A dynamic assessment method for groundwater pollution risk in chemical parks integrating multi-source heterogeneous data

Abstract

The present invention relates to the field of data processing technology, and in particular to a method for dynamically assessing groundwater pollution risks in chemical parks by integrating multi-source heterogeneous data. The method comprises the following steps: obtaining multi-source heterogeneous data from chemical parks and pre-processing the data to obtain pre-processed multi-source heterogeneous data; performing preliminary feature extraction on the pre-processed multi-source heterogeneous data and performing fusion processing to obtain fused data; constructing a bidirectionally coupled groundwater pollution migration simulation model based on the fused data, calculating groundwater flow rate and pollutant concentration, and constructing a three-dimensional concentration distribution of pollutants; constructing a weighted heterogeneous graph based on the three-dimensional concentration distribution of pollutants, identifying pollution propagation paths, calculating path potential scores, and calculating pollution risk indexes to dynamically assess pollution risks. The method solves the technical problems of traditional groundwater pollution risk assessment methods in that they are not accurate enough in processing multi-source heterogeneous data from chemical parks, their analysis is not comprehensive enough, and their adaptability is poor.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method for dynamically assessing groundwater pollution risks in a chemical park by integrating multi-source heterogeneous data. Background Art

[0002] The monitoring, assessment and control of groundwater pollution require a large amount of data support. Traditional pollution assessment methods mostly rely on point monitoring data, lack a comprehensive understanding of the dynamic migration of pollutants in groundwater bodies, and have great limitations. Especially in chemical parks with multiple pollution sources and a wide variety of pollutants, traditional pollution monitoring and assessment methods often cannot fully reflect the migration paths and diffusion trends of pollutants, resulting in inaccurate and delayed assessment results, which in turn affects the effectiveness of decision-making.

[0003] Currently, research on groundwater pollution assessment technology focuses primarily on source identification and concentration monitoring of single pollutants. However, due to the interactions between pollutants, single-pollutant models appear insufficient. In particular, in pollutant diffusion simulations, traditional models often overlook the feedback effect of pollutant concentration on groundwater flow velocity, resulting in inaccurate predictions of pollutant migration paths and concentration distributions. Furthermore, the migration of pollutants in groundwater is influenced not only by the hydraulic head gradient but also by multiple factors, including groundwater flow velocity, soil permeability, and the pollutant's own characteristics. Due to the dynamic changes in these factors in real-world environments, the static assumptions of traditional models often fail to accurately predict the migration of groundwater pollution.

[0004] In summary, traditional groundwater pollution risk assessment methods still have technical problems such as inaccurate processing of multi-source heterogeneous data in chemical parks, incomplete analysis, and poor adaptability. Summary of the Invention

[0005] The present invention provides a dynamic assessment method for groundwater pollution risk in chemical parks by integrating multi-source heterogeneous data, so as to solve the technical problems of traditional groundwater pollution risk assessment methods in processing multi-source heterogeneous data in chemical parks inaccurately, analyzing them incompletely, and having poor adaptability.

[0006] The present invention provides a method for dynamically assessing groundwater pollution risk in a chemical park by integrating multi-source heterogeneous data, specifically including the following technical solutions:

[0007] A method for dynamic assessment of groundwater pollution risk in a chemical park by integrating multi-source heterogeneous data includes the following steps:

[0008] S1. Obtain multi-source heterogeneous data from the chemical park and preprocess it to obtain preprocessed multi-source heterogeneous data; perform preliminary feature extraction on the preprocessed multi-source heterogeneous data to obtain preliminary feature data; fuse the preliminary feature data to obtain fused data; based on the fused data, construct a bidirectionally coupled groundwater pollution migration simulation model, calculate groundwater flow velocity and pollutant concentration, and construct a three-dimensional pollutant concentration distribution;

[0009] S2. Based on the three-dimensional concentration distribution of pollutants, a weighted heterogeneous graph is constructed and the pollution propagation path is identified. For the pollution propagation path, the path potential score and pollution risk index are calculated to dynamically assess the pollution risk.

[0010] Preferably, the S1 specifically includes:

[0011] Performing feature dimensionality reduction processing on the preliminary feature data to obtain feature data after dimensionality reduction; performing mapping processing on the feature data after dimensionality reduction to obtain mapped feature data; performing fusion processing on the mapped feature data and normalizing it to obtain fused data.

[0012] Preferably, the S1 specifically includes:

[0013] The bidirectionally coupled groundwater pollution migration simulation model calculates the head gradient based on the head data in the fused data, and introduces the feedback term of the pollutant concentration gradient in the fused data to the head gradient to obtain the groundwater flow velocity.

[0014] Preferably, the S1 specifically includes:

[0015] Based on the groundwater flow velocity and pollutant concentration, the convection term, diffusion term and degradation term are constructed, and combined with the periodic disturbance term to construct the pollutant migration equation; the pollutant migration equation is discretized by spatial central difference and time forward difference, and iterative calculation is performed to solve the pollutant migration equation, obtain the pollutant concentration, and construct the three-dimensional concentration distribution of pollutants.

[0016] Preferably, the S2 specifically includes:

[0017] Each spatial position of the three-dimensional concentration distribution of pollutants is taken as a node. Based on the spatial distance between nodes and the pollutant concentration of the nodes, and introducing the spatial attenuation factor, the edge weight is obtained, a weighted heterogeneous graph is constructed, and the pollution propagation path is identified.

[0018] Preferably, the S2 specifically includes:

[0019] The node state is propagated on the weighted heterogeneous graph to obtain the final node state; based on the final node state and combined with the edge weight, the path potential score of the pollution propagation path is calculated.

[0020] Preferably, the S2 specifically includes:

[0021] The pollution transmission path with the highest path potential score is selected as the main pollution path. Based on the pollutant concentration at the end node of the main pollution path, the exposure factor, time period disturbance factor and external intervention adjustment item are introduced to calculate the pollution risk index.

[0022] Preferably, the S2 specifically includes:

[0023] The exposure factor is obtained through a simulation fitting algorithm based on the frequency of human activities, groundwater consumption and resource sensitivity level in the area affected by the main pollution path.

[0024] Preferably, the S2 specifically includes:

[0025] Compare the pollution risk index with the preset risk threshold, divide the risk level intervals, and realize dynamic assessment of pollution risk.

[0026] The beneficial effects of the technical solution of the present invention are:

[0027] 1. By introducing a feedback term based on Darcy's law regarding the pollutant concentration gradient's effect on the hydraulic head gradient, a bidirectionally coupled expression for the reverse effect of the pollutant concentration gradient on groundwater flow velocity was constructed. This makes the pollutant concentration distribution not only a result of the flow field but also a regulatory factor in its evolution. This method dynamically responds to the reshaping of local hydraulic paths by concentration hotspots, realistically simulating the complex physical processes of "backflow" or "diversion" in high-concentration areas at pollution sources, and enhancing the credibility of simulations of pollution migration paths. This structural breakthrough significantly improves the authenticity and spatiotemporal continuity of pollution evolution predictions.

[0028] 2. A weighted heterogeneous graph modeling approach is introduced, abstracting underground space grid nodes into graph nodes. Multiple attributes, such as pollutant concentration, geological permeability, and spatial distance, are integrated, and node states are propagated using a graph neural network to calculate path potential scores. This process dynamically extracts high-risk primary paths within pollution transmission pathways, eliminating the reliance on artificially defined paths or fixed flow direction assumptions in pollution transmission path analysis, thereby improving the objectivity and dynamic adaptability of path identification.

[0029] 3. By incorporating exposure factors into the potential path score of the main pollution pathway, pollutant concentration, migration distance, and periodic disturbances, a risk assessment formula is generated that combines multiple nonlinear functions, including exponential, cosine, logarithmic, power, and square root functions, to calculate the pollution risk index. This risk assessment formula not only captures the direct threat posed by pollutant concentration to sensitive target nodes but also simulates the interactive effects of multiple factors, including human activity, resource importance, and management intervention. This unified measurement of exposure behavior and pollution effects overcomes the limitations of traditional risk models that rely on single-factor modeling. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 This is a flow chart of a method for dynamically assessing groundwater pollution risks in chemical parks by integrating multi-source heterogeneous data as described in the present invention. DETAILED DESCRIPTION

[0031] In order to further illustrate the technical means and effects adopted by the present invention to achieve the predetermined purpose of the invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0032] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0033] The following describes in detail a specific scheme of a method for dynamically assessing groundwater pollution risk in a chemical park by integrating multi-source heterogeneous data provided by the present invention with reference to the accompanying drawings.

[0034] Refer to the attached Figure 1 , which shows a flow chart of a method for dynamically assessing groundwater pollution risk in a chemical park by integrating multi-source heterogeneous data, provided by one embodiment of the present invention. The method includes the following steps:

[0035] S1. Obtain multi-source heterogeneous data from the chemical park and preprocess it to obtain preprocessed multi-source heterogeneous data; perform preliminary feature extraction on the preprocessed multi-source heterogeneous data to obtain preliminary feature data; fuse the preliminary feature data to obtain fused data; based on the fused data, construct a bidirectionally coupled groundwater pollution migration simulation model, calculate groundwater flow velocity and pollutant concentration, and construct a three-dimensional pollutant concentration distribution;

[0036] Data is collected from chemical parks through data collection equipment such as sensors, remote sensing satellites, and drones to obtain multi-source heterogeneous data of chemical parks. The multi-source heterogeneous data of chemical parks include sensor data, remote sensing data, water quality data, geological and hydrological data, and pollution source data. The sensor data include online sensor data such as pH value, conductivity, temperature, turbidity, etc. of groundwater monitoring points in the chemical park, which exist in the form of time series and have a high frequency (such as sampling once every minute or hour). The remote sensing data is image data collected by remote sensing satellites or drones, which contains information data such as surface temperature, soil moisture, vegetation index, etc., and exists in the form of grids. Each pixel represents the environmental characteristics of a certain geographical location; the water quality data includes pollutant concentration information, such as ammonia nitrogen, total phosphorus, etc., and the sampling frequency is low (such as once a month); the geological and hydrological data include information data such as groundwater level and soil permeability; the pollution source data includes wastewater discharge volume, pollutant type, emission description, pollution record, etc.; the multi-source heterogeneous data of the chemical park are preprocessed to obtain preprocessed multi-source heterogeneous data; the preprocessing process includes missing value filling, denoising, time synchronization, data space mapping and rasterization, as well as standardization and normalization. The methods used are technical means well known to those skilled in the art and will not be described in detail here;

[0037] Furthermore, preliminary feature extraction is performed on the pre-processed multi-source heterogeneous data through the existing modal encoder to obtain preliminary feature data; specifically, for data of different modalities, preliminary feature extraction is performed using a dedicated modal encoder: such as using a convolutional neural network to process remote sensing data, using a long short-term memory network to encode text information (such as emission descriptions in pollution source data), using a fully connected network to process time series data (such as sensor data, water quality data), etc.; the preliminary feature data is subjected to feature dimensionality reduction processing using a method such as principal component analysis to obtain reduced-dimensional feature data; the reduced-dimensional feature data is mapped through linear transformation and nonlinear activation function to obtain mapped feature data; the mapped feature data is fused through existing fusion technology (such as tensor decomposition fusion technology) and normalized to eliminate the dimension to obtain fused feature data, i.e., fused data; the methods adopted in the above-mentioned preliminary feature extraction and fusion process are technical means well known to those skilled in the art and will not be elaborated here;

[0038] Furthermore, based on the fused data, a bidirectionally coupled groundwater pollution migration simulation model was constructed to obtain groundwater flow velocity and pollutant concentration. Unlike traditional groundwater pollution migration models, which usually only use head gradient to drive flow while ignoring the feedback effect of pollutant concentration gradient itself on the flow path, the bidirectionally coupled groundwater pollution migration simulation model combines Darcy's law with the pollutant migration model, introduces the feedback term of pollutant concentration gradient on head gradient, and constructs the groundwater flow velocity calculation formula:

[0039] ,

[0040] in, is the groundwater velocity, which is a vector, indicating the time and spatial location The speed and direction of groundwater flow, the spatial position includes the horizontal axis direction, the vertical axis direction, and the vertical direction. ; The permeability coefficient is the permeability of the rock and soil medium to the flow of groundwater, reflecting the difficulty of groundwater flow in the soil or rock. It varies with the type of stratum (such as sand, clay, etc.). For example, the reference value range of the permeability coefficient of sand is to ; It is at the moment and spatial location The hydraulic head gradient at represents the spatial gradient of the groundwater level change, which is calculated by the hydraulic head data in the fused data. The calculation method is a technical means well known to those skilled in the art and will not be described in detail here. It is the feedback coefficient of pollutant concentration on groundwater flow velocity, which indicates the degree of influence of pollutant concentration gradient on groundwater flow velocity, and reflects the feedback effect of pollutant concentration on the physical or chemical reaction of groundwater flow. It is obtained by experimental fitting method, and the reference value range is to ; is the pollutant concentration gradient, indicating that at time and spatial location The spatial change rate of pollutant concentration at a location. The pollutant concentration gradient directly affects the diffusion and migration of pollutants in groundwater. It depends on the specific pollutant and is obtained from the fused data.

[0041] Based on the groundwater flow rate and pollutant concentration, the convection term, diffusion term and degradation term are constructed, and combined with the periodic disturbance term, the pollutant migration equation is constructed:

[0042] ,

[0043] in, It is at the moment and spatial location Pollutant concentration The instantaneous rate of change is used to describe the dynamic changes of pollutant concentrations in groundwater over time; is the diffusion coefficient of pollutants in groundwater, which indicates the diffusion capacity of pollutants in water, that is, the diffusion rate of pollutants in water through molecular diffusion and other means. It is calculated based on the water properties and pollutant types in the fused data using a groundwater diffusion model. The water properties include porosity, permeability, saturation, dynamic viscosity, density, conductivity, resistivity, etc. It is the pollutant degradation rate coefficient, which is used to describe the degradation, reaction or transformation rate of pollutants in groundwater (such as biodegradation, chemical degradation, etc.). It determines the decay rate of pollutant concentration over time and depends on the nature of the pollutant and environmental factors. It is obtained through literature review and the reference value range is to ; It is the periodic disturbance coefficient, which is related to environmental factors (such as precipitation, temperature changes, etc.), and reflects the periodic impact of external environmental changes (such as rainfall, temperature changes) on the concentration of groundwater pollutants. It is obtained based on the environmental data in the fused data and the existing climate model prediction. The reference value range is to ; It is a time function of periodic disturbances, which is used to describe the periodic impact of external factors (such as precipitation, seasonal changes, etc.) on pollutant concentrations, which changes periodically over time. is the disturbance period; represents the convection term, which indicates the convection process of pollutants flowing with groundwater; is the diffusion term, which represents the concentration change of pollutants in groundwater due to molecular diffusion or turbulent diffusion; The Laplace operator representing the pollutant concentration represents the spatial second-order derivative of the pollutant concentration and is used to describe the diffusion behavior of pollution in three-dimensional space; is the degradation term, which represents the degradation process of the pollutant; is a periodic disturbance term that describes the impact of external periodic factors (such as precipitation and temperature) on the concentration of groundwater pollutants. The diffusion model and climate model are both well-known technical means to those skilled in the art and will not be described in detail here.

[0044] Furthermore, the above pollutant migration equation is discretized by spatial central difference and time forward difference through three-dimensional spatial grid and explicit Euler method, and iterative calculation is performed to solve the pollutant migration equation, obtain the pollutant concentration, and construct the three-dimensional concentration distribution of pollutants. The time step in the iterative process is set according to the groundwater flow velocity and the pollutant diffusion coefficient based on the expert experience method; the solution method of the pollutant migration equation is a technical means well known to those skilled in the art and will not be described in detail here.

[0045] S2. Based on the three-dimensional concentration distribution of pollutants, a weighted heterogeneous graph is constructed and the pollution propagation path is identified. For the pollution propagation path, the path potential score and pollution risk index are calculated to dynamically assess the pollution risk.

[0046] Based on the three-dimensional concentration distribution of pollutants, a weighted heterogeneous graph is constructed to identify the pollution propagation path, that is, to identify the possible paths from the pollution source node (the node with the highest pollutant concentration) to the sensitive target node (judged according to expert experience, such as water wells, farmland, and residential areas); the weighted heterogeneous graph , taking each spatial location as a node (including properties such as pollutant concentration, geological permeability, and pollutant type), and Respectively represent the set of vertices and edges of the weighted heterogeneous graph, and use Represents any node, each edge Represents the hydraulic connection between two spatial locations; combining the physical accessibility of the spatial diffusion path, the driving force of the pollutant migration concentration, and the filtering capacity of the geological permeability, the edge weight calculation formula is constructed:

[0047] ,

[0048] in, Is a connecting node and nodes The edge weights between nodes reflect the pollution from Propagate to nodes likelihood or priority; It is the spatial attenuation factor, which is used to control the influence of spatial distance on the weight of pollution propagation path. It affects the attenuation degree of pollutants when they propagate in water. It is determined based on Darcy's law and the physical characteristics of pollutant diffusion, and the reference value is 0.01 to 0.1. is a node and nodes The spatial distance between the two points, that is, the Euclidean distance between the two points. The calculation method of the Euclidean distance is a technical means well known to those skilled in the art and will not be described in detail here; and Represents nodes respectively and nodes Pollutant concentration at the site; and Represents nodes respectively and nodes The local permeability coefficient of the formation describes the ability of groundwater to penetrate different soil or rock layers. Get, the value range is to ; Used to describe the physical accessibility of spatial diffusion paths; It represents the bidirectional pollution potential, which is used to describe the driving force of pollutant migration concentration; Used to describe the filtration capacity of geological permeability. The above formula considers the superposition effect of pollutant concentration and the attenuation effect of medium permeability through weighted processing to ensure that high-concentration and high-permeability paths have higher scores;

[0049] Furthermore, the weighted heterogeneous graph is input into the graph neural network to obtain the final node state; the graph neural network is a technical means well known to those skilled in the art. The number of layers of the graph neural network is determined according to the specific application scenario and will not be elaborated here; based on the final node state, the path potential score of the pollution propagation path is calculated. , the specific formula is:

[0050] ,

[0051] in, It is the pollution propagation path, which represents the path sequence starting from the pollution source node, along the groundwater flow direction, and reaching a sensitive target node (such as a water well, farmland, residential area, etc.), and is composed of several adjacent edges; is a pair of adjacent nodes in the pollution propagation path; is a node In the graph neural network The final node state obtained after layer propagation; is a node The L2 norm of the final node state, indicating the node The expression intensity of the comprehensive pollution situation, The larger the value, the more significant the node is affected by pollution; is a node In the graph neural network The final node state obtained after layer propagation; is a node The L2 norm of the final node state, indicating the node The expression intensity of the comprehensive pollution situation; It is the propagation attenuation adjustment factor, which is used to control the degree of adjustment of the potential score by the distance of the pollution propagation path in the pollution propagation. It is determined according to the expert experience method, and the reference value range is ; It is a nonlinear penalty term for the spatial distance of the pollution propagation path, which uses a nonlinear function to alleviate the extreme penalty that may be caused by long paths in the path potential score;

[0052] Furthermore, the path potential scores of all possible pollution transmission paths are The maximum value is taken and used for subsequent assessment of the pollution risk index. Specifically, the pollution transmission path with the highest path potential score is selected as the main pollution path. The risk assessment formula is constructed by combining the frequency of human activities in the area affected by the main pollution path, groundwater consumption, and resource sensitivity level. The risk assessment formula introduces multiple nonlinear perturbation terms to improve the ability to express actual situations. The pollution risk index is calculated according to the risk assessment formula. The specific formula is as follows:

[0053] ,

[0054] in, It's time Pollution risk index; It's time The pollutant concentration of the sensitive target node is determined by the pollutant concentration of the end node of the main pollution path; It is the path distance influencing factor of pollutant concentration. The longer the path, the faster the pollutant spreads and the greater the pollution risk index. It is determined based on expert experience and the reference value range is 0.005~0.05; The exposure risk amplification adjustment coefficient is used to amplify or weaken the role of the exposure factor in the risk assessment formula. It is determined based on expert experience and has a reference value range of 0.1 to 5. It's time The length of the migration path from the pollution source node to the sensitive target node is determined by the total length of the pollution main path; It's time The exposure factor is derived based on the frequency of human activities, groundwater consumption, and resource sensitivity level in the area affected by the main pollution path, using existing simulation fitting algorithms such as polynomial regression. The frequency of human activities in the area affected by the path is obtained through statistical analysis, the groundwater consumption in the area affected by the path is obtained through a water meter monitoring system, and the resource sensitivity level in the area affected by the path is determined based on expert experience. is the frequency of environmental periodic disturbances, which is determined according to the specific application scenario and is not limited here; It is an external intervention adjustment item, determined according to the expert experience method, with a reference value range of 0.1 to 10. The larger the value of the external intervention adjustment item, the stronger the management intervention. is the response amplification index, which is determined based on expert experience and has a reference value range of 1.0 to 3.0. The larger the response amplification index, the more sensitive the pollution risk assessment. The combined effects of pollution source concentration and pollutant propagation path length (distance) are taken into account; It represents the exposure intensity factor, which combines the exposure factor and the exposure risk amplification adjustment coefficient to describe the intensity of pollution exposure; is the time period disturbance factor, which takes into account the time period disturbance during the pollution exposure process; is the environmental response moderator, which controls the environmental moderation effect by logarithmically transforming the external intervention moderator; It is a combination of the exposure intensity factor and the time period disturbance factor, reflecting the original intensity and periodic disturbance of pollution exposure; Dynamically integrate changes in pollutant exposure with actual external intervention effects through periodic changes (such as seasonal fluctuations) and adjustments to external interventions;

[0055] Finally, the pollution risk index is compared with the risk threshold preset based on expert experience, and the risk level intervals are divided to achieve dynamic assessment of pollution risk and provide on-site early warning:

[0056] when When , it indicates low risk;

[0057] when When , it means medium risk;

[0058] when When , it indicates high risk;

[0059] when When , it indicates extremely high risk;

[0060] in, 、 and Represent low risk threshold, medium risk threshold and high risk threshold respectively.

[0061] In summary, a dynamic assessment method for groundwater pollution risk in chemical parks that integrates multi-source heterogeneous data has been completed.

[0062] The order in which the embodiments of the invention are presented is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0063] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.

[0064] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the scope of protection of the present invention.

Claims

1. A method for dynamic assessment of groundwater pollution risk in chemical parks by integrating multi-source heterogeneous data, characterized in that: The following steps are involved: S1. Acquire multi-source heterogeneous data from the chemical park and preprocess it to obtain preprocessed multi-source heterogeneous data; Perform preliminary feature extraction on the preprocessed multi-source heterogeneous data to obtain preliminary feature data; perform fusion processing on the preliminary feature data to obtain fused data; Based on the hydraulic head data in the fused data, the hydraulic head gradient is calculated, and the feedback term of the pollutant concentration gradient in the fused data on the hydraulic head gradient is introduced to construct a two-way coupled groundwater pollution migration simulation model and calculate the groundwater flow rate; Based on groundwater flow velocity and pollutant concentration, convection term, diffusion term and degradation term are constructed, and combined with periodic disturbance term, the pollutant migration equation is constructed; The pollutant migration equation is discretized by spatial central difference and time forward difference, and iterative calculation is performed to solve the pollutant migration equation, obtain the pollutant concentration, and construct the three-dimensional concentration distribution of the pollutant; S2. Based on the three-dimensional concentration distribution of pollutants, construct a weighted heterogeneous graph and identify pollution propagation paths; For the pollution transmission path, the path potential score is calculated, and the pollution risk index is calculated to dynamically evaluate the pollution risk.

2. A method for dynamic assessment of groundwater pollution risk in a chemical park by integrating multi-source heterogeneous data according to claim 1, characterized in that: Said S1 specifically includes: Performing feature dimensionality reduction processing on the preliminary feature data to obtain feature data after dimensionality reduction; performing mapping processing on the feature data after dimensionality reduction to obtain mapped feature data; performing fusion processing on the mapped feature data and normalizing it to obtain fused data.

3. The method for dynamic assessment of groundwater pollution risk in a chemical park by integrating multi-source heterogeneous data according to claim 1 is characterized in that: Said S2 specifically includes: Each spatial position of the three-dimensional concentration distribution of pollutants is taken as a node. Based on the spatial distance between nodes and the pollutant concentration of the nodes, and introducing the spatial attenuation factor, the edge weight is obtained, a weighted heterogeneous graph is constructed, and the pollution propagation path is identified.

4. A method for dynamic assessment of groundwater pollution risk in a chemical park by integrating multi-source heterogeneous data according to claim 3, characterized in that: Said S2 specifically includes: The node state is propagated on the weighted heterogeneous graph to obtain the final node state; based on the final node state and combined with the edge weight, the path potential score of the pollution propagation path is calculated.

5. A method for dynamic assessment of groundwater pollution risk in a chemical park by integrating multi-source heterogeneous data according to claim 4, characterized in that: Said S2 specifically includes: The pollution transmission path with the highest path potential score is selected as the main pollution path. Based on the pollutant concentration at the end node of the main pollution path, the exposure factor, time period disturbance factor and external intervention adjustment item are introduced to calculate the pollution risk index.

6. A method for dynamic assessment of groundwater pollution risk in a chemical park by integrating multi-source heterogeneous data according to claim 5, characterized in that: Said S2 specifically includes: The exposure factor is obtained through a simulation fitting algorithm based on the frequency of human activities, groundwater consumption and resource sensitivity level in the area affected by the main pollution path.

7. A method for dynamic assessment of groundwater pollution risk in a chemical park by integrating multi-source heterogeneous data according to claim 6, characterized in that: Said S2 specifically includes: Compare the pollution risk index with the preset risk threshold, divide the risk level intervals, and realize dynamic assessment of pollution risk.

Images