Lead-zinc mining area particulate heavy metal migration simulation method and system
By establishing a mining area metal migration simulation model and combined with rainfall and soil data, the problem of granular heavy metal cadmium migration simulation in the lead -zinc mining area has been solved, and the accurate simulation of cadmium migration is achieved, and an important analysis basis is provided for pollution soil management.
Patent Information
- Application Number
- CN202411389238.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-30
- Publication Date
- 2025-01-24
Smart Images
Figure CN119358387A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of environmental engineering technology, and more particularly to a method and system for simulating the migration of particulate heavy metals in a lead-zinc mining area. Background Art
[0002] Understanding the migration and transformation process of heavy metal ions in soil is important for soil pollution control. Different types of heavy metals have different solubility, reactivity, and adsorption capacity, which will lead to differences in runoff distribution and sediment enrichment behavior of heavy metals.
[0003] The modeling of pollutant migration in a watershed can be divided into two stages: the land surface process stage and the river process stage. For heavy metal cadmium in soil, accurate simulation of the land surface process stage is the key to source prevention and control. In the land surface process stage, the content of particulate cadmium that migrates with sediment is difficult to calculate and needs to be simulated in the model.
[0004] Therefore, how to simulate the migration of particulate heavy metal cadmium is an urgent problem that technicians in this field need to solve. Summary of the invention
[0005] In view of this, the present invention provides a method and system for simulating the migration of particulate heavy metals in a lead-zinc mining area. By establishing a metal migration simulation model for the mining area and combining rainfall and soil data, the migration simulation of particulate heavy metal cadmium is realized.
[0006] In order to achieve the above object, the present invention adopts the following technical solution:
[0007] A method for simulating the migration of particulate heavy metals in a lead-zinc mining area, comprising:
[0008] Establishing a mining area metal migration simulation model, dividing the mining area metal migration simulation model into grids, performing multi-layer processing on each grid, randomly sampling the grids after multi-layer processing, and obtaining soil data;
[0009] The particulate cadmium content is predicted based on the pre-trained content prediction model combined with the soil data, and a multivariate regression analysis is performed according to the predicted value of the particulate cadmium content to obtain the particulate cadmium enrichment coefficient in the sediment;
[0010] Calculate sediment erosion based on the grid-based sediment loss area, sediment content, and total rainfall;
[0011] The migration amount of particulate cadmium in each grid area is calculated based on the sediment erosion amount and the particulate cadmium enrichment coefficient;
[0012] Based on the migration amount of particulate cadmium, the grid is processed by a sequential simulation method to obtain three-dimensional surface data. The three-dimensional surface data is alternately expressed using GRID and TIN data formats to update the mining area metal migration simulation model to form a migration simulation three-dimensional model.
[0013] Preferably, the acquisition of soil data specifically includes: randomly sampling the grid using a diagonal sampling method, a quincuncial sampling method, a chessboard sampling method or an S-shaped sampling method to obtain a plurality of sampling points; gridding the sampling points according to a set grid unit size to obtain a multi-layer fine grid; and collecting soil data based on the multi-layer fine grid.
[0014] Preferably, the pre-trained content prediction model specifically includes:
[0015] Constructing at least two content prediction sub-models, and using historical soil data to train each of the content prediction sub-models to obtain trained content prediction sub-models;
[0016] The influence values of the historical soil data are respectively input into the trained content prediction sub-models to obtain multiple prediction results, and the prediction results are coupled based on the Bayesian algorithm to construct the content prediction model.
[0017] Preferably, the constructing of the content prediction model specifically includes:
[0018] The particulate cadmium content in each prediction result is divided into N numerical intervals, each interval corresponding to a category; the prediction accuracy of the content prediction sub-model in a certain category is calculated, and a corresponding regional weight is assigned to each content prediction sub-model according to the accuracy, and the regional weight and the prediction result are weighted and summed to construct the content prediction model.
[0019] Preferably, the weighted summation of the regional weight and the prediction result specifically includes: obtaining the minimum and maximum particulate cadmium contents in the prediction result, specifying the number of categories to be divided, dividing the particulate cadmium content into intervals, and each interval corresponds to a particulate cadmium content category, and the formula is as follows:
[0020]
[0021] Where d is the length of each interval, P max and P min are the maximum and minimum values of the particle cadmium content, respectively, and N is the number of specified categories;
[0022] Import the influence values of historical soil data into the content prediction sub-model again to obtain the category corresponding to the prediction result, compare the category of the prediction result with the category of the actual value, calculate the probability P(A) of each prediction category appearing in the interval length d, and the probability of falling into each prediction category when the prediction is accurate is P(A|accurate);
[0023]
[0024] Among them, n A is the number of times the predicted value falls within interval A, and n is the total number of predictions;
[0025]
[0026] Among them, n 准确 (A) is the number of times the predicted value falls within the A interval and the prediction is accurate, n 准确 is the total number of accurate predictions; P(A|accurate) of each content prediction sub-model is calculated;
[0027] According to the prediction accuracy of each sub-model in different intervals, a corresponding weighted value is assigned to each sub-model. The weighted value distribution formula is as follows:
[0028]
[0029] Among them, w i is the weighted value of the i-th content prediction sub-model, and the content prediction model is constructed according to the weighted value:
[0030]
[0031] Among them, P i is the predicted value of the i-th content prediction sub-model, P is the predicted value of the comprehensive particulate cadmium content, and M is the number of prediction sub-models.
[0032] Preferably, the calculation of sediment erosion specifically includes:
[0033] sed=SHW 0 R;
[0034] Determine the amount of sediment erosion in the area where each grid is located within a preset period of time, where sed is the amount of sediment erosion, S is the grid area, H is the sediment content, and W 0 is the total rainfall of the grid, and R is the correction factor.
[0035] Preferably, obtaining the migration amount of particulate cadmium within each grid area specifically includes:
[0036] Q s =k s P s S s·sed×10 -3 ;
[0037] Among them, Q s is the migration amount of particulate cadmium in the eroded sediment of grid s, sed is the amount of sediment erosion, ks is the enrichment coefficient of particulate cadmium in grid s, P s is the predicted value of the comprehensive particulate cadmium content in the soil of grid s, S s is the grid area of grid s.
[0038] Preferably, the processing of the grid by using a sequential simulation method based on the migration amount of the particulate cadmium specifically includes processing multiple layers of fine grids by using a sequential simulation method to obtain a continuous surface of the migration amount of the particulate cadmium:
[0039] Select a random position X in the fine grid and determine all nearest neighbor positions within the set search radius of position X;
[0040] Based on the variogram model, a predicted value of X and the estimated standard deviation at X are obtained as a linear weighted combination of T selected points;
[0041] Using the cumulative normal distribution of the mean value D and standard deviation SD of the random position X predicted value of the migration of particulate cadmium at point T, a random variable is selected and used as the estimate of X;
[0042] Next, another random position X is selected in the fine grid, and the random variables of the random points of the fine grid are obtained according to the above method, and the three-dimensional surface data of the migration amount of particulate cadmium are obtained by repeating the process in sequence.
[0043] A simulation system for the migration of particulate heavy metals in a lead-zinc mining area, comprising:
[0044] A soil data acquisition module is used to establish a metal migration simulation model for a mining area, divide the metal migration simulation model into grids, perform multi-layer processing on each grid, randomly sample the grids after multi-layer processing, and acquire soil data;
[0045] An enrichment coefficient analysis module predicts the content of particulate cadmium based on a pre-trained content prediction model combined with the soil data, and performs a multivariate regression analysis based on the predicted value of the particulate cadmium content to obtain the enrichment coefficient of particulate cadmium in the sediment;
[0046] The erosion calculation module calculates the sediment erosion based on the sediment loss area, sediment content, and total rainfall of the grid;
[0047] A migration amount calculation module is used to calculate the migration amount of particulate cadmium in each grid area based on the sediment erosion amount and the particulate cadmium enrichment coefficient;
[0048] The data updating module processes the grid by a sequential simulation method based on the migration amount of the particulate cadmium to obtain three-dimensional surface data, and uses GRID and TIN data formats to alternately express the three-dimensional surface data to update the metal migration simulation model of the mining area to form a migration simulation three-dimensional model.
[0049] Through the above technical scheme, it can be known that compared with the prior art, the present invention discloses a method and system for simulating the migration of particulate heavy metals in a lead-zinc mining area. By establishing a metal migration simulation model in a mining area, grid division, multi-layer processing, sampling and obtaining soil data, a series of steps, the complex environment of the mining area is fully considered, and a solid data basis is provided for subsequent analysis; the content prediction model is pre-trained and constructed in combination with a variety of methods, such as constructing multiple sub-models and coupling based on the Bayesian algorithm, which can improve the accuracy of the prediction of particulate cadmium content. The numerical interval is divided and the weight is assigned according to the accuracy, which further enhances the reliability of the prediction; the sediment loss area, sediment content, total rainfall and correction coefficient of the grid are used to calculate the sediment erosion amount, and the main factors affecting the sediment erosion are fully considered. The migration amount of particulate cadmium in each grid area is calculated in combination with the sediment erosion amount and the particulate cadmium enrichment coefficient, which can accurately reflect the migration of cadmium in the mining area; the grid is sampled by diagonal sampling method, plum blossom sampling method, chessboard sampling method or S-shaped sampling method, which can ensure the representativeness and comprehensiveness of the sampling. BRIEF DESCRIPTION OF THE DRAWINGS
[0050] 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 embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.
[0051] Figure 1 A flow chart of the method steps provided by the present invention;
[0052] Figure 2 A schematic diagram of the structure provided by the present invention. DETAILED DESCRIPTION
[0053] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. 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 creative work are within the scope of protection of the present invention.
[0054] The embodiment of the present invention discloses a method for simulating the migration of heavy metals in particle state in a lead-zinc mining area. Figure 1As shown, including:
[0055] Establish a metal migration simulation model in the mining area, divide the metal migration simulation model into grids (the number of grids is determined according to different research purposes), perform multi-layer processing on each grid, randomly sample the grids after multi-layer processing, and obtain soil data;
[0056] The particulate cadmium content was predicted based on the pre-trained content prediction model combined with soil data, and the particulate cadmium enrichment coefficient in sediment was obtained by multivariate regression analysis based on the predicted value of the particulate cadmium content.
[0057] Calculate sediment erosion based on the grid-based sediment loss area, sediment content, and total rainfall;
[0058] The migration amount of particulate cadmium in each grid area was calculated based on the amount of sediment erosion and the enrichment coefficient of particulate cadmium;
[0059] Based on the migration amount of particulate cadmium, the grid is processed by sequential simulation method to obtain three-dimensional surface data. The three-dimensional surface data are alternately expressed using GRID and TIN data formats to update the metal migration simulation model of the mining area and form a three-dimensional migration simulation model.
[0060] In a specific embodiment, the soil data includes soil pH value, soil organic matter content, soil particle size, and sediment content.
[0061] In a specific embodiment, it also includes performing a normality test on the soil data of the randomly sampled grid. If the data is normally distributed, multi-layer processing is continued. If the data is not normally distributed, data transformation is performed, such as taking logarithm, sine or cosine.
[0062] In a specific embodiment, obtaining soil data specifically includes: randomly sampling the grid using a diagonal sampling method, a plum blossom sampling method, a chessboard sampling method or an S-shaped sampling method to obtain a plurality of sampling points; gridding the sampling points according to a set grid unit size to obtain a multi-layer fine grid; collecting soil data based on the multi-layer fine grid; in order to enhance the improvement of long-distance effects and short-range variability, performing multi-layer process processing on the randomly sampled grid, starting from a relatively coarse grid, gradually generating finer and finer grids, and finally forming a fine grid.
[0063] For each randomly sampled grid, the multi-layer processing process involves two aspects. One is the sampling method. There are many sampling methods, and the sampling methods are different in different situations. They are: 1. Diagonal sampling method: suitable for sewage irrigation plots, sampling at the central points of each diagonal equally divided; 2. Plum blossom sampling method: suitable for plots with small area, flat terrain and uniform soil; 3. Chessboard sampling method: suitable for plots with medium area, flat terrain, basically complete terrain and uneven soil; 4. S-shaped or snake sampling method: suitable for plots with small area, uneven terrain, uneven soil and more sampling points; 5. Grid sampling method, for the grid processing of sampling points, the grid processing has a typical matrix structure, and the computer calculation saves time; the grid cell size (Cell Size), also known as analysis resolution, can be set manually according to the purpose of research. For example, when performing fusion analysis with satellite images, it is generally consistent with the resolution of satellite images, such as the 25m*25m resolution of remote sensing TM images. The spatial analysis of raster data is performed on the basis of each raster unit. If the raster unit is too large, the accuracy of the analysis result will be reduced. If the unit is too small, a large amount of data will be generated and the calculation speed will be reduced. Therefore, it is necessary to choose a suitable unit size.
[0064] In a specific embodiment, the pre-trained content prediction model specifically includes:
[0065] constructing at least two content prediction sub-models, and using historical soil data to train each content prediction sub-model to obtain a trained content prediction sub-model;
[0066] The influence values of historical soil data are input into each trained content prediction sub-model to obtain multiple prediction results. The prediction results are coupled based on the Bayesian algorithm to construct a content prediction model.
[0067] In a specific embodiment, constructing a content prediction model specifically includes:
[0068] The particulate cadmium content in each prediction result is divided into N numerical intervals, each interval corresponding to a category; the prediction accuracy of the content prediction sub-model in a certain category is calculated, and a corresponding regional weight is assigned to each content prediction sub-model according to the accuracy. The regional weight and the prediction result are weighted and summed to construct a content prediction model.
[0069] The content prediction sub-model includes but is not limited to two or more of BP neural network, case reasoning method, SVM algorithm and decision tree method.
[0070] In a specific embodiment, the weighted sum of the regional weight and the prediction result specifically includes: obtaining the minimum and maximum particulate cadmium content in the prediction result, specifying the number of categories to be divided, dividing the particulate cadmium content into intervals, and each interval corresponds to a particulate cadmium content category, and the formula is as follows:
[0071]
[0072] Where d is the length of each interval, P max and P min are the maximum and minimum values of the particle cadmium content, respectively, and N is the number of specified categories;
[0073] Import the influence values of historical soil data into the content prediction sub-model again to obtain the category corresponding to the prediction result, compare the category of the prediction result with the category of the actual value, calculate the probability P(A) of each prediction category appearing in the interval length d, and the probability of falling into each prediction category when the prediction is accurate is P(A|accurate);
[0074]
[0075] Among them, n A is the number of times the predicted value falls within interval A, and n is the total number of predictions;
[0076]
[0077] Among them, n 准确 (A) is the number of times the predicted value falls within the A interval and the prediction is accurate, n 准确 is the total number of accurate predictions; P(A|accurate) of each content prediction sub-model is calculated;
[0078] According to the prediction accuracy of each sub-model in different intervals, a corresponding weighted value is assigned to each sub-model. The weighted value distribution formula is as follows:
[0079]
[0080] Among them, w i is the weighted value of the i-th content prediction sub-model, and the content prediction model is constructed according to the weighted value:
[0081]
[0082] Among them, P i is the prediction value of the i-th content prediction sub-model, P is the prediction value of the comprehensive particulate cadmium content, and M is the number of prediction sub-models. Taking the above four as an example, i = 1, 2, 3, and 4, respectively referring to the prediction value of BP neural network, the prediction value of SVM, the prediction value of case reasoning, and the prediction value of decision tree.
[0083] In a specific embodiment, performing a multivariate regression analysis based on the predicted value of the particulate cadmium content to obtain the particulate cadmium enrichment coefficient in the sediment includes using field runoff plots and indoor control experiment monitoring data, and determining the quantitative relationship between the sediment enrichment coefficient of surface cadmium and multiple main controlling factors through multivariate statistical regression analysis.
[0084] In a specific embodiment, calculating the amount of sediment erosion specifically includes:
[0085] sed=SHW 0 R;
[0086] Determine the amount of sediment erosion in each grid area within a preset period of time, where sed is the amount of sediment erosion, S is the grid area, H is the sediment content, and W is the amount of sediment erosion. 0 is the total rainfall of the grid, and R is the correction factor.
[0087] Among them, the correction coefficient is the surface runoff coefficient of the area where the production and construction project is located. The surface runoff coefficient comprehensively reflects the impact of natural geographical elements in the basin on the precipitation-runoff relationship. The surface runoff coefficient is determined by technical personnel in this field based on the actual land conditions in the area where the production and construction project is located and after consulting the surface runoff coefficient empirical value table.
[0088] In a specific embodiment, the total rainfall calculation of the grid specifically includes: obtaining image vectors of grid monitoring images in outdoor environments, and calculating the average value of image vectors of grid monitoring images in each outdoor environment to obtain an average image vector;
[0089] The average image vector is input into a pre-trained convolutional neural network model, and the convolutional neural network model outputs a predicted label for the rainfall period, wherein the convolutional neural network model is trained by outdoor environment monitoring sample images annotated with rainfall labels;
[0090] Determine whether the rainfall label of the rainfall period is the same as the predicted label of the rainfall period;
[0091] If the rainfall label of the rainfall period is different from the predicted label of the rainfall period, the compensation coefficient of the rainfall period is obtained according to a preset compensation coefficient corresponding table, wherein the compensation coefficient corresponding table is used to record the corresponding relationship between the rainfall label of the rainfall period and the predicted label and the compensation coefficient;
[0092] If the rainfall label of the rainfall period is the same as the predicted label of the rainfall period, the compensation coefficient of the rainfall period is determined to be 1;
[0093] The total rainfall in the area where the grid is located within the preset period is calculated according to the following formula:
[0094] W 0 =∑kp j ;
[0095] Among them, W 0 is the total rainfall of the grid, k is the compensation coefficient of the rainfall period, p j is the rainfall in rainfall period j.
[0096] In one embodiment, the compensation coefficient table corresponds to the following Table 1:
[0097] Table 1 Rainfall compensation coefficient
[0098]
[0099] In a specific embodiment, obtaining the migration amount of particulate cadmium within each grid area specifically includes:
[0100] Q s =k s P s S s ·sed×10 -3 ;
[0101] Among them, Q s is the migration amount of particulate cadmium in the eroded sediment of grid s, sed is the amount of sediment erosion, ks is the enrichment coefficient of particulate cadmium in grid s, P s is the predicted value of the comprehensive particulate cadmium content in the soil of grid s, S s is the grid area of grid s.
[0102] In a specific embodiment, the grid is processed by a sequential simulation method based on the migration amount of particulate cadmium, which specifically includes processing multiple layers of fine grids by a sequential simulation method to obtain a continuous surface of the migration amount of particulate cadmium:
[0103] Select a random position X in the fine grid and determine all nearest neighbor positions within the set search radius of position X;
[0104] Based on the variogram model, a predicted value of X and the estimated standard deviation at X are obtained as a linear weighted combination of T selected points;
[0105] Using the cumulative normal distribution of the mean value D and standard deviation SD of the random position X predicted value of the migration of particulate cadmium at point T, a random variable is selected and used as the estimate of X;
[0106] Next, another random position X is selected in the fine grid, and the random variables of the random points of the fine grid are obtained according to the above method, and the three-dimensional surface data of the migration amount of particulate cadmium are obtained by repeating the process in sequence.
[0107] Screening the variogram model, the screening variogram model is mainly based on curve fitting for computer automatic processing and calculation. The variogram model mainly includes exponential model, Gaussian model or spherical model and other models.
[0108] Based on random walk, local search and conditional estimation, sequential simulation of particle cadmium migration at random sampling points is realized to form a seamless surface.
[0109] Sequential simulation integrates various information by stochastically modeling the properties of particulate cadmium migration, and integrates the correlation and uncertainty of this information into the model, emphasizing that the probability model is the overall probability characteristic of the action and result. In the sequential simulation process, the simulated value conditions of the measured points need to be converted to the measured values, so the simulated values of the measured points are equal to the measured values, which is suitable for quantitatively characterizing the heterogeneity and uncertainty of particulate cadmium migration. With the increase of the number of simulations, the sequential simulation describes the distribution of soil parameter values in the entire simulation area in more detail, the ratio of nugget value to base value gradually increases, and the range gradually approaches the measured data.
[0110] For a random position X in the fine grid, all the nearest neighbor positions within the defined search radius at position X are determined; the selection of the nearest neighbor position is exactly what is required by the autocorrelation characteristics of the regionalized variable of particulate cadmium migration. Autocorrelation refers to the correlation characteristics of the particulate cadmium migration itself as the distance changes.
[0111] The obtained variogram model provides a predicted value at X and an estimated standard deviation at X as a linear weighted combination of T selected points, and uses the cumulative normal distribution of mean D and standard deviation SD to select a random value of the random variable as an estimate of X. The linear weighted combination of T selected points can be considered as an indicator, and the mean D is the average of the predicted values of the random position X of the migration of particulate cadmium at N points.
[0112] Then, randomly walk to another unvisited grid position, and continue to determine the random variables according to the definition of the random position of the fine grid and the variogram model mentioned above until all nodes of the fine grid are visited. Repeat this process for multiple layers of fine grids, and each time the result is added to the input data point as the source data point value for random walk simulation. In this repeated process, the random walk method reduces the risk of false appearance when searching the grid and achieves the stability of model operation.
[0113] For regional variables such as the migration of particulate cadmium, stationarity is one of the goals in the simulation and modeling process of raster spatial data, and the stability model promotes the gradual promotion of model application. In the random walk search process, there are two types of stationarity: one is mean stationarity, which assumes that the mean is constant and independent of position; the other is second-order stationarity related to the covariance function and intrinsic stationarity related to the semivariogram function. Second-order stationarity assumes that the covariance of any two points with the same distance and direction is the same, and the covariance is only related to the values of the two points and has nothing to do with their positions.
[0114] Random simulation can generate numerous realizations, each of which shows the same pattern but in different ways. In the univariate distribution model, the uncertainty of the random variable is statistically calculated through the series of results. Similarly, the output results of a series of random simulations show the characteristics of uncertainty. There are specific correlation rules in the random process, which can be used to predict and estimate the uncertainty of the prediction results. With the help of sequential simulation, the pattern of uncertainty is expressed in the form of random probability. As for intrinsic stationarity, the intrinsic stationarity assumption of random sample data shows that the variance of any two points at the same distance and direction is the same. This stationarity is necessary for the stability of the variance function model of random sampling points and the determination of the variance function type, or stationarity is the prerequisite for establishing a variance model.
[0115] Gaussian sequential simulation is a basic simulation method that overcomes the smoothing effect of the Kriging method. This simulation requires first defining the joint probability model of all grid point attribute values. The joint distribution is defined as:
[0116] F(z 1 ,z 2 ,z 3 ,…,z N )=Pr(Z(u 1 )≤z 1 ,…,Z(u N )≤z N );
[0117] u 1 ,u 2 ,u 3 ,…,u N Z(u 1 ) 1 ;
[0118] Among them, z 1 ,z 2 ,z 3 ,…,z N For point u 1,u 2 ,u 3 ,…,u N The measured value; Pr is the probability; Z(u 1 ) is u 1 The main idea of the sequential Gaussian simulation conditional method is to sequentially find the conditional cumulative distribution function of each grid node along a random path and obtain the simulation value from the conditional cumulative distribution function.
[0119] A simulation system for the migration of heavy metals in particle form in lead-zinc mining areas, such as Figure 2 As shown, including:
[0120] The soil data acquisition module establishes a metal migration simulation model for the mining area, divides the metal migration simulation model into grids, performs multi-layer processing on each grid, and randomly samples the grids after multi-layer processing to obtain soil data;
[0121] The enrichment coefficient analysis module predicts the content of particulate cadmium based on the pre-trained content prediction model combined with soil data, and performs multivariate regression analysis based on the predicted value of the particulate cadmium content to obtain the enrichment coefficient of particulate cadmium in sediment;
[0122] The erosion calculation module calculates the sediment erosion based on the sediment loss area, sediment content, and total rainfall of the grid;
[0123] The migration amount calculation module calculates the migration amount of particulate cadmium in each grid area based on the sediment erosion amount and the particulate cadmium enrichment coefficient;
[0124] The data update module processes the grid using a sequential simulation method based on the migration amount of particulate cadmium to obtain three-dimensional surface data. The three-dimensional surface data is alternately expressed using GRID and TIN data formats to update the metal migration simulation model of the mining area and form a three-dimensional migration simulation model.
[0125] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.
[0126] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A method for simulating the migration of particulate heavy metals in lead-zinc mining areas, characterized in that: include: Establishing a mining area metal migration simulation model, dividing the mining area metal migration simulation model into grids, performing multi-layer processing on each grid, randomly sampling the grids after multi-layer processing, and obtaining soil data; The particulate cadmium content is predicted based on the pre-trained content prediction model combined with the soil data, and a multivariate regression analysis is performed according to the predicted value of the particulate cadmium content to obtain the particulate cadmium enrichment coefficient in the sediment; Calculate sediment erosion based on the grid-based sediment loss area, sediment content, and total rainfall; The migration amount of particulate cadmium in each grid area is calculated based on the sediment erosion amount and the particulate cadmium enrichment coefficient; Based on the migration amount of particulate cadmium, the grid is processed by a sequential simulation method to obtain three-dimensional surface data. The three-dimensional surface data is alternately expressed using GRID and TIN data formats to update the mining area metal migration simulation model to form a migration simulation three-dimensional model.
2. The method for simulating the migration of particulate heavy metals in a lead-zinc mining area according to claim 1, characterized in that: The method of obtaining soil data specifically includes: randomly sampling the grid using a diagonal sampling method, a quincunx sampling method, a chessboard sampling method or an S-shaped sampling method to obtain a plurality of sampling points; gridding the sampling points according to a set grid unit size to obtain a multi-layer fine grid; and collecting soil data based on the multi-layer fine grid.
3. The method for simulating the migration of particulate heavy metals in a lead-zinc mining area according to claim 1, characterized in that: The pre-trained content prediction model specifically includes: Constructing at least two content prediction sub-models, and using historical soil data to train each of the content prediction sub-models to obtain trained content prediction sub-models; The influence values of the historical soil data are respectively input into the trained content prediction sub-models to obtain multiple prediction results, and the prediction results are coupled based on the Bayesian algorithm to construct the content prediction model.
4. The method for simulating the migration of particulate heavy metals in a lead-zinc mining area according to claim 3, characterized in that: The constructing of the content prediction model specifically includes: The particulate cadmium content in each prediction result is divided into N numerical intervals, each interval corresponding to a category; the prediction accuracy of the content prediction sub-model in a certain category is calculated, and a corresponding regional weight is assigned to each content prediction sub-model according to the accuracy, and the regional weight and the prediction result are weighted and summed to construct the content prediction model.
5. A method for simulating the migration of particulate heavy metals in a lead-zinc mining area according to claim 4, characterized in that: The weighted summation of the regional weight and the prediction result specifically includes: obtaining the minimum and maximum particulate cadmium contents in the prediction result, specifying the number of categories to be divided, dividing the particulate cadmium content into intervals, and each interval corresponds to a particulate cadmium content category, and the formula is as follows: Where d is the length of each interval, P max and P min are the maximum and minimum values of the particle cadmium content, respectively, and N is the number of specified categories; Import the influence values of historical soil data into the content prediction sub-model again to obtain the category corresponding to the prediction result, compare the category of the prediction result with the category of the actual value, calculate the probability P(A) of each prediction category appearing in the interval length d, and the probability of falling into each prediction category when the prediction is accurate is P(A|accurate); Among them, n A is the number of times the predicted value falls within interval A, and n is the total number of predictions; Among them, n 准确 (A) is the number of times the predicted value falls within the A interval and the prediction is accurate, n 准确 is the total number of accurate predictions; P(A|accurate) of each content prediction sub-model is calculated; According to the prediction accuracy of each sub-model in different intervals, a corresponding weighted value is assigned to each sub-model. The weighted value distribution formula is as follows: Among them, w i is the weighted value of the i-th content prediction sub-model, and the content prediction model is constructed according to the weighted value: P i ′=w i ·P i Among them, P i is the predicted value of the i-th content prediction sub-model, P is the predicted value of the comprehensive particulate cadmium content, and M is the number of prediction sub-models.
6. The method for simulating the migration of particulate heavy metals in a lead-zinc mining area according to claim 1, characterized in that: The calculation of sediment erosion specifically includes: sed=SHW0R; The amount of sediment erosion in the area where each grid is located within a preset time period is determined, wherein sed is the amount of sediment erosion, S is the grid area, H is the sediment content, W0 is the total rainfall of the grid, and R is the correction coefficient.
7. The method for simulating the migration of particulate heavy metals in a lead-zinc mining area according to claim 1, characterized in that: The step of obtaining the migration amount of particulate cadmium within each grid area specifically includes: Q s =k s P s S s ·sed×10 -3 ; Among them, Q s is the migration amount of particulate cadmium in the eroded sediment of grid s, sed is the amount of sediment erosion, ks is the enrichment coefficient of particulate cadmium in grid s, P s is the predicted value of the comprehensive particulate cadmium content in the soil of grid s, S s is the grid area of grid s.
8. The method for simulating the migration of particulate heavy metals in a lead-zinc mining area according to claim 2, characterized in that: The processing of the grid by using a sequential simulation method based on the migration amount of the particulate cadmium specifically includes processing multiple layers of fine grids by using a sequential simulation method to obtain a continuous surface of the migration amount of the particulate cadmium: Select a random position X in the fine grid and determine all nearest neighbor positions within the set search radius of position X; Based on the variogram model, a predicted value of X and the estimated standard deviation at X are obtained as a linear weighted combination of T selected points; Using the cumulative normal distribution of the mean value D and standard deviation SD of the random position X predicted value of the migration of particulate cadmium at point T, a random variable is selected and used as the estimate of X; Next, another random position X is selected in the fine grid, and the random variables of the random points of the fine grid are obtained according to the above method, and the three-dimensional surface data of the migration amount of particulate cadmium are obtained by repeating the process in sequence.
9. A system for simulating the migration of heavy metals in a lead-zinc mining area in a particle state, using a method for simulating the migration of heavy metals in a lead-zinc mining area in a particle state as claimed in any one of claims 1 to 8, characterized in that: include: A soil data acquisition module is used to establish a metal migration simulation model for a mining area, divide the metal migration simulation model into grids, perform multi-layer processing on each grid, randomly sample the grids after multi-layer processing, and acquire soil data; An enrichment coefficient analysis module predicts the content of particulate cadmium based on a pre-trained content prediction model combined with the soil data, and performs a multivariate regression analysis based on the predicted value of the particulate cadmium content to obtain the enrichment coefficient of particulate cadmium in the sediment; The erosion calculation module calculates the sediment erosion based on the sediment loss area, sediment content, and total rainfall of the grid; A migration amount calculation module is used to calculate the migration amount of particulate cadmium in each grid area based on the sediment erosion amount and the particulate cadmium enrichment coefficient; The data updating module processes the grid by a sequential simulation method based on the migration amount of the particulate cadmium to obtain three-dimensional surface data, and uses GRID and TIN data formats to alternately express the three-dimensional surface data to update the metal migration simulation model of the mining area to form a migration simulation three-dimensional model.