient value;

[0148] If the compliance line segment length ratio coefficient of the Lorenz curve of the PRB wall permeability coefficient of a preset position node is not greater than the preset coefficient value, the corresponding preset position node of the PRB wall is defined as an abnormal node of the PRB wall permeability coefficient;

[0149] A PRB wall abnormality assessment report and early warning information are generated according to abnormal nodes of the PRB wall permeability coefficient, and the parameters of the corresponding machine learning model are adjusted and optimized in combination with the PRB wall abnormality assessment report.

[0150] The length of the curve segment in the permeability coefficient range domain is divided by the length of the curve segment outside the range domain to obtain the proportional coefficient of the compliance segment length of the Lorenz curve of the PRB wall permeability coefficient. This coefficient reflects the proportional relationship that the permeability coefficient of each preset position node conforms to the preset range. Compare the proportional coefficient with the preset coefficient value. If the proportional coefficient of a preset position node is not greater than the preset coefficient value, the node is defined as an abnormal node of the PRB wall permeability coefficient. This judgment can accurately find out the location nodes where permeability coefficient anomalies may exist, which may indicate that the PRB wall has problems such as blockage, structural damage or failure of reactive fillers at this location. Generate a PRB wall abnormality assessment report and generate early warning information based on the determined abnormal nodes of the PRB wall permeability coefficient. This report and early warning information can promptly inform relevant personnel of possible problems with the PRB wall. At the same time, the parameters of the corresponding machine learning model are adjusted and optimized in combination with the abnormal assessment report. This means that the machine learning model is improved through feedback data in the actual governance process, so that the model can more accurately predict the design parameters of the PRB wall, improve the accuracy and reliability of the model, and thus provide better support for subsequent PRB wall design and governance.

[0151] In general, the Lorentz curve fitting algorithm is used to visualize and quantify the permeability coefficient data, accurately identify the location nodes where problems may exist, and the generated abnormal assessment reports and early warning information help to timely discover and deal with potential problems of PRB walls. At the same time, the machine learning model is optimized using actual governance data, which improves the model's predictive ability and realizes a virtuous cycle from feedback of actual governance effects to model optimization, thereby improving the overall effect of PRB wall design and governance and ensuring that the governance accuracy requirements of the target site are met.

[0152] In addition, after collecting the three-dimensional hydrogeological parameters, drilling data and historical monitoring data of the target site, the three-dimensional hydrogeological parameters, drilling data and historical monitoring data of the target site are denoised based on the decision tree algorithm:

[0153] Construct a comprehensive data set including three-dimensional hydrogeological parameters, borehole data and historical monitoring data, standardize each data item in the data set, and clarify the attribute type and value range of each data item; based on the prior knowledge of the hydrogeological system of the target site and the preliminary statistical analysis of the data, determine the initial attribute set for building the decision tree. This attribute set should cover parameters that have a key impact on data quality and characteristics, such as formation permeability and water level variation in three-dimensional hydrogeological parameters, borehole depth and core integrity in borehole data, and pollutant concentration change trends in historical monitoring data;

[0154] Next, based on evaluation indicators such as information gain and Gini coefficient, the most discriminative attribute is selected from the initial attribute set as the root node of the decision tree, and the data set is divided into several sub-data sets according to different values ​​of the attribute; for each sub-data set, the attribute selection and data set division operations are recursively repeated to build a decision tree structure; in the process of tree construction, a preset stop criterion is set, such as when the number of samples in the sub-data set is lower than a preset threshold, the purity of the sub-data set reaches a predetermined standard (for example, the proportion of a certain category of data exceeds a set ratio), or the depth of the decision tree reaches a preset maximum depth, the splitting is stopped;

[0155] After the decision tree is built, each leaf node of the decision tree is traversed; for the data samples in the leaf nodes, the similarity between them and other samples in the leaf nodes is analyzed; if the difference in key attributes between a certain sample and most of the samples in the leaf nodes exceeds the preset similarity threshold, the sample is determined to be noise data. The similarity evaluation here can be based on distance measurement methods such as Euclidean distance and Mahalanobis distance, combined with the physical meaning of the data for comprehensive judgment;

[0156] Finally, the samples judged as noise are removed from the original data set to obtain the denoised three-dimensional hydrogeological parameters, drilling data and historical monitoring data.

[0157] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A groundwater pollution intelligent responsive permeable reaction wall design system, characterized in that: The design system includes: Constructing a site information database, collecting three-dimensional hydrogeological parameters and drilling data of the target site, and using the drilling data to generate a three-dimensional geological spatial distribution map and flow field distribution of the target site according to spatial interpolation information, and importing the three-dimensional hydrogeological parameters, three-dimensional geological spatial distribution map and flow field distribution of the target site into the site information database; Constructing a pollutant information library, obtaining characteristic information of various pollutants, and importing the characteristic information of various pollutants into the pollutant information library; constructing a filler information library, obtaining characteristic information of various reactive fillers, and importing the characteristic information of various reactive fillers into the filler information library; Constructing a pairing information library, collecting historical monitoring data, pollution source information, and pollutant migration patterns of the target site, determining available reaction fillers for various pollutants based on the historical monitoring data, pollution source information, and pollutant migration patterns of the target site, and importing the available reaction fillers for various pollutants into the pairing information library; Constructing a machine learning model, and importing characteristic information of various pollutants and available reaction fillers into the machine learning model for training, thereby obtaining a trained machine learning model; Acquire the three-dimensional hydrogeological parameters, three-dimensional geological spatial distribution map and flow field distribution of the target site in the site information database, and construct a three-dimensional finite difference model of the target site based on the three-dimensional hydrogeological parameters, three-dimensional geological spatial distribution map and flow field distribution of the target site in combination with the finite difference method; Obtain pollutant information of the target site, import the pollutant information of the target site into the trained machine learning model, and obtain the PRB wall design plan of the target site; import the PRB wall design plan of the target site into the three-dimensional finite difference model to obtain the final PRB wall design plan of the target site.

2. The intelligent responsive permeable reactive wall design system for groundwater pollution according to claim 1 is characterized in that: Using the drilling data, a three-dimensional geological spatial distribution map of the target site is generated according to the spatial interpolation information, specifically: Acquire drilling data of each borehole in the target site, the drilling data including coordinate information of the drilling location, geological information, hydraulic conductivity, water level data, porosity data, rock layer thickness data and stratum lithology; Determine a three-dimensional coordinate system according to the scope of the target site, and locate the coordinates of each borehole in the coordinate system; Taking the stratum lithology and thickness data of the drilling location as the basic data point, starting from the bottom layer, for each stratum, the stratum lithology and thickness of each three-dimensional coordinate point in the target site are calculated according to the Kriging interpolation method; In the horizontal direction, according to the plane position relationship and thickness change of the boreholes, the interpolation range is gradually expanded to determine the distribution boundary of the same stratum on the plane; In the vertical direction, according to the up-down relationship and thickness changes of the strata in different boreholes, the continuous distribution of the strata in the vertical direction is constructed by interpolation; For geological structure data, including faults and folds, they are incorporated into the interpolation process as constraints. If there is dislocation of the strata on both sides of the fault, the dislocation is calibrated in three-dimensional space according to the coordinates of the fault position; the bending shape of the fold is constructed in three-dimensional space according to the corresponding shape data in the drilling data; By continuously integrating the interpolation results of various strata and geological structures, a three-dimensional geological spatial distribution map of the target site is eventually formed.

3. The intelligent responsive permeable reaction wall design system for groundwater pollution according to claim 2 is characterized in that: Using the borehole data, the flow field distribution of the target site is generated according to the spatial interpolation information, specifically: Using the three-dimensional coordinate system of the target site as a framework, each borehole is positioned according to its coordinates, and the water level data in the borehole is used as a feature basis. Based on the inverse distance weighted interpolation method, the water level value of each three-dimensional coordinate point in the target site is obtained, thereby obtaining the spatial distribution of the water level; Based on the hydraulic conductivity data and Darcy's law, the hydraulic gradient of each three-dimensional coordinate point is calculated based on the water level distribution. In the calculation process, the influence of the anisotropy of the formation on the hydraulic conductivity is combined to calculate the hydraulic gradient according to the characteristics of different formations. Using the hydraulic conductivity and the calculated hydraulic gradient, the groundwater velocity vector is calculated again according to Darcy's law to determine the direction and velocity of groundwater flow in the corresponding stratum; The flow field distribution of the target site is constructed based on the direction and speed of groundwater flow in each stratum and the spatial distribution of water level.

4. The intelligent responsive permeable reactive wall design system for groundwater pollution according to claim 1 is characterized in that: Based on the historical monitoring data of the target site, pollution source information and pollutant migration rules, the available reactive fillers for various pollutants are determined, specifically: Obtain historical monitoring data and pollution source information of the target site; Determine all pollutants that have appeared in the target site based on the historical monitoring data, and determine the emission intensity and emission mode of various pollutants that have appeared in the target site from the pollution source information; Obtaining the migration rules of various pollutants that have appeared in the target site, and determining the diffusion characteristics of various pollutants that have appeared in the target site according to the migration rules; wherein the diffusion characteristics include diffusion coefficients and diffusion directions; and obtaining characteristic information of various pollutants that have appeared in the target site from the pollutant information database; Generate characteristic text information of various pollutants that have appeared in the target site based on the characteristic information, emission intensity, emission mode and dispersion characteristics of various pollutants that have appeared in the target site; Acquire characteristic information of each reaction filler in the filler information library, introduce a local sensitive attention mechanism, and analyze the attention score between the characteristic information of each reaction filler and the characteristic text information of various pollutants that have appeared in the target site according to the local sensitive attention mechanism; The attention scores between the characteristic information of each reactive filler and the characteristic text information of various pollutants that have appeared in the target site are compared and analyzed with the preset score threshold; If the attention score between the characteristic information of a certain reaction filler and the characteristic text information of a certain pollutant that has appeared in the target site is greater than a preset score threshold, the reaction filler is marked as an available reaction filler for the corresponding pollutant; After the attention scores between the characteristic information of each reaction filler and the characteristic text information of various pollutants that have appeared in the target site are compared and analyzed with the preset score threshold, the available reaction fillers of various pollutants that have appeared in the target site are obtained.

5. The intelligent responsive permeable reactive wall design system for groundwater pollution according to claim 1 is characterized in that: According to the three-dimensional hydrogeological parameters, three-dimensional geological spatial distribution map and flow field distribution of the target site, a three-dimensional finite difference model of the target site is constructed in combination with the finite difference method. Specifically: According to the three-dimensional geological spatial distribution map of the target site, the spatial layout, thickness variation, interlayer contact relationship and geological structural characteristics of different strata are obtained to determine the spatial framework and stratum boundary conditions of the model; And according to the flow field distribution of the target site, the spatial distribution of groundwater level, the magnitude and direction of hydraulic gradient and the dynamic characteristics of water flow are obtained; According to the geometric shape, size, spatial structure, stratum boundary conditions, groundwater level spatial distribution, hydraulic gradient size and direction, and water flow dynamic characteristics of the target site, the target site is divided into three-dimensional spatial grids to obtain a number of grid nodes; For each grid node, the partial differential equation describing groundwater movement is discretized using the finite difference method to obtain the equivalent value of the source-sink intensity of each grid node. At the same time, the zero flux boundary conditions of each grid node are determined based on the three-dimensional geological spatial distribution map and flow field distribution. Substitute the equivalent value of the source-sink intensity of each grid node into the corresponding position in the discrete equation, and set the boundary constraints of the discrete equation according to the zero flux boundary conditions determined by the three-dimensional geological spatial distribution map and the flow field distribution; Then, the source and sink terms in the discrete equations are adjusted according to the source and sink information in the flow field distribution, and the discrete equations of all grid nodes are combined according to the three-dimensional spatial coordinate relationship to construct a three-dimensional finite difference model of the target site.

6. The intelligent responsive permeable reactive wall design system for groundwater pollution according to claim 1 is characterized in that: Construct a machine learning model, import the characteristic information of various pollutants and available reaction fillers into the machine learning model for training, and obtain a trained machine learning model, specifically: Acquiring characteristic information of various pollutants in the pollutant information library, and acquiring available reactive fillers of various pollutants in the pairing information library; Search the data network based on the characteristic information of various pollutants and available reactive fillers to obtain PRB wall design solutions for treating various pollutants; The characteristic information of various pollutants and available reactive fillers are used as input feature nodes, and the PRB wall design scheme for treating various pollutants is used as output feature nodes; a mapping relationship topology diagram is constructed according to the input feature nodes and the output feature nodes; Building a machine learning model based on a graph neural network, converting the connection relationship between the input feature nodes and the output feature nodes in the mapping relationship topology graph into a data structure recognizable by the machine learning model, encoding the information of the nodes and edges in the mapping relationship topology graph according to the graph general relationship and embedding it into the parameter space of the machine learning model for learning and training; After each training cycle is completed, the feature contribution of the machine learning model is obtained. If the feature contribution of the machine learning model gradually increases and is greater than a preset threshold, the training is completed and the trained machine learning model is output.

7. The intelligent responsive permeable reactive wall design system for groundwater pollution according to claim 1 is characterized in that: Obtain pollutant information of the target site, import the pollutant information of the target site into the trained machine learning model, and obtain the PRB wall design scheme of the target site; import the PRB wall design scheme of the target site into the three-dimensional finite difference model to obtain the final PRB wall design scheme of the target site, specifically: Obtaining pollutant information of the target site, importing the pollutant information of the target site into the trained machine learning model, and obtaining a PRB wall design plan of the target site; the PRB wall design plan includes a PRB wall design thickness, position, and filler; Importing the PRB wall design of the target site into the three-dimensional finite difference model for evaluation, and obtaining the numerical simulation results of the permeable reactive wall; Define target conditions based on minimizing pollutant concentrations and maximizing PRB life, and set constraints based on geological conditions, cost budget, and construction feasibility; Determine whether the numerical simulation results of the permeable reaction wall meet the target conditions and the constraint conditions; if the numerical simulation results of the permeable reaction wall meet the target conditions and the constraint conditions, the PRB wall design scheme is directly output as the final PRB wall design scheme; Otherwise, a genetic algorithm is introduced, and an iterative optimization cycle is set, and the PRB wall design scheme is adjusted and optimized according to the genetic algorithm. After each iterative optimization cycle is completed, an optimized PRB wall design scheme is obtained; The optimized PRB wall design scheme is input into the three-dimensional finite difference model for evaluation to obtain the numerical simulation results of the permeable reaction wall; Until the numerical simulation results of the permeable reaction wall of the optimized PRB wall design scheme meet the target conditions and constraint conditions, the iteration is stopped, and the optimized PRB wall is output as the final PRB wall design scheme.

8. The intelligent responsive permeable reactive wall design system for groundwater pollution according to claim 1 is characterized by: The three-dimensional hydrogeological parameters include strata, water levels, horizontal / vertical permeability coefficients, groundwater flow rates, porosity, and water storage rates.

9. The intelligent responsive permeable reactive wall design system for groundwater pollution according to claim 1 is characterized by: The characteristic information of the reaction filler includes material, ratio, size, specific surface area and reaction rate constant; the characteristic information of the pollutant includes the negative logarithm of the acid dissociation constant and the molecular weight.

Images