Heavy metal pollution assessment method for industrial land

Abstract

The invention discloses a heavy metal pollution assessment method for industrial land in the technical field of heavy metal pollution assessment, which comprises the following steps: determining the sampling point density of a prediction area by adopting an improved space Kriging interpolation method to obtain the number of sampling points per hectare; constructing a weight coefficient model by using the multi-dimensional pollution index; establishing a soil and underground water double-medium migration evolution model; constructing a heavy metal pollution factor diffusion prediction system based on an LSTM neural network; and the operation target prediction system outputs prediction data of heavy metal pollution diffusion in a certain period in the future. The density of the sampling points is determined by adopting the improved space Kriging interpolation method, the variation characteristics of the space can be reflected more accurately, the weight coefficient model is constructed by utilizing the multi-dimensional pollution index, the weight distribution of the heavy metal pollution factor can be more reasonable, and by establishing the soil and groundwater double-medium migration evolution model, the heavy metal pollution factor can be more accurately distributed. The migration and transformation process of pollutants in soil and underground water can be comprehensively simulated.

Description

Technical Field

[0001] The present invention relates to the technical field of heavy metal pollution assessment, in particular to a method for assessing heavy metal pollution in industrial land. Background Art

[0002] The 2014 National Soil Pollution Survey Bulletin pointed out that in heavily polluting enterprises and their surrounding areas, industrial wastelands, industrial parks, centralized solid waste treatment and disposal sites, and in typical industrial and mining sites and their surrounding soils, pollution exceeding standards accounted for 36.3%, 34.9%, 29.4%, 21.3%, and 33.4%, respectively; heavy metal pollution was particularly serious, involving cadmium, lead, copper, arsenic, zinc, mercury, chromium, and other heavy metals.

[0003] The accumulation and excessive levels of heavy metals pose a direct threat to ecosystems and biodiversity. By monitoring heavy metal concentrations on industrial land, we can identify contaminated areas and implement appropriate soil remediation plans to prevent further contamination.

[0004] Most existing methods for assessing heavy metal pollution in land assess the degree of heavy metal pollution in the land by setting up sampling points and using the weighted average method to calculate the pollution index. On the one hand, the existing assessment methods lack quantitative representation of spatial heterogeneity, making it difficult to quantitatively measure and analyze unevenly distributed heavy metal pollution. At the same time, the existing assessment methods tend to ignore the dynamic characteristics of heavy metal migration and diffusion, and lack consideration of the material transfer between the soil and groundwater coupling system, resulting in insufficient assessment accuracy. Summary of the Invention

[0005] The purpose of the present invention is to provide a method for assessing heavy metal pollution in industrial land. By adopting an improved spatial kriging interpolation method to determine the sampling point density, the number of sampling points per hectare is obtained, which can more accurately reflect the spatial variation characteristics and improve sampling efficiency. By using a multidimensional pollution index to construct a weight coefficient model, multiple pollution factors can be comprehensively considered, and the weight distribution of heavy metal pollution factors can be made more reasonable. In conjunction with the establishment of a soil and groundwater dual-medium migration and evolution model, the migration and transformation process of pollutants in soil and groundwater can be comprehensively simulated, providing data support for pollution control.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] In a first aspect, the present invention provides a method for assessing heavy metal pollution in industrial land, comprising:

[0008] The improved spatial kriging interpolation method is used to determine the sampling point density of the prediction area and obtain the number of sampling points per hectare;

[0009] Use multi-dimensional pollution index to build a weight coefficient model;

[0010] Establish a dual-medium migration and evolution model of soil and groundwater;

[0011] Construct a heavy metal pollution factor diffusion prediction system based on LSTM neural network;

[0012] The target prediction system is run to output the prediction data of heavy metal pollution diffusion in a certain period in the future.

[0013] By using the improved spatial Kriging interpolation method to determine the sampling point density and obtain the number of sampling points per hectare, the spatial variation characteristics can be more accurately reflected and the sampling efficiency can be improved. By using the multidimensional pollution index to construct a weight coefficient model, various pollution factors can be comprehensively considered, making the weight distribution of heavy metal pollution factors more reasonable. By establishing a dual-medium migration and evolution model of soil and groundwater, the migration and transformation process of pollutants in soil and groundwater can be fully simulated, providing data support for pollution control. By constructing a heavy metal pollution factor diffusion prediction system based on the LSTM neural network, the diffusion trend of heavy metal pollution factors can be accurately predicted. By running the target prediction system to output the predicted data of heavy metal pollution diffusion in a certain period in the future, it can provide a data basis for environmental protection and governance, which is conducive to reducing environmental pollution and protecting the ecological environment.

[0014] As a further solution of the present invention, the improved spatial kriging interpolation method is used to determine the sampling point density of the prediction area, and the number of sampling points per hectare is obtained, which includes:

[0015] D=ceil[(A / 1000)×(1+0.5σ 2 / μ)];

[0016] N = ceil(A*D);

[0017] Where, D is the number of sampling points per hectare, ceil is the rounding function, A is the site area (㎡), σ 2 / μ is the dispersion coefficient of historical data, where σ is the standard deviation of historical data, μ is the mean value of historical data, and N is the number of sampling points per hectare.

[0018] By using the improved spatial Kriging interpolation method to determine the sampling point density, the number of sampling points per hectare is obtained. This not only provides a reasonable way to determine the sampling point density, but also reduces the subjectivity of traditional empirical judgment through numerical calculations and improves the accuracy of sampling. In addition, by combining the degree of discreteness of historical data, the distribution of sampling points can be made more consistent with the actual situation, thereby improving the accuracy of the sampling results.

[0019] As a further solution of the present invention, the method of constructing a weight coefficient model using a multi-dimensional pollution index includes:

[0020] Construct a dynamic weighted comprehensive pollution index model;

[0021] Calculate the weight coefficient of each pollution index and use the weight normalization method to obtain the weight coefficient of each heavy metal pollution factor;

[0022] The weight coefficients of various heavy metal pollution factors are input into the dynamic weight comprehensive pollution index model to obtain the comprehensive pollution index of heavy metal pollution factors.

[0023] By using a multidimensional pollution index to construct a weight coefficient model to comprehensively calculate the impact of different heavy metal pollution factors, the accuracy and reliability of environmental pollution assessment can be improved. By constructing a dynamic weight comprehensive pollution index model, the weight coefficient of each pollution factor can be flexibly adjusted according to the pollution status under different time and space conditions, thereby more accurately reflecting the actual situation of environmental pollution and improving the accuracy of environmental pollution assessment. By comprehensively considering the impact of different pollution factors, the one-sidedness of single pollution factor assessment can be avoided.

[0024] As a further solution of the present invention, the construction of a dynamic weighted comprehensive pollution index model includes:

[0025] DWCP I(t)=Σ[w' i (t)×(C i / C 0i ) k ] / Σw' i (t); The dynamic weight comprehensive pollution index model is used to measure the weight of the i-th heavy metal pollution factor at time t.

[0026] Where DWCP I is the dynamic weighted comprehensive pollution index of heavy metal pollution factors, w' i (t) is the weight coefficient of heavy metal pollution factor, C i is the current pollution concentration of the i-th heavy metal pollution factor, C 0i is the initial pollution concentration of the i-th heavy metal pollution factor, and k is the index parameter; the initial pollution concentration is obtained from historical data or initial data of a certain period, and the index parameter is the rate of change of the pollution index of the heavy metal pollution factor within the specified period.

[0027] By constructing a dynamic weight comprehensive pollution index model and introducing a weight coefficient that changes dynamically over time, the actual impact of heavy metal pollution factors can be reflected more accurately. At the same time, by combining the current concentration of heavy metal pollution factors with their historical change trends, decision-making and data support can be provided for environmental protection strategies for industrial land. By setting a weight coefficient that can be adjusted dynamically, the accuracy of assessing the comprehensive impact of heavy metal pollution factors can be improved, avoiding the assessment bias caused by fixed weights in traditional methods.

[0028] As a further solution of the present invention, the weight coefficients of each pollution index are calculated, and the weight coefficients of each heavy metal pollution factor are obtained by using a weight normalization method, including:

[0029] The migration-toxicity coupling model was used to calculate the weight coefficients of various pollution indices.

[0030] As a further solution of the present invention, the weight coefficients of each pollution index calculated using the migration-toxicity coupling model include:

[0031] w i (t) = α·K ow +β·(v g ·t) / d+γ·LD50 -1 ;

[0032] Where K ow is the octanol-water partition coefficient, v g is the groundwater seepage velocity, d is the groundwater depth, LD50 is the median lethal dose, α, β, γ are fitting coefficients;

[0033] Calculate the normalized weight value w' of each heavy metal pollution factor i (t).

[0034] As a further solution of the present invention, the value ranges of α, β and γ are all (0-1).

[0035] As a further solution of the present invention, the establishment of a soil and groundwater dual-medium migration evolution model includes:

[0036]

[0037] Where, is the rate of change of the concentration of heavy metal pollution factor C with time t, D eff is the effective diffusion coefficient of heavy metal pollution factors, is the Laplace operator, is the second-order spatial derivative of concentration C, v is the groundwater flow velocity vector, is the gradient of concentration C, λ is the fitting coefficient, S(x, y, t) represents the source and sink terms of heavy metal pollution factors, the second-order spatial derivative represents the concentration change caused by the diffusion of heavy metal pollution factors, the gradient is used to represent the rate of change and direction of the concentration of heavy metal pollution factors in space, the source term of heavy metal pollution factors represents the generation of heavy metal pollution factors at the target position and target time, and the sink term of heavy metal pollution factors represents the disappearance of heavy metal pollution factors at the target position and target time.

[0038] As a further solution of the present invention, a heavy metal pollution factor diffusion prediction system is constructed based on an LSTM neural network, including:

[0039]

[0040] Where L is the overall predicted value of all heavy metal pollution factors involved in the calculation, 1 / N is the predicted concentration of the overall heavy metal pollution factors, and y t is the concentration value of the heavy metal pollution factor actually detected at t, is the predicted value of the tth actual detected metal pollution factor, 0.2 is the regularization coefficient, is the gradient operator, is the prediction error, is the gradient of the prediction error, is the mean square error, The predicted concentration values ​​of the first to T-th heavy metal pollution factors are summed up; the mean square error is used to measure the square error between the predicted value and the true value. The role of the square is to amplify the error value, which is conducive to focusing on the heavy metal pollution factors with larger errors in the model. The gradient is used to represent the rate of change of the error in space, and the regularization coefficient is used to control the weight of the gradient term in the loss function.

[0041] As a further solution of the present invention, the operation target prediction system outputs prediction data of heavy metal pollution diffusion in a certain period in the future, including:

[0042] Based on historical data and actual monitoring information of heavy metal pollution factors, the time length is input into the target prediction system;

[0043] Output the predicted data of heavy metal pollution spread in a certain period of time in the future.

[0044] By building an operation target prediction system, we can use historical data and real-time monitoring information, and combine it with advanced prediction algorithms to predict the future spread of heavy metal pollution. By predicting the spread of heavy metal pollution in a certain period of time in the future, we can provide timely warnings and take countermeasures to effectively reduce the impact of heavy metal pollution on the environment and human health.

[0045] In a second aspect, a system is also provided, which adopts the method for heavy metal pollution assessment of industrial land as described in the above scheme, and the system includes a spatial sampling module, a dynamic weight calculator, a migration model solver, a diffusion prediction module and a risk visualization interface;

[0046] The spatial sampling module is used to determine the sampling point density of the prediction area using the improved spatial Kriging interpolation method to obtain the number of sampling points per hectare;

[0047] The dynamic weight calculator is used to construct a weight coefficient model using the multi-dimensional pollution index;

[0048] The migration model solver is used to establish a dual-medium migration evolution model of soil and groundwater;

[0049] The diffusion prediction module is used to build a heavy metal pollution factor diffusion prediction system based on the LSTM neural network;

[0050] The risk visualization interface is used to run the target prediction system to output the predicted data of heavy metal pollution diffusion in a certain period of time in the future.

[0051] Compared with the prior art, the present invention has the following beneficial effects:

[0052] 1. The present invention determines the density of sampling points and obtains the number of sampling points per hectare, which can more accurately reflect the spatial variation characteristics and improve sampling efficiency. By using a multidimensional pollution index to construct a weight coefficient model, it can comprehensively consider multiple pollution factors and make the weight distribution of heavy metal pollution factors more reasonable. In conjunction with the establishment of a soil and groundwater dual-medium migration and evolution model, it can comprehensively simulate the migration and transformation process of pollutants in soil and groundwater, providing more reliable data support for pollution control.

[0053] 2. The present invention determines the density of sampling points by adopting an improved spatial kriging interpolation method, which not only provides a reasonable method for determining the density of sampling points, but also reduces the subjectivity of traditional empirical judgment through numerical calculation, thereby improving the accuracy of sampling. In addition, by combining the discrete degree of historical data, the distribution of sampling points can be made more consistent with the actual situation, thereby improving the accuracy of the sampling results.

[0054] 3. The present invention constructs a dynamic weight comprehensive pollution index model and introduces a weight coefficient that changes dynamically over time, which can more accurately reflect the actual impact of heavy metal pollution factors. At the same time, by combining the current concentration of heavy metal pollution factors with their historical change trends, it can provide decision-making and data support for environmental protection strategies for industrial land. By setting a weight coefficient that can be dynamically adjusted, it can improve the accuracy of assessing the comprehensive impact of heavy metal pollution factors and avoid the assessment bias caused by fixed weights in traditional methods.

[0055] 4. By constructing an operation target prediction system, the present invention can utilize historical data and real-time monitoring information, and combine it with advanced prediction algorithms to predict the future spread of heavy metal pollution. By predicting the spread of heavy metal pollution in a certain period of time in the future, timely warnings can be issued and countermeasures can be taken to effectively reduce the impact of heavy metal pollution on the environment and human health. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 is a flow chart of the method steps of the present invention;

[0057] Figure 2 This is a system module connection diagram of the present invention. DETAILED DESCRIPTION

[0058] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0059] Example,

[0060] See also Figure 1 In an embodiment of the present invention, a method for assessing heavy metal pollution in industrial land includes:

[0061] S1: Use the improved spatial kriging interpolation method to determine the sampling point density of the prediction area and obtain the number of sampling points per hectare;

[0062] S2: Use the multi-dimensional pollution index to construct a weight coefficient model;

[0063] S3: Establish a dual-medium migration evolution model of soil and groundwater;

[0064] S4: Construct a heavy metal pollution factor diffusion prediction system based on LSTM neural network;

[0065] S5: Run the target prediction system to output the prediction data of heavy metal pollution diffusion in a certain period in the future.

[0066] By using the improved spatial Kriging interpolation method to determine the sampling point density and obtain the number of sampling points per hectare, the spatial variation characteristics can be more accurately reflected and the sampling efficiency can be improved. By using the multidimensional pollution index to construct a weight coefficient model, various pollution factors can be comprehensively considered, making the weight distribution of heavy metal pollution factors more reasonable. By establishing a dual-medium migration and evolution model of soil and groundwater, the migration and transformation process of pollutants in soil and groundwater can be fully simulated, providing data support for pollution control. By constructing a heavy metal pollution factor diffusion prediction system based on the LSTM neural network, the diffusion trend of heavy metal pollution factors can be accurately predicted. By running the target prediction system to output the predicted data of heavy metal pollution diffusion in a certain period in the future, it can provide a data basis for environmental protection and governance, which is conducive to reducing environmental pollution and protecting the ecological environment.

[0067] Preferably, the improved spatial kriging interpolation method is used to determine the sampling point density of the prediction area, and obtaining the number of sampling points per hectare includes:

[0068] D=ceil[(A / 1000)×(1+0.5σ 2 / μ)];

[0069] N = ceil(A*D);

[0070] Where, D is the number of sampling points per hectare, ceil is the rounding function, A is the site area (㎡), σ 2 / μ is the dispersion coefficient of historical data, where σ is the standard deviation of historical data, μ is the mean value of historical data, and N is the number of sampling points per hectare.

[0071] By using the improved spatial Kriging interpolation method to determine the sampling point density, the number of sampling points per hectare is obtained. This not only provides a reasonable way to determine the sampling point density, but also reduces the subjectivity of traditional empirical judgment through numerical calculations and improves the accuracy of sampling. In addition, by combining the degree of discreteness of historical data, the distribution of sampling points can be made more consistent with the actual situation, thereby improving the accuracy of the sampling results.

[0072] Preferably, the method of constructing a weight coefficient model using a multi-dimensional pollution index includes:

[0073] Construct a dynamic weighted comprehensive pollution index model;

[0074] Calculate the weight coefficient of each pollution index and use the weight normalization method to obtain the weight coefficient of each heavy metal pollution factor;

[0075] The weight coefficients of various heavy metal pollution factors are input into the dynamic weight comprehensive pollution index model to obtain the comprehensive pollution index of heavy metal pollution factors.

[0076] By using a multidimensional pollution index to construct a weight coefficient model to comprehensively calculate the impact of different heavy metal pollution factors, the accuracy and reliability of environmental pollution assessment can be improved. By constructing a dynamic weight comprehensive pollution index model, the weight coefficient of each pollution factor can be flexibly adjusted according to the pollution status under different time and space conditions, thereby more accurately reflecting the actual situation of environmental pollution and improving the accuracy of environmental pollution assessment. By comprehensively considering the impact of different pollution factors, the one-sidedness of single pollution factor assessment can be avoided.

[0077] Preferably, the construction of a dynamic weighted comprehensive pollution index model includes:

[0078] DWCPI(t)=Σ[w' i (t)×(C i / C 0i ) k ] / Σw' i (t); The dynamic weight comprehensive pollution index model is used to measure the weight of the i-th heavy metal pollution factor at time t.

[0079] Where DWCPI is the dynamic weighted comprehensive pollution index of heavy metal pollution factors, w' i (t) is the weight coefficient of heavy metal pollution factor, C i is the current pollution concentration of the i-th heavy metal pollution factor, C 0i is the initial pollution concentration of the i-th heavy metal pollution factor, k is the index parameter; the initial pollution concentration is obtained from historical data or initial data of a certain period, the index parameter is the rate of change of the pollution index of the heavy metal pollution factor within the specified period, and the initial pollution concentration is the background value of the heavy metal pollution factor.

[0080] By constructing a dynamic weight comprehensive pollution index model and introducing a weight coefficient that changes dynamically over time, the actual impact of heavy metal pollution factors can be reflected more accurately. At the same time, by combining the current concentration of heavy metal pollution factors with their historical change trends, decision-making and data support can be provided for environmental protection strategies for industrial land. By setting a weight coefficient that can be adjusted dynamically, the accuracy of assessing the comprehensive impact of heavy metal pollution factors can be improved, avoiding the assessment bias caused by fixed weights in traditional methods.

[0081] Preferably, the weight coefficients of the various pollution indices are calculated, and the weight coefficients of the various heavy metal pollution factors are obtained by using a weight normalization method, including:

[0082] The migration-toxicity coupling model was used to calculate the weight coefficients of various pollution indices.

[0083] Preferably, the weight coefficients of each pollution index calculated using the migration-toxicity coupling model include:

[0084] w i (t) = α·K ow +β·(v g ·t) / d+γ·LD50 -1 ;

[0085] Where K ow is the octanol-water partition coefficient, v g is the groundwater seepage velocity, d is the groundwater depth, LD50 is the median lethal dose, α, β, γ are fitting coefficients;

[0086] Calculate the normalized weight value w' of each heavy metal pollution factor i (t).

[0087] Preferably, the value ranges of α, β and γ are all (0-1).

[0088] Preferably, the establishment of a soil and groundwater dual-medium migration evolution model includes:

[0089]

[0090] Where, is the rate of change of the concentration of heavy metal pollution factor C with time t, D eff is the effective diffusion coefficient of heavy metal pollution factors, is the Laplace operator, is the second-order spatial derivative of concentration C, v is the groundwater flow velocity vector, is the gradient of concentration C, λ is the fitting coefficient, S(x, y, t) represents the source and sink terms of heavy metal pollution factors, the second-order spatial derivative represents the concentration change caused by the diffusion of heavy metal pollution factors, the gradient is used to represent the rate of change and direction of the concentration of heavy metal pollution factors in space, the source term of heavy metal pollution factors represents the generation of heavy metal pollution factors at the target position and target time, and the sink term of heavy metal pollution factors represents the disappearance of heavy metal pollution factors at the target position and target time.

[0091] Preferably, a heavy metal pollution factor diffusion prediction system is constructed based on an LSTM neural network, including:

[0092]

[0093] Where L is the overall predicted value of all heavy metal pollution factors involved in the calculation, 1 / N is the predicted concentration of the overall heavy metal pollution factors, and y t is the concentration value of the heavy metal pollution factor actually detected at t, is the predicted value of the tth actual detected metal pollution factor, 0.2 is the regularization coefficient, is the gradient operator, is the prediction error, is the gradient of the prediction error, is the mean square error, The predicted concentration values ​​of the first to T-th heavy metal pollution factors are summed up; the mean square error is used to measure the square error between the predicted value and the true value. The role of the square is to amplify the error value, which is conducive to focusing on the heavy metal pollution factors with larger errors in the model. The gradient is used to represent the rate of change of the error in space, and the regularization coefficient is used to control the weight of the gradient term in the loss function.

[0094] Preferably, the operation target prediction system outputs prediction data of heavy metal pollution diffusion in a certain period in the future, including:

[0095] Based on historical data and actual monitoring information of heavy metal pollution factors, the time length is input into the target prediction system;

[0096] Output the predicted data of heavy metal pollution spread in a certain period of time in the future.

[0097] By building an operation target prediction system, we can use historical data and real-time monitoring information, and combine it with advanced prediction algorithms to predict the future spread of heavy metal pollution. By predicting the spread of heavy metal pollution in a certain period of time in the future, we can provide timely warnings and take countermeasures to effectively reduce the impact of heavy metal pollution on the environment and human health.

[0098] See also Figure 2 In an embodiment of the present invention, a system adopts the heavy metal pollution assessment method for industrial land as described in the above scheme, and the system includes a spatial sampling module, a dynamic weight calculator, a migration model solver, a diffusion prediction module and a risk visualization interface;

[0099] The spatial sampling module is used to determine the sampling point density of the prediction area using the improved spatial Kriging interpolation method to obtain the number of sampling points per hectare;

[0100] The dynamic weight calculator is used to construct a weight coefficient model using the multi-dimensional pollution index;

[0101] The migration model solver is used to establish a dual-medium migration evolution model of soil and groundwater;

[0102] The diffusion prediction module is used to build a heavy metal pollution factor diffusion prediction system based on the LSTM neural network;

[0103] The risk visualization interface is used to run the target prediction system to output the predicted data of heavy metal pollution diffusion in a certain period of time in the future.

[0104] Input the site parameters of the prediction area into the spatial sampling module. The site parameters are A = 85000m 2=8.5 hectares, historical σ 2 / μ=0.38, historical σ 2 / μ is obtained based on the heavy metal pollution factor monitoring data of the past five years, and further calculation results in D=42;

[0105] Set the site parameter to A = 85000m 2 Substitute into the density sampling formula: D = ceil[(A / 1000)×(1+0.5σ 2 / μ)], the result D = 11 points / hectare, substitute the result D into N = ceil(A*D) = 94;

[0106] Therefore, the theoretical number of sampling points in the prediction area is calculated to be 94. In actual implementation, it is rounded to 42 key points after adjustment. The table of sampling heavy metal pollution factor concentrations is as follows Table 1:

[0107]

[0108]

[0109] Weight calculation parameters, taking t = 10 years:

[0110] The calculated values ​​of the weight coefficients of the pollution index of the sampled heavy metal pollution factors are shown in Table 2:

[0111]

[0112] Substitute the parameters of As, Cd and Hg in Table 2 into the weight coefficient calculation formula:

[0113] w i (t) = α·K ow +β·(v g ·t) / d+γ·LD50 -1 , we get:

[0114] w As (t)=0.4·0.01+0.3·(0.2·10·365) / 3+0.3·15 -1 =73.024;

[0115] w Cd (t)=0.4·0.8+0.3·(0.2·10·365) / 3+0.3·5 -1 =73.38;

[0116] w Hg (t)=0.4·3.2+0.3·(0.2·10·365) / 3+0.3·0.3 -1 =24.18;

[0117] The normalized weights are calculated using the weight normalization method:

[0118] The normalized weight value of As is: w' As (t) = w As (t) / [w As (t)+w Cd (t)+w Hg (t)]=73.024 / (73.024+73.38+24.18)=0.428;

[0119] The normalized weight value of Cd is: w' Cd (t) = w Cd (t) / [w As (t)+w Cd (t)+w Hg (t)]=73.38 / (73.024+73.38+24.18)=0.430;

[0120] The normalized weight value of Hg is: w' Hg (t) = w Hg (t) / [w As (t)+w Cd (t)+w Hg (t)]=24.18 / (73.024+73.38+24.18)=0.142.

[0121] Effective diffusion coefficient D of heavy metal pollution factors eff =1.2·10 -6 m 2 / s;

[0122] Seepage velocity v = 0.2 m / d = 2.315·10 -6 m / s;

[0123] Pollution source intensity S(x,y,t)=28.5mg / kg / year.

[0124] The target area was divided into 50 × 50 grids using the finite element method discretization method, with a spatial step size of Δx = Δy = 3 m;

[0125] The time step is taken as Δt = 30 days;

[0126] The implicit Euler method is used to solve the partial differential equations and the solution of the soil and groundwater dual-medium migration evolution model is obtained.

[0127] The concentration data of heavy metal pollution factors are input into the input layer of the diffusion prediction module, the gradient operator is set to 0.2, and the diffusion prediction results of heavy metal pollution factors in the next 10 years are output through the heavy metal pollution factor diffusion prediction system.

[0128] By combining the weight values ​​of heavy metal pollution factors with the prediction results of the diffusion of heavy metal pollution factors to assess the degree of heavy metal pollution in the target area, the accuracy of heavy metal pollution assessment can be improved.

[0129] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solution and inventive concept of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. A method for assessing heavy metal pollution in industrial land, characterized in that: include: The improved spatial kriging interpolation method is used to determine the sampling point density of the prediction area and obtain the number of sampling points per hectare; Detect and obtain the concentration matrix of several heavy metal factors and construct a three-dimensional pollution cloud map; Use multi-dimensional pollution index to build a weight coefficient model; Establish a dual-medium migration and evolution model of soil and groundwater; Construct a heavy metal pollution factor diffusion prediction system based on LSTM neural network; The target prediction system is run to output the prediction data of heavy metal pollution diffusion in a certain period in the future.

2. A method for assessing heavy metal pollution in industrial land according to claim 1, characterized in that: The improved spatial kriging interpolation method is used to determine the sampling point density of the prediction area, and the number of sampling points per hectare is obtained, which includes: D=ceil[(A / 1000)×(1+0.5σ 2 / μ)]; N = ceil(A*D); Where, D is the number of sampling points per hectare, ceil is the rounding function, A is the site area (㎡), σ 2 / μ is the dispersion coefficient of historical data, where σ is the standard deviation of historical data, μ is the mean value of historical data, and N is the number of sampling points per hectare.

3. The method for heavy metal pollution assessment for industrial land according to claim 1, characterized in that: The method of constructing a weight coefficient model using a multi-dimensional pollution index includes: Construct a dynamic weighted comprehensive pollution index model; Calculate the weight coefficient of each pollution index and use the weight normalization method to obtain the weight coefficient of each heavy metal pollution factor; The weight coefficients of various heavy metal pollution factors are input into the dynamic weight comprehensive pollution index model to obtain the comprehensive pollution index of heavy metal pollution factors.

4. The method for heavy metal pollution assessment for industrial land according to claim 3, characterized in that: The construction of the dynamic weight comprehensive pollution index model includes: DWCPI(t)=Σ[w’ i (t)×(C i / C 0i ) k ] / Σw’ i (t); Where DWCPI is the dynamic weighted comprehensive pollution index of heavy metal pollution factors, w' i (t) is the normalized weight value of heavy metal pollution factor, C i is the current pollution concentration of the i-th heavy metal pollution factor, C 0i is the initial pollution concentration of the i-th heavy metal pollution factor, and k is the exponential parameter.

5. The method for heavy metal pollution assessment for industrial land according to claim 4, characterized in that: Calculate the weight coefficient of each pollution index and use the weight normalization method to obtain the weight coefficient of each heavy metal pollution factor, including: The migration-toxicity coupling model was used to calculate the weight coefficients of various pollution indices.

6. The method for heavy metal pollution assessment for industrial land according to claim 5, characterized in that: The weight coefficients of various pollution indices are calculated using the migration-toxicity coupling model, including: w i (t)=α·K ow +β·(v g ·t) / d+γ·LD50 -1 ; Where K ow is the octanol-water partition coefficient, v g is the groundwater seepage velocity, d is the groundwater depth, LD50 is the median lethal dose, α, β, γ are fitting coefficients; Calculate the normalized weight value w' of each heavy metal pollution factor i (t).

7. The method for heavy metal pollution assessment for industrial land according to claim 6, characterized in that: The value ranges of α, β and γ are all (0-1).

8. The method for heavy metal pollution assessment for industrial land according to claim 1, characterized in that: The establishment of the soil and groundwater dual-medium migration evolution model includes: Where, is the rate of change of the concentration of heavy metal pollution factor C with time t, D eff is the effective diffusion coefficient of heavy metal pollution factors, is the Laplace operator, is the second-order spatial derivative of concentration C, v is the groundwater flow velocity vector, is the gradient of concentration C, λ is the fitting coefficient, and S(x, y, t) represents the source and sink terms of heavy metal pollution factors.

9. The method for heavy metal pollution assessment for industrial land according to claim 1, characterized in that: A heavy metal pollution factor diffusion prediction system is constructed based on the LSTM neural network, including: Where L is the overall predicted value of all heavy metal pollution factors involved in the calculation, 1 / N is the predicted concentration of the overall heavy metal pollution factors, and y t is the concentration value of the heavy metal pollution factor actually detected at t, is the predicted value of the tth actual detected metal pollution factor, 0.2 is the regularization coefficient, is the gradient operator, is the prediction error, is the gradient of the prediction error, is the mean square error, The sum of the predicted concentration values ​​of the first to Tth heavy metal pollution factors.

10. The method for heavy metal pollution assessment for industrial land according to claim 1, characterized in that: The operation target prediction system outputs prediction data of heavy metal pollution diffusion in a certain period in the future, including: Based on historical data and actual monitoring information of heavy metal pollution factors, the time length is input into the target prediction system; Output the predicted data of heavy metal pollution spread in a certain period of time in the future.

Images