Soil-groundwater pollution risk threshold dynamic adjustment method and system based on environmental model

Abstract

The invention relates to the technical field of soil monitoring, in particular to a soil-groundwater pollution risk threshold dynamic adjustment method and system based on an environment model, and the method comprises the steps: collecting soil and groundwater monitoring data, constructing an environment model containing the soil pollution concentration, the groundwater pollution concentration and the pollutant migration rate, obtaining comprehensive pollution characteristic data, extracting main characteristics, and determining the contribution degree of each characteristic to risk assessment; calculating the initial weight of each feature according to the contribution degree, and obtaining a weight distribution scheme; calculating pollution risk values of the soil and the underground water in combination with the pollution load data and the weight values, and preliminarily setting a risk threshold value; monitoring pollutant concentration fluctuation data in real time, predicting a pollutant migration trend through a time sequence analysis algorithm, and obtaining a risk change prediction result; and combining the risk change prediction result and the environmental capacity constraint optimization risk threshold to obtain a dynamically adjusted risk threshold. According to the invention, accurate extraction of pollution characteristics and dynamic adjustment of weights are realized.

Description

Technical Field

[0001] This invention relates to the field of soil monitoring technology, and in particular to a method and system for dynamically adjusting soil-groundwater pollution risk thresholds based on an environmental model. Background Technology

[0002] Soil and groundwater pollution risk assessment is a crucial area in environmental management, directly impacting ecological security and human health. Accurate and scientific assessment of pollution risks not only provides a basis for pollution prevention and control but also effectively guides resource protection and governance decisions. Currently, while many assessment methods can identify pollution sources and potential risks, most rely on fixed parameters or single-scenario analysis, making them ill-suited to complex and ever-changing environmental conditions. These methods often fall short, especially when facing differences in hydrogeological conditions across regions, the dynamic nature of pollutant migration, and the impact of human activities on the environment. The core challenge lies in how to comprehensively consider multidimensional factors and achieve dynamic adjustments. The primary issue is constructing a quantitative assessment system that simultaneously encompasses multiple indicators such as water quality, pollution load, and environmental capacity. However, these indicators are difficult to quantify uniformly in practical applications due to the complexity of environmental conditions. For example, in an industrial zone, fluctuations in groundwater pollutant concentrations may lead to underestimating the risk using traditional methods, resulting in delayed remediation measures. Quantified indicators require dynamic adjustment of risk thresholds through real-time updated parameters and weights, but existing technologies lack flexible weight allocation mechanisms and cannot optimize thresholds promptly based on changes in pollutant type or receptor sensitivity. This makes it difficult for risk assessments to provide accurate decision support in the face of sudden pollution incidents or long-term environmental changes. Therefore, how to construct a comprehensive, multi-dimensional quantitative indicator and dynamically adjust risk thresholds has become a key issue in soil and groundwater pollution risk management. Summary of the Invention

[0003] The purpose of this invention is to provide a method and system for dynamically adjusting soil-groundwater pollution risk thresholds based on an environmental model, which enables accurate extraction of pollution characteristics, dynamic adjustment of weights, and further, real-time risk assessment.

[0004] To achieve the above objectives, on the one hand, the present invention provides a method for dynamically adjusting soil-groundwater pollution risk thresholds based on an environmental model, comprising:

[0005] Collect soil and groundwater monitoring data, construct an environmental model that includes soil pollution concentration, groundwater pollution concentration and pollutant migration rate, and obtain comprehensive pollution characteristic data;

[0006] Based on the comprehensive pollution characteristic data, the main characteristics of soil pollution, groundwater pollution, and pollutant migration are extracted, and the contribution of each characteristic to the risk assessment is determined.

[0007] The weight values ​​of soil pollution, groundwater pollution, and pollutant migration are calculated based on the contribution, and a weight allocation scheme for the multidimensional indicators is obtained.

[0008] By combining pollution load data and the aforementioned weight values, the pollution risk values ​​of soil and groundwater are calculated, and preliminary risk thresholds are set.

[0009] Real-time monitoring of pollutant concentration fluctuations, prediction of pollutant migration trends through time series analysis algorithms, and acquisition of risk change prediction results;

[0010] The risk threshold is optimized by combining the risk change prediction results and environmental capacity constraints to obtain a dynamically adjusted risk threshold.

[0011] On the other hand, the present invention also provides a dynamic adjustment system for soil-groundwater pollution risk thresholds based on an environmental model, comprising:

[0012] The data acquisition module is used to collect soil and groundwater monitoring data, construct an environmental model that includes soil pollution concentration, groundwater pollution concentration and pollutant migration rate, and obtain comprehensive pollution characteristic data.

[0013] The feature extraction module is used to extract the main features of soil pollution, groundwater pollution and pollutant migration based on the comprehensive pollution feature data, and to determine the contribution of each feature to the risk assessment.

[0014] The weight allocation module is used to calculate the weight values ​​of soil pollution, groundwater pollution and pollutant migration based on the contribution, and to obtain the weight allocation scheme of the multi-dimensional indicators.

[0015] The threshold setting module is used to calculate the pollution risk value of soil and groundwater by combining the pollution load data and the weight value, and to initially set the risk threshold.

[0016] The risk prediction module is used to monitor pollutant concentration fluctuation data in real time, predict pollutant migration trends through time series analysis algorithms, and obtain risk change prediction results.

[0017] The threshold adjustment module is used to optimize the risk threshold by combining the risk change prediction results and environmental capacity constraints, and obtain the dynamically adjusted risk threshold.

[0018] The beneficial effects of this invention are as follows:

[0019] This invention addresses the complex operational scenarios of soil and groundwater pollution monitoring, feature extraction, and dynamic risk assessment, solving the challenges of integrating multi-dimensional pollution data, quantifying feature contributions, and dynamically optimizing risk thresholds. It achieves accurate extraction of pollution features and dynamic adjustment of weights, further enabling real-time risk assessment. This significantly improves the scientific rigor and response efficiency of soil and groundwater pollution management, providing an intelligent solution for environmental risk prevention and control. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a flowchart of a method for dynamically adjusting soil-groundwater pollution risk thresholds based on an environmental model, according to an embodiment of the present invention. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0024] On the one hand, this embodiment provides a method for dynamically adjusting soil-groundwater pollution risk thresholds based on an environmental model, such as... Figure 1 As shown, it includes:

[0025] Collect soil and groundwater monitoring data, construct an environmental model that includes soil pollution concentration, groundwater pollution concentration and pollutant migration rate, and obtain comprehensive pollution characteristic data;

[0026] Based on the comprehensive pollution characteristic data, the main characteristics of soil pollution, groundwater pollution, and pollutant migration are extracted, and the contribution of each characteristic to the risk assessment is determined.

[0027] The weight values ​​of soil pollution, groundwater pollution, and pollutant migration are calculated based on the contribution, and a weight allocation scheme for the multidimensional indicators is obtained.

[0028] By combining pollution load data and the aforementioned weight values, the pollution risk values ​​of soil and groundwater are calculated, and preliminary risk thresholds are set.

[0029] Real-time monitoring of pollutant concentration fluctuations, prediction of pollutant migration trends through time series analysis algorithms, and acquisition of risk change prediction results;

[0030] The risk threshold is optimized by combining the risk change prediction results and environmental capacity constraints to obtain a dynamically adjusted risk threshold.

[0031] Specifically, this embodiment addresses the complex operational scenarios of soil and groundwater pollution monitoring, feature extraction, and dynamic risk assessment, solving the challenges of integrating multi-dimensional pollution data, quantifying feature contributions, and dynamically optimizing risk thresholds. This embodiment achieves accurate extraction of pollution features and dynamic adjustment of weights, further enabling real-time risk assessment. It significantly improves the scientific rigor and response efficiency of soil and groundwater pollution management, providing an intelligent solution for environmental risk prevention and control.

[0032] Furthermore, soil and groundwater monitoring data were collected to construct an environmental model that includes soil pollution concentration, groundwater pollution concentration, and pollutant migration rate, thereby obtaining comprehensive pollution characteristic data, including:

[0033] A sensor network was used to collect soil pollution concentration, groundwater pollution concentration, and hydrogeological parameters from multiple data sources to obtain a monitoring dataset.

[0034] Missing values ​​were filled and outliers were removed from the monitoring dataset. The monitoring dataset was then fused using the Kalman filter algorithm to generate a multi-source fused dataset.

[0035] Based on the multi-source fusion dataset, an environmental model including soil pollution concentration, groundwater pollution concentration, and pollutant migration rate is constructed by combining soil characteristics and groundwater flow field to obtain a comprehensive index dataset.

[0036] Specifically, this embodiment acquires soil and groundwater monitoring data through a sensor network and sampling equipment. Assuming 100 soil samples and 50 groundwater samples are collected in a certain area, pollutants such as benzo[a]pyrene (BaP) and heavy metal cadmium (Cd) are detected. The concentration range of BaP in the soil is 0.1-5.0 mg / kg, and the concentration of cadmium is 0.5-10.0 mg / kg. The concentration of BaP in the groundwater is 0.01-0.5 μg / L, and the concentration of cadmium is 1.0-50.0 μg / L. The data is stored in CSV format, including spatial information such as latitude and longitude, and depth (0-50 cm for soil, 1-10 m for groundwater). Combined with hydrogeological parameters, the permeability coefficient (K=10) is obtained using geological survey data. -4Hydrogeological parameters (m / s), porosity (n = 0.3), and groundwater velocity (v = 0.1 m / d) were spatially interpolated using ArcGIS software. A continuous hydrogeological parameter distribution map was generated using the Kriging method, with interpolation errors controlled within 5%. A multi-source data fusion algorithm was applied, using Kalman filtering to fuse soil and groundwater pollutant concentration data. Assuming an initial covariance matrix P0 = 0.1, process noise Q = 0.01, and observation noise R = 0.05, iterative calculations yielded a fused concentration distribution dataset, reducing the BaP fusion concentration error to 3%. A multi-dimensional index system was constructed, including soil pollution concentration, groundwater pollution concentration, and pollutant migration rate. The migration rate was calculated using Darcy's law, yielding a BaP migration rate of approximately 0.0033 m / d. Principal component analysis (PCA) was used to extract comprehensive pollution characteristics, assuming three principal components explain 80% of the variance, generating a comprehensive dataset containing pollution concentration and migration rate.

[0037] Furthermore, based on the comprehensive pollution characteristic data, the main characteristics of soil pollution, groundwater pollution, and pollutant migration are extracted, and the contribution of each characteristic to the risk assessment is determined, including:

[0038] After normalizing the comprehensive pollution characteristic data, independent component analysis and the ReliefF algorithm are used to extract features and take the intersection to obtain the main features;

[0039] After calculating the contribution of each major feature using the random forest algorithm, the Delphi algorithm is used for correction to obtain the final contribution. If the contribution is higher than the preset contribution threshold, the corresponding major feature is retained to obtain the high contribution feature set.

[0040] Specifically, this embodiment uses a standardization method for the comprehensive pollution feature dataset to convert each variable into a dimensionless value with a mean of 0 and a standard deviation of 1, in order to eliminate dimensional differences. The formula is z = (x - μ) / σ, where x is the original value, μ is the mean, and σ is the standard deviation.

[0041] Independent Component Analysis (ICA) and the ReliefF algorithm are used to extract features and then take their intersection. Examples of obtaining the main features are shown below.

[0042] (1) Independent Component Analysis (ICA) Extraction of Features:

[0043] Input: A 10×8 data matrix consisting of 8 normalized indicators;

[0044] Operation: The FastICA algorithm was used to decompose the components into three independent components (IC1, IC2, IC3). The top three indices with the highest absolute values ​​of the component loadings are as follows:

[0045] IC1 (absolute load): Soil benzene concentration (0.89), groundwater benzene concentration (0.82), groundwater flow velocity (0.76);

[0046] IC2 (absolute load): soil cadmium concentration (0.91), groundwater cadmium concentration (0.87), soil permeability coefficient (0.63);

[0047] IC3 (absolute load): soil lead concentration (0.85), groundwater pH (0.72), soil permeability coefficient (0.58);

[0048] Screening results: Indicators with absolute load values ​​> 0.6 were selected to obtain the ICA feature set: {soil benzene, groundwater benzene, groundwater flow velocity, soil cadmium, groundwater cadmium, soil permeability coefficient, soil lead, groundwater pH value}.

[0049] (2) Feature extraction using the ReliefF algorithm:

[0050] Input: 8 normalized indicators + sample risk level labels (low / medium / high risk, based on historical data);

[0051] Operation: Calculate the importance score for each feature (range 0-1):

[0052] Soil benzene concentration: 0.92, groundwater benzene concentration: 0.88, groundwater flow velocity: 0.85;

[0053] Soil cadmium concentration: 0.89, groundwater cadmium concentration: 0.83, soil permeability coefficient: 0.76;

[0054] Soil lead concentration: 0.61; Groundwater pH: 0.52;

[0055] Screening results: Indicators with scores > 0.6 were obtained, and the ReliefF feature set was obtained: {soil benzene, groundwater benzene, groundwater flow velocity, soil cadmium, groundwater cadmium, soil permeability coefficient, soil lead};

[0056] Feature intersection: The intersection of the extraction results of the two algorithms is the main feature set: {soil benzene concentration, groundwater benzene concentration, groundwater flow velocity, soil cadmium concentration, groundwater cadmium concentration, soil permeability coefficient, soil lead concentration}.

[0057] After calculating the contribution of each major feature using the Random Forest algorithm, the Delphi algorithm is used for correction to obtain the final contribution, which includes:

[0058] Random forest calculates initial contribution: Input the main features + sample risk level labels, construct a random forest model with several decision trees, and output the feature importance (Gini coefficient);

[0059] Delphi algorithm correction: Independent scoring (1-10 points) by soil-groundwater pollution assessment experts, calculation of average score and coordination coefficient, and output of final contribution after fusion.

[0060] Furthermore, based on the stated contribution, weight values ​​for soil pollution, groundwater pollution, and pollutant migration are calculated to obtain a weight allocation scheme for the multidimensional indicators, including:

[0061] A decision matrix is ​​constructed based on the contribution, and the weight values ​​of soil pollution, groundwater pollution and pollutant migration are calculated. If the deviation of the weight value exceeds the preset weight threshold, the feature contribution of the decision matrix is ​​adjusted iteratively to obtain the converged weight value.

[0062] The weight allocation ratios of soil pollution, groundwater pollution, and pollutant migration are determined based on the converged weight values, and the dominant pollution factors are extracted from the weight allocation ratios.

[0063] The pollutant migration rate is corrected using the dominant pollutant factor to obtain an adjusted pollutant migration rate distribution. Based on the adjusted pollutant migration rate distribution and weight allocation ratio, a weight allocation scheme for multidimensional indicators is generated.

[0064] Furthermore, after obtaining the weight allocation scheme of the multidimensional indicators, the method further includes:

[0065] Monitor the real-time changes of data for each key feature and calculate the magnitude of the changes. If the magnitude of the changes exceeds a preset threshold, it is marked as an abnormal data point, and a set of abnormal data points is obtained.

[0066] The grey relational analysis is used to calculate the degree of correlation between each indicator and abnormal data points. Based on the degree of correlation and the historical weight allocation scheme, new weight values ​​are calculated, the weight allocation scheme is updated, and the sliding window method is used to smooth the weight changes to obtain dynamically adjusted weight values.

[0067] Specifically, for example, if at a certain moment the groundwater level is 5.2 meters, the soil moisture content is 30%, and the pollutant concentration is 0.8 mg / L, the data is transmitted to a cloud database via the MQTT protocol and stored in JSON format: {"water_level":5.2,"soil_moisture":30,"pollutant":0.8}. Thresholds are set: groundwater level change ±0.5 meters, soil moisture content change ±5%, and pollutant concentration change ±0.2 mg / L. Assuming the data for the next hour is {"water_level":5.8,"soil_moisture":34,"pollutant":1.1}, the changes are calculated as follows: groundwater level change 0.6 meters (>0.5), soil moisture content change 4% (<5%), and pollutant concentration change 0.3 mg / L (>0.2). Since the groundwater level and pollutant concentration exceed the thresholds, a weight adjustment is triggered. Grey relational analysis is used to calculate the degree of correlation between each indicator and abnormal data points. The data of each indicator corresponding to the abnormal data points are used as reference sequences, and the historical data or data under normal conditions of each monitoring indicator are used as comparison sequences. Grey relational analysis is used to calculate the correlation coefficient between each comparison sequence and the reference sequence at each time point. Based on the correlation coefficient at each time point, the correlation degree between each comparison sequence and the reference sequence is calculated. The calculated correlation degree is normalized to obtain new weight values.

[0068] As another optional embodiment, the pollution risk values ​​of soil and groundwater are calculated by combining pollution load data and weight values, and a preliminary risk threshold is set as follows: Assume that pollution monitoring points in a certain area collect pollutant data, including a cadmium concentration of 2.5 mg / kg, an arsenic concentration of 15 mg / kg, and a nitrate concentration of 20 mg / L in groundwater. The weights are adjusted based on pollutant toxicity and environmental sensitivity, with initial weights of 0.4 for cadmium, 0.3 for arsenic, and 0.2 for nitrate. The adjusted weights are 0.45, 0.35, and 0.2, respectively. First, a fuzzy evaluation matrix is ​​constructed, standardizing pollutant concentrations into membership degrees. A cadmium concentration of 2.5 mg / kg corresponds to the soil quality standard (upper limit 5 mg / kg), with a membership degree of 2.5 / 5 = 0.5; an arsenic concentration of 15 mg / kg (standard 20 mg / kg) has a membership degree of 15 / 20 = 0.75; and a nitrate concentration of 20 mg / L (standard 50 mg / L) has a membership degree of 20 / 50 = 0.4. The membership vector R = [0.5, 0.75, 0.4] is formed. Next, the fuzzy comprehensive evaluation algorithm is applied, and the weight vector W = [0.45, 0.35, 0.2] is used in matrix operations with the membership matrix R to obtain the comprehensive evaluation value B.

[0069] B=W·R=0.45×0.5+0.35×0.75+0.2×0.4=0.225+0.2625+0.08=0.5675.

[0070] Furthermore, by monitoring pollutant concentration fluctuations in real time and using time series analysis algorithms to predict pollutant migration trends, risk change prediction results are obtained, including:

[0071] Obtain structured pollutant concentration data and extract time series features, calculate fluctuation amplitude, and if the fluctuation amplitude exceeds a preset fluctuation threshold, trigger an anomaly detection mechanism to determine an abnormal concentration event.

[0072] Based on the aforementioned abnormal concentration events, the ARIMA model is used to analyze time series data, predict pollutant migration trends, and obtain trend prediction results.

[0073] By combining the trend prediction results with the mapping of pollutant migration paths using a geographic information system, spatial distribution characteristics are obtained, and a random forest model is used to assess risk changes and obtain the risk probability distribution.

[0074] Based on the risk probability distribution, dynamically updated data on pollutant migration risk is generated, and the risk change prediction results are obtained.

[0075] Specifically, this embodiment uses a sensor network to collect PM2.5 concentration data in the air in real time, assuming data is collected once per minute. The data is stored in a cloud database in the format of timestamp and concentration value (unit: μg / m³). 3 For example, the PM2.5 concentration sequence from 10:00 to 10:10 on a certain day is [45.2, 46.1, 47.8, 50.3, 55.7, 60.2, 62.5, 58.9, 54.3, 49.8]. First, the concentration fluctuation amplitude is calculated. The mean μ = 53.08 and the standard deviation σ ≈ 5.77 are calculated using the standard deviation formula. The preset threshold is that the standard deviation exceeds 5. The current fluctuation amplitude of 5.77 exceeds the threshold, triggering the prediction process. Next, the ARIMA(2,1,1) time series analysis algorithm is used for trend prediction. p = 2 (autoregression order), d = 1 (difference order), and q = 1 (moving average order) are selected, and the model parameters are fitted using the least squares method. Based on historical 10-minute data, ARIMA predicts the PM2.5 concentration trend for the next 30 minutes, obtaining the predicted value sequence [48.5, 47.2, 46.0, ...], showing that the concentration is gradually decreasing. Risk change prediction sets a risk threshold (e.g., PM2.5 > 50 is high risk), and combines it with a prediction sequence to determine that the probability of high risk within the next 10 minutes is 20%, generating an alarm message and pushing it to the management platform. Simultaneously, the prediction results are combined with wind speed and direction data (e.g., wind speed 2 m / s, wind direction northeast) to infer that pollutants may migrate towards the northeast of the park, generating a migration path heat map.

[0076] Furthermore, by combining the risk change prediction results and environmental capacity constraints, the risk threshold is optimized to obtain a dynamically adjusted risk threshold, including:

[0077] The environmental capacity constraint is determined, and the risk change prediction result is compared and analyzed with the environmental capacity constraint. The risk threshold is adjusted according to the analysis result. The adjustment of the risk threshold adopts a multi-objective optimization algorithm, which sets the upper and lower limits of the risk threshold, takes the risk control effect and economic feasibility as the objective function, and uses the risk change prediction result and environmental capacity constraint as the constraint conditions. The optimal risk threshold is found through selection, crossover and mutation of genetic algorithm.

[0078] A consistency check is performed on the adjusted risk threshold. If the check passes, the adjusted risk threshold is output. If the check fails, the multi-objective optimization algorithm is re-executed to find the optimal risk threshold again until the check passes, and the final risk threshold is output.

[0079] Specifically, in this embodiment, when optimizing the risk threshold by combining the risk change prediction results and environmental capacity constraints, the environmental capacity constraint is first clarified. Environmental capacity refers to the maximum amount of pollutants that a certain environmental unit can accommodate, which can be calculated based on environmental dynamics models and ecotoxicological data. For example, the environmental capacity of soil considers the soil's adsorption capacity and microbial degradation capacity, while the environmental capacity of groundwater considers groundwater runoff and dilution capacity. The risk change prediction results are compared and analyzed with the environmental capacity constraints. If the predicted risk value may exceed the environmental capacity constraint in the future, it indicates that the initially set risk threshold is too high and needs to be reduced. If the predicted risk value is much lower than the environmental capacity constraint, and considering factors such as economic costs, the risk threshold can be appropriately increased.

[0080] Environmental capacity constraints are determined by regional pollution carrying capacity. For example, assuming the maximum environmental capacity for PM2.5 in a certain area is 50 μg / m³. 3 The current measured value is 40 μg / m 3 The remaining capacity is 10 μg / m³. 3 The constraint is that the risk index is positively correlated with PM2.5 concentration; for every 1 unit increase in the risk index, PM2.5 concentration increases by 0.5 μg / m³. 3 Therefore, a predicted risk index of 26 corresponds to an increase of 13 μg / m³ in PM2.5. 3 If the capacity constraint is exceeded, the threshold needs to be optimized.

[0081] A multi-objective optimization algorithm, such as a genetic algorithm, is employed to optimize the risk threshold with the objective functions of risk control effectiveness (keeping the risk value within environmental capacity constraints) and economic feasibility (reducing governance costs). During the optimization process, upper and lower limits for the risk threshold are set, and the predicted risk changes and environmental capacity constraints are used as constraints. The optimal risk threshold is then found through selection, crossover, and mutation operations using a genetic algorithm.

[0082] After obtaining the dynamically adjusted risk threshold, it is validated. The adjusted threshold is applied to actual pollution monitoring data to observe whether the risk assessment results are reasonable and accurately reflect changes in pollution risk. If the validation results meet the requirements, they are determined as the final dynamically adjusted risk threshold; if they do not meet the requirements, the process is repeated until a satisfactory result is obtained. Simultaneously, a dynamic adjustment mechanism for the risk threshold is established, periodically (e.g., quarterly or semi-annually) to readjust the risk threshold based on new monitoring data and forecast results to adapt to changes in pollution conditions.

[0083] On the other hand, this embodiment also provides a dynamic adjustment system for soil-groundwater pollution risk thresholds based on an environmental model, including:

[0084] The data acquisition module is used to collect soil and groundwater monitoring data, construct an environmental model that includes soil pollution concentration, groundwater pollution concentration and pollutant migration rate, and obtain comprehensive pollution characteristic data.

[0085] The feature extraction module is used to extract the main features of soil pollution, groundwater pollution and pollutant migration based on the comprehensive pollution feature data, and to determine the contribution of each feature to the risk assessment.

[0086] The weight allocation module is used to calculate the weight values ​​of soil pollution, groundwater pollution and pollutant migration based on the contribution, and to obtain the weight allocation scheme of the multi-dimensional indicators.

[0087] The threshold setting module is used to calculate the pollution risk value of soil and groundwater by combining the pollution load data and the weight value, and to initially set the risk threshold.

[0088] The risk prediction module is used to monitor pollutant concentration fluctuation data in real time, predict pollutant migration trends through time series analysis algorithms, and obtain risk change prediction results.

[0089] The threshold adjustment module is used to optimize the risk threshold by combining the risk change prediction results and environmental capacity constraints, and obtain the dynamically adjusted risk threshold.

[0090] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made by those skilled in the art to the technical solutions of the present invention without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A method for dynamically adjusting soil-groundwater pollution risk thresholds based on an environmental model, characterized in that, include: Collect soil and groundwater monitoring data, construct an environmental model that includes soil pollution concentration, groundwater pollution concentration and pollutant migration rate, and obtain comprehensive pollution characteristic data; Based on the comprehensive pollution characteristic data, the main characteristics of soil pollution, groundwater pollution, and pollutant migration are extracted, and the contribution of each characteristic to the risk assessment is determined. The weight values ​​of soil pollution, groundwater pollution, and pollutant migration are calculated based on the contribution, and a weight allocation scheme for the multidimensional indicators is obtained. By combining pollution load data and the aforementioned weight values, the pollution risk values ​​of soil and groundwater are calculated, and preliminary risk thresholds are set. Real-time monitoring of pollutant concentration fluctuations, prediction of pollutant migration trends through time series analysis algorithms, and acquisition of risk change prediction results; The risk threshold is optimized by combining the risk change prediction results and environmental capacity constraints to obtain a dynamically adjusted risk threshold.

2. The method for dynamically adjusting soil-groundwater pollution risk thresholds based on an environmental model according to claim 1, characterized in that, Collect soil and groundwater monitoring data, construct an environmental model including soil pollution concentration, groundwater pollution concentration, and pollutant migration rate, and obtain comprehensive pollution characteristic data, including: A sensor network was used to collect soil pollution concentration, groundwater pollution concentration, and hydrogeological parameters from multiple data sources to obtain a monitoring dataset. Missing values ​​were filled and outliers were removed from the monitoring dataset. The monitoring dataset was then fused using the Kalman filter algorithm to generate a multi-source fused dataset. Based on the multi-source fusion dataset, an environmental model including soil pollution concentration, groundwater pollution concentration, and pollutant migration rate is constructed by combining soil characteristics and groundwater flow field to obtain a comprehensive index dataset.

3. The method for dynamically adjusting soil-groundwater pollution risk thresholds based on environmental models according to claim 1, characterized in that, Based on the comprehensive pollution characteristic data, the main characteristics of soil pollution, groundwater pollution, and pollutant migration are extracted, and the contribution of each characteristic to the risk assessment is determined, including: After normalizing the comprehensive pollution characteristic data, independent component analysis and the ReliefF algorithm are used to extract features and take the intersection to obtain the main features; After calculating the contribution of each major feature using the random forest algorithm, the Delphi algorithm is used for correction to obtain the final contribution. If the contribution is higher than the preset contribution threshold, the corresponding major feature is retained to obtain the high contribution feature set.

4. The method for dynamically adjusting soil-groundwater pollution risk thresholds based on environmental models according to claim 1, characterized in that, Based on the stated contribution, the weight values ​​for soil pollution, groundwater pollution, and pollutant migration are calculated to obtain a weight allocation scheme for the multidimensional indicators, including: A decision matrix is ​​constructed based on the contribution, and the weight values ​​of soil pollution, groundwater pollution and pollutant migration are calculated. If the deviation of the weight value exceeds the preset weight threshold, the feature contribution of the decision matrix is ​​adjusted iteratively to obtain the converged weight value. The weight allocation ratios of soil pollution, groundwater pollution, and pollutant migration are determined based on the converged weight values, and the dominant pollution factors are extracted from the weight allocation ratios. The pollutant migration rate is corrected using the dominant pollutant factor to obtain an adjusted pollutant migration rate distribution. Based on the adjusted pollutant migration rate distribution and weight allocation ratio, a weight allocation scheme for multidimensional indicators is generated.

5. The method for dynamically adjusting soil-groundwater pollution risk thresholds based on an environmental model according to claim 4, characterized in that, After obtaining the weight allocation scheme of the multidimensional indicators, the method further includes: Monitor the real-time changes of data for each key feature and calculate the magnitude of the changes. If the magnitude of the changes exceeds a preset threshold, it is marked as an abnormal data point, and a set of abnormal data points is obtained. The grey relational analysis is used to calculate the degree of correlation between each indicator and abnormal data points. Based on the degree of correlation and the historical weight allocation scheme, new weight values ​​are calculated, the weight allocation scheme is updated, and the sliding window method is used to smooth the weight changes to obtain dynamically adjusted weight values.

6. The method for dynamically adjusting soil-groundwater pollution risk thresholds based on environmental models according to claim 1, characterized in that, Real-time monitoring of pollutant concentration fluctuations, prediction of pollutant migration trends using time series analysis algorithms, and acquisition of risk change prediction results, including: Obtain structured pollutant concentration data and extract time series features, calculate fluctuation amplitude, and if the fluctuation amplitude exceeds a preset fluctuation threshold, trigger an anomaly detection mechanism to determine an abnormal concentration event. Based on the aforementioned abnormal concentration events, the ARIMA model is used to analyze time series data, predict pollutant migration trends, and obtain trend prediction results. By combining the trend prediction results with the mapping of pollutant migration paths using a geographic information system, spatial distribution characteristics are obtained, and a random forest model is used to assess risk changes and obtain the risk probability distribution. Based on the risk probability distribution, dynamically updated data on pollutant migration risk is generated, and the risk change prediction results are obtained.

7. The method for dynamically adjusting soil-groundwater pollution risk thresholds based on environmental models according to claim 1, characterized in that, By combining the risk change prediction results and environmental capacity constraints, the risk threshold is optimized to obtain a dynamically adjusted risk threshold, including: The environmental capacity constraint is determined, and the risk change prediction result is compared and analyzed with the environmental capacity constraint. The risk threshold is adjusted according to the analysis result. The adjustment of the risk threshold adopts a multi-objective optimization algorithm, which sets the upper and lower limits of the risk threshold, takes the risk control effect and economic feasibility as the objective function, and uses the risk change prediction result and environmental capacity constraint as the constraint conditions. The optimal risk threshold is found through selection, crossover and mutation of genetic algorithm. A consistency check is performed on the adjusted risk threshold. If the check passes, the adjusted risk threshold is output. If the check fails, the multi-objective optimization algorithm is re-executed to find the optimal risk threshold again until the check passes, and the final risk threshold is output.

8. A dynamic adjustment system for soil-groundwater pollution risk thresholds based on an environmental model, characterized in that, include: The data acquisition module is used to collect soil and groundwater monitoring data, construct an environmental model that includes soil pollution concentration, groundwater pollution concentration and pollutant migration rate, and obtain comprehensive pollution characteristic data. The feature extraction module is used to extract the main features of soil pollution, groundwater pollution and pollutant migration based on the comprehensive pollution feature data, and to determine the contribution of each feature to the risk assessment. The weight allocation module is used to calculate the weight values ​​of soil pollution, groundwater pollution and pollutant migration based on the contribution, and to obtain the weight allocation scheme of the multi-dimensional indicators. The threshold setting module is used to calculate the pollution risk value of soil and groundwater by combining the pollution load data and the weight value, and to initially set the risk threshold. The risk prediction module is used to monitor pollutant concentration fluctuation data in real time, predict pollutant migration trends through time series analysis algorithms, and obtain risk change prediction results. The threshold adjustment module is used to optimize the risk threshold by combining the risk change prediction results and environmental capacity constraints, and obtain the dynamically adjusted risk threshold.

Images