Pollutant responsibility division method, device, equipment and medium for complex industrial site

Abstract

The invention discloses a pollutant responsibility division method and device for a complex industrial site, equipment and a medium, and relates to the technical field of environmental governance, and the method comprises the steps: obtaining multi-source data in a target land parcel, and constructing a fingerprint index to recognize a potential pollution source type, then, a migration path constraint condition is established in combination with a site hydrogeological structure and historical production activity information; the data and conditions are input into a preset pollution contribution calculation model, pollution contribution proportions of all potential pollution sources in different space areas and depth layers are output, then a pollution responsibility division result is generated, the problem of multi-source pollution responsibility division in a complex industrial site is solved, the accuracy and scientificity of responsibility division are improved, and the responsibility division efficiency is improved. Powerful support is provided for pollution treatment and environment restoration.

Description

Technical Field

[0001] This invention relates to the field of environmental governance technology, and in particular to a method, apparatus, equipment and medium for determining pollutant liability in complex industrial sites. Background Technology

[0002] Currently, the allocation of pollution liability for complex industrial sites mainly relies on traditional pollution source analysis methods, such as mathematical statistical models (e.g., PMF models), pollutant fingerprinting techniques (i.e., isotope analysis), and migration process simulations. While these methods can identify the characteristics and migration paths of pollution sources to some extent, they have significant limitations when dealing with industrial sites that have multiple sources and complex histories. For example, traditional methods struggle to handle pollution signals from intertwined internal and external sources, cannot effectively distinguish the contribution ratios of different pollution sources, and have shortcomings in spatial resolution and temporal scale.

[0003] However, existing technologies still have many shortcomings in their application in complex sites. First, the ambiguity of source identification makes it difficult to accurately analyze the feature contributions of each potential source; second, the uncertainty of the migration process makes it difficult to confirm the rationality of responsibility association; and finally, the bottleneck of responsibility quantification limits the accurate delineation of dominant pollution sources. These problems make it difficult for existing technologies to meet the refined requirements of responsibility delineation in complex industrial sites in practical applications, especially in scenarios with spatial overlap of multiple pollution sources and complex historical changes, where the accuracy and reliability of responsibility delineation are difficult to guarantee.

[0004] Therefore, how to achieve accurate pollution source identification, migration path verification, and responsibility quantification in complex industrial sites has become an urgent problem to be solved. Summary of the Invention

[0005] The main objective of this application is to provide a method, apparatus, equipment, and medium for assigning pollutant liability in complex industrial sites, aiming to solve the technical problem of how to improve the accuracy of quantifying pollution liability in complex industrial sites.

[0006] To achieve the above objectives, this application proposes a method for allocating pollutant liability in complex industrial sites, including: Acquire multi-source data within the target site, including soil physicochemical data, spatial covariate data, pollutant concentration data, geographic data, and resistivity anomaly area information; Fingerprint indicators are constructed based on the multi-source data to identify potential pollution source types; Establish migration path constraints by combining information on the site's hydrogeological structure and historical production activities; The multi-source data, the fingerprint index, and the migration path constraints are input into a preset pollution contribution calculation model, which outputs the pollution contribution ratio of each potential pollution source type in different spatial regions and depth layers. The preset pollution contribution calculation model embeds a two-dimensional partitioning module of spatial region and depth layer. The two-dimensional partitioning module divides the target plot into different spatial regions according to a preset planar grid and into different depth layers according to a preset soil layer thickness. The preset planar grid is set according to the pollution complexity of the plot, and the preset soil layer thickness is set according to the soil sampling depth interval. The pollution responsibility allocation results are generated based on the pollution contribution ratio.

[0007] In one embodiment, the step of inputting the multi-source data, the fingerprint index, and the migration path constraints into a preset pollution contribution calculation model, and outputting the pollution contribution ratio of each potential pollution source type in different spatial regions and depth layers, includes: The correlation data between soil porosity and permeability coefficient are extracted from the multi-source data. Combined with the groundwater flow direction information in the migration path constraints, a coupling weight factor of medium permeability-water flow orientation is constructed. The coupling weight factor is used to correct the differences in pollutant migration efficiency in different spatial regions. Based on the association weights of characteristic pollutants in the fingerprint index, and combined with the time period of historical production activities, a correction coefficient for source strength-time accumulation is calculated, wherein the correction coefficient is used to quantify the time decay effect of emission behavior of different potential pollution source types. The coupling weight factor, the correction coefficient, the multi-source data, and the fingerprint index are dimensionally fused to form an enhanced input dataset; The enhanced input dataset is input into a preset pollution contribution calculation model, which outputs the pollution contribution ratio of each potential pollution source type in the corresponding spatial region and depth layer.

[0008] In one embodiment, the step of dimensionally fusing the coupling weight factor, the correction coefficient, the multi-source data, and the fingerprint index to form an enhanced input dataset includes: Extract pollutant concentration data from the multi-source data, group them according to different spatial regions and depth layers, and obtain the pollutant concentration subsets corresponding to each region-depth layer; The coupling weight factor and the soil porosity data of the corresponding region-depth layer are multiplied element by element to obtain the medium characteristic data, which is used to reflect the site medium adaptability of pollution migration. The correction coefficient and the correlation weight of each characteristic pollutant in the fingerprint index are weighted and summed to obtain fingerprint feature data, which is used to highlight the historical emission cumulative effect of different pollution sources. The pollutant concentration subset, the medium feature data, and the fingerprint feature data are dimensionally aligned to obtain aligned pollutant concentration subset, aligned medium feature data, and aligned fingerprint feature data. The aligned pollutant concentration subset, aligned medium feature data, and aligned fingerprint feature data are weighted and calculated based on the regional pollution sensitivity coefficient to obtain the weighted pollutant concentration subset, weighted medium feature data, and weighted fingerprint feature data. The regional pollution sensitivity coefficient is set according to whether the corresponding region is a historical production core area and whether there are underground pipeline leakage points. The weighted pollutant concentration subset, the weighted medium feature data, and the weighted fingerprint feature data are concatenated to form an enhanced input dataset.

[0009] In one embodiment, the step of inputting the enhanced input dataset into a preset pollution contribution calculation model and outputting the pollution contribution ratio of each potential pollution source type in the corresponding spatial region and depth layer includes: The coefficient of variation of pollutant concentration in multi-source data is extracted by the two-dimensional partitioning module of the preset pollution contribution calculation model. The pollution complexity of a site is determined based on the pollutant concentration variation coefficient. In areas with high pollution complexity, the grid side length of the preset planar grid is reduced to a preset short side length, while in areas with low pollution complexity, the preset long side length is maintained. The preset short side length is set according to the minimum identification unit of the high pollution variation area, and the preset long side length is set according to the grid density of the conventional pollution survey. The thickness of the basic soil layer is determined by the dual-dimensional segmentation module based on the soil sampling depth interval. For depth segments with continuous pollution halos, the preset soil layer thickness is reduced to half of the thickness of the basic soil layer. For depth segments with a sharp drop in pollution concentration, the preset soil layer thickness is increased to twice the thickness of the basic soil layer. In the dual-dimensional partitioning module, a grid-soil layer association index table is established to record the coordinate mapping relationship between each spatial region grid and the corresponding depth layer; The data of the same grid-soil layer unit in the enhanced input dataset are integrated into independent samples and index labels are embedded in the independent samples to obtain sample data with index labels, wherein the index labels include spatial region number and depth layer number; The sample data with index labels is input into a preset pollution contribution calculation model. The division rules of the two-dimensional partitioning module are called through the index labels, and the migration path constraints of the corresponding region-depth layer are loaded for calculation to obtain the pollution contribution ratio. Based on the pollution contribution ratio, the corresponding spatial regions and depth layers are back-mapped according to the index labels, and the pollution contribution ratios of each potential pollution source type in different spatial regions and depth layers of the target plot are summarized.

[0010] In one embodiment, the step of constructing fingerprint indicators based on the multi-source data to identify potential pollution source types includes: Soil physicochemical data, pollutant concentration data, and spatial covariate data are extracted from the multi-source data to screen out pollutant combinations corresponding to industrial production activities, wherein the pollutant combinations include chromium-nickel-zinc combinations and lead-petroleum hydrocarbon combinations. The concentration ratio of pollutants in each pollutant combination is calculated based on the constant material ratio of different industrial production activities. The pollutant combination whose concentration ratio fluctuation is lower than a preset fluctuation threshold is identified as a candidate fingerprint feature. The candidate fingerprint features are spatially anchored by combining the resistivity anomaly region information of the multi-source data to obtain the spatial anchoring result; The candidate fingerprint features are verified for temporal validity based on the historical production activity period and pollutant degradation cycle to obtain valid candidate fingerprint features. The valid candidate fingerprint features are de-mixed to obtain the de-mixed fingerprint features; The demixed fingerprint features are compared with a preset industrial pollution source fingerprint database to calculate the feature matching degree; By combining the spatial anchoring results and feature matching degree, the potential pollution source types in each region are determined.

[0011] In one embodiment, the step of establishing migration path constraints by combining site hydrogeological structure and historical production activity information includes: The core parameters of the site's hydrogeological structure are extracted, and a three-dimensional hydrogeological model is constructed. The core parameters include groundwater flow velocity, aquifer thickness, aquitard depth, and lithological permeability coefficient. Based on the three-dimensional hydrogeological model combined with historical production activity information, the location of pollution sources, emission methods, and pollutant outflow periods are inferred to obtain the theoretical migration range; Based on resistivity anomaly area information from multi-source data, underground hidden passages are identified, and the routes along these underground hidden passages are set as priority migration paths. Calculate the pollutant concentration at each sampling point and the distance attenuation coefficient of the pollution source within the theoretical migration range, and obtain the spatial correlation verification results based on the distance attenuation coefficient and the three-dimensional hydrogeological model; Based on the theoretical migration range, preferred migration paths, and spatial correlation verification results, valid migration paths are confirmed, and migration path constraints are established, including spatial range, time window, and migration channel.

[0012] In one embodiment, the step of generating pollution liability allocation results based on the pollution contribution ratio includes: The pollution contribution ratios of each potential pollution source type are aggregated according to a spatial grid to generate a three-dimensional responsibility heat map; Regions in the responsibility heatmap whose contribution ratio is higher than a preset high confidence threshold are identified and marked as dominant responsibility regions. Regions in the responsibility heatmap whose contribution ratio is lower than a preset low confidence threshold are identified and marked as non-responsibility regions. Identify areas in the responsibility heatmap where the contribution ratio is higher than the low confidence threshold and lower than the high confidence threshold, and mark them as areas of undetermined responsibility. Output the pollution liability allocation results, which include spatial location, pollution source type, pollution contribution ratio, and liability level.

[0013] Furthermore, to achieve the above objectives, this application also proposes a pollutant liability allocation device for complex industrial sites, the device comprising: The acquisition module is used to acquire multi-source data within the target plot, including soil physicochemical data, spatial covariate data, pollutant concentration data, geographic data, and resistivity anomaly area information; The identification module is used to construct fingerprint indicators based on the multi-source data to identify the types of potential pollution sources; The module is used to establish migration path constraints by combining information on the site's hydrogeological structure and historical production activities. The calculation module is used to input the multi-source data, the fingerprint index, and the migration path constraints into a preset pollution contribution calculation model, and output the pollution contribution ratio of each potential pollution source type in different spatial regions and depth layers. The preset pollution contribution calculation model embeds a two-dimensional partitioning module of spatial region and depth layer. The two-dimensional partitioning module divides the target plot into different spatial regions according to a preset planar grid and into different depth layers according to a preset soil layer thickness. The preset planar grid is set according to the pollution complexity of the plot, and the preset soil layer thickness is set according to the soil sampling depth interval. The results module is used to generate pollution liability allocation results based on the pollution contribution ratio.

[0014] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable medium, on which a computer program is stored, which, when executed by a processor, implements the steps of the pollutant liability delineation method for complex industrial sites as described above.

[0015] In addition, to achieve the above objectives, this application also provides a computer program product, which includes a computer program that, when executed by a processor, implements the steps of the pollutant liability delineation method for complex industrial sites as described above.

[0016] This application acquires multi-source data within the target site and constructs fingerprint indicators to identify potential pollution source types. Then, it establishes migration path constraints by combining the site's hydrogeological structure and historical production activity information. These data and conditions are input into a pre-defined pollution contribution calculation model, which outputs the pollution contribution ratio of each potential pollution source in different spatial regions and depth layers, thereby generating pollution liability allocation results. This solves the challenge of multi-source pollution liability allocation in complex industrial sites, improves the accuracy and scientific rigor of liability allocation, and provides strong support for pollution control and environmental remediation. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart illustrating the first embodiment of the method for allocating pollutant liability in complex industrial sites according to this application; Figure 2 This is a flowchart illustrating the second embodiment of the pollutant liability allocation method for complex industrial sites in this application. Figure 3 This is a schematic diagram of the modular structure of the pollutant liability allocation device for complex industrial sites, which is the first embodiment of the pollutant liability allocation method for complex industrial sites in this application.

[0019] Figure 4 This is a schematic diagram of the equipment structure of the hardware operating environment involved in the pollutant liability division method for complex industrial sites in this application embodiment.

[0020] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0021] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0022] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0023] Currently, the allocation of pollution liability for complex industrial sites mainly relies on traditional pollution source analysis methods, such as mathematical statistical models (e.g., PMF models), pollutant fingerprinting techniques (i.e., isotope analysis), and migration process simulations. While these methods can identify the characteristics and migration paths of pollution sources to some extent, they have significant limitations when dealing with industrial sites that have multiple sources and complex histories. For example, traditional methods struggle to handle pollution signals from intertwined internal and external sources, cannot effectively distinguish the contribution ratios of different pollution sources, and have shortcomings in spatial resolution and temporal scale.

[0024] However, existing technologies still have many shortcomings in their application in complex sites. First, the ambiguity of source identification makes it difficult to accurately analyze the feature contributions of each potential source; second, the uncertainty of the migration process makes it difficult to confirm the rationality of responsibility association; and finally, the bottleneck of responsibility quantification limits the accurate delineation of dominant pollution sources. These problems make it difficult for existing technologies to meet the refined requirements of responsibility delineation in complex industrial sites in practical applications, especially in scenarios with spatial overlap of multiple pollution sources and complex historical changes, where the accuracy and reliability of responsibility delineation are difficult to guarantee.

[0025] Based on the above, this application provides a method for allocating pollutant liability in complex industrial sites, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the pollutant liability allocation method for complex industrial sites according to this application. In this embodiment, the pollutant liability allocation method for complex industrial sites includes steps S10 to S50: Step S10: Obtain multi-source data within the target plot.

[0026] It should be noted that multi-source data refers to a collection of various data related to the pollution status of a target site, obtained from different collection channels and analytical dimensions. Multi-source data includes soil physicochemical data, spatial covariate data, pollutant concentration data, geographic data, and resistivity anomaly area information. Soil physicochemical data refers to data reflecting the physical and chemical properties of soil, providing crucial information for describing soil characteristics and directly influencing the migration and occurrence of pollutants in the soil, including pH, organic matter content, porosity, and permeability coefficient. Spatial covariate data refers to auxiliary data related to the spatial attributes of different locations within the target site, reflecting the differences in environmental conditions related to pollution at different spatial locations, including distances from potential pollution sources, distances from pipelines, and elevations. Pollutant concentration data refers to the content data of various pollutants within the target site obtained through professional testing methods, including specific content information for pollutants such as cadmium, lead, copper, nickel, zinc, chromium, hexavalent chromium, arsenic, mercury, and petroleum hydrocarbons. Geographic data refers to data related to the geographic features and geological structure of the target site, including topographic data, geological structure data, and high-density electrical resistivity tomography (EDT) data. Information on resistivity anomalies refers to information about areas obtained through resistivity detection in geophysical exploration that show significant differences in resistivity values ​​compared to normal areas surrounding the target site. Such areas may be associated with underground pollution distribution, hidden channels, etc.

[0027] Specifically, firstly, a comprehensive on-site investigation and sampling of the target site is conducted. Soil samples are collected and their physicochemical properties are measured, such as pH value, organic matter content, porosity, and permeability coefficient. These physicochemical data reflect the basic characteristics of the soil and the migration conditions of pollutants. Next, spatial covariate data are collected, including the distance and elevation of the sampling point from potential pollution sources, as well as its relative location to surrounding facilities. This data helps analyze the spatial distribution patterns and migration paths of pollutants. Then, the concentration of pollutants in the soil samples is measured, covering common pollutants such as heavy metals and petroleum hydrocarbons. Concentration data is crucial for identifying pollution sources and assessing the degree of pollution. Furthermore, geographic data of the site, such as digital elevation models (DEMs) and land use types, is acquired. This data provides geographic context for the spatial location and migration analysis of pollution sources. Finally, geophysical exploration techniques, such as resistivity imaging, are used to identify resistivity anomalies within the site. These anomalies may indicate hidden underground channels or areas of concentrated pollution, providing important clues for accurate identification of pollution sources and confirmation of migration paths. By integrating these multi-source data, the pollution status and environmental characteristics of the target site can be comprehensively and accurately reflected, laying a solid foundation for subsequent pollution source analysis and responsibility delineation.

[0028] Step S20: Construct fingerprint indicators based on multi-source data to identify potential pollution source types.

[0029] It should be noted that fingerprint indicators refer to a set of distinctive information extracted from multi-source data that characterizes the unique features of a specific pollution source. These indicators consist of pollutant combinations, concentration ratios, and correlation weights, and are the core basis for distinguishing different pollution sources. Potential pollution source types refer to various categories of pollution sources that may cause pollution to the target site, inferred from fingerprint indicators and multi-source data. These include pollution sources from electroplating plants, waste recycling stations, auto repair shops, and concrete mixing plants.

[0030] Further, step S20 includes: extracting soil physicochemical data, pollutant concentration data, and spatial covariate data from multi-source data; screening out pollutant combinations corresponding to industrial production activities, including chromium-nickel-zinc combinations and lead-petroleum hydrocarbon combinations; calculating the concentration ratio of pollutants within each pollutant combination based on the constant material ratio of different industrial production activities; identifying the characteristics of pollutant combinations whose concentration ratio fluctuations are lower than a preset fluctuation threshold as candidate fingerprint features; spatially anchoring the candidate fingerprint features based on resistivity anomaly area information from multi-source data to obtain spatial anchoring results; verifying the temporal validity of the candidate fingerprint features based on the historical production activity period and pollutant degradation cycle to obtain valid candidate fingerprint features; unmixing the valid candidate fingerprint features to obtain unmixed fingerprint features; comparing the unmixed fingerprint features with a preset industrial pollution source fingerprint database to calculate the feature matching degree; and determining the potential pollution source types for each region based on the spatial anchoring results and feature matching degree.

[0031] It should be noted that material ratio constancy refers to the characteristic that the proportions of various materials used in a specific industrial production process remain relatively stable over a long period of production. This stability is reflected in the combination and concentration ratio of the pollutants generated. Concentration ratio refers to the proportional relationship between the concentrations of two or more pollutants in a pollutant combination. Compared to the concentration of a single pollutant, it better reflects the unique characteristics of the pollution source and is more stable. The preset fluctuation threshold refers to a pre-set critical standard for judging whether the concentration ratio is stable. When the fluctuation range of the concentration ratio is lower than this standard, it indicates that the ratio is stable and can be used as a candidate for fingerprint features. Candidate fingerprint features refer to pollutant combination features whose concentration ratio fluctuation range is lower than the preset fluctuation threshold. These are preliminary feature information with the potential to distinguish pollution sources. Spatial anchoring refers to the correlation analysis between candidate fingerprint features and resistivity anomaly area information to strengthen the correspondence between the two, clarify the spatial distribution pattern of candidate fingerprint features, and improve the targeting of features. Temporal validity verification refers to the verification process of judging whether candidate fingerprint features conform to temporal logic by combining the historical production activity period and the pollutant degradation cycle, used to eliminate invalid features without corresponding production activity support. The pollutant degradation cycle refers to the time required for pollutants to gradually decompose and reduce their concentration to the environmental background level through physical, chemical, or biological processes in the natural environment. Valid candidate fingerprint features refer to candidate fingerprint features that, after time-series validity verification, are confirmed to match historical production activity periods and have practical significance for pollution source tracing. Unmixing refers to using professional methods such as factor analysis to separate the intertwined fingerprint signals of different pollution sources, eliminate mutual interference, and extract the pure fingerprint features of a single pollution source. The pre-established industrial pollution source fingerprint database refers to a pre-established database containing characteristic fingerprint information of various common industrial pollution sources, storing standard information such as pollutant combinations and concentration ratios corresponding to different industrial types. Feature matching degree refers to the similarity between the unmixed fingerprint features and the fingerprints of various pollution sources in the pre-established industrial pollution source fingerprint database; it is a key quantitative indicator for determining the type of potential pollution sources.

[0032] Specifically, firstly, soil physicochemical data, pollutant concentration data, and spatial covariate data are extracted from the collected multi-source data. By analyzing these data, pollutant combinations corresponding to industrial production activities are screened. For example, soil physicochemical data, pollutant concentration data, and spatial covariate data are extracted from the multi-source data of the target site. Based on the chromium passivation, nickel plating, and zinc plating processes in electroplating production, a chromium-nickel-zinc combination is screened. Based on the waste battery processing and mechanical parts dismantling processes in waste recycling, a lead-petroleum hydrocarbon combination is screened. Next, based on the constant ratio of materials such as chromium anhydride and nickel sulfate in the electroplating process, the concentration ratios of chromium to nickel, nickel, and zinc, and nickel to zinc are calculated. Then, based on the stability of the dismantling ratio of waste batteries and mechanical parts in waste recycling, the concentration ratio of lead to petroleum hydrocarbons is calculated. A preset fluctuation threshold is set, and the above concentration ratio characteristics with fluctuation amplitudes below this threshold are identified as candidate fingerprint features. Subsequently, combining the resistivity anomaly information of the target site, it was found that the distribution range of chromium-nickel-zinc related candidate fingerprint features highly overlapped with the resistivity anomaly areas around the electroplating workshop, and the lead-petroleum hydrocarbon related candidate fingerprint features corresponded to the resistivity anomaly areas of the waste recycling and dismantling area, thus obtaining spatial anchoring results. Next, based on the production period of the electroplating plant from 1999 to 2017, the operation period of the waste recycling station from 2017 to 2020, and the degradation cycles of chromium, nickel, zinc, lead, and petroleum hydrocarbons, candidate fingerprint features that exceeded the degradation cycle and had no corresponding production activity support were eliminated, resulting in valid candidate fingerprint features. This verification process ensured that the identified fingerprint features matched historical production activities in time, eliminating misjudgments caused by natural degradation or other irrelevant factors, thus obtaining valid candidate fingerprint features. Then, factor analysis was used to unmix the valid candidate fingerprint features, successfully separating the fingerprint signals of the electroplating source and the waste recycling station source, resulting in unmixed fingerprint features. Demixing further improves the purity and representativeness of fingerprint features, making them closer to the true characteristics of a single pollution source. Next, the demixed fingerprint features are compared with a pre-set fingerprint database of industrial pollution sources such as electroplating plants and waste recycling stations. The results show that the chromium-nickel-zinc fingerprint features have a very high matching degree with the electroplating plant fingerprints, and the lead-petroleum hydrocarbon fingerprint features have a very high matching degree with the waste recycling station fingerprints. The pre-set industrial pollution source fingerprint database contains typical fingerprint features of known pollution sources; comparison allows for a quantitative assessment of the similarity between the demixed fingerprint features and known pollution sources. Finally, combining the spatial anchoring results and feature matching degree, the potential pollution source type in the area surrounding the electroplating workshop of the target site is determined to be an electroplating plant pollution source, and the potential pollution source type in the area surrounding the waste recycling and dismantling area is determined to be a waste recycling station pollution source. Spatial anchoring provides geographical location information of the pollution source, while feature matching degree provides type information. The combination of the two can accurately identify the potential pollution source types in different areas, providing a scientific basis for subsequent responsibility allocation and pollution control.

[0033] Step S30: Establish migration path constraints by combining information on the site's hydrogeological structure and historical production activities.

[0034] It should be noted that the site's hydrogeological structure refers to the geological structural characteristics of the target site and its surrounding area, including the groundwater burial state, flow characteristics, and lithological distribution. Core parameters include groundwater flow velocity, aquifer thickness, aquitard depth, and lithological permeability coefficient. Historical production activity information refers to detailed information related to various industrial production and operation activities carried out on the target site in the past, reflecting the source and emission characteristics of pollution, including the location of pollution sources, the time period of pollutant outflow, and specific emission methods. Migration path constraints refer to the set of limiting rules set to clarify the effective migration path of pollutants from the pollution source to the receiver, used to standardize the analysis scope and logic of migration paths. Migration path constraints include spatial scope, time window, and migration channel. Spatial scope refers to the specific spatial area covered by the effective migration path, clearly defining the spatial boundary of pollutant migration. Time window refers to the time interval corresponding to the pollutant's migration from the pollution source to different locations along the effective migration path. Migration channel refers to the specific path traversed by the pollutant during migration, including preferred migration paths and other regular migration channels.

[0035] Further, step S30 includes: extracting the core parameters of the site's hydrogeological structure and constructing a three-dimensional hydrogeological model; inferring the theoretical migration range based on the location, emission method, and outflow time of pollution sources in conjunction with historical production activity information from the three-dimensional hydrogeological model; identifying underground hidden channels based on resistivity anomaly area information from multi-source data and setting the underground hidden channels as priority migration paths; calculating the distance attenuation coefficient between the pollutant concentration at each sampling point and the pollution source within the theoretical migration range, and obtaining spatial correlation verification results based on the distance attenuation coefficient and the three-dimensional hydrogeological model; and confirming the effective migration path and establishing migration path constraints by combining the theoretical migration range, priority migration path, and spatial correlation verification results.

[0036] It should be noted that the theoretical migration range refers to the spatial area that pollutants may migrate to at different times, inferred from the groundwater movement patterns and historical production activity information based on a three-dimensional hydrogeological model. Hidden underground channels refer to underground passages that have not been discovered through conventional exploration and can facilitate pollutant migration; common examples include geological fissures and abandoned underground pipelines. Preferred migration paths refer to the paths that pollutants preferentially choose during migration due to their low resistance and high efficiency; these paths are typically designated along the routes of hidden underground channels.

[0037] Specifically, through site hydrogeological surveys, core parameters such as groundwater flow velocity, aquifer thickness, aquitard depth, and lithological permeability coefficient were extracted for the target site. These parameters were used to construct a three-dimensional hydrogeological model of the target site. These parameters allow for accurate simulation of groundwater flow fields and pollutant migration paths, providing a scientific basis for subsequent responsibility delineation. Next, based on the three-dimensional hydrogeological model and combined with historical production activity information, the specific locations, pollutant discharge methods, and pollutant outflow periods of the electroplating plant (1999-2017) and the waste recycling station (2017-2020) were identified. By analyzing groundwater movement patterns and pollutant diffusion characteristics, the theoretical migration range of pollutants emitted from the two pollution sources during different periods was deduced. This reasoning process, combining historical data and model simulation, predicts the possible diffusion range of pollutants, providing a theoretical foundation for subsequent migration path analysis. Next, resistivity anomaly information of the target site was retrieved. Based on this information, hidden channels such as abandoned underground pipelines were identified, and the routes along these hidden channels were designated as priority migration paths. Areas with high permeability, such as cracks, were considered priority migration routes for pollutants. Designating these underground hidden channels as priority migration paths more accurately reflects the actual migration of pollutants. Then, the pollutant concentrations at 64 sampling points in the target site were calculated, along with the distance attenuation coefficients of the electroplating plant and waste recycling station within the theoretical migration range. Verification using a three-dimensional hydrogeological model revealed that the distance attenuation coefficients at some sampling points conformed to the laws of groundwater flow and pollutant migration, while others did not. This led to spatial correlation verification results. This verification process ensured the rationality and accuracy of the migration paths. Finally, by integrating the theoretical migration range, priority migration paths, and spatial correlation verification results, the migration paths of pollutants from the electroplating plant and waste recycling station to each effective sampling point were confirmed. Migration path constraints, including clearly defined spatial ranges, time windows, and migration channels, were established, clarifying the specific paths and ranges of pollutant migration and providing clear spatial and temporal basis for pollution liability delineation. Through this series of steps, the migration paths of pollutants can be accurately identified and verified, providing scientific support for the delineation of responsibility in complex industrial sites.

[0038] Step S40: Input multi-source data, fingerprint indicators and migration path constraints into the preset pollution contribution calculation model, and output the pollution contribution ratio of each potential pollution source type in different spatial regions and depth layers.

[0039] It should be noted that the preset pollution contribution calculation model refers to a model used to quantify the degree of pollution contribution from different pollution sources. It is a machine learning model specifically adapted to the pollution characteristics of complex industrial sites. The preset pollution contribution calculation model embeds a two-dimensional spatial region-depth layer partitioning module. This module divides the target site into different spatial regions according to a preset planar grid and into different depth layers according to a preset soil layer thickness. The preset planar grid is set based on the site's pollution complexity, and the preset soil layer thickness is set based on the soil sampling depth interval. Site pollution complexity refers to the overall complexity of pollution within the target site, determined by multiple factors such as the number of pollution types, the magnitude of pollution concentration variation, and the uniformity of pollution distribution. The soil sampling depth interval refers to the vertical distance between two adjacent sampling points when collecting soil samples. It is a fundamental parameter reflecting the vertical pollution distribution characteristics of the site and directly determines the initial setting of the preset soil layer thickness. The pollution contribution ratio refers to the proportion of contribution of each potential pollution source type to the pollution in different spatial regions and depth layers of the target site. It is a core quantitative data calculated by the model based on multi-source data, fingerprint indicators, and migration path constraints, and directly determines the proportional basis for responsibility allocation.

[0040] Specifically, the acquired multi-source data, constructed fingerprint indicators, and established migration path constraints are integrated and input into a pre-defined pollution contribution calculation model. Multi-source data provides basic environmental information about the site, fingerprint indicators clarify the characteristics of potential pollution sources, and migration path constraints limit the possible paths and ranges of pollutant migration. The combination of these three provides the model with comprehensive and accurate input. Next, based on this input data, the model uses internal algorithms to quantitatively analyze the pollution contribution of different types of potential pollution sources within the site. The model considers the distribution of each potential pollution source in different spatial regions and depth layers, combining the feature matching degree of the fingerprint indicators and the effectiveness of the migration paths to calculate the pollution contribution ratio of each potential pollution source in each region and depth layer. This process utilizes machine learning or statistical analysis methods, capable of handling complex multi-source data and nonlinear relationships, ensuring the scientific validity and accuracy of the results. Finally, the model outputs the pollution contribution ratio of each type of potential pollution source in different spatial regions and depth layers. These ratios are presented in numerical form, intuitively reflecting the specific contribution of each pollution source within the site. In this way, the responsibility of each pollution source in site pollution can be clearly defined, providing a quantitative basis for subsequent environmental remediation and responsibility determination, and achieving accurate division of responsibility for pollution in complex industrial sites.

[0041] Step S50: Generate pollution liability allocation results based on the proportion of pollution contribution.

[0042] It should be noted that the pollution liability delineation result refers to the final result that clarifies the scope, proportion, and level of pollution liability of each responsible party at different spatial locations and depths of the target site. It is an important basis for the allocation of pollution control funds and the resolution of liability disputes. The pollution liability delineation result includes spatial location, pollution source type, pollution contribution ratio, and level of liability.

[0043] Further, step S50 includes: aggregating the pollution contribution ratios of each potential pollution source type according to a spatial grid to generate a three-dimensional responsibility heatmap; identifying areas in the responsibility heatmap where the contribution ratio is higher than a preset high confidence threshold and marking them as dominant responsibility areas; identifying areas in the responsibility heatmap where the contribution ratio is lower than a preset low confidence threshold and marking them as non-responsibility areas; identifying areas in the responsibility heatmap where the contribution ratio is higher than the low confidence threshold and lower than the high confidence threshold and marking them as areas with pending responsibility; and outputting the pollution responsibility allocation results.

[0044] It should be noted that the spatial grid refers to the independent horizontal spatial units formed after the target plot is divided into a preset planar grid by a two-dimensional partitioning module. Each unit corresponds to a specific geographical range and number. The preset high confidence threshold refers to a pre-set critical value for the proportion of pollution contribution used to determine primary responsibility, set based on site pollution characteristics and liability determination requirements. The preset low confidence threshold refers to a pre-set critical value for the proportion of pollution contribution used to determine no direct responsibility. It is a quantitative standard for defining that pollution in a certain area is not significantly related to the corresponding pollution source; below this value, the responsibility for that pollution source is excluded. The primary responsibility area refers to the spatial-depth area where the proportion of pollution contribution is higher than the preset high confidence threshold. The pollution cause in this area is strongly correlated with the corresponding potential pollution source type, and the pollution source bears the primary pollution responsibility. The non-responsibility area refers to the spatial-depth area where the proportion of pollution contribution is lower than the preset low confidence threshold. The pollution in this area is not significantly related to the corresponding potential pollution source type, and the pollution source does not need to bear pollution responsibility in this area. The area of ​​pending responsibility refers to the spatial-depth region where the proportion of pollution contribution is between the preset low confidence threshold and the preset high confidence threshold. The correlation between the pollution in this area and the corresponding pollution source is insufficient, and the responsibility cannot be clearly assigned at this time. Further verification or negotiation is required to define the responsibility.

[0045] Specifically, firstly, based on the spatial grid divided by the site, the pollution contribution ratios of two types of pollution sources—electroplating plants and waste recycling stations—at various depth layers were collected and integrated to generate a three-dimensional responsibility heat map that intuitively reflects the distribution of pollution responsibility. This heat map spatially displays the distribution of pollution contribution in different areas, while also reflecting the vertical distribution of pollution at different depth layers, providing a visualization tool for intuitively displaying pollution responsibility.

[0046] Next, after setting preset high-confidence and low-confidence thresholds, areas with pollution contribution ratios higher than the high-confidence threshold, such as the area surrounding the electroplating workshop and along deep pipelines, were identified from the heat map and marked as the electroplating plant's primary responsibility area. The area surrounding the waste recycling and dismantling area was marked as the waste recycling station's primary responsibility area. Areas with pollution contribution ratios lower than the low-confidence threshold, such as the edges of land parcels, were identified and marked as non-responsibility areas. Areas between the two thresholds were marked as areas with pending responsibility. The high-confidence threshold was set based on statistical analysis of pollution contribution ratios and actual site conditions, ensuring a high degree of certainty regarding pollution responsibility in these areas and providing a clear basis for determining the responsibility of the primary pollution source. The low-confidence threshold was set to exclude areas with minimal or negligible pollution contribution, avoiding unnecessary disputes and improving the efficiency and accuracy of responsibility allocation. The pollution responsibility in areas with pending responsibility is relatively ambiguous, requiring further investigation and analysis to clarify responsibility attribution, providing a transitional area for subsequent refined responsibility allocation.

[0047] Finally, the labeling results from all regions were integrated to output pollution liability delineation results. These results clearly define the spatial grid number, depth layer interval, corresponding pollution source type, specific pollution contribution ratio (e.g., electroplating plants contribute approximately 80% to deep lead), and the responsibility levels of primary and pending responsibility for each region. This information provides clear criteria for liability delineation for environmental management departments and relevant responsible parties, helping to facilitate the smooth progress of pollution control work and achieve a scientific and accurate delineation of pollution liability for complex industrial sites.

[0048] This embodiment acquires multi-source data within the target site and constructs fingerprint indicators to identify potential pollution source types. Then, it establishes migration path constraints by combining the site's hydrogeological structure and historical production activity information. These data and conditions are input into a pre-defined pollution contribution calculation model, which outputs the pollution contribution ratio of each potential pollution source in different spatial regions and depth layers, thereby generating pollution liability allocation results. This solves the problem of multi-source pollution liability allocation in complex industrial sites, improves the accuracy and scientific rigor of liability allocation, and provides strong support for pollution control and environmental remediation.

[0049] Based on the first embodiment of this application, in the second embodiment of this application, the content that is the same as or similar to that in Embodiment 1 above can be referred to the above description, and will not be repeated hereafter. Based on this, please refer to... Figure 2 The method for assigning responsibility for pollutants in complex industrial sites, step S40, further includes steps S201 to S204: Step S201: Extract the correlation data between soil porosity and permeability coefficient from multi-source data, and combine it with the groundwater flow direction information in the migration path constraint to construct a coupling weight factor of medium permeability-water flow orientation.

[0050] It should be noted that soil void ratio refers to the ratio of the total volume of pores in soil to the total volume of solid particles in the soil. It is a core physical indicator characterizing the looseness of soil. A higher void ratio indicates more developed soil pores, which directly affects the infiltration and migration rate of pollutants in the soil medium. Permeability coefficient is an indicator characterizing the ability of soil or rock strata to allow groundwater and dissolved pollutants to permeate through. Its value reflects the ease of permeation; a higher permeability coefficient indicates higher pollutant migration efficiency in that area. Correlation data refers to combined data formed by matching soil void ratio and permeability coefficient one-to-one according to the spatial location of the target plot. It can reflect the correspondence and synergistic variation patterns of soil void ratio and permeability coefficient in different spatial regions. Coupling weighting factors are used to correct for differences in pollutant migration efficiency in different spatial regions.

[0051] Specifically, from the multi-source soil physicochemical data of this site, soil porosity and permeability coefficient data at 64 sampling points and at various depths were extracted. These data were then matched according to the spatial coordinates of each sampling point to form spatial correlation data between soil porosity and permeability coefficient. Groundwater flow direction information explicitly defined in the migration path constraints of this site was retrieved. The groundwater in this site generally flows from northwest to southeast. Higher values ​​were assigned to the northwest-facing area, and lower values ​​were assigned to the southeast-facing area. Based on the porosity-permeability coefficient correlation data for each spatial region... Calculations showed that the permeability quantification value of the medium around the electroplating workshop was significantly higher due to the development of soil porosity and high permeability coefficient, while the permeability quantification value of the medium in the edge area of ​​the plot was lower. The permeability quantification value of the medium and the vectorized groundwater flow assignment results were calculated by weighted fusion algorithm to construct the medium permeability-water flow orientation coupling weight factor of the plot. The coupling weight factor value of the water-facing area around the electroplating workshop was the highest, which was used to correct the higher pollutant migration efficiency in this area. The coupling weight factor value of the water-facing area on the southeast side of the plot was lower, which was used to correct the lower pollutant migration efficiency in this area.

[0052] Step S202: Based on the association weights of characteristic pollutants in the fingerprint index and combined with the time period length of historical production activities, calculate the correction coefficient of source strength-time accumulation.

[0053] It should be noted that the association weight of characteristic pollutants refers to the quantitative weight value assigned to each characteristic pollutant in the fingerprint index. This weight is determined based on the identifiability of the characteristic pollutant in identifying pollution sources and the closeness of its association with corresponding production activities. A higher weight value indicates a more significant role for the characteristic pollutant in characterizing the pollution source. The duration of historical production activities refers to the continuous time span of industrial production activities corresponding to each potential pollution source type within the target site, i.e., the duration from the start to the cessation of production activities. This duration directly affects the accumulation of pollutants within the site. The correction coefficient is used to quantify the time decay effect of emission behavior for different potential pollution source types.

[0054] Specifically, firstly, the association weights of chromium-nickel-zinc characteristic pollutants corresponding to the electroplating plant and the association weights of lead-petroleum hydrocarbon characteristic pollutants corresponding to the waste recycling station were extracted from the fingerprint indicators. The comprehensive source strength quantification values ​​of the electroplating plant and the waste recycling station were then summarized, with the electroplating plant having a higher comprehensive source strength quantification value. The historical production activities of the site were reviewed. The electroplating plant operated from 1999 to 2017, a period of 18 years, while the waste recycling station operated from 2017 to 2020, a period of 3 years. The lengths of the two periods were standardized and quantified. The comprehensive source strength quantification values ​​of the electroplating plant and the waste recycling station were weighted with their respective standardized production period lengths. The calculated source strength-time cumulative correction coefficient of the electroplating plant was much higher than that of the waste recycling station. This quantifies that the electroplating plant, due to its longer production period, has a more significant time decay effect on pollutants emitted, while the waste recycling station, due to its shorter production period, has a relatively weaker time decay effect.

[0055] Step S203 involves dimensional fusion of the coupling weight factor, correction coefficient, multi-source data, and fingerprint index to form an enhanced input dataset.

[0056] It should be noted that the enhanced input dataset refers to a dataset that integrates all core features of coupling weight factors, correction coefficients, multi-source data, and fingerprint indicators after dimensionality fusion. Compared with single data types, it contains richer information such as pollution migration, source strength time series, site medium, and pollution source characteristics, which can provide more comprehensive and accurate input support for model calculation.

[0057] Further, step S203 includes: extracting pollutant concentration data from multi-source data, grouping them according to different spatial regions and depth layers to obtain pollutant concentration subsets corresponding to each region-depth layer; multiplying the coupling weight factor and the soil porosity ratio data of the corresponding region-depth layer element-by-element to obtain media feature data; weighting and summing the correlation weights of each characteristic pollutant in the correction coefficient and fingerprint index to obtain fingerprint feature data; dimensionally aligning the pollutant concentration subsets, media feature data, and fingerprint feature data to obtain aligned pollutant concentration subsets, aligned media feature data, and aligned fingerprint feature data; weighting the aligned pollutant concentration subsets, aligned media feature data, and aligned fingerprint feature data based on the regional pollution sensitivity coefficient to obtain weighted pollutant concentration subsets, weighted media feature data, and weighted fingerprint feature data; and concatenating the weighted pollutant concentration subsets, weighted media feature data, and weighted fingerprint feature data to form an enhanced input dataset.

[0058] It should be noted that the pollutant concentration subset refers to the set of pollutant concentration data corresponding to a single spatial area and a single depth layer combination unit of the target site, extracted from the pollutant concentration data of multi-source data. Media characteristic data refers to the quantitative data obtained by multiplying the coupling weight factor and soil porosity data element-by-element. It can characterize the adaptation relationship between soil media characteristics and pollutant migration efficiency in different regions and depth layers. Media characteristic data is used to reflect the site media adaptability of pollution migration. Fingerprint characteristic data refers to the quantitative data obtained by weighted summation of correction coefficients and characteristic pollutant association weights. Fingerprint characteristic data is used to highlight the historical emission cumulative effect of different pollution sources. The historical emission cumulative effect refers to the pollution impact effect formed by the continuous deposition and accumulation of pollutants within the site during the continuous emission period of each potential pollution source. The longer the emission period and the higher the characteristic pollutant association weight, the more significant this effect. The regional pollution sensitivity coefficient refers to the quantitative weight value set according to the pollution risk characteristics of each area of ​​the target plot. It is set according to whether the corresponding area is a historical production core area and whether there are underground pipeline leakage points. The closer to the historical production core area, the greater the regional pollution sensitivity coefficient. The closer to the underground pipeline leakage point, the greater the regional pollution sensitivity coefficient.

[0059] Specifically, firstly, the concentration data of pollutants such as cadmium, lead, and chromium were extracted from the multi-source data of the site. These were then grouped according to spatial grids and depth layers to obtain a subset of pollutant concentrations corresponding to each grid-depth layer unit. This grouping method can more precisely reflect the distribution characteristics of pollutants within the site, providing basic data for subsequent analysis. Next, the soil porosity data of each unit was extracted and multiplied element-by-element with the previously constructed media permeability-water flow orientation coupling weighting factor to obtain media characteristic data that reflects the site's media adaptability. The media characteristic data values ​​of the grid units surrounding the electroplating workshop were significantly higher. The coupling weighting factor reflects the difference in pollutant migration efficiency, while the soil porosity data reflects the adaptability of the site's media. Through multiplication, the media characteristic data can quantify the ease of pollutant migration in different areas and depth layers, providing a basis for site media-related aspects of pollution liability delineation. Then, the source strength-time cumulative correction coefficient was weighted and summed with the characteristic pollutant association weights of chromium-nickel-zinc and lead-petroleum hydrocarbons to obtain fingerprint characteristic data. The fingerprint characteristic data corresponding to the electroplating plant had higher values ​​due to its longer production period. The correction coefficient considers the time decay effect of historical emissions, while the correlation weight of the fingerprint index reflects the characteristics of different pollution sources. The weighted sum of fingerprint feature data can highlight the cumulative effect of historical emissions from different pollution sources, providing key information for identifying major pollution sources. Subsequently, the pollutant concentration subset, media characteristic data, and fingerprint feature data are dimensionally aligned according to the grid-depth layer standard to ensure complete consistency in the two-dimensional division of the three types of data. The aligned data can more accurately reflect the interrelationships between various factors, providing a unified data format for subsequent comprehensive analysis. Next, high regional pollution sensitivity coefficients are set for historical production core areas such as electroplating workshops and waste dismantling areas, as well as areas with underground pipeline leakage points, while low coefficients are set for edge areas. The coefficients are weighted and calculated with the three types of aligned data to strengthen the data characteristics of high-sensitivity areas. Through weighted calculation, weighted pollutant concentration subsets, media characteristic data, and fingerprint feature data are obtained, which can more accurately reflect the pollution characteristics and responsibility attribution of each area. Finally, the weighted data from the three categories are concatenated according to feature dimensions. Each grid-depth layer unit integrates information such as pollution concentration, site medium, and pollution source characteristics, ultimately forming an enhanced input dataset for the target site. This enhanced input dataset integrates information on pollutant concentration, site medium suitability, and historical emission cumulative effects, providing a more comprehensive and accurate input for the pre-defined pollution contribution calculation model, thus helping to improve the accuracy and reliability of pollution liability delineation.

[0060] Step S204: Input the enhanced input dataset into the preset pollution contribution calculation model and output the pollution contribution ratio of each potential pollution source type in the corresponding spatial region and depth layer.

[0061] It should be noted that a spatial region refers to an independent horizontal spatial unit formed by a pre-defined planar grid based on the complexity of pollution within a site. Each unit corresponds to a clearly defined geographical area and serves as the spatial dimension for calculating the pollution contribution ratio. A depth layer refers to an independent vertical layer formed by a pre-defined soil layer thickness based on soil sampling depth intervals. Each layer corresponds to a clearly defined depth range and serves as the vertical dimension for calculating the pollution contribution ratio. The pollution contribution ratio refers to the proportion of contribution made by each type of potential pollution source to the pollution formation of a single spatial region—depth layer unit—within the target site. It is the core value for quantifying pollution source responsibility and directly reflects the degree of pollution impact of each pollution source on different areas and depths.

[0062] Further, step S204 includes: extracting the pollutant concentration variation coefficient from multi-source data through a two-dimensional segmentation module of a preset pollution contribution calculation model; determining the pollution complexity of a plot based on the pollutant concentration variation coefficient, wherein in areas with high pollution complexity, the grid side length of the preset planar grid is reduced to a preset short side length, while in areas with low pollution complexity, the preset long side length is maintained. The preset short side length is set based on the minimum identification unit of the high-variable pollution area, and the preset long side length is set based on the grid density of a conventional pollution survey; determining the thickness of the basic soil layer based on the soil sampling depth interval through the two-dimensional segmentation module, wherein for depth segments with continuous pollution halos, the preset soil layer thickness is reduced to half the thickness of the basic soil layer, and for depth segments where the pollution concentration drops sharply, the preset soil layer thickness is increased to twice the thickness of the basic soil layer; in In the dual-dimensional partitioning module, a grid-soil layer association index table is established to record the coordinate mapping relationship between each spatial region grid and the corresponding depth layer. Data from the same grid-soil layer unit in the enhanced input dataset are integrated into independent samples, and index labels are embedded in the independent samples to obtain index-labeled sample data, where the index labels include spatial region number and depth layer number. The index-labeled sample data is input into a preset pollution contribution calculation model. The partitioning rules of the dual-dimensional partitioning module are called through the index labels, and the migration path constraints of the corresponding region-depth layer are loaded for calculation to obtain the pollution contribution ratio. Based on the pollution contribution ratio, the corresponding spatial region and depth layer are back-mapped according to the index labels, and the pollution contribution ratio of each potential pollution source type in different spatial regions and depth layers of the target plot is summarized.

[0063] It should be noted that the pollutant concentration variation coefficient is a quantitative indicator characterizing the dispersion of pollutant concentrations within a target site. It is calculated as the ratio of the standard deviation to the average value of pollutant concentrations. A higher value indicates greater fluctuations in pollutant concentrations and more uneven pollution distribution in the area. Site pollution complexity is an overall indicator comprehensively reflecting the degree of pollutant concentration variation, the uniformity of pollution distribution, and the diversity of pollution types within a target site. The core criterion is the pollutant concentration variation coefficient; a higher coefficient indicates higher site pollution complexity. The preset planar grid refers to the basic grid structure pre-defined for the horizontal spatial division of the target site. It serves as the benchmark for dividing spatial areas, and its grid side length can be dynamically adjusted according to the site pollution complexity. The preset short side length refers to the planar grid side length set for areas with high pollution complexity. It is determined based on the smallest identification unit of highly variable pollution areas, enabling refined spatial division of highly complex pollution areas. The smallest identification unit of highly variable pollution areas refers to the smallest horizontal spatial unit capable of accurately identifying areas with rapidly changing pollutant concentrations. It is the smallest scale for dividing highly complex pollution areas into grids, ensuring that pollution details in this area can be effectively captured. The preset long side length refers to the side length of the planar grid set for areas with low pollution complexity. It is determined based on the grid density of conventional pollution surveys, improving the efficiency of spatial division while ensuring pollution identification effectiveness. Conventional pollution survey grid density refers to the grid layout density standard commonly used in industrial site pollution surveys. It serves as the basis for dividing low-pollution-complexity areas into grids and conforms to industry standard survey practices. A continuous pollution halo refers to an area where pollutants are continuously distributed within a certain vertical depth range of the target site, with no significant sudden drop in concentration and a continuous pollution range. An independent sample refers to a single data sample formed by integrating all feature data belonging to the same spatial area grid-depth layer combination unit from the enhanced input dataset. It contains all pollution-related feature information of that two-dimensional unit.

[0064] Specifically, firstly, the concentration variation coefficients of pollutants such as cadmium, lead, and chromium were extracted from the multi-source data of the site using the model's two-dimensional partitioning module. These variation coefficients reflect the spatial and depth-related variations in pollutant concentrations, providing a basis for subsequent grid partitioning and soil layer thickness adjustment.

[0065] Next, the area surrounding the electroplating workshop and the waste dismantling area was identified as having high pollution complexity, while the edge of the site was identified as having low pollution complexity. The pre-set planar grid side length for high-complexity areas was reduced to a pre-set short side length based on the minimum identification unit for high-variability pollution areas, while the pre-set long side length for low-complexity areas remained based on the grid density set for conventional pollution surveys. Then, the thickness of the basic soil layer was determined based on the soil sampling depth interval using a two-dimensional partitioning module. For the 3-6 meter continuous pollution halo depth range, the soil layer thickness was reduced to half the basic thickness; for the 8-10 meter depth range where pollution concentration drops sharply, the soil layer thickness was expanded to twice the basic thickness. This flexible adjustment of the soil layer thickness better adapts to the pollution characteristics at different depths. Subsequently, a grid-soil layer association index table was established in the two-dimensional partitioning module to record the coordinate mapping relationship between each spatial grid and the depth layer. This index table provides a clear spatial and depth correspondence for subsequent data integration and model calculations, ensuring data accuracy and consistency. Next, feature data from the same grid-soil unit in the enhanced input dataset are integrated into independent samples, and index labels containing spatial region numbers and depth layer numbers are embedded to obtain indexed sample data. These index labels, including spatial region numbers and depth layer numbers, provide clear regional and depth information for model calculation, facilitating the model's accurate invocation of corresponding partitioning rules and migration path constraints during computation. Then, this sample data is input into the extreme gradient boosting model, which invokes the two-dimensional partitioning rules through the index labels and loads the corresponding region-depth layer migration path constraints for calculation, obtaining the pollution contribution ratio of each sample. This process utilizes the model's two-dimensional partitioning capability and migration path constraints to ensure the scientific validity and accuracy of the calculation results. Finally, the pollution contribution ratio is back-mapped to the corresponding spatial region and depth layer according to the index labels, summarizing the pollution contribution ratios of electroplating plants and waste recycling stations in different spatial regions and depth layers of the site. For example, the electroplating plant contributes approximately 80% to lead pollution in the deep 3-6 meter layer. This reverse mapping process combines the quantitative results output by the model with specific spatial and depth locations, providing a clear and intuitive basis for the division of pollution liability and enabling a refined division of pollution liability for complex industrial sites.

[0066] This embodiment extracts the correlation data between soil porosity and permeability coefficient from multi-source data, and combines it with groundwater flow direction information in the migration path constraints to construct a coupling weight factor to correct for differences in pollutant migration efficiency in different spatial regions. Simultaneously, based on the correlation weights of characteristic pollutants in the fingerprint index and the duration of historical production activities, a source strength-time accumulation correction coefficient is calculated to quantify the time decay effect of emission behavior of different potential pollution source types. Through the dimensional fusion of the coupling weight factor, correction coefficient, multi-source data, and fingerprint index, an enhanced input dataset is formed and input into a preset pollution contribution calculation model, outputting the pollution contribution ratio of each potential pollution source type in the corresponding spatial region and depth layer. The refined construction of weight factors and correction coefficients significantly improves the accuracy of pollution source identification and the scientific nature of responsibility allocation, providing strong support for pollution control in complex industrial sites.

[0067] Based on the first embodiment of this application, this application also provides a device for determining pollutant liability in complex industrial sites. Please refer to... Figure 3 The device includes: The acquisition module 10 is used to acquire multi-source data within the target plot. The multi-source data includes soil physicochemical data, spatial covariate data, pollutant concentration data, geographic data, and resistivity anomaly area information.

[0068] The identification module 20 is used to construct fingerprint indicators based on multi-source data to identify potential pollution source types.

[0069] Module 30 is used to establish migration path constraints by combining information on the site's hydrogeological structure and historical production activities.

[0070] The calculation module 40 is used to input multi-source data, fingerprint indicators and migration path constraints into a preset pollution contribution calculation model, and output the pollution contribution ratio of each potential pollution source type in different spatial regions and depth layers. The preset pollution contribution calculation model embeds a two-dimensional partitioning module of spatial region-depth layer. The two-dimensional partitioning module divides the target plot into different spatial regions according to a preset planar grid and into different depth layers according to a preset soil layer thickness. The preset planar grid is set according to the pollution complexity of the plot, and the preset soil layer thickness is set according to the soil sampling depth interval.

[0071] Result module 50 is used to generate pollution liability allocation results based on the proportion of pollution contribution.

[0072] The pollutant liability allocation device for complex industrial sites provided in this application, employing the pollutant liability allocation method for complex industrial sites described in the above embodiments, can solve the technical problem of how to improve the accuracy of quantifying pollution liability for complex industrial sites. Compared with the prior art, the beneficial effects of the pollutant liability allocation device for complex industrial sites provided in this application are the same as those of the pollutant liability allocation method for complex industrial sites provided in the above embodiments, and other technical features in the pollutant liability allocation device for complex industrial sites are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0073] This application provides a pollutant liability delineation device for complex industrial sites. The pollutant liability delineation device for complex industrial sites includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the pollutant liability delineation method for complex industrial sites in the above embodiment 1.

[0074] The following is for reference. Figure 4 This document illustrates a structural schematic diagram of a contaminant liability apportionment device suitable for implementing embodiments of this application in complex industrial sites. The contaminant liability apportionment device for complex industrial sites in embodiments of this application may include, but is not limited to, mobile terminals such as laptops, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), and vehicle-mounted terminals (e.g., vehicle navigation terminals), as well as fixed terminals such as digital TVs and desktop computers. Figure 4 The pollutant liability delineation device shown for complex industrial sites is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments in this application.

[0075] like Figure 4As shown, the contaminant liability delineation equipment for complex industrial sites may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage device 1003 into a random access memory (RAM) 1004. The RAM 1004 also stores various programs and data required for the operation of the contaminant liability delineation equipment for complex industrial sites. The processing unit 1001, ROM 1002, and RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following can be connected to I / O interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. Communication device 1009 allows the contaminant liability allocation equipment in complex industrial sites to exchange data with other devices wirelessly or via wired communication. Although contaminant liability allocation equipment with various complex industrial sites is shown in the figures, it should be understood that implementation or possession of all shown is not required. More or fewer may be implemented alternatively.

[0076] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from ROM 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0077] The pollutant liability allocation device for complex industrial sites provided in this application, employing the pollutant liability allocation method for complex industrial sites described in the above embodiments, can solve the technical problem of how to improve the accuracy of quantifying pollution liability for complex industrial sites. Compared with the prior art, the beneficial effects of the pollutant liability allocation device for complex industrial sites provided in this application are the same as those of the pollutant liability allocation method for complex industrial sites provided in the above embodiments, and other technical features of the pollutant liability allocation device for complex industrial sites are the same as those disclosed in the previous embodiment method, and will not be repeated here.

[0078] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0079] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0080] This application provides a computer-readable medium having computer-readable program instructions (i.e., a computer program) stored thereon, the computer-readable program instructions being used to perform the pollutant liability delineation method for complex industrial sites in the above embodiments.

[0081] The computer-readable medium provided in this application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor devices, or any combination thereof. More specific examples of computer-readable media may include, but are not limited to: electrical connections with one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable medium may be any tangible medium containing or storing a program that can be executed by instructions, used by a device, or used in conjunction with it. The program code contained on the computer-readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0082] The aforementioned computer-readable medium may be included in a pollutant liability delineation device for complex industrial sites; or it may exist independently and not be installed in a pollutant liability delineation device for complex industrial sites.

[0083] The aforementioned computer-readable medium carries one or more programs that, when executed by a contaminant liability apportionment device for complex industrial sites, enable the device to write computer program code for performing the operations of this application in one or more programming languages ​​or a combination thereof. These programming languages ​​include object-oriented programming languages—such as Java and C++—and conventional procedural programming languages—such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0084] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of methods and computer program products according to various embodiments of this application. In this regard, all blocks in the flowcharts or block diagrams may represent a module, segment, or portion of code containing one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that all blocks in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using dedicated hardware-based implementations that perform the specified functions or operations, or using a combination of dedicated hardware and computer instructions.

[0085] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0086] The readable medium provided in this application is a computer-readable medium that stores computer-readable program instructions (i.e., a computer program) for executing the above-described method for allocating pollutant liability for complex industrial sites. This method addresses the technical problem of improving the accuracy of quantifying pollution liability for complex industrial sites. Compared to the prior art, the beneficial effects of the computer-readable medium provided in this application are the same as those of the pollutant liability allocation method for complex industrial sites provided in the above embodiments, and will not be elaborated upon here.

[0087] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the above-described method for assigning responsibility for pollutants at complex industrial sites.

[0088] The computer program product provided in this application solves the technical problem of how to improve the accuracy of quantifying pollution liability for complex industrial sites. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as those of the pollutant liability allocation method for complex industrial sites provided in the above embodiments, and will not be repeated here.

[0089] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A method for allocating responsibility for pollutants in complex industrial sites, characterized in that, The method includes: Acquire multi-source data within the target site, including soil physicochemical data, spatial covariate data, pollutant concentration data, geographic data, and resistivity anomaly area information; Fingerprint indicators are constructed based on the multi-source data to identify potential pollution source types; Establish migration path constraints by combining information on the site's hydrogeological structure and historical production activities; The multi-source data, the fingerprint index, and the migration path constraints are input into a preset pollution contribution calculation model, which outputs the pollution contribution ratio of each potential pollution source type in different spatial regions and depth layers. The preset pollution contribution calculation model embeds a two-dimensional partitioning module of spatial region and depth layer. The two-dimensional partitioning module divides the target plot into different spatial regions according to a preset planar grid and into different depth layers according to a preset soil layer thickness. The preset planar grid is set according to the pollution complexity of the plot, and the preset soil layer thickness is set according to the soil sampling depth interval. The pollution liability allocation results are generated based on the pollution contribution ratio.

2. The method as described in claim 1, characterized in that, The step of inputting the multi-source data, the fingerprint index, and the migration path constraints into a preset pollution contribution calculation model, and outputting the pollution contribution ratio of each potential pollution source type in different spatial regions and depth layers, includes: The correlation data between soil porosity and permeability coefficient are extracted from the multi-source data. Combined with the groundwater flow direction information in the migration path constraints, a coupling weight factor of medium permeability-water flow orientation is constructed. The coupling weight factor is used to correct the differences in pollutant migration efficiency in different spatial regions. Based on the association weights of characteristic pollutants in the fingerprint index, and combined with the time period of historical production activities, a correction coefficient for source strength-time accumulation is calculated, wherein the correction coefficient is used to quantify the time decay effect of emission behavior of different potential pollution source types. The coupling weight factor, the correction coefficient, the multi-source data, and the fingerprint index are dimensionally fused to form an enhanced input dataset; The enhanced input dataset is input into a preset pollution contribution calculation model, which outputs the pollution contribution ratio of each potential pollution source type in the corresponding spatial region and depth layer.

3. The method as described in claim 2, characterized in that, The step of dimensionally fusing the coupling weight factor, the correction coefficient, the multi-source data, and the fingerprint index to form an enhanced input dataset includes: Extract pollutant concentration data from the multi-source data, group them according to different spatial regions and depth layers, and obtain the pollutant concentration subsets corresponding to each region-depth layer; The coupling weight factor and the soil porosity data of the corresponding region-depth layer are multiplied element by element to obtain the medium characteristic data, which is used to reflect the site medium adaptability of pollution migration. The correction coefficient and the correlation weight of each characteristic pollutant in the fingerprint index are weighted and summed to obtain fingerprint feature data, which is used to highlight the historical emission cumulative effect of different pollution sources. The pollutant concentration subset, the medium feature data, and the fingerprint feature data are dimensionally aligned to obtain aligned pollutant concentration subset, aligned medium feature data, and aligned fingerprint feature data. The aligned pollutant concentration subset, aligned medium feature data, and aligned fingerprint feature data are weighted and calculated based on the regional pollution sensitivity coefficient to obtain the weighted pollutant concentration subset, weighted medium feature data, and weighted fingerprint feature data. The regional pollution sensitivity coefficient is set according to whether the corresponding region is a historical production core area and whether there are underground pipeline leakage points. The weighted pollutant concentration subset, the weighted medium feature data, and the weighted fingerprint feature data are concatenated to form an enhanced input dataset.

4. The method as described in claim 2, characterized in that, The step of inputting the enhanced input dataset into a preset pollution contribution calculation model and outputting the pollution contribution ratio of each potential pollution source type in the corresponding spatial region and depth layer includes: The coefficient of variation of pollutant concentration in multi-source data is extracted by the two-dimensional partitioning module of the preset pollution contribution calculation model. The pollution complexity of a site is determined based on the pollutant concentration variation coefficient. In areas with high pollution complexity, the grid side length of the preset planar grid is reduced to a preset short side length, while in areas with low pollution complexity, the preset long side length is maintained. The preset short side length is set according to the minimum identification unit of the high pollution variation area, and the preset long side length is set according to the grid density of the conventional pollution survey. The thickness of the basic soil layer is determined by the dual-dimensional segmentation module based on the soil sampling depth interval. For depth segments with continuous pollution halos, the preset soil layer thickness is reduced to half of the thickness of the basic soil layer. For depth segments with a sharp drop in pollution concentration, the preset soil layer thickness is increased to twice the thickness of the basic soil layer. In the dual-dimensional partitioning module, a grid-soil layer association index table is established to record the coordinate mapping relationship between each spatial region grid and the corresponding depth layer; The data of the same grid-soil layer unit in the enhanced input dataset are integrated into independent samples and index labels are embedded in the independent samples to obtain sample data with index labels, wherein the index labels include spatial region number and depth layer number; The sample data with index labels is input into a preset pollution contribution calculation model. The division rules of the two-dimensional partitioning module are called through the index labels, and the migration path constraints of the corresponding region-depth layer are loaded for calculation to obtain the pollution contribution ratio. Based on the pollution contribution ratio, the corresponding spatial regions and depth layers are back-mapped according to the index labels, and the pollution contribution ratios of each potential pollution source type in different spatial regions and depth layers of the target plot are summarized.

5. The method as described in claim 1, characterized in that, The step of constructing fingerprint indicators based on the multi-source data to identify potential pollution source types includes: Soil physicochemical data, pollutant concentration data, and spatial covariate data are extracted from the multi-source data to screen out pollutant combinations corresponding to industrial production activities, wherein the pollutant combinations include chromium-nickel-zinc combinations and lead-petroleum hydrocarbon combinations. The concentration ratio of pollutants in each pollutant combination is calculated based on the constant material ratio of different industrial production activities. The pollutant combination whose concentration ratio fluctuation is lower than a preset fluctuation threshold is identified as a candidate fingerprint feature. The candidate fingerprint features are spatially anchored by combining the resistivity anomaly region information of the multi-source data to obtain the spatial anchoring result; The candidate fingerprint features are verified for temporal validity based on the historical production activity period and pollutant degradation cycle to obtain valid candidate fingerprint features. The valid candidate fingerprint features are de-mixed to obtain the de-mixed fingerprint features; The demixed fingerprint features are compared with a preset industrial pollution source fingerprint database to calculate the feature matching degree; By combining the spatial anchoring results and feature matching degree, the potential pollution source types in each region are determined.

6. The method as described in claim 1, characterized in that, The steps for establishing migration path constraints by combining information on the site's hydrogeological structure and historical production activities include: The core parameters of the site's hydrogeological structure are extracted, and a three-dimensional hydrogeological model is constructed. The core parameters include groundwater flow velocity, aquifer thickness, aquitard depth, and lithological permeability coefficient. Based on the three-dimensional hydrogeological model combined with historical production activity information, the location of pollution sources, emission methods, and pollutant outflow periods are inferred to obtain the theoretical migration range; Based on resistivity anomaly area information from multi-source data, underground hidden passages are identified, and the routes along these underground hidden passages are set as priority migration paths. Calculate the pollutant concentration at each sampling point and the distance attenuation coefficient of the pollution source within the theoretical migration range, and obtain the spatial correlation verification results based on the distance attenuation coefficient and the three-dimensional hydrogeological model; Based on the theoretical migration range, preferred migration paths, and spatial correlation verification results, valid migration paths are confirmed, and migration path constraints are established, including spatial range, time window, and migration channel.

7. The method as described in claim 1, characterized in that, The step of generating pollution liability allocation results based on the pollution contribution ratio includes: The pollution contribution ratios of each potential pollution source type are aggregated according to a spatial grid to generate a three-dimensional responsibility heat map; Regions in the responsibility heatmap whose contribution ratio is higher than a preset high confidence threshold are identified and marked as dominant responsibility regions. Regions in the responsibility heatmap whose contribution ratio is lower than a preset low confidence threshold are identified and marked as non-responsibility regions. Identify areas in the responsibility heatmap where the contribution ratio is higher than the low confidence threshold and lower than the high confidence threshold, and mark them as areas of undetermined responsibility. Output the pollution liability allocation results, which include spatial location, pollution source type, pollution contribution ratio, and liability level.

8. A device for assigning responsibility for pollutants in complex industrial sites, characterized in that, The device includes: The acquisition module is used to acquire multi-source data within the target plot, including soil physicochemical data, spatial covariate data, pollutant concentration data, geographic data, and resistivity anomaly area information; The identification module is used to construct fingerprint indicators based on the multi-source data to identify the types of potential pollution sources; The module is used to establish migration path constraints by combining information on the site's hydrogeological structure and historical production activities. The calculation module is used to input the multi-source data, the fingerprint index, and the migration path constraints into a preset pollution contribution calculation model, and output the pollution contribution ratio of each potential pollution source type in different spatial regions and depth layers. The preset pollution contribution calculation model embeds a two-dimensional partitioning module of spatial region and depth layer. The two-dimensional partitioning module divides the target plot into different spatial regions according to a preset planar grid and into different depth layers according to a preset soil layer thickness. The preset planar grid is set according to the pollution complexity of the plot, and the preset soil layer thickness is set according to the soil sampling depth interval. The results module is used to generate pollution liability allocation results based on the pollution contribution ratio.

9. A device for assigning responsibility for pollutants in complex industrial sites, characterized in that, The device includes: a memory, a processor, and a pollutant liability assignment program for complex industrial sites stored in the memory and running on the processor, the pollutant liability assignment program for complex industrial sites being configured to implement the steps of the pollutant liability assignment method for complex industrial sites as described in any one of claims 1-7.

10. A storage medium, characterized in that, The storage medium stores a pollutant liability allocation program for complex industrial sites, which, when executed by a processor, implements the steps of the pollutant liability allocation method for complex industrial sites as described in any one of claims 1-7.

Images