Fluorine chemical industry park PFAS prediction system based on multi-medium dynamic distribution and collaborative verification method

Abstract

The invention discloses a fluorine chemical industry park PFAS prediction system based on multi-medium dynamic distribution and a collaborative verification method, and relates to the technical field of environmental monitoring and pollution control, and the system comprises a fluorine chemical industry data acquisition module, a fluorine chemical industry feature extraction module, a cross-medium association distribution module, a PFAS concentration risk prediction module and a PFAS visual display module. The fluorine chemical industry data acquisition module is used for acquiring and preprocessing chemical oxygen demand data and fluorine ion concentration data in a water body at a discharge outlet of a fluorine chemical industry park, and sampling and detecting PFAS concentration in the water body, sediments and aquatic organisms. By constructing the multi-medium dynamic distribution PFAS prediction system and the cooperative verification method, the PFAS pollution monitoring and early warning capability of the fluorine chemical industry park is effectively improved, accurate prediction of the PFAS concentration and rapid identification of the pollution source are realized, the detection cost is reduced, the real-time performance and the dynamic early warning level are improved, a scientific basis is provided for environmental pollution control, and the method is suitable for popularization and application. The development of the environment-friendly fluorine chemical industry is promoted.

Description

Technical Field

[0001] The present invention relates to the technical field of environmental monitoring and pollution control, and in particular to a PFAS prediction system and collaborative verification method for a fluorine chemical park based on multi-media dynamic allocation. Background Art

[0002] PFAS, or per- and polyfluoroalkyl substances, are a family of unnatural, synthetic organic compounds composed primarily of carbon and fluorine atoms. They possess exceptionally high thermal and chemical stability, leading to their widespread use in industrial and consumer applications, such as textiles, surfactants, food packaging, non-stick coatings, and firefighting foams. PFAS are not readily biodegradable and can accumulate in the environment and in organisms. Exposure to certain types of PFAS can have serious health consequences, including endocrine disruption, cancer, thyroid disease, and liver and kidney damage. PFAS also encompass a variety of specific compounds, such as perfluorooctane sulfonic acid (PFOS), perfluorooctanoic acid (PFOA), and perfluorohexane sulfonic acid (PFHEXA). These compounds are found at high concentrations and rates in the environment and exhibit varying toxic effects on organisms. Due to global concern regarding the long-term exposure and potential health risks of PFAS, many countries have implemented strict regulations and restrictions on PFAS. Therefore, understanding the properties, applications, and health impacts of PFAS is crucial for protecting human health and environmental safety.

[0003] In order to solve the current problems of PFAS, such as high detection cost, unrealized multi-media dynamic distribution prediction, and insufficient real-time and dynamic early warning capabilities, the existing technology mainly relies on traditional laboratory detection methods to monitor the distribution of PFAS in different environmental media through regular sampling and analysis. However, this method is not only time-consuming and labor-intensive, but also costly, and it is difficult to achieve real-time monitoring and dynamic early warning of PFAS pollution. In addition, due to the complex behavior of PFAS in the environment and the dynamic distribution between multiple media, traditional prediction methods often find it difficult to accurately predict its migration and transformation laws, which in turn makes it impossible to effectively guide PFAS pollution control and risk management. In order to overcome these limitations, a PFAS prediction system and collaborative verification method based on multi-media dynamic distribution for fluorine chemical parks are proposed. Summary of the Invention

[0004] The purpose of the present invention is to provide a PFAS prediction system and collaborative verification method for a fluorine chemical park based on multi-media dynamic allocation to solve the problems raised in the above background technology.

[0005] To solve the above technical problems, the technical solutions adopted by the present invention are as follows: First, a PFAS prediction system for a fluorine chemical park based on multi-media dynamic allocation includes a fluorine chemical data acquisition module, a fluorine chemical feature extraction module, a cross-media association allocation module, a PFAS concentration risk prediction module, and a PFAS visualization display module;

[0006] The fluorine chemical industry data acquisition module collects and pre-processes chemical oxygen demand data and fluoride ion concentration data in the water bodies at the fluorine chemical industry park discharge outlet, and samples and detects PFAS concentrations in water bodies, sediments and aquatic organisms;

[0007] The fluorine chemical feature extraction module extracts fluorine chemical features from the pre-processed chemical oxygen demand data and fluoride ion concentration data;

[0008] The cross-media correlation allocation module combines the extracted fluorine chemical characteristics to construct a cross-media correlation model and a multi-media dynamic allocation model to predict the concentration of PFAS in water, sediments and aquatic organisms;

[0009] The PFAS concentration risk prediction module analyzes whether the predicted PFAS concentration in each medium deviates from a preset reference value based on the predicted concentration of PFAS in water, sediment, and aquatic organisms;

[0010] The PFAS visualization display module visualizes the spatial distribution of PFAS concentrations in water, sediments, and aquatic organisms, and compares historical predictions with actual detection data.

[0011] Secondly, the PFAS prediction system and collaborative verification method for fluorine chemical parks based on multi-media dynamic allocation include the following steps:

[0012] S1. Collect and pre-process chemical oxygen demand data and fluoride ion concentration data from the fluorine chemical park discharge outlet, and sample and test PFAS concentrations in water, sediments, and aquatic organisms;

[0013] S2, extracting fluorine chemical characteristics based on pre-processed chemical oxygen demand data and fluoride ion concentration data;

[0014] S3. Combine the extracted fluorine chemical characteristics to construct a cross-media association model and a multi-media dynamic distribution model to predict the concentrations of PFAS in water, sediments, and aquatic organisms respectively;

[0015] S4. Based on historical data on PFAS concentrations in water, sediments, and aquatic organisms, preset reference values ​​for PFAS concentrations in each medium and compare them with the predicted concentrations of PFAS in water, sediments, and aquatic organisms to analyze whether there are any deviations from the reference values;

[0016] S5. Visualize the spatial distribution of PFAS concentrations in water, sediments, and aquatic organisms, and compare historical predictions with actual detection data.

[0017] A further improvement of the technical solution of the present invention is that in S1, the collection and preprocessing process of the chemical oxygen demand data and fluoride ion concentration data in the water body at the fluorine chemical park outlet includes:

[0018] An online COD analyzer using ultraviolet spectroscopy and a fluoride ion selective electrode sensor were deployed in parallel in the monitoring wells of the main channel of the fluorine chemical park's discharge outlet. The ultraviolet light source emitted a 254nm wavelength beam that penetrated the water sample, and the beam attenuation intensity was detected. Based on the Lambert-Beer law, chemical oxygen demand data was collected from the water body at the fluorine chemical park's discharge outlet. The fluoride ion selective membrane on the electrode surface of the fluoride ion selective electrode sensor produced a potential response with the fluoride ions in the water sample, and the fluoride ion concentration data was obtained using the Nernst equation.

[0019] The 3σ criterion was used to identify outliers in the hourly collected chemical oxygen demand (COD) and fluoride ion concentration (FID) data. Data points outside the range of ±3 times the standard deviation of the mean were marked as invalid and discarded. Missing data for no more than 2 consecutive hours were filled using linear interpolation of adjacent time points. The collected COD and FID data were scaled to the [0, 1] range using Min-Max normalization.

[0020] A further improvement of the technical solution of the present invention is that in S1, the process of sampling and detecting the concentration of PFAS in water, sediment and aquatic organisms includes:

[0021] Fully automatic water samplers, gravity-fed sediment samplers, and trawl-type biological samplers are deployed downstream of the fluorine chemical park's discharge outlet. Monthly, they collect samples of 1 liter of water, 500 grams of sediment, and 50-100 grams of individual aquatic organisms, respectively. PFAS are detected using a liquid chromatography-mass spectrometry instrument equipped with a solid-phase extraction cartridge and ion source. Prior to testing, water samples are fixed with nitric acid, sediments are ultrasonically extracted with acetonitrile, and biological samples are extracted with accelerated solvents.

[0022] In mass spectrometry detection, water samples, sediment samples and aquatic organism samples are separated by liquid chromatography and then enter the electrospray ion source to generate PFAS parent ions in positive ion mode. The multiple reaction monitoring mode is used, and a specific mass-to-charge ratio is set. The PFAS parent ion intensity is recorded by the mass spectrometer, and the PFAS concentration in water, sediment and aquatic organisms is calibrated based on the internal standard method.

[0023] A further improvement of the technical solution of the present invention is that in S2, the process of extracting fluorine chemical characteristics based on the pre-processed chemical oxygen demand data and fluoride ion concentration data includes:

[0024] Fluorine chemical characteristics include the dynamic accumulation index of chemical oxygen demand and the fluctuation rate of fluoride ion concentration;

[0025] Based on the pre-processed hourly chemical oxygen demand data, with a 24-hour sliding window period, the arithmetic mean and standard deviation of the chemical oxygen demand data within the window are calculated, and the dynamic cumulative index of chemical oxygen demand is obtained by adding the mean and twice the standard deviation;

[0026] The mean and standard deviation of the pretreated fluoride ion concentration data were calculated on a monthly basis, and the fluctuation rate of fluoride ion concentration was obtained using the coefficient of variation formula.

[0027] A further improvement of the technical solution of the present invention is that in S3, the process of constructing a cross-media correlation model to predict the concentration of PFAS in water includes:

[0028] A cross-media association model was constructed based on the long short-term memory network architecture. The extracted fluorine chemical characteristics were input into the input layer of the cross-media association model, and seasonal coding and real-time flow data were introduced.

[0029] The hidden layer neurons of the cross-media association model capture the dynamic correlation between fluorine chemical characteristics and PFAS concentrations in water bodies, output hidden states, and fuse seasonal coding and real-time flow data with the hidden states through the fully connected layer to output the predicted value of PFAS concentration in water bodies. The predicted value of PFAS concentration in water bodies is constrained according to the water-sediment-organism multi-media distribution law.

[0030] A further improvement of the technical solution of the present invention is that in S3, the process of constructing a multi-media dynamic distribution model to predict the concentration of PFAS in sediments and aquatic organisms includes:

[0031] The bioconcentration factor (BCF) was determined by fitting the ratio of PFAS concentrations in organisms and sediments in historical data. Combined with the predicted PFAS concentration in water bodies output by the cross-media correlation model, a multi-media dynamic distribution model was constructed based on the law of conservation of mass to calculate the PFAS concentrations in sediments and aquatic organisms.

[0032] A further improvement of the technical solution of the present invention is that: S4 specifically includes:

[0033] Based on the contribution of water, sediments, and aquatic organisms to ecological risks in historical data, historical data of equal length are analyzed, and the average values ​​of various PFAS concentrations in water, sediments, and aquatic organisms are combined to determine the reference values ​​of PFAS concentrations in water, sediments, and aquatic organisms;

[0034] Compare the predicted PFAS concentration value of each medium with its corresponding reference value, analyze the deviation rate of the PFAS concentration of each medium, and determine whether there is any abnormality in the predicted PFAS concentration value of each medium, thereby triggering independent warnings for each medium.

[0035] A further improvement of the technical solution of the present invention is that in S5, the process of visually displaying the spatial distribution of PFAS concentration and the comparison between the predicted value and the actual detection value includes:

[0036] Based on the coordinates of monitoring points in the fluorine chemical park and the predicted PFAS concentrations in corresponding water bodies, sediments, and aquatic organisms, the inverse distance weighted method was used. A hydrodynamic model was introduced to simulate the diffusion path of pollutants. The concentration field output by the hydrodynamic model was combined with inverse distance weighted interpolation to interpolate the discrete point concentration data into a continuous spatial distribution. The interpolated concentrations were mapped to color intensity according to the gradient, with red representing high concentrations, blue representing medium concentrations, and green representing low concentrations. This generated a heat map of the spatial distribution of PFAS concentrations in the park.

[0037] Use a line graph to show the comparison between historical predictions and actual detection data. Align the comprehensive risk index with the actual detection value, as well as the predicted and actual values ​​of water bodies, sediments and aquatic organisms on the monthly time axis. Connect the predicted values ​​of each month with a dotted line, with time on the horizontal axis and concentration on the vertical axis. Connect the sampling detection values ​​with a solid line and display them superimposed with the prediction curve.

[0038] Due to the adoption of the above technical solution, the present invention has the following technical advancements compared to the prior art:

[0039] 1. The present invention provides a PFAS prediction system and collaborative verification method for fluorine chemical parks based on multi-media dynamic allocation, which realizes the concentration prediction of PFAS in various environmental media, significantly improves the accuracy and real-time performance of the prediction, and provides a scientific basis for park environmental management.

[0040] 2. The present invention provides a PFAS prediction system and collaborative verification method for fluorine chemical parks based on multi-media dynamic allocation. By visually displaying the spatial distribution of PFAS concentrations and comparing predicted and actual detection data, it provides intuitive data support for environmental pollution control and governance in the park, and helps to formulate more scientific and reasonable pollution control strategies. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments described in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0042] Figure 1 Flowchart of the present invention. DETAILED DESCRIPTION

[0043] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0044] Examples, such as Figure 1 As shown, the present invention provides a PFAS prediction system for a fluorine chemical park based on multi-media dynamic allocation, including a fluorine chemical data acquisition module, a fluorine chemical feature extraction module, a cross-media association allocation module, a PFAS concentration risk prediction module, and a PFAS visualization display module;

[0045] Fluorine chemical data acquisition module, which collects and pre-processes chemical oxygen demand data and fluoride ion concentration data from the fluorine chemical park discharge outlet, and samples and detects PFAS concentrations in water, sediments, and aquatic organisms;

[0046] Fluorine chemical feature extraction module, which extracts fluorine chemical features from pre-processed chemical oxygen demand data and fluoride ion concentration data;

[0047] The cross-media correlation allocation module combines the extracted fluorine chemical characteristics to build a cross-media correlation model and a multi-media dynamic allocation model to predict the concentration of PFAS in water, sediments and aquatic organisms;

[0048] The PFAS concentration risk prediction module analyzes whether the predicted PFAS concentration in each medium deviates from the preset reference value based on the predicted concentration of PFAS in water, sediments, and aquatic organisms;

[0049] The PFAS visualization display module visualizes the spatial distribution of PFAS concentrations in water, sediments, and aquatic organisms, and compares historical predictions with actual detection data.

[0050] Example 2, as Figure 1 As shown, based on Example 1, the present invention also provides a technical solution: a PFAS prediction system and collaborative verification method for a fluorine chemical park based on multi-media dynamic allocation, comprising the following steps:

[0051] S1. Collect and pre-process the chemical oxygen demand data and fluoride ion concentration data in the water at the fluorine chemical park outlet, sample and detect the PFAS concentration in the water, sediment and aquatic organisms, and deploy an ultraviolet spectroscopy online COD analyzer and a fluoride ion selective electrode sensor in parallel in the monitoring well of the main channel of the fluorine chemical park outlet. The ultraviolet light source emits a 254nm wavelength light beam to penetrate the water sample and detect the attenuation intensity of the light beam. According to the Lambert-Beer law, collect the chemical oxygen demand data in the water at the fluorine chemical park outlet. The fluoride ion selective membrane on the electrode surface of the fluoride ion selective electrode sensor The fluoride ion concentration data is obtained by the Nernst equation. The UV spectroscopy online COD analyzer is located 10 meters, 50 meters, 100 meters and 200 meters downstream of the discharge port, and a fully automatic water quality sampler is deployed. The water is fully mixed and measured. The probe is inserted vertically 0.5 meters below the water surface. The data is automatically collected once an hour with a measurement range of 0-500 ng / g. The fluoride ion selective electrode sensor is automatically calibrated once every 24 hours with a standard sodium fluoride solution. The chemical oxygen demand data and fluoride ion concentration collected every hour are Data was collected and outliers were identified using the 3σ criterion. If a data point exceeded the mean ± 3 times the standard deviation, it was marked as invalid data and eliminated. For data missing for no more than 2 consecutive hours, the linear interpolation method of the adjacent time point data was used to fill the gap. The collected chemical oxygen demand data and fluoride ion concentration data were scaled to the [0,1] interval through Min-Max standardization. A fully automatic water quality sampler, a gravity column sediment sampler, and a trawl-type biological sampler were deployed downstream of the fluorine chemical park discharge outlet. 1 liter of water, 500 grams of sediment, and 50-100 grams of individual water were collected each month. Biological samples are tested for PFAS using a liquid chromatography-mass spectrometry instrument equipped with a solid phase extraction column and an ion source. Before testing, water samples are fixed with nitric acid, sediments are ultrasonically extracted with acetonitrile, and biological samples are extracted with accelerated solvents. During mass spectrometry testing, water samples, sediment samples, and aquatic organism samples are separated by liquid chromatography and then enter the electrospray ion source to generate PFAS parent ions in positive ion mode. Multiple reaction monitoring mode is used, a specific mass-to-charge ratio is set, and the PFAS parent ion intensity is recorded by the mass spectrometer. The PFAS concentrations in water, sediments, and aquatic organisms are calibrated based on the internal standard method.

[0052] S2. Extract fluorine chemical characteristics based on the pretreated chemical oxygen demand data and fluoride ion concentration data. Fluorine chemical characteristics include the dynamic accumulation index of chemical oxygen demand and the fluctuation rate of fluoride ion concentration. Based on the pretreated hourly chemical oxygen demand data, with a 24-hour sliding window period, calculate the arithmetic mean and standard deviation of the chemical oxygen demand data within the window, and add the mean to twice the standard deviation to obtain the dynamic accumulation index of chemical oxygen demand. For the pretreated fluoride ion concentration data, calculate its mean and standard deviation on a monthly basis, and obtain the fluctuation rate of fluoride ion concentration using the coefficient of variation formula;

[0053] S3. Combine the extracted fluorine chemical characteristics to construct a cross-media association model and a multi-media dynamic allocation model to predict the concentrations of PFAS in water, sediments and aquatic organisms respectively. The cross-media association model is constructed based on the long short-term memory network architecture. The extracted fluorine chemical characteristics are input into the input layer of the cross-media association model, and seasonal coding and real-time flow data are introduced. The hidden layer neurons of the cross-media association model capture the dynamic association between the fluorine chemical characteristics and the PFAS concentration in the water body and output the hidden state. The seasonal coding and real-time flow data are fused with the hidden state through the fully connected layer to output the predicted value of PFAS concentration in the water body. According to the multi-media allocation law of water-sediment-organism, the predicted value of PFAS concentration in the water body is constrained. The bioconcentration coefficient is determined by fitting the ratio of PFAS concentration in organisms and sediments in historical data. Combined with the predicted value of PFAS concentration in water body output by the cross-media association model, a multi-media dynamic allocation model is constructed based on the law of conservation of mass to calculate the PFAS concentration in sediments and the PFAS concentration in aquatic organisms.

[0054] S4. Based on the historical data of PFAS concentration values ​​in water, sediments and aquatic organisms, preset reference values ​​of PFAS concentrations in each medium, and compare them with the predicted values ​​of PFAS concentrations in water, sediments and aquatic organisms to analyze whether there are any deviations from the reference values. Based on the contribution of water, sediments and aquatic organisms to ecological risks in historical data, analyze historical data of equal time lengths, combine the average values ​​of various PFAS concentrations in water, sediments and aquatic organisms, determine the reference values ​​of PFAS concentrations in water, sediments and aquatic organisms, compare the predicted values ​​of PFAS concentrations in each medium with their corresponding reference values, analyze the deviation rates of PFAS concentrations in each medium, determine whether there are any abnormalities in the predicted values ​​of PFAS concentrations in each medium, and then trigger independent warnings for each medium;

[0055] S5. Visualize the spatial distribution of PFAS concentrations in water bodies, sediments, and aquatic organisms, and compare historical predictions with actual detection data. Based on the coordinates of the monitoring points in the fluorine chemical park and the predicted PFAS concentrations in the corresponding water bodies, sediments, and aquatic organisms, the inverse distance weighted method is used, and a hydrodynamic model is introduced to simulate the diffusion path of pollutants. The concentration field output by the hydrodynamic model is combined with inverse distance weighted interpolation to interpolate the discrete point concentration data into a continuous spatial distribution. The interpolated concentrations are mapped to color intensity according to the gradient, with red indicating high concentration, blue indicating medium concentration, and green indicating low concentration. A heat map of the spatial distribution of PFAS concentrations in the park is generated, and a line graph is used to display the comparison between historical predictions and actual detection data. The comprehensive risk index and actual detection values, as well as the predicted and actual values ​​of water bodies, sediments, and aquatic organisms are aligned on the monthly time axis. The predicted values ​​for each month are connected with a dotted line, with time on the horizontal axis and concentration on the vertical axis. The sample detection values ​​are connected with a solid line, and displayed superimposed with the prediction curve.

[0056] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. PFAS prediction system for fluorine chemical industry park based on multi-media dynamic allocation, characterized by: It includes fluorine chemical data acquisition module, fluorine chemical feature extraction module, cross-media association allocation module, PFAS concentration risk prediction module and PFAS visualization display module; The fluorine chemical industry data acquisition module collects and pre-processes chemical oxygen demand data and fluoride ion concentration data in the water bodies at the fluorine chemical industry park discharge outlet, and samples and detects PFAS concentrations in water bodies, sediments and aquatic organisms; The fluorine chemical feature extraction module extracts fluorine chemical features from the pre-processed chemical oxygen demand data and fluoride ion concentration data; The cross-media correlation allocation module combines the extracted fluorine chemical characteristics to construct a cross-media correlation model and a multi-media dynamic allocation model to predict the concentration of PFAS in water, sediments and aquatic organisms; The PFAS concentration risk prediction module analyzes whether the predicted PFAS concentration in each medium deviates from a preset reference value based on the predicted concentration of PFAS in water, sediment, and aquatic organisms; The PFAS visualization display module visualizes the spatial distribution of PFAS concentrations in water, sediments, and aquatic organisms, and compares historical predictions with actual detection data.

2. The PFAS collaborative verification method for fluorine chemical parks based on multi-media dynamic allocation is implemented based on the PFAS prediction system for fluorine chemical parks based on multi-media dynamic allocation according to claim 1, and is characterized in that: It consists of the following steps: S1. Collect and pre-process chemical oxygen demand data and fluoride ion concentration data from the fluorine chemical park discharge outlet, and sample and test PFAS concentrations in water, sediments, and aquatic organisms; S2, extracting fluorine chemical characteristics based on pre-processed chemical oxygen demand data and fluoride ion concentration data; S3. Combine the extracted fluorine chemical characteristics to construct a cross-media association model and a multi-media dynamic distribution model to predict the concentrations of PFAS in water, sediments, and aquatic organisms respectively; S4. Based on historical data on PFAS concentrations in water, sediments, and aquatic organisms, preset reference values ​​for PFAS concentrations in each medium and compare them with the predicted concentrations of PFAS in water, sediments, and aquatic organisms to analyze whether there are any deviations from the reference values; S5. Visualize the spatial distribution of PFAS concentrations in water, sediments, and aquatic organisms, and compare historical predictions with actual detection data.

3. The PFAS collaborative verification method for fluorine chemical industry parks based on multi-media dynamic allocation according to claim 2, characterized in that: In S1, the collection and preprocessing of chemical oxygen demand data and fluoride ion concentration data in the water body at the fluorine chemical industry park outlet includes: An online COD analyzer using ultraviolet spectroscopy and a fluoride ion selective electrode sensor were deployed in parallel in the monitoring wells of the main channel of the fluorine chemical park's discharge outlet. The ultraviolet light source emitted a 254nm wavelength beam that penetrated the water sample, and the beam attenuation intensity was detected. Based on the Lambert-Beer law, chemical oxygen demand data was collected from the water body at the fluorine chemical park's discharge outlet. The fluoride ion selective membrane on the electrode surface of the fluoride ion selective electrode sensor produced a potential response with the fluoride ions in the water sample, and the fluoride ion concentration data was obtained using the Nernst equation. The 3σ criterion was used to identify outliers in the hourly collected chemical oxygen demand (COD) and fluoride ion concentration (FID) data. Data points outside the range of ±3 times the standard deviation of the mean were marked as invalid and discarded. Missing data for no more than 2 consecutive hours were filled using linear interpolation of adjacent time points. The collected COD and FID data were scaled to the [0, 1] range using Min-Max normalization.

4. The PFAS collaborative verification method for fluorine chemical industry parks based on multi-media dynamic allocation according to claim 3, characterized in that: In S1, the process of sampling and detecting PFAS concentrations in water, sediments, and aquatic organisms includes: Fully automatic water samplers, gravity-fed sediment samplers, and trawl-type biological samplers are deployed downstream of the fluorine chemical park's discharge outlet. Monthly, they collect samples of 1 liter of water, 500 grams of sediment, and 50-100 grams of individual aquatic organisms, respectively. PFAS are detected using a liquid chromatography-mass spectrometry instrument equipped with a solid-phase extraction cartridge and ion source. Prior to testing, water samples are fixed with nitric acid, sediments are ultrasonically extracted with acetonitrile, and biological samples are extracted with accelerated solvents. In mass spectrometry detection, water samples, sediment samples and aquatic organism samples are separated by liquid chromatography and then enter the electrospray ion source to generate PFAS parent ions in positive ion mode. The multiple reaction monitoring mode is used, and a specific mass-to-charge ratio is set. The PFAS parent ion intensity is recorded by the mass spectrometer, and the PFAS concentration in water, sediment and aquatic organisms is calibrated based on the internal standard method.

5. The PFAS collaborative verification method for fluorine chemical industry parks based on multi-media dynamic allocation according to claim 4, characterized in that: In S2, the process of extracting fluorine chemical characteristics based on the pre-processed chemical oxygen demand data and fluoride ion concentration data includes: Fluorine chemical characteristics include the dynamic accumulation index of chemical oxygen demand and the fluctuation rate of fluoride ion concentration; Based on the pre-processed hourly chemical oxygen demand data, with a 24-hour sliding window period, the arithmetic mean and standard deviation of the chemical oxygen demand data within the window are calculated, and the dynamic cumulative index of chemical oxygen demand is obtained by adding the mean and twice the standard deviation; The mean and standard deviation of the pretreated fluoride ion concentration data were calculated on a monthly basis, and the fluctuation rate of fluoride ion concentration was obtained using the coefficient of variation formula.

6. The PFAS collaborative verification method for fluorine chemical parks based on multi-media dynamic allocation according to claim 5, characterized in that: In S3, the process of constructing a cross-media correlation model to predict the concentration of PFAS in water includes: A cross-media association model was constructed based on the long short-term memory network architecture. The extracted fluorine chemical characteristics were input into the input layer of the cross-media association model, and seasonal coding and real-time flow data were introduced. The hidden layer neurons of the cross-media association model capture the dynamic correlation between fluorine chemical characteristics and PFAS concentrations in water bodies, output hidden states, and fuse seasonal coding and real-time flow data with the hidden states through the fully connected layer to output the predicted value of PFAS concentration in water bodies. The predicted value of PFAS concentration in water bodies is constrained according to the water-sediment-organism multi-media distribution law.

7. The PFAS collaborative verification method for a fluorine chemical park based on multi-media dynamic allocation according to claim 6, characterized in that: In S3, the process of constructing a multi-media dynamic distribution model to predict the concentration of PFAS in sediments and aquatic organisms includes: The bioconcentration coefficient was determined by fitting the ratio of PFAS concentrations in organisms and sediments in historical data. Combined with the predicted PFAS concentration in water bodies output by the cross-media correlation model, a multi-media dynamic distribution model was constructed based on the law of conservation of mass to calculate the PFAS concentrations in sediments and aquatic organisms.

8. The PFAS collaborative verification method for fluorine chemical industry parks based on multi-media dynamic allocation according to claim 7, characterized in that: The S4 specifically includes: Based on the contribution of water, sediments, and aquatic organisms to ecological risks in historical data, historical data of equal length are analyzed, and the average values ​​of various PFAS concentrations in water, sediments, and aquatic organisms are combined to determine the reference values ​​of PFAS concentrations in water, sediments, and aquatic organisms; Compare the predicted PFAS concentration value of each medium with its corresponding reference value, analyze the deviation rate of the PFAS concentration of each medium, and determine whether there is any abnormality in the predicted PFAS concentration value of each medium, thereby triggering independent warnings for each medium.

9. The PFAS collaborative verification method for a fluorine chemical park based on multi-media dynamic allocation according to claim 8, characterized in that: In S5, the process of visually displaying the spatial distribution of PFAS concentrations and the comparison between the predicted values ​​and the actual detected values ​​includes: Based on the coordinates of monitoring points in the fluorine chemical park and the predicted PFAS concentrations in corresponding water bodies, sediments, and aquatic organisms, the inverse distance weighted method was used. A hydrodynamic model was introduced to simulate the diffusion path of pollutants. The concentration field output by the hydrodynamic model was combined with inverse distance weighted interpolation to interpolate the discrete point concentration data into a continuous spatial distribution. The interpolated concentrations were mapped to color intensity according to the gradient, with red representing high concentrations, blue representing medium concentrations, and green representing low concentrations. This generated a heat map of the spatial distribution of PFAS concentrations in the park. Use a line graph to show the comparison between historical predictions and actual detection data. Align the comprehensive risk index with the actual detection value, as well as the predicted and actual values ​​of water bodies, sediments and aquatic organisms on the monthly time axis. Connect the predicted values ​​of each month with a dotted line, with time on the horizontal axis and concentration on the vertical axis. Connect the sampling detection values ​​with a solid line and display them superimposed with the prediction curve.

Images