Method for rapidly tracing underground water pollution in chemical industry park

Abstract

The invention relates to the field of groundwater pollution tracing methods, in particular to a method for rapidly tracing groundwater pollution in a chemical industrial park, and solves the problems of high cost and technical threshold, poor operability and long tracing time of groundwater pollution tracing in the prior art. The method comprises the following steps: S1, collecting basic data; s2, defining a monitoring area; s3, data matching is carried out; obtaining a preliminary judgment result according to a data matching result; s4, secondary judgment is carried out, and the park groundwater pollution source is determined. Data acquisition can be carried out through the existing underground water environment monitoring well in the park, the cost is low, and the operation is simple and convenient; performing data analysis and matching on enterprise basic data and environment monitoring data to obtain a suspected pollution source enterprise list of the park; on the basis, the park pollution source is further locked, and park underground water pollution traceability work is completed.

Description

Technical Field

[0001] The present invention relates to the field of methods for tracing the source of groundwater pollution, and particularly to a method for rapidly tracing the source of groundwater pollution in a chemical industrial park. Background Art

[0002] Tracing the source of groundwater pollution is a method of studying the law of pollutant migration in groundwater and finding the source and degree of groundwater pollution by combining modern analytical detection techniques. Once the aquifer is contaminated, it is difficult and costly to repair and treat, and source prevention and control are the key to groundwater pollution repair and treatment. Therefore, identifying the pollution source is crucial for formulating groundwater pollution prevention and control strategies in chemical industrial parks.

[0003] Currently, common methods for tracing the source of groundwater pollution in chemical industrial parks include mathematical statistics methods, model optimization methods, and geochemical footprint methods, etc. These methods require a sufficient amount of basic data, detailed site information, professional simulation software, or specialized isotope detection to achieve the tracing of groundwater pollution in the park. This makes the cost and technical threshold of tracing the source of groundwater pollution in chemical industrial parks high, the operability poor, and it may cause the inability to carry out the tracing work in a timely manner after the groundwater pollution is discovered in the park, which is not conducive to the park to block the pollution source in a timely manner and prevent the further spread of groundwater pollution.

[0004] There is an urgent need for a new method for tracing the source of groundwater pollution that can solve the above problems. Summary of the Invention

[0005] A method for rapidly tracing the source of groundwater pollution in a chemical industrial park proposed by the present invention solves the problems of high cost, high technical threshold, poor operability, and long tracing time in tracing the source of groundwater pollution in the prior art.

[0006] The technical solution of the present invention is realized as follows: A method for rapidly tracing the source of groundwater pollution in a chemical industrial park includes the following steps: S1. Collect basic data: Collect enterprises and pollution-related information related to groundwater pollution in the chemical industrial park and establish a database; the database includes: enterprise situation in the park: enterprise list; enterprise environmental impact report form; enterprise completion environmental protection acceptance monitoring report, and the enterprise environmental impact report form includes enterprise characteristic over-standard factors and groundwater characteristic substances of the enterprise; park situation: pollutant emission declaration registration form of park enterprises; environmental statistical report form of park enterprises; hydrogeological investigation report of the area where the park is located; engineering geological investigation report of the park; general layout plan of the park and enterprises; park boundary range map; existing groundwater environmental monitoring well information in the park, and the monitoring well information includes: coordinates, well depth, water level, and monitoring data; investigation report on the groundwater environmental status of the park; records of groundwater environmental pollution accidents of park enterprises; the database is set on a server; the server also includes a client that can be communicatively connected; S2. Define the monitoring area: Define the monitoring scope in combination with the coordinates of existing groundwater environmental monitoring wells; the layout plan of the park and enterprises, the boundary scope map of the park, and the positions of groundwater environmental monitoring wells can be defined after being gridified through satellite images; S3. Conduct data matching: Regularly monitor the groundwater through existing groundwater environmental monitoring wells, upload the detection results, and match and save them in combination with the database in S1; based on the data matching results, obtain the preliminary judgment results; S31 The detection data is processed by the data acquisition module in the server; S32 Compare the detection data with the information in the database in the information processing module of the server; when the detection data exceeds the set threshold, the information processing module sends a trigger alarm instruction and sends the monitoring well information related to the detection data to the park management client; S33 After receiving the alarm instruction, the park management client gives an alarm prompt; S4. Based on the judgment result in S3, conduct a secondary judgment to determine the groundwater pollution source in the park; and according to the result of the secondary judgment, send a self-inspection alarm to the enterprise client where the polluted area is located, prompting the enterprise to conduct self-inspection and report the self-inspection result; at the same time, extract and send the pipeline water body for inspection; the inspection results can be sent to both the park management client and the enterprise client at the same time; S5. The enterprise quickly locates leaks according to the self-inspection alarm to achieve traceability; in the case that no pollutant leakage is found in the enterprise self-inspection, quickly lock the pollutant-emitting enterprise according to the comparison between the inspection result and the main pollutant detection result to achieve accurate traceability.

[0007] For a further technical solution, in step S32, the information processing module can also obtain the two monitoring wells with the highest correlation with the detection data of this monitoring well according to the correlation matrix, and re-define the monitoring area according to the positions of the three monitoring wells; and send the newly defined monitoring area to the client at the same time.

[0008] For a preferred technical solution, in step S32, while sending the newly defined monitoring area to the client at the same time, match the database, call the list of enterprises in this area, and preliminarily screen and match the enterprise list; and conduct a secondary screening through the Kendall rank correlation coefficient, calculate the correlation of the preliminarily screened and matched enterprise list in combination with the characteristics of groundwater characteristic substances, sort the matching indexes in order, and send the top three enterprise lists to the client.

[0009] Preferably, a correlation calculation model is set up; S41 Model establishment: Based on the correlation matrix and Kendall's rank correlation coefficient, a correlation evaluation model for groundwater pollution source tracing is established; S42 Model training: The groundwater environmental pollution accident records of park enterprises collected in the database; information such as the decision letters for park enterprises to order corrections of groundwater illegal acts are input into the model. The degree of correlation is sorted by the least squares method, and the correlation weight sorting is set; the correlation between monitoring wells, the characteristics of characteristic substances and the correlation and performance training between the concentration of characteristic substances and enterprises are carried out; S43 Model evaluation: Based on neural networks for intelligent training and learning, the data in the model training is inverted, and the training results and training process of the correlation evaluation model are evaluated and optimized.

[0010] Preferably, the setting of the correlation calculation model includes: S41. Model establishment: Based on the correlation matrix and Kendall's rank correlation coefficient, a correlation evaluation model for groundwater pollution source tracing is established; The Pearson correlation coefficient is used to calculate the correlation between the monitoring well data. The monitoring well and the th monitoring data are respectively and . The Pearson correlation coefficient calculation formula is: , where , ; is the number of groundwater monitoring wells, is the number of enterprises, and a matrix with a size of is constructed; the element in the matrix represents the correlation coefficient between the th data item and the th data item; For two variables and with observations , , …, ; First, and are sorted respectively to obtain their ranks and . The Kendall rank correlation coefficient calculation formula is: , where is all data pairs and in the number of pairs with the same sign as minus the number of pairs with different signs; Combine the Kendall rank correlation coefficient with the correlation matrix to form a comprehensive correlation evaluation index for screening enterprises and monitoring wells with a high degree of association with pollution incidents; S42. Model training: Input the records of groundwater environmental pollution accidents of park enterprises collected in the database; information such as the decision letters for park enterprises to order corrections of groundwater illegal acts into the model, sort the degree of correlation by the least squares method, and set the correlation weight sorting; train the correlation between monitoring wells, the characteristics of characteristic substances, and the correlation and performance between the concentration of characteristic substances and enterprises; Specifically, by minimizing the loss function and adjusting the model parameters, the model can accurately sort the correlation between monitoring wells, the characteristics of characteristic substances, and the correlation between the concentration of characteristic substances and enterprises, and construct the loss function: , where is the number of samples, the correlation prediction value is , and the actual correlation is marked as ; S43. Model evaluation: Based on neural network for intelligent training and learning, invert the data in the model training, and evaluate and optimize the training results and process of the correlation evaluation model. Specifically: (1) Neural network intelligent training and learning: Use a multi-layer perceptron neural network structure to perform intelligent training and learning on the correlation evaluation model, and divide the training data set into a training set, a validation set, and a test set; (2) Inversion and optimization: Invert the model training results based on the test set data, that is, compare and analyze the correlation results predicted by the model with the actual situation: Accuracy: ; Recall rate: ; F1 value: ; where is the true positive, is the false positive, is the false negative.

[0011] For a further technical solution, based on the list of the top three enterprises screened in step S32, conduct on-site inspections to confirm whether there are risks, detect the emissions of risk enterprises, and determine that the pollution source is the groundwater pollution source of the park.

[0012] Further technical solution: A real-time monitoring device may also be provided in the monitoring well. The real-time monitoring device includes a pH sensor; a heavy metal concentration sensor: electrochemical method: determining the concentration by measuring the potential difference between heavy metal ions in the solution and the electrode; a monitoring probe for characteristic organic pollutants: collecting pictures in the detection well in real time, transmitting them through communication to the server for picture processing, and monitoring the turbidity and other characteristic pollutants in real time.

[0013] Preferably, the real-time monitoring device is communicatively connected to the data acquisition module of the server to process the data; evaluating the data collected by the real-time monitoring device through a correlation evaluation model, optimizing the correlation degree and weight of the collected data; and regularly sending the correlation information to the client.

[0014] A method for rapid tracing of groundwater pollution in a chemical industrial park disclosed by the present invention can collect data through the existing groundwater environment monitoring wells in the park, with low cost and simple operation; analyze and match the enterprise basic data and environmental monitoring data to obtain a list of suspected pollution source enterprises in the park; on this basis, further lock the pollution sources in the park through the judgment of suspected pollution source monitoring data and on-site inspections, and complete the tracing of groundwater pollution in the park. Using this method makes the tracing of groundwater pollution in the chemical industrial park more efficient and less costly; by setting up a correlation evaluation system, further enhancing the correlation between the monitoring wells and the collected data and the pollution source enterprises, improving the accuracy of matching enterprises, narrowing the scope of secondary inspections, and improving the tracing efficiency; quickly locking the area through correlation determination during the tracing process, and processing through multiple channels such as manual verification in the park, self-inspection by enterprises, and comparison of detection results in parallel, improving the efficiency and accuracy of tracing, and narrowing the scope of pollution impact. Brief Description of the Drawings

[0015] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0016] Figure 1 : Flow schematic diagram of the present invention. Detailed Embodiments

[0017] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0018] Embodiment 1 A method for rapid tracing of groundwater pollution in a chemical industrial park. Data is collected through existing groundwater environmental monitoring wells, and the data and pollution-related data collected from enterprises and the park are entered into a database through a server; the monitoring areas containing monitoring well information divided are also stored and put into the database; a data collection module is set to process the collected data; comparison is carried out through a data processing module; a threshold is set. When the detected data exceeds the set threshold, the information processing module sends a trigger alarm instruction and sends the monitoring well information related to the detected data to the client; after receiving the alarm instruction, the client sends the alarm prompt and the monitoring area to the client together as a preliminary determination result; the staff can further lock the pollution source in the park and complete the tracing by conducting secondary monitoring data research and on-site inspection based on the information collected by the client. The specific steps are as follows: S1. Collect basic data: Collect enterprises and pollution-related information related to groundwater pollution in the chemical industrial park and establish a database; Enterprise situation in the park: enterprise list; enterprise environmental impact report form; enterprise completion environmental protection acceptance monitoring report, and the enterprise environmental impact report form includes enterprise characteristic over-standard factors and groundwater characteristic substances of the enterprise; Park situation: pollutant declaration registration form of park enterprises; environmental statistical report form of park enterprises; hydrogeological exploration report of the area where the park is located; engineering geological exploration report of the park; plane layout diagram of the park and enterprises; park boundary scope diagram; existing groundwater environmental monitoring well information in the park, and the monitoring well information includes: coordinates, well depth, water level and monitoring data; investigation report on the groundwater environmental status of the park; records of groundwater environmental pollution accidents of park enterprises; the database is set on the server; the server also includes a client that can be communicatively connected.

[0019] S2. Demarcate the monitoring area: Combine satellite images to demarcate the plane layout diagram of the park and enterprises, the park boundary scope diagram and the location of groundwater environmental monitoring wells after grid division, and each monitoring area contains at least one existing groundwater environmental monitoring well.

[0020] S3 performs data matching: Regularly monitor the groundwater through existing groundwater environment monitoring wells, upload the detection results, and perform matching and storage in combination with the database in S1; obtain a preliminary determination result based on the data matching result; S31 The detection data processes the data through the data acquisition module in the server; S32 Compare the detection data with the information in the database in the information processing module of the server; when the detection data exceeds the set threshold, the information processing module sends a trigger alarm instruction and sends the monitoring well information related to the detection data to the client; S33 After receiving the alarm instruction, the client gives an alarm prompt.

[0021] S4. Based on the determination result in S3, conduct a secondary determination, conduct a site inspection to confirm whether there is a risk, detect the emissions of risk enterprises, and determine that the pollution source is the groundwater pollution source in the park; and according to the result of the secondary determination, send a self-inspection alarm to the client of the enterprise where the polluted area is located, prompting the enterprise to conduct self-inspection and report the self-inspection result; at the same time, extract and send the pipeline water body for inspection; the inspection results of the inspection can be sent to the park management client and the enterprise client at the same time; S5. The enterprise quickly locates leaks according to the self-inspection alarm and realizes traceability; in the case that the enterprise does not find pollutant leakage during self-inspection, quickly lock the pollutant-emitting enterprise according to the comparison between the inspection result and the main pollutant detection result, and realize accurate traceability.

[0022] In this embodiment, by collecting complete enterprise and park information related to pollutants and using existing monitoring wells for data collection, complex detection means are not required, the operation is simple, and the cost is low.

[0023] Embodiment 2 On the basis of Embodiment 1, by setting a correlation matrix in the information processing module, obtain the two monitoring wells with the highest correlation with the detection data of this monitoring well, and redefine the monitoring area according to the positions of the three monitoring wells; and send the newly defined monitoring area to the client at the same time; while sending the newly defined monitoring area to the client at the same time, match the database, call the list of enterprises in this area, and preliminarily screen and match the list of enterprises; and conduct a secondary screening through the Kendall rank correlation coefficient, calculate the correlation of the preliminarily screened and matched list of enterprises in combination with the characteristics of groundwater characteristic substances, sort the matching indexes in order, and send the list of the top three enterprises to the client.

[0024] Taking the detection of an abnormal increase in the concentration of benzene series substances in a park monitoring well, with the benzene concentration reaching 0.08mg / L, as an example for description: S1. Collect basic data Collect detailed basic information of enterprises in the park: including enterprise name, unified social credit code, registered address, actual production address, legal representative, contact information, establishment time, production scale, etc. Enterprise environmental impact report form: Obtain the environmental impact report form of each enterprise, and extract the enterprise characteristic over-standard factors and groundwater characteristic substances. Park situation: Pollutant discharge declaration registration form: Collect the pollutant discharge declaration registration forms of park enterprises, and master information such as the types of pollutants declared by enterprises to the ecological environment department, the discharge volume, and the discharge destination. Environmental statistical report form: Obtain the environmental statistical report forms of park enterprises, and analyze the overall pollutant discharge trends and structures in the park, including the discharge totals of conventional pollutants such as chemical oxygen demand, ammonia nitrogen, sulfur dioxide, and nitrogen oxides, as well as characteristic pollutants such as benzene series and heavy metals. Hydrogeological exploration report: The hydrogeological exploration report of the area where the park is located shows Engineering geological exploration report: Record the engineering geological conditions of the park. Layout plan and boundary scope map: Draw a detailed layout plan of the park and enterprises, mark the positions of enterprise workshops, warehouses, wastewater treatment facilities, solid waste temporary storage sites, etc.; clarify the park boundary scope map and determine the boundary between the park and the surrounding environment. Groundwater environmental monitoring well information: and monitoring data (in addition to benzene concentration, it also includes concentration data of other pollutants). Environmental status investigation report: Sort out the groundwater environmental status investigation report of the park to understand information such as the groundwater environmental background value and historical pollution situation in the park. Pollution accident records and rectification decision letters: Record the groundwater environmental pollution accident records of park enterprises.

[0025] Store all the above data in a server equipped with a high-performance processor, and develop a Web-based client application for users such as park management departments and enterprises to access through the network to realize data query, upload, and download functions.

[0026] S2. Demarcate the monitoring area Use satellite images to grid the layout plan of the park and enterprises, the park boundary scope map, and the positions of groundwater environmental monitoring wells. Taking the monitoring wells with benzene concentration exceeding the standard as the center, combined with the groundwater flow direction, first demarcate a circular initial monitoring area. At the same time, considering the enterprise distribution and terrain characteristics in the park, appropriately adjust the area so that its boundary coincides with obvious geographical features such as enterprise boundaries and roads, and finally determine the number of enterprises covered by the monitoring area.

[0027] S3. Conduct data matching S31. Data processing: The data acquisition module in the server is written in Python language, and the pandas library is used to process the regularly monitored groundwater data.

[0028] S32. Data comparison and analysis: Compare the processed detection data with the information in the database. When the benzene concentration in the monitoring well exceeds the set threshold, the information processing module of the server immediately sends a trigger alarm instruction.

[0029] Correlation matrix calculation: Establish a correlation matrix between the groundwater monitoring well data and the enterprise emission data; the matrix elements represent the degree of correlation between the monitoring well and the enterprise. The Pearson correlation coefficient is used to calculate the degree of correlation, and the formula is: , where and are the monitoring well data and the enterprise emission data respectively, and are their respective means, is the number of data samples. Through calculation, find the two monitoring wells with the highest correlation with the monitoring well detection data.

[0030] Preliminary screening and secondary screening: According to the results of the correlation matrix, preliminarily screen out the list of enterprises that may be related to pollution. Then use the Kendall rank correlation coefficient for secondary screening, and the calculation formula is: , where S is the sum of the rank differences of the data pairs, is the number of data pairs. Combine the characteristics of groundwater characteristic substances (benzene series) for correlation calculation, sort the matching indexes from high to low, and finally determine the top three enterprises and send the list of these three enterprises to the client.

[0031] S33. Client alarm prompt: After receiving the alarm instruction, the client of the park management department and the relevant enterprise clients give alarm prompts in various ways. The client application pops up a red alarm window to display the alarm information and at the same time emits a beeping sound; in addition, an alarm text message is sent to the relevant person in charge via text message.

[0032] S4. Secondary determination The park environmental monitoring department organizes professional and technical personnel to carry out secondary determination work on the top three enterprises.

[0033] On-site investigation: Conduct a detailed on-site inspection of the three enterprises, check the key areas such as the production workshop, storage warehouse, and wastewater treatment facilities of the enterprises, check for any leakage, and verify the production ledger, pollutant emission records and other materials of the enterprises.

[0034] Pipeline water body extraction and submission for inspection: Extract the pipeline water body involving benzene series emissions from the enterprises. The water samples are collected using professional sampling bottles and protective agents such as copper sulfate are added to prevent microbial decomposition. The water samples are sent to a third-party testing agency for testing, and the testing items include the concentrations of benzene series substances such as benzene, toluene, and xylene.

[0035] Result analysis: The test results show that the benzene concentration in the pipeline water sample of one of the enterprises is much higher than the average levels of the other two enterprises and the park. Combining with the on-site investigation, it is preliminarily determined that this enterprise is the pollution source. The test results submitted for inspection are sent to the park management client and the client of this enterprise at the same time.

[0036] S5. Self-inspection and traceability of enterprises After receiving the self-inspection alarm, the enterprise immediately starts the internal investigation procedure.

[0037] Equipment inspection: Organize professional maintenance personnel to conduct a comprehensive inspection of production equipment, pipelines, valves, etc., and use equipment such as ultrasonic leak detectors and infrared thermal imagers to detect whether there are leakage points.

[0038] Ledger check: Check materials such as production ledgers, raw material usage records, and wastewater discharge records to check for data anomalies or illegal discharge behaviors.

[0039] Emergency treatment: If a leakage point is found, immediately take emergency measures to plug the leak, and clean and disinfect the leakage area.

[0040] If the enterprise's self-inspection does not find any pollutant leakage, the park environmental monitoring department conducts further analysis based on the test results submitted for inspection and the main pollutant test results. By comparing the production process and raw material components of the enterprise with the components and concentrations of pollutants in the monitoring wells, and combining with the groundwater flow model, it is finally determined that a certain reactor of the enterprise has benzene leakage due to aging of the seal, and it enters the groundwater through soil infiltration to achieve accurate traceability.

[0041] Through the determination of correlation, a newly demarcated monitoring area is obtained, and the determination range is initially screened and narrowed; through the calculation of the correlation of groundwater characteristic substances, the initially determined range is further narrowed, the range of secondary determination is reduced, and the cost is further reduced, and the labor input cost is reduced.

[0042] Example 3 On the basis of the above two examples, a correlation calculation model is preferably set up; it is used to illustrate the establishment, training and evaluation process of the correlation evaluation model for groundwater pollution traceability: 1. Construction of correlation matrix The correlation matrix is used to quantify the degree of association between data of different monitoring wells and between monitoring well data and enterprise emission data. Suppose there are groundwater monitoring wells and enterprises in the chemical industrial park, and a matrix with a size of is constructed . The element in the matrix represents the correlation coefficient between the th data item (which can be a monitoring well or an enterprise) and the th data item.

[0043] For the correlation between monitoring well data, the Pearson correlation coefficient is calculated. Let the monitoring well and the th monitoring data be and respectively. The calculation formula for the Pearson correlation coefficient is: where , .

[0044] For the correlation between monitoring well data and enterprise emission data, the Pearson correlation coefficient is also used. The enterprise emission data is regarded as another set of time series data for calculation. For example, if the emission data of a certain pollutant of enterprise is , and the corresponding pollutant monitoring data of monitoring well is , then the calculation method of their correlation coefficient is the same as above.

[0045] 2. Introduction of Kendall rank correlation coefficient The Kendall rank correlation coefficient is used to measure the rank correlation degree of two variables. In the source tracing of groundwater pollution, it is especially suitable for dealing with data situations with non-numerical or many repeated values, such as the association between enterprise violation records and pollution events in monitoring wells.

[0046] Let the two variables and have observations , , …, . First, sort and respectively to obtain their ranks and . The calculation formula for the Kendall rank correlation coefficient is: where is the number of pairs with the same sign minus the number of pairs with different signs among all data pairs and . and .

[0047] Combine the Kendall rank correlation coefficient with the correlation matrix to form a comprehensive correlation evaluation index for screening enterprises and monitoring wells with a high degree of association with pollution events.

[0048] 3. Model training S41. Data input Sort out and preprocess the records of groundwater environmental pollution accidents of park enterprises and the decision letters for ordering park enterprises to correct groundwater violations collected in the database, and convert them into a format suitable for model training.

[0049] For the pollution accident records, extract information such as the accident occurrence time, types of pollutants involved, pollution impact scope, treatment measures, etc.; for the decision letters on illegal acts, extract information such as the name of the illegal enterprise, illegal facts, punishment content, etc. At the same time, associate the corresponding monitoring well data and enterprise emission data to form a training dataset. S42. Least squares sorting Use the least squares method to sort the degree of correlation. Assume that the correlation prediction value output by the model is , and the actual correlation label (such as whether it is a pollution source enterprise) is , and construct a loss function ( is the number of samples).

[0050] By minimizing the loss function and adjusting the model parameters, the model can accurately sort the correlation between monitoring wells, the characteristics of characteristic substances, and the correlation between the concentration of characteristic substances and enterprises. For example, for the correlation scores between different enterprises and pollution monitoring wells, through least squares optimization, the real pollution source enterprises can obtain higher correlation rankings.

[0051] S43. Correlation weight setting According to the actual situation and historical data of the chemical industrial park, set weights for different types of correlations (such as the spatial correlation between monitoring wells, the concentration correlation between enterprise emissions and monitoring well data, the correlation degree between enterprise violation records and pollution incidents, etc.).

[0052] For example, if historical data shows that enterprise violation records have an important guiding role in pollution tracing, then assign a higher weight to the correlation between enterprise violations and pollution incidents calculated based on the Kendall rank correlation coefficient; if the proximity relationship between monitoring wells in space has a greater impact on pollution diffusion, then assign corresponding weights to the spatial correlation of monitoring wells. By continuously adjusting the weights, optimize the accuracy of the model for actual pollution tracing.

[0053] 4. Model evaluation (1) Neural network intelligent training and learning Use neural network structures such as multi-layer perceptron (MLP) to conduct intelligent training and learning on the correlation evaluation model. Divide the training dataset into a training set, a validation set, and a test set (such as dividing it in a ratio of 7:1:2).

[0054] During the training process, the neural network calculates the prediction results through forward propagation and updates the network parameters through backpropagation algorithms (such as gradient descent), continuously adjusting the model's prediction ability for correlation. At the same time, the validation set is used to evaluate the model in real time to avoid overfitting.

[0055] (2) Inversion and optimization Based on the test set data, the training results of the model are inverted, that is, the correlation results predicted by the model are compared and analyzed with the actual situation. Evaluation metrics such as the accuracy, recall rate, and F1 value of the model are calculated. For example: Accuracy:

[0056] Recall rate:

[0057] F1 value:

[0058] Among them, is the number of true positives (the number of correctly predicted pollution sources), is the number of false positives (the number of incorrectly predicted pollution sources), is the number of false negatives (the number of incorrectly predicted non-pollution sources).

[0059] According to the evaluation results, the training process and parameters of the correlation evaluation model are optimized, such as adjusting the number of layers and nodes of the neural network, resetting the correlation weights, improving the data preprocessing method, etc., continuously enhancing the accuracy and reliability of the model in groundwater pollution source tracing.

[0060] By establishing a correlation evaluation model for groundwater pollution source tracing, the correlation between monitoring wells, the characteristics of characteristic substances, and the correlation between the concentration of characteristic substances and enterprises can be effectively determined, greatly improving the accuracy of determining pollution source enterprises through monitoring well data; further narrowing the preliminary determination range, reducing the secondary determination range, and further reducing costs and labor input costs.

[0061] Example 4 The concentration of heavy metal ions in the solution is determined by measuring the potential difference between the heavy metal ions and the electrode through the heavy metal concentration sensor in the real-time monitoring device; The pictures in the detection well are collected in real time through the characteristic organic pollutant monitoring probe, transmitted through communication to the server for picture processing, and the turbidity and other characteristic pollutants are monitored in real time.

[0062] Based on the above three embodiments, real-time monitoring devices are set in the monitoring wells, and real-time monitoring devices are installed in the key monitoring wells in the park. pH value sensor: A high-precision industrial-grade pH value sensor is selected and installed 2 meters below the water surface of the monitoring well. The pH value data is automatically collected every 10 minutes and transmitted to the server through a 4G wireless communication module. When the pH value exceeds the normal range (6.5 - 8.5), the server immediately sends an alarm signal. Heavy metal concentration sensor: An electrochemical heavy metal concentration sensor is used, which can measure the concentrations of multiple heavy metal ions simultaneously. The sensor is installed in the monitoring well in an immersion manner, and the data is measured every 30 minutes. After the data is preliminarily processed by the built-in data processing module of the sensor, it is transmitted to the server through a wireless network. Characteristic organic pollutant monitoring probe: An HD camera is installed as the characteristic organic pollutant monitoring probe to collect pictures in the detection well in real time. The camera is fixed above the wellhead of the monitoring well through a bracket, facing the well interior, and a picture is taken every 5 minutes. The pictures are transmitted to the server through a 5G network, and the server uses a deep learning image recognition algorithm to process the pictures to monitor the turbidity and the presence of characteristic pollutants such as oil films and floating objects in real time. When an obvious oil film appears in the recognized picture, the server immediately sends an alarm message.

[0063] The data is processed by the data acquisition module of the server to which the real-time monitoring device is communicatively connected. The data collected by the real-time monitoring device is evaluated through a correlation evaluation model, the correlation degree and weight of the collected data are optimized to improve the accuracy of the correlation evaluation; and the correlation data is regularly sent to the client. When the collected data exceeds the set threshold, a real-time prompt is given to reduce the speed of pollutant spread and can also lock the pollution range faster, achieving the purpose of rapid traceability.

[0064] Embodiment 5 Based on the above four embodiments, through the detection results and correlation judgment, the polluted area is newly delimited, and relevant enterprises of pollutants are quickly screened through correlation; through multi-channel synchronous parallel processing of manual verification in the park, self-inspection by enterprises and comparison with detection results, the efficiency and accuracy of traceability are improved, and the scope of pollution impact is reduced.

[0065] Of course, without departing from the spirit and essence of the present invention, those skilled in the art should be able to make various corresponding changes and deformations according to the present invention, but these corresponding changes and deformations should all fall within the protection scope of the appended claims of the present invention.

Claims

1. A method for rapid tracing of groundwater pollution in chemical industrial parks, characterized in that: It includes the following steps: S1. Collect basic data: Collect information on enterprises related to groundwater pollution and pollution-related information in the chemical industrial park, and establish a database; S2. Demarcate the monitoring area: Demarcate the monitoring scope in combination with the coordinates of existing groundwater environmental monitoring wells; The layout plan of the park and enterprises, the boundary scope map of the park, and the positions of groundwater environmental monitoring wells are demarcated after being gridified through satellite images; S3. Conduct data matching: Regularly monitor groundwater through existing groundwater environmental monitoring wells, upload the detection results, and perform matching and storage in combination with the database in S1; According to the data matching results, obtain the preliminary judgment results; S4. Based on the judgment results in S3, conduct a secondary judgment to determine the groundwater pollution sources in the park; And according to the results of the secondary judgment, send a self-inspection alarm to the enterprise client of the area determined to be polluted, prompting the enterprise to conduct self-inspection and report the self-inspection results; At the same time, extract and send the pipeline water body for inspection; The inspection results of the inspection can be sent to the park management client and the enterprise client at the same time; S5. The enterprise quickly locates leaks according to the self-inspection alarm to achieve traceability; In the case that no pollutant leakage is found in the enterprise self-inspection, according to the comparison between the inspection results and the main pollutant detection results, quickly lock the pollutant-emitting enterprise to achieve accurate traceability.

2. The method according to claim 1, wherein The database in S1 includes: Situation of park enterprises: Enterprise list; Enterprise environmental impact report form; Enterprise completion environmental protection acceptance monitoring report, and the enterprise environmental impact report form includes enterprise characteristic over-standard factors and groundwater characteristic substances of the enterprise; Situation of the park: Declaration registration form for pollutants discharged by park enterprises; Environmental statistical statements of park enterprises; Hydrogeological investigation report of the area where the park is located; Engineering geological investigation report of the park; Layout plan of the park and enterprises; Boundary scope map of the park; Information on existing groundwater environmental monitoring wells in the park, The monitoring well information includes: coordinates, well depth, water level, and monitoring data; Investigation report on the groundwater environmental status of the park; Records of groundwater environmental pollution accidents of park enterprises; The database is set on the server; The server also includes a client that can be communicatively connected.

3. The method according to claim 1 or 2, characterized in that, Step S3 also includes: S31. The detection data is processed by the data acquisition module in the server; S32. Compare the detection data with the information in the database in the information processing module of the server; When the detection data exceeds the set threshold, the information processing module sends a trigger alarm instruction and sends the monitoring well information related to the detection data to the park management client; S33. After the park management client receives the alarm instruction, it gives an alarm prompt.

4. The method according to claim 3, characterized in that: In step S32, the information processing module obtains the two monitoring wells with the highest correlation with the detection data of this monitoring well according to the correlation matrix, and demarcates a new monitoring area according to the positions of the three monitoring wells; And send the newly demarcated monitoring area to the park management client at the same time.

5. The method according to claim 3, wherein: In step S32, while the newly demarcated monitoring area is sent to the park management client, the database is matched to call the list of enterprises in this area and preliminarily screen and match the enterprise list; and a secondary screening is carried out through the Kendall rank correlation coefficient. The correlation between the preliminarily screened and matched enterprise list and the characteristics of groundwater characteristic substances is calculated, the matching indexes are sorted in order, and the list of the top three enterprises is sent to the park management client and the clients of the top three enterprises.

6. The method according to claim 4 or 5, characterized in that: Set up a correlation calculation model; The setting of the correlation calculation model includes: S41. Model establishment: Based on the correlation matrix and the Kendall rank correlation coefficient, establish a correlation evaluation model for groundwater pollution source tracing; The Pearson correlation coefficient is used to calculate the correlation between the data of monitoring wells. The monitoring well and the monitoring well 's sub-monitoring data are respectively and , and the calculation formula of the Pearson correlation coefficient is: ， Among them , ; is the number of groundwater monitoring wells, is the number of enterprises, and a matrix of size is constructed ; The element in the matrix represents the correlation coefficient between the th data item and the th data item; Two variables and have observations , , …, ; First, sort and separately to obtain their ranks and . The calculation formula for the Kendall rank correlation coefficient is as follows: ， Among them is for all data pairs and in the number of pairs with the same sign as minus the number of pairs with different signs; Combine the Kendall rank correlation coefficient with the correlation matrix to form a comprehensive correlation evaluation index for screening enterprises and monitoring wells with a high degree of association with pollution incidents; S42. Model training: Input the records of groundwater environmental pollution accidents of park enterprises collected in the database; the decision letters for park enterprises to order the correction of groundwater illegal acts into the model, sort the degree of correlation by the least squares method, and set the correlation weight sorting; train the correlation between monitoring wells, the characteristics of characteristic substances and the correlation and performance between the concentration of characteristic substances and enterprises; Specifically, by minimizing the loss function and adjusting the model parameters, the model can accurately sort the correlation between monitoring wells, the characteristics of characteristic substances and the correlation between the concentration of characteristic substances and enterprises, and construct the loss function: ， wherein, is the number of samples, and the correlation prediction value is , and the actual correlation label is ; S43. Model evaluation: Based on neural network for intelligent training and learning, invert the data in the model training, and evaluate and optimize the training results and process of the correlation evaluation model. Specifically: (1) Neural network intelligent training and learning: Use the multi-layer perceptron neural network structure to carry out intelligent training and learning on the correlation evaluation model, and divide the training data set into a training set, a validation set and a test set; (2) Inversion and optimization: Invert the model training results based on the test set data, that is, compare and analyze the correlation results predicted by the model with the actual situation: Accuracy: ; Recall rate: ; F1 value: ; Among them, is a true positive example, is a false positive example, is a false negative example.

7. The method according to claim 6, characterized in that: Based on the list of the top three enterprises screened in step S32, confirm whether there is a risk through on-site inspection, detect the emissions of risk enterprises, and determine that the pollution source is the groundwater pollution source of the park.

8. The method according to claim 7, characterized in that: A real-time monitoring device is arranged in the monitoring well, and the real-time monitoring device includes a pH value sensor, a heavy metal concentration sensor and a characteristic organic pollutant monitoring probe; Measure the potential difference between the heavy metal ions and the electrode in the solution through the heavy metal concentration sensor in the real-time monitoring device to determine its concentration; Collect and detect the pictures in the detection well in real time through the characteristic organic pollutant monitoring probe, and communicate and transmit them to the server for picture processing to monitor the turbidity and other characteristic pollutant conditions in real time.

9. The method according to claim 8, wherein: The real-time monitoring device is communicatively connected to the data acquisition module of the server to process the data; evaluate the data collected by the real-time monitoring device through the correlation evaluation model, optimize the correlation degree and weight of the collected data; and regularly send the correlation information to the park management client.

Images