Soil polycyclic aromatic hydrocarbon migration flux internet of things-gis spatial modeling analysis method and system

Abstract

The application relates to a soil polycyclic aromatic hydrocarbon migration flux Internet of Things-GIS spatial modeling analysis method and system, which comprises the following steps: collecting data in real time through Beidou satellite and LoRa dual-mode transmission technology, and outputting a standardized data set; based on the standardized data set, the sources and contribution rates of PAHs are analyzed through a characteristic ratio method and an APCS-MLR model, adsorption and diffusion coefficients are determined through indoor experiments, a migration rate equation is established, and key parameters are calculated; a spatial database is constructed in ArcGIS, a GIS model is constructed by fusing a physical mechanism and a random forest model through Bayesian optimization based on the standardized data set and the key parameters, PAHs migration simulation is realized, and the GIS model is constructed; real-time monitoring data are input into the GIS model, the spatiotemporal distribution of PAHs in a future preset time period is predicted, a migration probability graph and three-dimensional vertical migration characteristics are output; and based on the migration probability graph and the three-dimensional vertical migration characteristics, a risk grade is divided by integrating a Nemerow index, a toxicity equivalent and a lifetime carcinogenic risk.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of space modeling, in particular to a soil polycyclic aromatic hydrocarbon (PAHs) migration flux Internet of Things (IoT)-GIS space modeling analysis method and system. BACKGROUND

[0002] PAHs are persistent organic pollutants formed by two or more benzene rings, which have strong carcinogenic, teratogenic and mutagenic properties. The accumulation and migration of PAHs in soil pose a serious threat to the ecological environment and human health. As the main "sink" of PAHs in the environment, once contaminated, soil is difficult to recover naturally due to its stable physicochemical properties and slow degradation. Accurate understanding of the migration of PAHs in soil is the core prerequisite for pollution tracing, risk control and remediation.

[0003] PAHs migration is influenced by soil physicochemical properties, meteorological conditions, pollution sources (coal combustion, traffic emissions), etc. Traditional monitoring relies on laboratory analysis, which involves collecting soil samples (mainly 0~20 cm surface soil), and using gas chromatography-mass spectrometry (GC-MS) to determine PAHs concentration.

[0004] According to the above-mentioned technology, the inventors believe that traditional research relies on static monitoring and local analysis, which is not convenient for migration simulation and risk warning, affecting the prediction accuracy of pollution sources, migration trend prediction or health risk warning, leading to passive response in pollution control.

[0005] Based on this, the application provides a soil polycyclic aromatic hydrocarbon migration flux IoT-GIS space modeling analysis method and system. SUMMARY

[0006] In order to improve the problem that traditional research relies on static monitoring and local analysis, which is not convenient for migration simulation and risk warning, affecting the prediction accuracy of pollution sources, migration trend prediction or health risk warning, leading to passive response in pollution control, the application provides a soil polycyclic aromatic hydrocarbon migration flux IoT-GIS space modeling analysis method and system.

[0007] In a first aspect, the application provides a soil polycyclic aromatic hydrocarbon migration flux IoT-GIS space modeling analysis method, which adopts the following technical solution: comprising:

[0008] Real-time collection of PAHs concentration data, environmental factor data and remote sensing image data, and use of Bayesian data fusion algorithm for spatio-temporal matching and outlier removal of multi-source data, output of standardized data set;

[0009] Based on the standardized dataset, the source and contribution rate of PAHs were analyzed by the eigenvalue method and the APCS-MLR model. Combined with indoor experiments to determine the adsorption and diffusion coefficients, a migration rate equation was established and key parameters were calculated.

[0010] A spatial database is constructed in ArcGIS. Based on the standardized dataset and the key parameters, the physical mechanism and the random forest model are integrated, and the migration simulation of PAHs is achieved through Bayesian optimization to construct a GIS model.

[0011] The real-time monitoring data is input into the GIS model to predict the spatiotemporal distribution of PAHs within a preset time period in the future, and outputs a migration probability map and three-dimensional vertical migration features.

[0012] Based on the migration probability map and the three-dimensional vertical migration features, a joint risk index is constructed by integrating the Nemerow index, toxicity equivalent, and lifetime carcinogenic risk, and risk levels are classified.

[0013] Preferably, the real-time acquisition of PAHs concentration data, environmental factor data, and remote sensing image data, and the use of a Bayesian data fusion algorithm to perform spatiotemporal matching and outlier removal on the multi-source data, outputting a standardized dataset, including:

[0014] Based on PAHs-specific fluorescence spectroscopy sensors, soil physicochemical sensors, and meteorological sensors, concentration data of 21 PAHs and data of 17 environmental factors were collected. Sensor data were transmitted via LoRa wireless communication, and remote sensing image data and pollution source emission inventories were received to form the original dataset.

[0015] The original dataset is time-synchronized and spatially registered, and then input into ArcGIS to generate a spatial unit attribute table of a 100 m × 100 m grid.

[0016] Outliers are removed using the 3σ criterion, the effective sample size is retained, and a spatiotemporally aligned dataset is output.

[0017] Using the spatiotemporally aligned dataset as input, a Bayesian fusion model is constructed: sensor data is assigned high weights, remote sensing inversion data is assigned medium weights, and the posterior probability distribution is calculated iteratively using the Markov chain Monte Carlo algorithm. The fused standardized dataset is then output, which includes the concentrations of 21 PAHs, environmental factor values, and data reliability for each grid cell.

[0018] Preferably, based on the standardized dataset, the source and contribution rate of PAHs are analyzed using the eigenvalue ratio method and the APCS-MLR model. Combined with indoor experimental measurements of adsorption and diffusion coefficients, a migration rate equation is established, and key parameters are calculated, including:

[0019] According to the concentration of 21 PAHs monomers in the standardized data set, a characteristic ratio is calculated, and the main pollution source type and spatial distribution characteristics are preliminarily identified;

[0020] Using the APCS-MLR model, KMO test and Bartlett spherical test are performed on the PAHs concentration data, the principal components with a characteristic ratio greater than are extracted, and the principal component loading matrix is converted into absolute factor scores to represent the contribution intensity of different pollution sources;

[0021] A regression equation is established with the total concentration of PAHs as the dependent variable and APCS as the independent variable, and the contribution rate of each principal component to the pollution source is calculated;

[0022] Based on the standardized data set and the pollution source contribution rate, a multivariate linear regression is used to take the PAHs migration rate as the dependent variable and the environmental factors as the independent variable, and a PAHs migration rate equation is constructed by combining the adsorption capacity and the diffusion coefficient to calculate the degradation coefficient and the migration rate;

[0023] The pollution source contribution rate, the adsorption capacity, the diffusion coefficient, the migration rate and the degradation coefficient are integrated to generate the key parameters that meet the GIS modeling.

[0024] Preferably, the spatial database is constructed in ArcGIS, based on the standardized data set and the key parameters, the physical mechanism and the random forest model are fused, the PAHs migration simulation is realized through Bayesian optimization, and the GIS model is constructed, including:

[0025] The spatial database is constructed in ArcGIS, the standardized data set and the key parameters are input, the terrain and land use elements are integrated, and the spatial database of 100 m x 100 m grid is generated;

[0026] The physical mechanism model is constructed, based on the migration rate, the diffusion coefficient and the degradation coefficient in the key parameters, the vertical and horizontal migration of PAHs in soil is simulated based on the hydrodynamic-adsorption desorption coupling equation, and the spatial and temporal distribution of the physical simulation of PAHs concentration is output;

[0027] The random forest model is constructed, the environmental factors and PAHs concentration in the standardized data set are input, the data-driven prediction is trained, and the data-driven PAHs concentration prediction result is output;

[0028] The physical mechanism model and the random forest model are fused through the Bayesian optimization algorithm to obtain a coupled model, and the data-driven prediction is constrained by the physical simulation result, and the fused PAHs migration simulation result is output;

[0029] The PAHs migration simulation result is input into ArcGIS, and after cross-validation, the spatial expression is performed to generate the GIS model containing the PAHs migration path and the concentration value line.

[0030] Preferably, the physical mechanism model and the random forest model are fused by the Bayesian optimization algorithm to obtain a coupled model, and the data-driven prediction is constrained by the physical simulation result, and the fused PAHs migration simulation result is output, including:

[0031] The physical simulation result output by the physical mechanism model, the data-driven prediction result output by the random forest model and the verification set data are input, the physical simulation result and the data-driven prediction result are one-to-one corresponding according to the grid unit, the initial weight of the physical simulation result and the data-driven prediction result is given, and a fusion input matrix with initial weight is constructed;

[0032] The fusion input matrix and the verification set data are input, the target function is defined as the determination coefficient of the measured PAHs concentration of the verification set and the fused prediction concentration, and the constraint condition is set as the determination coefficient greater than 0.85, and an optimization framework containing the target function and the constraint condition is formed;

[0033] Based on the optimization framework and the fusion input matrix, the Bayesian optimization algorithm is adopted, the weight is iteratively optimized by the Gaussian process proxy model, the determination coefficient is initialized to be calculated, if the determination coefficient is less than 0.85, 10 groups of weight combinations are randomly sampled to calculate the determination coefficient, the optimal sampling point is selected by using the expectation improvement function to adjust the weight, and the determination coefficient is not less than 0.85 in the continuous 5 rounds of iteration, and the optimal weight is output.

[0034] Based on the optimal weight, the physical simulation result and the data-driven prediction result, the PAHs concentration of each grid unit is calculated according to the fusion concentration = physical simulation result initial weight x physical simulation concentration + data-driven prediction result initial weight x data-driven prediction concentration, the migration path and the rate parameter in the physical simulation result are synchronously integrated, and the fused PAHs migration simulation result is output.

[0035] Preferably, the real-time monitoring data is input into the GIS model, the spatio-temporal distribution of PAHs in a future preset time period is predicted, and the migration probability graph and the three-dimensional vertical migration characteristics are output, including:

[0036] The real-time monitoring data is input into the GIS model, the future preset time period and the time granularity are set based on the user demand, the key parameters are loaded in the GIS model, and the prediction space range is determined as all grid units.

[0037] The migration rate equation in the coupling model and real-time meteorological data are called by a GIS model to simulate horizontal and vertical migration of PAHs, and output spatial and temporal distribution data of PAHs in a 100 m grid unit in a preset future period of time;

[0038] Based on the predicted spatial and temporal distribution data of the concentration of PAHs, the probability of the concentration of each grid unit exceeding a risk threshold is calculated, a color gradient probability map is generated by using an ArcGIS spatial analysis tool, and a high-risk aggregation area is marked;

[0039] PAHs concentration data in soil layers are extracted from the predicted data, the migration depth of PAHs in each soil layer is calculated, and a vertical profile feature map is generated.

[0040] Preferably, based on the migration probability map and the three-dimensional vertical migration feature, a joint risk index is constructed by integrating the Nemerow index, the toxic equivalent and the lifetime cancer risk, and a risk level is divided, including:

[0041] Based on the PAHs concentration data in the migration probability map and the concentration distribution of soil layers in the vertical migration feature map, the Nemerow index, the toxic equivalent and the lifetime cancer risk are calculated, and 100 m grid unit data of the three single risk indexes, namely the Nemerow index, the toxic equivalent and the lifetime cancer risk, are output;

[0042] Based on risk contribution analysis, the Nemerow index weight, the toxic equivalent weight and the lifetime cancer risk weight are given, and the sum of the weights is ensured to be 1, and a weight allocation scheme is output;

[0043] Based on the single risk index and the weight, the JRI value of each grid unit is calculated, wherein the Nemerow index, the toxic equivalent and the lifetime cancer risk are normalized, and JRI spatial distribution data of a 100 m grid unit are output;

[0044] Based on the frequency distribution of the JRI value, the risk is divided into three levels by using a natural breakpoint method, a high-risk area is marked by using ArcGIS, and a risk level spatial distribution map and a grading standard are output.

[0045] In a second aspect, an application discloses a soil polycyclic aromatic hydrocarbon migration flux Internet of Things-GIS spatial modeling analysis device, which adopts the following technical scheme and comprises:

[0046] A data acquisition module is configured to acquire PAHs concentration data, environmental factor data and remote sensing image data in real time, perform spatiotemporal matching and outlier elimination on multi-source data by using a Bayesian data fusion algorithm, and output a standardized data set;

[0047] An indoor detection module is configured to analyze sources and contribution rates of PAHs by a characteristic ratio method and an APCS-MLR model based on the standardized data set, and to establish a migration rate equation by combining with adsorption and diffusion coefficients determined by indoor experiments to calculate key parameters;

[0048] A model coupling module is configured to construct a spatial database in ArcGIS, to fuse physical mechanisms and random forest models based on the standardized data set and the key parameters, to realize PAHs migration simulation by Bayesian optimization, and to construct a GIS model.

[0049] A distribution calculation module is configured to input real-time monitoring data into the GIS model, to predict spatiotemporal distribution of PAHs in a future preset time period, and to output a migration probability map and three-dimensional vertical migration characteristics.

[0050] A risk level module is configured to integrate a Nemerow index, a toxicity equivalent, and a lifetime carcinogenic risk based on the migration probability map and the three-dimensional vertical migration characteristics, to construct a joint risk index, and to divide a risk level.

[0051] In a third aspect, the present application further provides a control device, which comprises:

[0052] The control device comprises a memory and a processor, and the memory stores a computer program capable of being loaded and executed by the processor to implement the soil PAHs migration flux IoT-GIS spatial modeling analysis method.

[0053] In a fourth aspect, the present application further provides a computer readable storage medium storing a computer program capable of being loaded and executed by a processor to implement the soil PAHs migration flux IoT-GIS spatial modeling analysis method.

[0054] In summary, in this application, the Internet of Things sensor is used in combination with Beidou / LoRa dual-mode transmission, and a Bayesian fusion algorithm is used to process multi-source data to generate a high-resolution (100m grid) standardized data set, solving the traditional monitoring time and space lag problem. Secondly, through the feature ratio method and APCS-MLR model to analyze the source and contribution rate of PAHs, combined with indoor experiment to measure the adsorption / diffusion coefficient, to establish the migration rate equation, to quantify the key parameters, and to break through the limitations of traditional source analysis qualitative rough; Then through the fusion of physical mechanism model and random forest model, and the Bayesian optimization to build GIS migration simulation model, realize the high-precision simulation of PAHs horizontal and vertical migration, overcome the defects of single model physical process loss or data-driven unconstrained, finally based on the model to predict the future time and space distribution of PAHs in a certain period of time, output migration probability map and three-dimensional vertical characteristics, and integrate the Nemerow index, toxicity equivalent, and carcinogenic risk to build the joint risk index (JRI), and divide the risk into three levels, so as to facilitate the migration simulation and risk warning, and the precise prediction of pollution source, prediction of migration trend or health risk warning, so as to facilitate the pollution control. BRIEF DESCRIPTION OF DRAWINGS

[0055] Figure 1 It is a process schematic diagram of a soil polycyclic aromatic hydrocarbon migration flux Internet of Things-GIS spatial modeling analysis method.

[0056] Figure 2 It is a structural block diagram of a soil polycyclic aromatic hydrocarbon migration flux Internet of Things-GIS spatial modeling analysis device. DETAILED DESCRIPTION

[0057] The following Figure 1 - Figure 2 The application is further described in detail.

[0058] The application relates to the technical field of soil pollution monitoring and risk assessment, and specifically discloses a soil polycyclic aromatic hydrocarbon (PAHs) migration flux Internet of Things-GIS spatial modeling analysis method. As a typical persistent organic pollutant, the accumulation and migration of polycyclic aromatic hydrocarbons in soil pose a serious threat to the ecological environment and human health, especially in industrial-intensive areas and energy bases, PAHs pollution presents the characteristics of "high concentration, strong heterogeneity and difficult degradation", and traditional monitoring methods have limitations such as low spatial and temporal resolution, rough source analysis, and fragmented modeling, making it difficult to achieve precise control.

[0059] To solve the above problems, the application integrates real-time sensing of Internet of Things, multi-source data fusion, mechanism-data coupled modeling and GIS spatial analysis technology to build an integrated technology system of "real-time monitoring-source analysis-dynamic simulation-risk early warning": first, the Internet of Things sensor network with Beidou / LoRa dual-mode transmission is used to collect PAHs concentration and environmental factor data, and a high-resolution standardized data set is generated by combining Bayesian algorithm to fuse multi-source information; second, the characteristic ratio method and APCS-MLR model are used to quantitatively analyze the contribution rate of pollution sources, and the migration key parameters are determined by indoor experiment to establish the migration rate equation; then, the GIS migration simulation model is built by fusing the physical mechanism model and the random forest model and optimized by Bayesian to realize the dynamic prediction of the spatio-temporal distribution of PAHs; finally, the joint risk index is built by integrating the pollution degree, toxicity intensity and health risk, and the risk level is divided and early warning.

[0060] With reference to Figure 1 The embodiments of the application at least include steps S10 to S50.

[0061] S10, real-time collection of PAHs concentration data, environmental factor data and remote sensing image data, and spatio-temporal matching and outlier elimination of multi-source data by using Bayesian data fusion algorithm, output of standardized data set.

[0062] S20, based on the standardized data set, analysis of PAHs sources and their contribution rates by the characteristic ratio method and APCS-MLR model, determination of adsorption and diffusion coefficients by indoor experiment, establishment of migration rate equation, and calculation of key parameters.

[0063] S30, construction of spatial database in ArcGIS, fusion of physical mechanism and random forest model based on standardized data set and key parameters, realization of PAHs migration simulation by Bayesian optimization, and construction of GIS model.

[0064] S40, input of real-time monitoring data into the GIS model, prediction of the spatio-temporal distribution of PAHs in the future preset time period, output of migration probability map and three-dimensional vertical migration characteristics.

[0065] S50, based on the migration probability map and three-dimensional vertical migration characteristics, integration of Nemerow index, toxicity equivalent and lifetime cancer risk to build a joint risk index, and division of risk level.

[0066] Specifically, the concentration of PAHs, environmental factors and remote sensing data are collected by Internet of Things sensors, and a standardized data set is generated by Bayesian fusion to solve the time and space lag problem of traditional monitoring. Secondly, the characteristic ratio method and APCS-MLR model are combined to analyze the contribution rate of pollution sources, and the adsorption / diffusion coefficient is determined by indoor experiment to establish the migration rate equation and quantify the key parameters. Then, the physical mechanism and random forest model are fused in ArcGIS, and the Bayesian optimization is used to build a high-precision GIS migration simulation model. Based on the model, the future spatio-temporal distribution of PAHs is predicted, and the migration probability map and three-dimensional vertical characteristics are output. Finally, the joint risk index is constructed by integrating the Nemerow index, toxicity equivalent and carcinogenic risk, and the three-level risk is divided to realize the linkage warning.

[0067] In some embodiments, step S10 specifically comprises the following steps: collecting 21 PAHs concentration data and 17 environmental factor data according to PAHs special fluorescence spectrum sensor, soil physical and chemical sensor and meteorological sensor, and transmitting sensor data through LoRa wireless communication, receiving remote sensing image data and pollution source emission list to form an original data set; time synchronization and spatial registration are performed on the original data set, and the spatial unit attribute table of 100 m x 100 m grid is generated in ArcGIS; 3σ criterion is used to remove outliers and retain effective sample size, and the time and space aligned data set is output; the Bayesian fusion model is constructed with the time and space aligned data set as input: high weight is given to sensor data, and medium weight is given to remote sensing inversion data; the posterior probability distribution is calculated by Markov chain Monte Carlo algorithm, and the fused standardized data set is output, which contains 21 PAHs concentration, environmental factor value and data reliability of each grid unit.

[0068] Specifically, first, 21 PAHs concentrations and 17 environmental factors are collected by PAHs fluorescence spectrum, soil physical and chemical, and meteorological sensors, and combined with LoRa transmission, remote sensing image and pollution source list to form an original data; after time synchronization and spatial registration, a 100 m grid attribute table is generated, and 3σ criterion is used to remove outliers to obtain time and space aligned data; finally, the Bayesian fusion model (sensor data high weight, remote sensing data medium weight) is used, the posterior probability is calculated by Markov chain Monte Carlo algorithm, and the standardized data set containing 21 PAHs concentration, environmental factor value and reliability of each grid is output, solving the problem of inconsistent time and space of multi-source data and uneven quality, providing high-precision input for subsequent modeling.

[0069] In some embodiments, step S20 specifically comprises the following steps: calculating the characteristic ratios according to the concentrations of 21 PAHs monomers in the standardized data set, preliminarily identifying the main pollution source type and spatial distribution characteristics; using the APCS-MLR model, performing KMO test and Bartlett spherical test on the PAHs concentration data, extracting the principal components with a characteristic ratio greater than 1, and converting the principal component loading matrix into absolute factor scores to represent the contribution intensity of different pollution sources; establishing a regression equation with the total concentration of PAHs as the dependent variable and APCS as the independent variable, and calculating the pollution source contribution rate corresponding to each principal component; based on the standardized data set and the pollution source contribution rate, using multiple linear regression, taking the PAHs migration rate as the dependent variable and the environmental factors as the independent variables, combining the adsorption capacity and the diffusion coefficient to construct the PAHs migration rate equation, and calculating the degradation coefficient and the migration rate; integrating the pollution source contribution rate, the adsorption capacity, the diffusion coefficient, the migration rate and the degradation coefficient to generate key parameters that meet the GIS modeling requirements.

[0070] The formula of the PAHs migration rate equation is:

[0071] v = 0.12 x wind speed + 0.08 x organic matter content - 0.05 x slope + 0.03 x ln(Koc) - 0.02 x D;

[0072] wherein, v: PAHs migration rate, which comprehensively reflects the superposition effect of horizontal and vertical migration;

[0073] Wind speed: environmental factor, representing the driving effect of atmospheric deposition and surface runoff on the diffusion of PAHs;

[0074] Organic matter content: soil physicochemical parameter, reflecting the adsorption capacity of soil to PAHs;

[0075] Slope: topographic factor, affecting the speed of surface runoff;

[0076] Koc: organic carbon normalized adsorption coefficient, representing the binding strength of PAHs and soil organic matter;

[0077] DD: diffusion coefficient, reflecting the molecular diffusion ability of PAHs in soil pores;

[0078] Coefficients (0.12, 0.08, etc.): empirical parameters fitted by multiple linear regression, quantifying the contribution weight of each factor to the migration rate.

[0079] Specifically, the characteristic ratios were calculated using the concentrations of 21 PAHs monomers in the standardized data to preliminarily identify the types and spatial distribution of pollution sources; then, the principal components were extracted after KMO and Bartlett tests through the APCS-MLR model, and were converted into absolute factor scores and combined with the regression equation to quantitatively calculate the contribution rates of each pollution source; based on the contribution rates and environmental factors, the migration rate equation was constructed by using multivariate linear regression combined with the adsorption capacity (Koc) and diffusion capacity (D) to obtain the degradation coefficient (k) and migration rate (v); finally, the structured key parameter set was generated by integrating the pollution source contribution rate, Koc, D, v, and k, which provided both mechanism and data support for subsequent GIS migration simulation.

[0080] In some embodiments, step S30 specifically comprises the following steps: constructing a spatial database in ArcGIS, inputting the standardized data set and key parameters, integrating terrain and land use elements, and generating a spatial database of 100 m x 100 m grid; constructing a physical mechanism model, simulating the vertical and horizontal migration of PAHs in soil based on the migration rate, diffusion coefficient, and degradation coefficient in the key parameters, and based on the water power-adsorption desorption coupling equation, and outputting the spatial and temporal distribution of the physical simulation of PAHs concentration; constructing a random forest model, inputting the environmental factors and PAHs concentration in the standardized data set, training the data-driven prediction, and outputting the data-driven prediction results of PAHs concentration; fusing the physical mechanism model and the random forest model through the Bayesian optimization algorithm to obtain the coupled model, and constraining the data-driven prediction by the physical simulation results, and outputting the fused PAHs migration simulation results; inputting the PAHs migration simulation results into ArcGIS, spatializing the results after cross-validation, and generating a GIS model containing the PAHs migration path and concentration value line.

[0081] Further, step S30 further comprises the following steps: taking the physical simulation results output by the physical mechanism model, the data-driven prediction results output by the random forest model and the verification set data as inputs, taking the physical simulation results and the data-driven prediction results one by one according to the grid units, assigning initial weights to the physical simulation results and the data-driven prediction results, and constructing a fusion input matrix with the initial weights; taking the fusion input matrix and the verification set data as inputs, defining a target function as a determination coefficient of the measured PAHs concentration of the verification set and the fusion predicted concentration, and setting a constraint condition as the determination coefficient being greater than 0.85, to form an optimization framework containing the target function and the constraint condition; based on the optimization framework and the fusion input matrix, a Bayesian optimization algorithm is adopted to iteratively optimize the weights through a Gaussian process surrogate model, to initialize the calculation of the determination coefficient, and if the determination coefficient is less than 0.85, 10 groups of weight combinations are randomly sampled to calculate the determination coefficient, and the optimal sampling point is selected by using an expected improvement function to adjust the weights, until the determination coefficient is not less than 0.85 in five consecutive iterations, and the optimal weights are output; based on the optimal weights, the physical simulation results and the data-driven prediction results, the PAHs concentration of each grid unit is calculated according to the fusion concentration = the initial weight of the physical simulation result x the physical simulation concentration + the initial weight of the data-driven prediction result x the data-driven prediction concentration, and the migration path and the rate parameter in the physical simulation results are synchronously integrated, to output the fused PAHs migration simulation results.

[0082] Specifically, first, the standardized data set, the key parameters and the terrain and land use elements are integrated in ArcGIS to generate a 100m grid spatial database, providing a basic framework for modeling; then, a physical mechanism model is constructed, based on the migration rate, the diffusion coefficient and the degradation coefficient, to simulate the vertical (soil concentration gradient) and horizontal (diffusion direction / rate) migration of PAHs through the water power-adsorption desorption coupling equation, and to output the physical simulation concentration distribution; at the same time, a random forest model is constructed to train the data-driven prediction based on the environmental factors and the PAHs concentration, and to output the data-driven results. The core is to fuse the two models through Bayesian optimization: taking the physical simulation results, the data-driven results and the verification set data as inputs, constructing a fusion matrix with initial weights, defining the determination coefficient R 2 ≥0.85 as the target, and iteratively optimizing the weights (randomly sampling 10 groups of weights and selecting the optimal by using the expected improvement function) through the Gaussian process surrogate model, until R 2 meets the standard for five consecutive times, and the optimal weights are output; the fusion concentration is calculated according to the optimal weights, the migration path and the rate parameter are integrated, and the PAHs migration simulation results are obtained; finally, the results are input into ArcGIS, spatially expressed after cross-validation, to generate a GIS model containing the migration path and the concentration contour, and to realize high-precision migration simulation.

[0083] In some embodiments, step S40 specifically includes the following steps: inputting real-time monitoring data into a GIS model, setting a future preset time period and time granularity based on user needs, loading key parameters into the GIS model, and determining the prediction spatial range as all grid cells; simulating the horizontal and vertical migration of PAHs by calling the migration rate equation in the coupled model and real-time meteorological data through the GIS model, and outputting the spatiotemporal distribution data of PAHs in 100 m grid cells within the future preset time period; calculating the probability that the concentration of each grid cell exceeds the risk threshold based on the predicted spatiotemporal distribution data of PAHs concentration, generating a color gradient probability map through ArcGIS spatial analysis tools, and marking high-risk cluster areas; extracting soil PAHs concentration data from the prediction data, calculating the proportion of PAHs migration depth in each soil layer, and generating a vertical profile feature map.

[0084] The process involves generating a color gradient probability map using ArcGIS spatial analysis tools to mark high-risk cluster areas. This "color gradient probability map" is essentially a concrete implementation of the "migration probability map." Furthermore, it involves extracting PAHs concentration data from the predicted data, calculating the proportion of PAHs migration depth in each soil layer, and generating a vertical profile feature map. This "vertical profile feature map" is a visualization of the "three-dimensional vertical migration features."

[0085] Specifically, real-time monitoring data is input into the GIS model. The future prediction time period (e.g., 7-90 days) and time granularity (daily / weekly average) are set according to user needs. Key parameters (migration rate, diffusion coefficient, etc.) are loaded, and all 100m grid cells are determined as the prediction spatial range. The GIS model calls the migration rate equation in the coupled model and real-time meteorological data (wind speed, rainfall) to simulate the horizontal migration (diffusion direction / rate) and vertical migration (soil concentration gradient) of PAHs, outputting the spatiotemporal distribution data of PAHs in each grid within the future time period. Based on the prediction data, the probability of concentration exceeding the risk threshold is calculated, and a color gradient migration probability map is generated using ArcGIS, marking high-risk cluster areas. Soil concentration data is extracted, the proportion of each soil layer and migration depth are calculated, and a vertical profile feature map is generated to intuitively display the vertical migration pattern of PAHs, providing a spatiotemporal dynamic basis for risk warning.

[0086] In some embodiments, step S50 specifically comprises the following steps: based on the PAHs concentration data in the migration probability map and the soil layer concentration distribution in the vertical migration characteristic map, the Nemerow index, the toxicity equivalent, and the lifetime cancer risk are calculated, and 100 m grid unit data of the three single risk indexes of the Nemerow index, the toxicity equivalent, and the lifetime cancer risk are output; based on the risk contribution degree analysis, the Nemerow index weight, the toxicity equivalent weight, and the lifetime cancer risk weight are given, and the sum of the weights is ensured to be 1, and a weight allocation scheme is output; based on the single risk index and the weight, the JRI value of each grid unit is calculated, wherein the Nemerow index, the toxicity equivalent, and the lifetime cancer risk are all normalized, and the JRI spatial distribution data of the 100 m grid unit is output; based on the JRI value frequency distribution, the natural breakpoint method is used to divide the risk into three levels, and the high-risk area is labeled through ArcGIS, and a risk level spatial distribution map and a grading standard are output.

[0087] Specifically, based on the PAHs concentration in the migration probability map and the soil layer distribution in the vertical migration characteristic map, the Nemerow index (pollution degree), the toxicity equivalent (BaPeq, toxicity intensity), and the lifetime cancer risk (ILCRs, health impact) are calculated, and 100 m grid single risk data are output; through risk contribution degree analysis, the three index weights (such as Nemerow 0.4, toxicity equivalent 0.3, and cancer risk 0.3) are given, and the sum of the weights is ensured to be 1; after the single index is normalized, the joint risk index (JRI) of each grid is calculated according to the weight, and the JRI spatial distribution is output; based on the JRI frequency distribution, the natural breakpoint method is used to divide the risk into low (<0.3), medium (0.3-1.0), and high (>1.0) three levels, the high-risk area is labeled through ArcGIS, and a risk level map and a standard are output, realizing migration-risk linkage evaluation, and providing a decision basis for precise management and control.

[0088] The implementation principle of the IoT-GIS spatial modeling and analysis method for soil polycyclic aromatic hydrocarbon (PAH) migration flux in this application embodiment is as follows: First, it utilizes IoT sensors and BeiDou / LoRa dual-mode transmission, combined with a Bayesian fusion algorithm to process multi-source data, generating a high-resolution (100m grid) standardized dataset to solve the problem of spatiotemporal lag in traditional monitoring. Second, it analyzes the sources and contribution rates of PAHs using the eigenvalue ratio method and the APCS-MLR model, and establishes a migration rate equation by combining indoor experimental measurements of adsorption / diffusion coefficients, quantifying key parameters and overcoming the limitations of traditional qualitative and coarse source apportionment. Finally, it integrates physical mechanisms... The model, along with the random forest model, is used to construct a GIS migration simulation model through Bayesian optimization. This model achieves high-precision simulation of the horizontal and vertical migration of PAHs, overcoming the shortcomings of single models that lack physical processes or data-driven constraints. Finally, based on the model, the spatiotemporal distribution of PAHs is predicted over a certain period of time in the future, and the migration probability map and three-dimensional vertical features are output. Furthermore, the Nemerow index, toxicity equivalent, and carcinogenic risk are integrated to construct a joint risk index (JRI), which is divided into three levels of risk. This facilitates migration simulation and risk warning, enabling accurate prediction of pollution sources, migration trends, or health risks, thus facilitating pollution control.

[0089] Figure 1 This is a flowchart illustrating an IoT-GIS spatial modeling and analysis method for soil polycyclic aromatic hydrocarbon migration flux in one embodiment. It should be understood that, although... Figure 1 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows; unless explicitly stated otherwise, there is no strict order requirement for the execution of these steps, and they can be executed in other orders; and Figure 1 At least some of the steps in the process may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

[0090] Based on the same technical concept, referring to Figure 2 This application also provides a soil polycyclic aromatic hydrocarbon migration flux IoT-GIS spatial modeling and analysis device, which adopts the following technical solution. The device includes:

[0091] The data acquisition module is used to collect PAHs concentration data, environmental factor data and remote sensing image data in real time, and uses Bayesian data fusion algorithm to perform spatiotemporal matching and outlier removal on multi-source data, and output a standardized dataset.

[0092] An indoor detection module is configured to analyze sources and contribution rates of PAHs based on a standardized data set by a characteristic ratio method and an APCS-MLR model, and to calculate key parameters by establishing a migration rate equation based on adsorption and diffusion coefficients determined by indoor experiments;

[0093] A model coupling module is configured to construct a spatial database in ArcGIS, to fuse physical mechanisms and a random forest model based on the standardized data set and the key parameters, to realize PAHs migration simulation by Bayesian optimization, and to construct a GIS model;

[0094] A distribution calculation module is configured to input real-time monitoring data into the GIS model, to predict the spatiotemporal distribution of PAHs in a future preset time period, and to output a migration probability map and three-dimensional vertical migration characteristics;

[0095] A risk level module is configured to integrate a Nemerow index, a toxicity equivalent, and a lifetime cancer risk based on the migration probability map and the three-dimensional vertical migration characteristics, to construct a joint risk index, and to divide a risk level.

[0096] In some embodiments, the data collection module is specifically configured to collect 21 PAHs concentration data and 17 environmental factor data according to a PAHs special fluorescence spectrum sensor, a soil physicochemical sensor, and a meteorological sensor, to transmit sensor data by LoRa wireless communication, to receive remote sensing image data and a pollution source emission list, and to form an original data set;

[0097] The original data set is subjected to time synchronization and spatial registration, and is input into ArcGIS to generate a spatial unit attribute table of a 100 m x 100 m grid;

[0098] Anomalies are removed by a 3σ criterion, and effective sample sizes are retained, and a spatiotemporally aligned data set is output;

[0099] The spatiotemporally aligned data set is taken as input to construct a Bayesian fusion model: high weight is given to sensor data, and medium weight is given to remote sensing inversion data, posterior probability distribution is iteratively calculated by a Markov chain Monte Carlo algorithm, a fused standardized data set is output, and the standardized data set includes 21 PAHs concentrations, environmental factor values, and data reliability of each grid unit.

[0100] In some embodiments, the indoor detection module is specifically configured to calculate characteristic ratios according to 21 PAHs monomer concentrations in the standardized data set, and to preliminarily identify main pollution source types and spatial distribution characteristics;

[0101] The APCS-MLR model is used to perform KMO and Bartlett spherical tests on the PAHs concentration data, to extract principal components with a characteristic ratio greater than 1, and to convert a principal component loading matrix into absolute factor scores to represent contribution intensities of different pollution sources;

[0102] A regression equation is established with the total concentration of PAHs as the dependent variable and APCS as the independent variable to calculate the contribution rate of each principal component to the pollution source;

[0103] Based on the standardized data set and the contribution rate of the pollution source, a multivariate linear regression is used to establish a PAHs migration rate equation with the PAHs migration rate as the dependent variable and the environmental factors as the independent variable, and the characteristic adsorption capacity and the characteristic diffusion coefficient are combined to calculate the degradation coefficient and the migration rate;

[0104] The contribution rate of the pollution source, the characteristic adsorption capacity, the characteristic diffusion coefficient, the migration rate and the degradation coefficient are integrated to generate key parameters that meet the GIS modeling requirements.

[0105] In some embodiments, the model coupling module is specifically used to build a spatial database in ArcGIS, input the standardized data set and the key parameters, integrate the terrain and land use elements, and generate a spatial database of 100 m x 100 m grid;

[0106] A physical mechanism model is constructed, and based on the migration rate, the characteristic diffusion coefficient and the degradation coefficient in the key parameters, the vertical and horizontal migration of PAHs in soil is simulated based on the water power-adsorption desorption coupling equation, and the spatial and temporal distribution of the physical simulation of PAHs concentration is output;

[0107] A random forest model is constructed, and the environmental factors and PAHs concentration in the standardized data set are input to train the data-driven prediction, and the data-driven PAHs concentration prediction result is output;

[0108] The physical mechanism model and the random forest model are fused through the Bayesian optimization algorithm to obtain a coupled model, and the data-driven prediction is constrained by the physical simulation result, and the fused PAHs migration simulation result is output;

[0109] The PAHs migration simulation result is input into ArcGIS, and after cross-validation, it is spatially expressed to generate a GIS model containing PAHs migration path and concentration contour line.

[0110] In some embodiments, the model coupling module is also used to input the physical simulation result output by the physical mechanism model, the data-driven prediction result output by the random forest model and the validation set data, to correspond the physical simulation result and the data-driven prediction result one by one according to the grid unit, to assign the physical simulation result and the data-driven prediction result initial weights, and to construct a fusion input matrix with initial weights;

[0111] With the fusion input matrix and the validation set data as input, the target function is defined as the determination coefficient of the measured PAHs concentration of the validation set and the fusion prediction concentration, and the constraint condition is set as the determination coefficient being greater than 0.85, forming an optimization framework containing the target function and the constraint condition;

[0112] Based on the optimization framework and the fused input matrix, a Bayesian optimization algorithm is used to iteratively optimize the weights through a Gaussian process surrogate model, initialize the calculation of the determination coefficient, if the determination coefficient is less than 0.85, randomly sample 10 groups of weight combinations to calculate the determination coefficient, use the expected improvement function to select the optimal sampling point to adjust the weight, until the determination coefficient is not less than 0.85 for 5 consecutive iterations, and output the optimal weight;

[0113] Based on the optimal weight, the physical simulation result and the data-driven prediction result, the PAHs concentration of each grid unit is calculated according to the fusion concentration = physical simulation result initial weight x physical simulation concentration + data-driven prediction result initial weight x data-driven prediction concentration, and the migration path and rate parameter in the physical simulation result are integrated synchronously, and the fused PAHs migration simulation result is output.

[0114] In some embodiments, the distribution calculation module is specifically configured to input real-time monitoring data into a GIS model, set a future preset time period and a time granularity based on user demand, load key parameters in the GIS model, and determine a prediction space range as all grid units;

[0115] The migration rate equation in the coupling model and real-time meteorological data are called through the GIS model to simulate the horizontal and vertical migration of PAHs, and the spatio-temporal distribution data of PAHs in the 100 m grid unit in the future preset time period is output;

[0116] Based on the predicted spatio-temporal distribution data of PAHs concentration, the probability of the concentration of each grid unit exceeding the risk threshold is calculated, a color gradient probability map is generated through an ArcGIS spatial analysis tool, and a high-risk aggregation area is marked;

[0117] The soil layer PAHs concentration data is extracted from the prediction data, the migration depth of PAHs in each soil layer is calculated based on the proportion of the soil layer, and a vertical profile feature map is generated.

[0118] In some embodiments, the risk level module is specifically configured to calculate the Nemerow index, the toxic equivalent and the lifetime cancer risk based on the PAHs concentration data in the migration probability map and the soil layer concentration distribution in the vertical migration feature map, and output the 100 m grid unit data of the three single risk indexes of the Nemerow index, the toxic equivalent and the lifetime cancer risk;

[0119] Based on the risk contribution analysis, the Nemerow index weight, the toxic equivalent weight and the lifetime cancer risk weight are assigned, and the sum of the weights is ensured to be 1, and a weight allocation scheme is output;

[0120] Based on the single risk index and the weight, the JRI value of each grid unit is calculated, wherein the Nemerow index, the toxic equivalent and the lifetime cancer risk are normalized, and the JRI spatial distribution data of the 100 m grid unit is output;

[0121] Based on the frequency distribution of JRI values, the risk is divided into three levels by using the natural breakpoint method, and the high-risk area is labeled by ArcGIS, and the spatial distribution map of risk level and grading standard are output.

[0122] The embodiment of the application further discloses a control device.

[0123] Specifically, the control device comprises a memory and a processor, and the memory stores a computer program capable of being loaded and executed by the processor to perform the soil polycyclic aromatic hydrocarbon migration flux Internet of Things-GIS spatial modeling analysis method.

[0124] The embodiment of the application further discloses a computer readable storage medium.

[0125] Specifically, the computer readable storage medium stores a computer program capable of being loaded and executed by the processor to perform the soil polycyclic aromatic hydrocarbon migration flux Internet of Things-GIS spatial modeling analysis method, and the computer readable storage medium comprises various storage program codes, for example, a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0126] The above are preferred embodiments of the application, and are not intended to limit the protection scope of the application, therefore: equivalent changes made according to the structure, shape, principle of the application should be covered within the protection scope of the application.

Claims

1. A soil polycyclic aromatic hydrocarbon migration flux Internet of Things-GIS spatial modeling analysis method, characterized in that, The method comprises the following steps: Real-time acquisition of PAHs concentration data, environmental factor data and remote sensing image data, and use of Bayesian data fusion algorithm for spatio-temporal matching and abnormal value elimination of multi-source data, to output a standardized data set; Based on the standardized data set, the sources and contribution rates of PAHs are analyzed by the characteristic ratio method and the APCS-MLR model, and the key parameters are calculated by establishing the migration rate equation combined with the determination of the adsorption and diffusion coefficients in the laboratory; In ArcGIS, a spatial database is constructed, and based on the standardized data set and the key parameters, the physical mechanism and the random forest model are fused, and the PAHs migration simulation is realized through Bayesian optimization to construct a GIS model; Real-time monitoring data are input into the GIS model to predict the spatio-temporal distribution of PAHs in a future preset time period, and a migration probability map and a three-dimensional vertical migration characteristic are output; Based on the migration probability map and the three-dimensional vertical migration characteristic, the joint risk index is constructed by integrating the Nemerow index, the toxic equivalent and the lifetime carcinogenic risk, and the risk level is divided.

2. The soil polycyclic aromatic hydrocarbon migration flux Internet of Things-GIS spatial modeling analysis method according to claim 1, characterized in that, The real-time acquisition of PAHs concentration data, environmental factor data and remote sensing image data, and the use of Bayesian data fusion algorithm for spatio-temporal matching and abnormal value elimination of multi-source data, to output a standardized data set, comprises: According to the PAHs special fluorescence spectrum sensor, the soil physicochemical sensor and the meteorological sensor, 21 kinds of PAHs concentration data and 17 kinds of environmental factor data are collected, and the sensor data are transmitted through LoRa wireless communication, and the remote sensing image data and the pollution source emission list are received to form an original data set; The original data set is time-synchronized and space-registered, and the spatial unit attribute table of 100 m x 100 m grid is generated in ArcGIS; Abnormal values are eliminated by using 3σ criterion, and effective sample size is retained, to output a spatio-temporally aligned data set; The Bayesian fusion model is constructed with the spatio-temporally aligned data set as input: high weight is given to the sensor data, and medium weight is given to the remote sensing inversion data, the posterior probability distribution is iteratively calculated by Markov chain Monte Carlo algorithm, and the standardized data set after fusion is output, which contains 21 kinds of PAHs concentration, environmental factor value and data reliability of each grid unit.

3. The soil polycyclic aromatic hydrocarbon migration flux Internet of Things-GIS spatial modeling analysis method according to claim 1, characterized in that, The standardized data set is used to analyze the sources and contribution rates of PAHs by the characteristic ratio method and the APCS-MLR model, and the key parameters are calculated by establishing the migration rate equation combined with the determination of the adsorption and diffusion coefficients in the laboratory, which comprises: According to the 21 kinds of PAHs monomer concentration in the standardized data set, the characteristic ratio is calculated to preliminarily identify the main pollution source type and spatial distribution characteristic; The KMO test and Bartlett sphere test are performed on the PAHs concentration data by using the APCS-MLR model, the principal components with characteristic ratio greater than are extracted, and the principal component loading matrix is converted into absolute factor score to represent the contribution strength of different pollution sources; The regression equation is established with PAHs total concentration as dependent variable and APCS as independent variable to calculate the contribution rate of each principal component to the pollution source; Based on the standardized dataset and the pollution source contribution rate, a multivariate linear regression is adopted to take the PAHs migration rate as the dependent variable and the environmental factors as the independent variable, and a PAHs migration rate equation is constructed by combining the representation of adsorption capacity and the representation of diffusion coefficient to calculate the degradation coefficient and the migration rate; The pollution source contribution rate, the representation of adsorption capacity, the representation of diffusion coefficient, the migration rate and the degradation coefficient are integrated to generate the key parameters meeting the GIS modeling.

4. The soil polycyclic aromatic hydrocarbon migration flux Internet of Things-GIS spatial modeling analysis method according to claim 3, characterized in that, The spatial database is constructed in ArcGIS, the standardized dataset and the key parameters are input, the topography and land use elements are integrated, and the spatial database of 100 m*100 m grid is generated; The physical mechanism model is constructed, the migration rate, the representation of diffusion coefficient and the degradation coefficient in the key parameters are used to simulate the vertical and horizontal migration of PAHs in soil based on the water power-adsorption desorption coupling equation, the spatial and temporal distribution of the physical simulation of PAHs concentration is output, and the physical simulation result is obtained; The random forest model is constructed, the environmental factors and PAHs concentration in the standardized dataset are input, the data-driven prediction is trained, and the data-driven PAHs concentration prediction result is output; The physical mechanism model and the random forest model are fused by the Bayesian optimization algorithm to obtain a coupled model, the data-driven prediction is constrained by the physical simulation result, and the fused PAHs migration simulation result is output; The PAHs migration simulation result is input into ArcGIS, and the spatial expression is performed after cross-validation to generate the GIS model containing the PAHs migration path and the concentration value line. The physical mechanism model and the random forest model are fused by the Bayesian optimization algorithm to obtain a coupled model, the data-driven prediction is constrained by the physical simulation result, and the fused PAHs migration simulation result is output, including:

5. The soil PAHs migration fluxes IoT-GIS spatial modeling analysis method according to claim 4, characterized in that, The physical simulation result output by the physical mechanism model, the data-driven prediction result output by the random forest model and the validation set data are input, the physical simulation result and the data-driven prediction result are one-to-one corresponding according to the grid unit, the physical simulation result and the data-driven prediction result are assigned with initial weights, and a fusion input matrix with initial weights is constructed; The fusion input matrix and the validation set data are input, the target function is defined as the determination coefficient of the measured PAHs concentration of the validation set and the fusion prediction concentration, and the constraint condition is set as the determination coefficient being greater than 0.85 to form an optimization framework containing the target function and the constraint condition; ​ Based on the optimization framework and the fusion input matrix, a Bayesian optimization algorithm is adopted, a Gaussian process surrogate model is used to iteratively optimize the weight, the determination coefficient is calculated, if the determination coefficient is less than 0.85, 10 groups of weight combinations are randomly sampled to calculate the determination coefficient, the optimal sampling point is selected by using the expected improvement function to adjust the weight, and the determination coefficient is not less than 0.85 in five consecutive iterations, and the optimal weight is output; Based on the optimal weight, the physical simulation result and the data-driven prediction result, the PAHs concentration of each grid unit is calculated according to the fusion concentration = physical simulation result initial weight x physical simulation concentration + data-driven prediction result initial weight x data-driven prediction concentration, the migration path and rate parameter in the physical simulation result are synchronously integrated, and the fused PAHs migration simulation result is output.

6. The soil PAHs migration fluxes IoT-GIS spatial modeling analysis method according to claim 5, characterized in that, The real-time monitoring data is input into the GIS model, the PAHs spatiotemporal distribution in a future preset time period is predicted, and a migration probability graph and a three-dimensional vertical migration feature are output, including: The real-time monitoring data is input into the GIS model, the future preset time period and the time granularity are set based on user demand, the key parameters are loaded in the GIS model, and the prediction space range is determined as all grid units; The migration rate equation in the coupling model and real-time meteorological data are called by the GIS model, the horizontal migration and vertical migration of PAHs are simulated, and the PAHs spatiotemporal distribution data of 100 m grid units in the future preset time period are output; Based on the predicted PAHs concentration spatiotemporal distribution data, the probability that the concentration of each grid unit exceeds the risk threshold is calculated, a color gradient probability graph is generated by using an ArcGIS spatial analysis tool, and a high-risk aggregation area is marked; The soil layer PAHs concentration data are extracted from the prediction data, the migration depth of PAHs in each soil layer is calculated based on the proportion of the soil layer, and a vertical profile feature graph is generated.

7. The soil PAHs migration fluxes IoT-GIS spatial modeling analysis method according to claim 6, characterized in that, Based on the migration probability graph and the three-dimensional vertical migration feature, the Nemerow index, the toxicity equivalent and the lifetime carcinogenic risk are integrated, the joint risk index is constructed, and the risk level is divided, including: Based on the PAHs concentration data in the migration probability graph and the soil layer concentration distribution in the vertical profile feature graph, the Nemerow index, the toxicity equivalent and the lifetime carcinogenic risk are calculated, and the 100 m grid unit data of the three single risk indexes, i.e., the Nemerow index, the toxicity equivalent and the lifetime carcinogenic risk, are output; Based on the risk contribution degree analysis, the Nemerow index weight, the toxicity equivalent weight and the lifetime carcinogenic risk weight are given, the sum of the weights is ensured to be 1, and the weight allocation scheme is output; Based on the single risk index and the weight, the JRI value of each grid unit is calculated, wherein the Nemerow index, the toxicity equivalent and the lifetime carcinogenic risk are normalized, and the JRI spatiotemporal distribution data of 100 m grid units are output; Based on the JRI value frequency distribution, the risk is divided into three levels by using the natural breakpoint method, the high-risk area is marked by using ArcGIS, and the risk level spatiotemporal distribution graph and the grading standard are output.

8. A soil polycyclic aromatic hydrocarbon migration flux Internet of Things-GIS spatial modeling analysis device, characterized in that, The device comprises: The data acquisition module is configured to collect PAHs concentration data, environmental factor data and remote sensing image data in real time, and perform spatio-temporal matching and abnormal value elimination on the multi-source data by using a Bayesian data fusion algorithm, and output a standardized data set; The indoor detection module is configured to analyze PAHs sources and contribution rates by using a characteristic ratio method and an APCS-MLR model based on the standardized data set, combine with adsorption and diffusion coefficients determined by indoor experiments, establish a migration rate equation, and calculate key parameters; The model coupling module is configured to construct a spatial database in ArcGIS, fuse a physical mechanism and a random forest model based on the standardized data set and the key parameters, realize PAHs migration simulation by using Bayesian optimization, and construct a GIS model; The distribution calculation module is configured to input real-time monitoring data into the GIS model, predict PAHs spatio-temporal distribution in a future preset time period, and output a migration probability map and three-dimensional vertical migration characteristics. The risk grade module is configured to integrate a Nemerow index, a toxicity equivalent and a lifetime carcinogenic risk based on the migration probability map and the three-dimensional vertical migration characteristics, construct a joint risk index, and divide a risk grade.

9. A control device, characterized by The device comprises: a memory and a processor, the memory storing a computer program capable of being loaded and executed by the processor to perform the method of any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, a memory storing a computer program capable of being loaded and executed by the processor to perform the method of any one of claims 1 to 7.

Images