CN115081310A - Method for predicting biological accessibility of mining and metallurgy sites - Google Patents

Abstract

The invention discloses a method for predicting the biological accessibility of a mining and smelting site, which relates to the technical field of mining and smelting, and comprises the following steps: s1, collecting soil characteristics, heavy metal components and heavy metal biological accessibility data in a research area through a collection module, and determining biological accessibility of soil at a sampling point, S2, according to the soil characteristics data, the heavy metal components and the heavy metal biological accessibility data, with the biological accessibility as a response variable, the soil characteristics and the heavy metal components as a prediction variable, establishing a random forest regression model, and the mining and metallurgical site biological accessibility prediction method can accurately obtain non-linear relation data between soil heavy metal biological effectiveness and the soil characteristics and the heavy metal components, and the constructed random forest model has good generalization capability, can predict biological accessibility of cadmium, lead and arsenic in different sites, and reduces influence of model prediction on prediction due to site differences.

Description

A Prediction Method for Bioavailability of Mining and Metallurgical Sites

technical field

The invention relates to the technical fields of mining and smelting, in particular to a method for predicting the biological accessibility of a mining and smelting site.

Background technique

With the rapid development of industrial activities such as mining and smelting, various heavy metals are released into the soil, resulting in heavy metal pollution of the soil environment. According to the first national soil pollution survey in China, the soil environmental pollution is relatively serious, and 34.9% of the industrial and mining abandoned land samples exceeded the Chinese soil environmental quality standards. Among them, three heavy metals, cadmium, lead and arsenic, have been concerned with diseases such as cardiovascular disease, nervous system damage and cancer. In order to better manage and control heavy metal pollution in soil environment, some countries in the world have issued general assessment criteria for soil environment, including various chemicals such as cadmium, lead and arsenic. Among these general evaluation criteria, in most cases, the total concentration of heavy metals is considered in the model, which means that these soil environmental standards do not take into account the bioavailability of heavy metals in soil. The restoration and reuse of industrial and mining land is very important.

The bioavailability of cadmium, lead, and arsenic was determined by comparing the accumulation of metals in animal tissues or urine with soluble reference substances such as sodium arsenate (NaH2AsO4) and lead acetate (Pb(AC)2) and cadmium chloride (CdCl2). Since bioavailability is determined by expensive and ethically controversial animal experiments, many in vitro gastrointestinal phase simulations are widely used as surrogate methods for assessing the bioavailability of metals and have been validated in animal experiments. The ratio of heavy metals to total content extracted from in vitro simulated gastrointestinal phase experiments was defined as bioaccessibility. The bioavailability of heavy metals determined by different in vitro gastrointestinal phase simulation methods is related to animal experiments. Therefore, it is necessary to use suitable in vitro gastrointestinal phase simulation methods to evaluate the bioavailability of heavy metals in soil. Accurate prediction models for bioaccessibility are created for specific sites, but model predictions vary from site to site, it is often difficult to apply a model designed for one site to another, and soil properties at sites can vary widely, which is extremely The earth limits the feasibility of using one model on two or more different sites.

SUMMARY OF THE INVENTION

The purpose of the present invention is to solve at least one of the technical problems existing in the prior art, and to provide a method for predicting the bioaccessibility of mining and metallurgy sites, which can solve the problem that the soil properties of the sites may be very different, which greatly limits the two The question of the feasibility of using a model for one or more different sites.

In order to achieve the above purpose, the present invention provides the following technical solution: a method for predicting the bioavailability of a mining and metallurgical site, comprising the following steps: the S1 collects soil properties and heavy metal components, heavy metal bioavailability in the research area through a collection module. availability data and determine the biological accessibility of the soil at the sampling site;

S2 establishes a random forest regression model through the modeling module based on soil property data, heavy metal components and heavy metal bioavailability data, with bioaccessibility as the response variable, soil properties and heavy metal components as the predictor variables;

S3 uses the calculation module, according to the above measurement data, with the bio-accessibility of heavy metals in the soil as the response variable, the soil properties and heavy metal components as the predictor variables, and the random bootstrap sampling method is used to construct a decision tree to establish a random forest regression model. ; Among them, bioaccessibility refers to the ratio of the heavy metal content extracted from the in vitro gastrointestinal phase simulation experiment to the heavy metal content in the soil;

S4 determines the number of predictors and the number of decision trees randomly sampled from each decision tree in the random forest regression model;

S5 calculates the contribution rate of predictors to the bioaccessibility of heavy metals based on the shap value according to the random forest regression model.

Preferably, the accuracy of the random forest regression model is estimated by using the five-fold cross-check method through the S2.

Preferably, the contribution rate of the predictor to the bioavailability of heavy metals is calculated according to a random forest regression model, and the contribution rate of the predictor to the bioavailability of heavy metals is calculated based on the shap value.

Preferably, according to the soil characteristics, heavy metal components, and heavy metal bioavailability data, the S2 modeling module uses the heavy metal bioavailability as a response variable, and uses soil characteristics and heavy metal components as predictions to establish a random forest. Regression model; in which, bioaccessibility refers to the ratio of the heavy metal content extracted from the in vitro gastrointestinal phase simulation experiment to the heavy metal content in the soil.

Preferably, the calculation module in S3 is used to calculate the contribution rate of geochemical factors to the bioavailability of heavy metals according to a random forest regression model.

Preferably, the memory is used to store the program, and the processor is used to load the program to execute the estimation method of soil heavy metal bioavailability.

Compared with the prior art, the beneficial effects of the present invention are:

(1) The method for predicting the bioavailability of mining and metallurgical sites can more accurately obtain the non-linear relationship data between soil heavy metal bioavailability and soil properties and heavy metal components, and the constructed random forest model has better performance The generalization ability of the model can predict the bioavailability of cadmium, lead, and arsenic in different sites, reduce the impact of model predictions on predictions that vary from location to location, and expand the scope of application.

Description of drawings

Below in conjunction with accompanying drawing and embodiment, the present invention is further described:

1 is a schematic flow chart of a method for predicting biological accessibility of a mining and metallurgical site according to the present invention;

2 is a schematic diagram of the importance ranking based on SHAP value calculation using a random forest model using soil properties and heavy metal forms as predictors in the present invention;

FIG. 3 is a schematic diagram of the importance ranking based on SHAP value calculation in the random forest model using soil properties as predictors according to the present invention.

Figure 4 is a schematic diagram of the composition and in vitro parameters of the UBM, IVG and PBET bioavailability assays of the present invention.

Figure 5 is the bioavailability of heavy metals in the gastric phase of the random forest (RF) model of the present invention. (a) Soil properties and heavy metal species as predictors. (b) Schematic diagram of soil properties as predictors.

FIG. 6 shows the importance of predicting variables to model performance by calculating feature importance by shap value in the present invention, and the analysis result.

Detailed ways

This part will describe the specific embodiments of the present invention in detail, and the preferred embodiments of the present invention are shown in the accompanying drawings. Each technical feature and overall technical solution of the invention should not be construed as limiting the protection scope of the invention.

In the description of the present invention, it should be understood that the azimuth description, such as the azimuth or position relationship indicated by up, down, front, rear, left, right, etc., is based on the azimuth or position relationship shown in the drawings, only In order to facilitate the description of the present invention and simplify the description, it is not indicated or implied that the indicated device or element must have a particular orientation, be constructed and operated in a particular orientation, and therefore should not be construed as limiting the present invention.

In the description of the present invention, greater than, less than, exceeding, etc. are understood as not including this number, and above, below, within, etc. are understood as including this number. If it is described that the first and the second are only for the purpose of distinguishing technical features, it cannot be understood as indicating or implying relative importance, or indicating the number of the indicated technical features or the order of the indicated technical features. relation.

In the description of the present invention, unless otherwise clearly defined, words such as setting, installation, connection should be understood in a broad sense, and those skilled in the art can reasonably determine the specific meanings of the above words in the present invention in combination with the specific content of the technical solution.

1-6, the present invention provides a technical solution: a method for predicting the biological accessibility of mining and metallurgy sites, including the following steps: S1 collects soil properties and heavy metal components, heavy metal biological Accessibility data and determination of soil bioaccessibility at sampling sites;

S2 uses the modeling module to establish a random forest regression model based on soil property data, heavy metal components and heavy metal bioaccessibility data, with bioaccessibility as a response variable, soil properties and heavy metal components as predictor variables, and a five-fold crossover model is used. The test method is used to estimate the accuracy of the random forest regression model. Based on the data of soil properties, heavy metal components, and heavy metal bioavailability, the modeling module takes heavy metal bioavailability as the response variable and soil properties and heavy metal components as predictions. Random forest regression model; in which, bioaccessibility refers to the ratio of the heavy metal content extracted from the in vitro gastrointestinal phase simulation experiment to the heavy metal content in the soil;

S3 uses the calculation module, according to the above measurement data, with the bio-accessibility of heavy metals in the soil as the response variable, the soil properties and heavy metal components as the predictor variables, and the random bootstrap sampling method is used to construct a decision tree to establish a random forest regression model. ; Among them, bioaccessibility refers to the ratio of the heavy metal content extracted from the in vitro gastrointestinal phase simulation experiment to the heavy metal content in the soil, and the calculation module is used to calculate the contribution rate of geochemical factors to the bioavailability of heavy metals according to the random forest regression model;

S4 determines the number of predictors and the number of decision trees randomly sampled from each decision tree in the random forest regression model;

S5 calculates the contribution rate of predictors to the bioaccessibility of heavy metals based on the shap value according to the random forest regression model.

The memory is used for storing the program, and the processor is used for loading the program to execute the estimation method of soil heavy metal bioavailability.

First, sample sampling: 33 soils from different contaminated sites in mining and smelting areas were collected from eight provinces in China. Each soil sample consisted of three soil cores (0-20 cm deep). Found in Table S1 in the Supplementary Information, all soils were air-dried at room temperature and then passed through a 10-mesh sieve to remove debris and pebbles for physical and chemical analysis of soil properties. A portion of the sample was ground in an agate mortar and sieved to less than 250 microns for total elemental and bioavailability analysis.

The detailed physicochemical properties of the soil were then measured. Soil pH and conductivity were measured with a pH meter and a conductivity meter at 1:2.5 and 1:5 mass:volume soil and water suspensions, soil total organic carbon (TOC) and total carbon ( TC) was measured by a total organic carbon analyzer (TOC, ASI-5000A), and the air-dried soil samples were leached with 1.66cmol/L Co(NH3)6CL3 solution to determine soil cation exchange capacity (CEC), soil dissolved total nitrogen (TN) was extracted with 0.5 mol-L-1K2SO4 at 1:5 (mass:volume) and measured with a total organic carbon analyzer (TOC, ASI-5000A, Shimazu) using a laser analyzer (Mastersizerm 3000, Morvern, UK) Soil texture (sand, silt, clay; volume percent) was determined to determine the soil particle size distribution, and to determine the total concentration of metals in the soil, a wet digestion microwave oven (Milestone MLS 1200Mega) was used to add a mixture of HNO3:HCL:HF, ratio As 2:1:1 (v:v:v), the solution was dried on a hot plate to completely remove HF, and the residue was dissolved in nitric acid. The resulting solutions were analyzed for metal content by ICP-MS and ICP-OES.

The optimized BCR three-step sequential extraction method was used to extract four components of heavy metals from soils at different sampling points. The method is divided into four stages: (1) exchangeable metals soluble in water or weakly linked to carbonates, obtained with 0.11 mol/L acidic acetic acid; (2) metals attached to iron and manganese oxides, With 0.1 mol/L hydroxylamine hydrochloride at pH 1.5, (3) metals combined with organics and sulfides, first oxidize the residue of step 2 with hydrogen peroxide, then with 1 mol/L ammonium acetate at pH 2 Oxidation proceeds, (4) residue state, obtained by decomposing the residue of step 3 with aqua regia and HF.

Three in vitro assays were employed to assess the bioavailability of HMS in soil, which were validated by animal models as the best in vitro method for predicting bioavailability, composition and analysis in the gastric phase (GP) and intestinal phase (IP), respectively. The parameters are shown in Figure 4.

Then, regression prediction: the bioaccessibility of soil cadmium measured by PBET method, soil lead measured by UBM method, and soil arsenic measured by IVG method were used as the response variables of random forest model. The component is used as a predictor to train a single regression tree, the number of trees is set to 200, a single regression tree trained with 200 lesson trees is combined, the test data is used for testing, and the random forest model obtained by the final combination is used for regression prediction. The importance of predictor variables to the bioavailability of different heavy metals and the interaction of predictor variables and response variables are shown in Figure 5.

After preprocessing logarithmic transformation of the target pollutant data set in the study, it was found that the fitting accuracy of the random forest model for the three heavy metals was high, such as Cd (R2CV=0.92, RMSEcv=0.30), Pb (R2CV= 0.3, RMSEcv=0.39) and As (R2CV=0.810, RMSEcv=0.23).

In order to gain a more comprehensive understanding of the salient features that affect the bioaccessibility of heavy metals, we used the feature importance calculated by the shap value to characterize the importance of the predictor variables to the model performance. The analysis results are shown in Figure 6.

The feature importance calculated based on the shap value shows that HMStotal, F1, F2 and F123 are always the more important features in the random forest model, although their importance rankings are slightly different in different heavy metal prediction models. Specifically, Cd In the prediction model, Cdtotal and F2 are the first and second most important features, followed by F123 and F1. The more important features in the lead prediction model are the same as the Cd prediction model, but there are differences in their importance ranking, and the ranking is F123> F2>Pbtotal>F1, and in the As-GP prediction model, F123 and Astotal were the two most important features, followed by EC and F1. In the bioavailability model that only used soil properties as predictors to predict Cd, the most important feature was Cdtotal, and the more important features were EC, Eh, and pH. It indicated that F1, F2 and F123 in the heavy metal components were the key factors affecting the bioavailability of heavy metals. In addition, electrical conductivity in soil is also a key factor affecting the bioaccessibility of arsenic in soil.

The nonlinear relationship data between soil heavy metal bioavailability and soil properties and heavy metal components can be obtained more accurately. And the constructed random forest model has good generalization ability, and can predict the concentrations of cadmium, lead and arsenic in different sites. Bioavailability.

The embodiments of the present invention have been described in detail above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned embodiments. Within the scope of knowledge possessed by those of ordinary skill in the technical field, various modifications can be made without departing from the purpose of the present invention. kind of change.

Claims (6)

1. A method for predicting the accessibility of organisms in a mining and smelting site is characterized by comprising the following steps: the method comprises the following steps: the S1 collects soil characteristics, heavy metal components and heavy metal biological accessibility data in the research area through the collection module, and determines the biological accessibility of the soil of the sampling point;

s2, establishing a random forest regression model by using the biological accessibility as a response variable and the soil characteristics and the heavy metal components as prediction variables according to the soil characteristic data, the heavy metal components and the heavy metal biological accessibility data through a modeling module;

s3, according to the measured data, using biological accessibility of heavy metals in soil as response variables and soil characteristics and heavy metal components as prediction variables through a calculation module, and constructing a decision tree by adopting a random bootstrap sampling method to establish a random forest regression model; wherein, the biological accessibility refers to the ratio of the content of heavy metal extracted in an in vitro gastrointestinal phase simulation experiment to the content of heavy metal in soil;

s4, determining the number of the prediction variables and the number of the decision trees randomly sampled by each decision tree in the random forest regression model;

s5, according to the random forest regression model, calculating the contribution rate of the prediction factors to the accessibility of the heavy metal organisms based on the shape value.

2. A method of predicting the bioassays of mining and metallurgical sites according to claim 1, wherein said method comprises: and estimating the precision of the random forest regression model by adopting a five-fold cross-checking method through the S2.

3. A method of predicting the bioassays of mining and metallurgical sites according to claim 1, wherein said method comprises: and calculating the contribution rate of the prediction factor to the heavy metal biological effectiveness according to the random forest regression model, and calculating the contribution rate of the prediction factor to the heavy metal biological effectiveness based on the shape value.

4. A method of predicting the bioassays of mining and metallurgical sites according to claim 1, wherein said method comprises: establishing a random forest regression model by the S2 modeling module according to the soil characteristics, the heavy metal components and the heavy metal biological accessibility data, taking the heavy metal biological accessibility as a response variable and taking the soil characteristics and the heavy metal components as predictions; wherein, the biological accessibility refers to the ratio of the content of heavy metal extracted in an in vitro gastrointestinal phase simulation experiment to the content of heavy metal in soil.

5. A method of predicting the bioassays of mining and metallurgical sites according to claim 1, wherein said method comprises: and calculating the contribution rate of the geochemical factors to the effectiveness of the heavy metal organisms according to a random forest regression model by the calculating module in the S3.

6. A method of predicting the bioassays of mining and metallurgical sites according to claim 1, wherein said method comprises: the through memory is used for storing programs, and the processor is used for loading the programs to execute the estimation method of the bioavailability of the soil heavy metal.

Images