Tracing method and system for atmospheric pollution in chemical industry park

Abstract

The invention discloses a tracing method and system for atmospheric pollution in a chemical industry park, and relates to the technical field of tracing of atmospheric pollution. The method comprises the following steps: determining at least one first pollution source according to an enterprise-pollutant-raw material incidence matrix and at least one pollutant in the chemical industry park; and according to the real-time pollutant concentration data and a Gaussian diffusion model simulation result, screening out an enterprise with a relatively high fitting degree with the real-time pollutant concentration data, and according to a screening result, determining at least one second pollution source. And determining at least one third pollution source according to the position of the at least one pollution source. And determining at least one fourth pollution source in the high-concentration change area. And determining a pollution traceability result. According to the invention, multi-source data are fused to trace the air pollution, the comprehensiveness and timeliness of the data are ensured, and the accuracy of pollution traceability can be improved.

Description

Technical Field

[0001] This invention relates to the field of air pollution source tracing technology, and in particular to a method for tracing air pollution in a chemical industrial park, a system for tracing air pollution in a chemical industrial park, a computer device, and a readable storage medium. Background Technology

[0002] The chemical industrial park is home to a variety of chemical enterprises, resulting in a concentration of hazardous sources. These sources interact with each other and with vulnerable targets, making accidents prone to serious consequences and chain reactions, thus highlighting safety issues.

[0003] Currently, the main management and control measures for chemical industrial parks are achieved by constructing a regional risk assessment indicator system. This system typically involves building an indicator system based on five aspects: personnel, machinery, materials, methods, and environment. A discriminant matrix is ​​established using the analytic hierarchy process (AHP), and the risk level of the chemical industrial park is calculated by layering these components. Additionally, a standardized safety system is constructed to reflect the safe operation status of the chemical industrial park. Assessment dimensions include the park's organizational structure, planning and construction, risk management, laws, regulations and management systems, safety education and training, emergency management systems, accident response, occupational hazard management, and inspection and assessment.

[0004] However, the existing risk warning models for chemical industrial parks are mostly composed of monitoring data from sensors or surveillance videos, as well as business data from information platforms. When the components of the model change, the safety risk index or level is obtained based on a simple algorithm. In essence, it is still a comprehensive prediction of a certain area, but it cannot locate specific risks or predict the trend of risk changes.

[0005] Current methods for tracing the source of air pollution in chemical industrial parks mainly rely on single data sources (such as sensor or video surveillance data), lacking deep fusion of multi-source data. This leads to inaccurate pollution source location and an inability to predict pollution trends in real time. Furthermore, existing models are mostly static, making it difficult to dynamically reflect the diffusion and changes of pollutants. Summary of the Invention

[0006] The technical problem to be solved by this invention is that the current related technologies lack in-depth fusion of multi-source data, which leads to inaccurate location of pollution sources and inability to predict pollution trends in real time.

[0007] To address the aforementioned shortcomings of existing technologies, the following solutions are provided:

[0008] In a first aspect, this invention provides a method for tracing the source of air pollution in a chemical industrial park. The method includes: S1. Real-time monitoring of atmospheric composition data and meteorological data in and around the chemical industrial park; determining at least one pollutant based on the atmospheric composition data of the chemical industrial park and surrounding areas; and determining at least one first pollution source based on the enterprise-pollutant-raw material correlation matrix within the chemical industrial park and the at least one pollutant. The enterprise-pollutant-raw material correlation matrix is ​​pre-generated and includes the correspondence between the enterprise identifiers in the chemical industrial park, the raw materials used in the production process of each enterprise, and the pollutants produced. S2. Obtaining real-time pollutant concentration data based on the real-time monitored atmospheric composition data of the chemical industrial park and surrounding areas. Selecting enterprises with a high degree of fit to the real-time pollutant concentration data based on the real-time pollutant concentration data and the simulation results of the Gaussian diffusion model, and determining at least one second pollution source based on the selection results. S3. In response to a pollution alarm event, acquiring pollution alarm data, and determining the location of at least one pollution source based on the pollution alarm data and the autoregressive moving average model, and determining at least one third pollution source based on the location of the at least one pollution source. The pollution alarm data consists of real-time pollutant concentration data distribution data and meteorological data within a first preset time period before and after the alarm time. S4. Obtain pollutant concentration data from each monitoring station based on atmospheric composition data of the chemical industrial park and surrounding areas. Determine areas of high concentration variation based on pollutant concentration data from each monitoring station, and identify at least one fourth pollution source within these areas. S5. Determine pollution source tracing results based on at least one first pollution source, at least one second pollution source, at least one third pollution source, at least one fourth pollution source, and the enterprise-pollutant-raw material correlation matrix within the chemical industrial park. Pollution source tracing results include at least one tracing enterprise, at least one pollutant, and at least one tracing raw material.

[0009] Optionally, S1 includes: S11. Real-time monitoring of atmospheric composition and meteorological data in and around the chemical industrial park using a multi-level monitoring network. The multi-level monitoring network includes ground-based fixed stations, mobile monitoring vehicles, and equipment mounted on unmanned aerial vehicles (UAVs). S12. Extracting raw material information, chemical reaction formulas, and production process parameters from the production data systems of various enterprises within the chemical industrial park, and combining this with a chemical engineering knowledge base to analyze potential emission factors at each production stage, constructing an enterprise-pollutant-raw material correlation matrix. S13. In the event of a pollution event occurring in the first area, identifying at least one pollutant based on the atmospheric composition data of the first area. The first area is an area within the chemical industrial park. S14. Identifying at least one primary pollution source based on at least one pollutant and the enterprise-pollutant-raw material correlation matrix.

[0010] Optionally, S13 includes: in the event of a pollution event in the first region, obtaining pollutant concentration data of all enterprises in the first region based on atmospheric composition data of the first region, performing dimensionality reduction processing on the pollutant concentration data of all enterprises in the first region using principal component analysis, and identifying at least one pollutant.

[0011] Optionally, S2 includes: S21. Based on the environmental impact assessment reports and discharge permits of each enterprise in the chemical industrial park, and combined with the real-time production conditions of each enterprise, calculate the theoretical emission intensity of various pollutants of each enterprise using material balance and emission coefficient methods to obtain the dynamic emission inventory of each enterprise. S22. Based on the dynamic emission inventory of each enterprise, obtain simulated data on the distribution of pollutant concentrations of various pollutants of each enterprise in the chemical industrial park and surrounding areas using Gaussian diffusion models and meteorological data. S23. Obtain real-time pollutant concentration data based on real-time monitoring of atmospheric composition data in the chemical industrial park and surrounding areas, compare the simulated data on the distribution of pollutant concentrations of various pollutants of each enterprise in the chemical industrial park and surrounding areas with the real-time pollutant concentration data using Euclidean distance, and select enterprises with a high degree of fit between the simulated data and the real-time pollutant concentration data of various pollutants of each enterprise in the chemical industrial park and surrounding areas based on the comparison results, and list them as potential pollution sources. At least one secondary pollution source includes a potential pollution source.

[0012] Optionally, S22 includes: S221, determining the coordinates of each enterprise and the wind speed data at each enterprise's coordinates; S222, calculating the lateral diffusion parameters and vertical diffusion parameters at each enterprise's coordinates; S223, determining the source strength of each enterprise based on its dynamic emission inventory; and S224, substituting the source strength, wind speed data, lateral diffusion parameters, and vertical diffusion parameters of each enterprise into the Gaussian formula to obtain simulated data on the concentration distribution of various pollutants from each enterprise in the chemical industrial park and surrounding areas.

[0013] Optionally, S3 includes: S31, determining a pollution alarm event in response to the concentration of volatile organic compounds in and around the chemical industrial park meeting a first preset condition. The first preset condition includes: within three consecutive preset monitoring periods, the concentration of a certain pollutant in the real-time pollutant concentration data is greater than a preset concentration threshold and exceeds the limit value by N times. N is a positive integer. S32, acquiring pollution alarm data. S33, modeling the pollution alarm data based on an autoregressive moving average model, analyzing the potential causes before the sudden increase in pollutant concentration, determining the location of at least one pollution source, and determining at least one third pollution source based on the location of the at least one pollution source.

[0014] Optionally, S33 includes: S331, preprocessing the pollution alarm data. This includes: performing a stationarity test on the real-time pollutant concentration data distribution data within a first preset time period before and after the alarm time; and eliminating seasonal trends by performing first-order differencing on the wind speed sequence in the meteorological data within the first preset time period before and after the alarm time. S332, using the Akaike information content criterion, traversing the candidate model set through a grid search method to determine the optimal model parameters. The optimal model parameters are the model parameters of the model with the smallest Akaike information content criterion value in the candidate model set. S333, constructing an autoregressive moving average model based on the optimal model parameters, fitting the real-time pollutant concentration data distribution data within the first preset time period before and after the alarm time to the autoregressive moving average model, calculating the fitted residual sequence, and identifying at least one potential outlier based on the fitted residual sequence. A potential outlier is a time point in the residual sequence that exceeds G times the standard deviation. G is a positive integer. S334. Upon identifying any potential anomaly, check the wind direction data near the timestamp corresponding to the potential anomaly. In response to a wind direction shift greater than a first preset angle near the timestamp corresponding to the potential anomaly, output the azimuth angle of the suspected source based on the dominant wind direction angle recorded at the time of the wind direction shift. The location of at least one pollution source is determined based on the azimuth angle of the suspected source.

[0015] Optionally, S4 includes: S41. Obtaining pollutant concentration data for each monitoring station based on atmospheric composition data of the chemical industrial park and surrounding areas. Using a clustering analysis algorithm, clustering analysis is performed on each monitoring station according to the similarity of pollutant concentration changes, dividing each monitoring station into multiple clusters, and identifying local high-concentration change areas in the chemical industrial park. At least one suspected emission source is identified based on enterprises located in local high-concentration change areas within the chemical industrial park. S42. Using a Lagrange particle diffusion model, the transport trajectory of pollutants from each of the at least one suspected emission source to each monitoring station is simulated, obtaining simulated data for each monitoring station. The simulated data for each monitoring station is fitted with the actual data for each monitoring station, and at least one fourth pollution source is identified from the at least one suspected emission source based on the fitting results.

[0016] Optionally, S42 includes: S421, using a Lagrange particle diffusion model to simulate the movement of pollutant particles from each of at least one suspected emission source, transported via the wind field to each monitoring station, generating a simulated concentration time series. S422, comparing the spatiotemporal variation characteristics of the simulated concentration time series with the actual data from each monitoring station, and constructing an objective function. S423, using a Bayesian optimization method to optimize the parameters of the Lagrange particle diffusion model so that the objective function converges to a preset threshold, and obtaining the optimized parameters of the Lagrange particle diffusion model. S424, based on the optimized parameters of the Lagrange particle diffusion model, determining at least one pollution source location from at least one suspected emission source with a confidence level higher than a preset confidence threshold. At least one fourth pollution source includes at least one pollution source location with a confidence level higher than a preset confidence threshold.

[0017] Optionally, S5 includes: S51, taking the first M results from at least one first pollution source, at least one second pollution source, at least one third pollution source, and at least one fourth pollution source respectively to obtain M first pollution sources, M second pollution sources, M third pollution sources, and M fourth pollution sources. S52, performing an intersection operation on the M first pollution sources, M second pollution sources, M third pollution sources, and M fourth pollution sources, with at least one traceable enterprise including the result of the intersection operation. S53, determining at least one traceable raw material based on the enterprise identifier of at least one traceable enterprise, at least one pollutant, and the enterprise-pollutant-raw material association matrix.

[0018] Secondly, this invention provides a source tracing system for air pollution in chemical industrial parks, comprising a first source tracing module, a second source tracing module, a third source tracing module, a fourth source tracing module, and a source tracing result determination module. The first source tracing module is configured to: monitor atmospheric composition data and meteorological data of the chemical industrial park and its surrounding areas in real time; determine at least one pollutant based on the atmospheric composition data of the chemical industrial park and its surrounding areas; and determine at least one first pollution source based on the enterprise-pollutant-raw material correlation matrix within the chemical industrial park and the at least one pollutant. The enterprise-pollutant-raw material correlation matrix is ​​pre-generated and includes the correspondence between the enterprise identifiers in the chemical industrial park, the raw materials used in the production process of each enterprise, and the pollutants produced. The second source tracing module is configured to: obtain real-time pollutant concentration data based on the real-time monitored atmospheric composition data of the chemical industrial park and its surrounding areas; screen enterprises with a high degree of fit to the real-time pollutant concentration data based on the real-time pollutant concentration data and the simulation results of the Gaussian diffusion model; and determine at least one second pollution source based on the screening results. The third source tracing module is configured to: in response to a pollution alarm event, obtain pollution alarm data; determine the location of at least one pollution source based on the pollution alarm data and the autoregressive moving average model; and determine at least one third pollution source based on the location of the at least one pollution source. The pollution alarm data consists of real-time pollutant concentration distribution data and meteorological data within a first preset time period before and after the alarm time. The fourth source tracing module is configured to: acquire pollutant concentration data from each monitoring station based on atmospheric composition data of the chemical industrial park and surrounding areas; determine high-concentration variation areas based on the pollutant concentration data from each monitoring station; and identify at least one fourth pollution source within these high-concentration variation areas. The source tracing result determination module is configured to: determine the pollution source tracing result based on at least one first pollution source, at least one second pollution source, at least one third pollution source, at least one fourth pollution source, and the enterprise-pollutant-raw material correlation matrix within the chemical industrial park. The pollution source tracing result includes at least one source enterprise, at least one pollutant, and at least one source raw material.

[0019] Thirdly, the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor runs the computer program stored in the memory, the processor executes the above-mentioned method for tracing the source of air pollution in chemical industrial parks.

[0020] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the processor executes the aforementioned method for tracing the source of air pollution in a chemical industrial park.

[0021] The present invention provides a method and system for tracing the source of air pollution in chemical industrial parks, which constructs a comprehensive and multi-level air pollution monitoring network for chemical industrial parks, integrates multi-source data, and uses multiple methods for source tracing to ensure the comprehensiveness and timeliness of data and improve the accuracy of pollution source tracing. Attached Figure Description

[0022] Figure 1 This is a flowchart illustrating a method for tracing the source of air pollution in a chemical industrial park, as described in an embodiment of the present invention.

[0023] Figure 2 This is a flowchart illustrating another method for tracing the source of air pollution in a chemical industrial park, as described in an embodiment of the present invention.

[0024] Figure 3 This is a schematic diagram illustrating the multi-level monitoring network collecting pollutant data and uploading the data through a communication network in an embodiment of the present invention;

[0025] Figure 4 This is a flowchart illustrating another method for tracing the source of air pollution in a chemical industrial park, as described in an embodiment of the present invention.

[0026] Figure 5 This is an example diagram showing the simulated data of pollutant concentration distribution of various enterprises in the chemical industrial park and surrounding areas, as described in an embodiment of the present invention.

[0027] Figure 6 This is a flowchart illustrating another method for tracing the source of air pollution in a chemical industrial park, as described in an embodiment of the present invention.

[0028] Figure 7 This is a flowchart illustrating another method for tracing the source of air pollution in a chemical industrial park, as described in an embodiment of the present invention.

[0029] Figure 8 This is a flowchart illustrating another method for tracing the source of air pollution in a chemical industrial park, as described in an embodiment of the present invention.

[0030] Figure 9 This is a flowchart illustrating another method for tracing the source of air pollution in a chemical industrial park, as described in an embodiment of the present invention.

[0031] Figure 10 This is a flowchart illustrating another method for tracing the source of air pollution in a chemical industrial park, as described in an embodiment of the present invention.

[0032] Figure 11 This is an example diagram of the concentration change simulation and trend graph based on the Lagrange particle diffusion model in an embodiment of the present invention;

[0033] Figure 12 This is a flowchart illustrating another method for tracing the source of air pollution in a chemical industrial park, as described in an embodiment of the present invention.

[0034] Figure 13 This is an architectural diagram of a source tracing system for air pollution in a chemical industrial park, as described in an embodiment of the present invention.

[0035] Figure 14 This is a structural diagram of a computer device according to an embodiment of the present invention. Detailed Implementation

[0036] To enable those skilled in the art to better understand the technical solution of the present invention, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0037] It is understood that the specific embodiments and accompanying drawings described herein are merely for explaining the invention and are not intended to limit the invention.

[0038] It is understood that, without conflict, the various embodiments and features in the embodiments of the present invention can be combined with each other.

[0039] It is understood that, for ease of description, only the parts related to the present invention are shown in the accompanying drawings, while the parts unrelated to the present invention are not shown in the drawings.

[0040] It is understood that each unit or module involved in the embodiments of the present invention may correspond to only one entity structure, or may be composed of multiple entity structures, or multiple units or modules may be integrated into one entity structure.

[0041] It is understood that, without conflict, the functions and steps marked in the flowcharts and block diagrams of this invention may occur in a different order than that marked in the accompanying drawings.

[0042] It is understood that the flowcharts and block diagrams of this invention illustrate the possible architecture, functions, and operations of systems, apparatuses, devices, and methods according to various embodiments of this invention. Each block in the flowchart or block diagram may represent a unit, module, program segment, or code, containing executable instructions for implementing the specified function. Furthermore, each block or combination of blocks in the block diagram and flowchart can be implemented using a hardware-based system to achieve the specified function, or using a combination of hardware and computer instructions.

[0043] It is understood that the units and modules involved in the embodiments of the present invention can be implemented by software or by hardware. For example, the units and modules can be located in a processor.

[0044] Embodiments of the present invention provide a method for tracing the source of air pollution in a chemical industrial park, such as... Figure 1 As shown, the method includes S1 to S5.

[0045] S1. Monitor atmospheric composition data and meteorological data of the chemical industrial park and its surrounding areas in real time, identify at least one pollutant based on the atmospheric composition data of the chemical industrial park and its surrounding areas, and identify at least one primary pollution source based on the enterprise-pollutant-raw material correlation matrix within the chemical industrial park and at least one pollutant.

[0046] In S1, the enterprise-pollutant-raw material association matrix is ​​pre-generated. The enterprise-pollutant-raw material association matrix includes the identification of each enterprise in the chemical industrial park, the correspondence between the raw materials used in the production process of each enterprise and the pollutants produced.

[0047] In some embodiments, such as Figure 2 As shown, the implementation methods of S1 may include S11 to S14.

[0048] S11. Utilize a multi-level monitoring network to monitor atmospheric composition data and meteorological data of the chemical industrial park and surrounding areas in real time.

[0049] In S11, the multi-level monitoring network includes ground fixed stations, mobile monitoring vehicles, and equipment carried by unmanned aerial vehicles.

[0050] For example, such as Figure 3 As shown, a multi-level monitoring network (ground fixed stations, mobile monitoring vehicles, and equipment carried by drones) can be used to comprehensively collect data on the composition of air pollutants in and around the park, covering conventional pollutants (sulfur dioxide, nitrogen oxides, etc.) and characteristic pollutants (volatile organic compounds and heavy metal compounds produced by specific chemical processes, etc.). At the same time, corresponding meteorological data (wind speed, wind direction, temperature, humidity, and air pressure) can be collected.

[0051] S12. Extract raw material information, chemical reaction formulas, and production process parameters from the production data systems of various enterprises in the chemical industrial park. Combine this with a chemical engineering knowledge base to analyze potential emission factors in each production stage and construct an enterprise-pollutant-raw material correlation matrix.

[0052] For example, raw material information, chemical reaction formulas, and production process parameters can be extracted from the production data system of enterprises in the industrial park. Combined with a chemical engineering knowledge base, potential pollutants emitted at each production stage can be analyzed to construct a pollutant-production raw material (i.e., raw material) correlation matrix. For instance, a chemical synthesis reaction uses chlorine-containing raw materials. According to the knowledge base analysis, it may produce characteristic pollutants such as hydrogen chloride. The corresponding relationships are marked in the matrix, as shown in Table 1.

[0053] Table 1. Example of the Enterprise-Pollutant-Raw Material Relationship Matrix

[0054] enterprise Types of pollutants Types and quantities of raw materials Production process parameters Company A VOCs Toluene: 500 kg / day Reaction temperature: 80℃ Company B <![CDATA[SO2]]> Sulfur: 200 kg / day Reaction pressure: 1 MPa Company C <![CDATA[NO x ]]> Ammonia water: 300 kg / day Reaction time: 4 hours Company D PM2.5 Coal: 1000 kg / day Combustion efficiency: 85%

[0055] In Table 1, VOCs refers to volatile organic compounds, SO2 is the chemical formula for sulfur dioxide, and NO... x Nitrogen oxides are a general term for nitrogen oxides, mainly referring to gaseous mixtures composed of nitrogen and oxygen elements such as nitric oxide (NO) and nitrogen dioxide (NO2). PM2.5 is short for fine particulate matter, which refers to suspended particles in ambient air with an aerodynamic diameter of ≤2.5 micrometers. These particles are not gases, but mixtures of solid or liquid particles, and may adsorb gaseous pollutants such as sulfur dioxide and nitrogen oxides.

[0056] S13. In the event of a pollution incident in the first region, identify at least one pollutant and, based on the at least one pollutant and the enterprise-pollutant-raw material correlation matrix, identify at least one raw material.

[0057] In S13, the first area is a region within a chemical industrial park.

[0058] In some embodiments, the implementation method of S13 may include: in the event of a pollution event in the first region, obtaining the concentration data of pollutants of all enterprises in the first region based on the atmospheric composition data of the first region, performing dimensionality reduction processing on the concentration data of pollutants of all enterprises in the first region using principal component analysis, and identifying at least one pollutant.

[0059] Understandably, when a pollution incident occurs in the target area (the first area), a comparison is first performed in the enterprise-pollutant-raw material correlation matrix to identify which enterprises have pollutants consistent with those causing the pollution incident. Then, based on the pollutant principal component analysis (PCA) algorithm for all enterprises in the target area, the large amount of collected pollutant concentration data is dimensionality-reduced to extract the main pollutants. PCA calculates eigenvalues ​​and eigenvectors through the covariance matrix, selecting the top few principal components with a cumulative contribution rate exceeding a certain threshold (e.g., 85%) to represent key pollutants, reducing the complexity of subsequent source tracing calculations and focusing on pollutants related to the core pollution source.

[0060] S14. Identify at least one primary pollution source based on at least one pollutant and the enterprise-pollutant-raw material correlation matrix.

[0061] Understandably, once at least one pollutant is identified, the company that produces each pollutant in the company-pollutant-raw material association matrix can be queried. The company found can be identified as the first pollution source, thus obtaining at least one first pollution source.

[0062] S2. Obtain real-time pollutant concentration data based on real-time monitoring of atmospheric composition data in and around the chemical industrial park. Based on the real-time pollutant concentration data and Gaussian diffusion model simulation results, select enterprises with a high degree of fit to the real-time pollutant concentration data, and determine at least one secondary pollution source based on the selection results.

[0063] In some embodiments, such as Figure 4 As shown, the implementation methods of S2 may include S21 to S23.

[0064] S21. Based on the environmental impact assessment reports and pollutant discharge permits of each enterprise in the chemical industrial park, and combined with the real-time production conditions of each enterprise in the chemical industrial park, the theoretical emission intensity of various pollutants of each enterprise is calculated using the material balance and emission coefficient method, and the dynamic emission inventory of each enterprise is obtained.

[0065] For example, based on the environmental impact assessment reports and pollutant discharge permits of enterprises in the industrial park, and combined with real-time production conditions (equipment operating time, load rate, and raw material consumption rate), the material balance and emission coefficient method can be used to calculate the theoretical emission intensity of various pollutants for each enterprise. For instance, given the raw material input, conversion rate, and corresponding pollutant emission coefficients of a chemical enterprise producing a certain product, the emission of a specific pollutant per unit time can be calculated in real time to form a dynamic emission inventory.

[0066] For example, it can be achieved in the following way:

[0067] ① Data Input Layer: Real-time access to the reactor temperature (±0.5℃ accuracy), raw material flow meter data (kg / min), and exhaust gas online monitoring instrument (VOCs concentration at the ppm level) of the enterprise's distributed control system (DCS).

[0068] ② For the chlorobenzene production section, a dynamic model is established based on the reaction formula C6H6 + Cl2 → C6H5Cl + HCl: Emissions = Raw material chlorine dosage × Conversion rate (92%) × By-product coefficient (0.08) × Real-time equipment load rate; When the DCS detects an abnormal increase of 5℃ in the reaction temperature, it automatically triggers emission coefficient correction (+15% deviation compensation). Material balance calculation engine.

[0069] ③ Output emission inventory: Generate an emission inventory in JavaScript key-value pair (JSON) data format, including company name (company ID), latitude and longitude, pollutant type, instantaneous emission intensity (g / s), and confidence score (0-1).

[0070] S22. Based on the dynamic emission inventories of each enterprise, simulated data on the concentration distribution of various pollutants in the chemical industrial park and surrounding areas are obtained using Gaussian diffusion models and meteorological data.

[0071] For example, a Gaussian diffusion model can be used to consider the impact of meteorological conditions (wind direction, wind speed, atmospheric stability) on pollutant transport and diffusion, simulating the concentration distribution of pollutants emitted by various enterprises in the industrial park and surrounding areas. In the model, based on the Pasquill-Gifford diffusion curve method (also known as the Pasquill-Gifford diffusion curve method, or simply the PG diffusion curve method), combined with the geographical location of the enterprises and their emission inventories, the predicted concentrations of pollutants at different downwind distances are calculated, obtaining simulated data on the concentration distribution of various pollutants from each enterprise in the chemical industrial park and surrounding areas. Figure 5 As shown.

[0072] In some embodiments, such as Figure 6 As shown, the implementation method of S22 may include S221 to S224.

[0073] S221. Determine the coordinates of each enterprise and, based on meteorological data of the chemical industrial park and surrounding areas, determine the wind speed data at the coordinates of each enterprise.

[0074] S222. Calculate the lateral diffusion parameters and vertical diffusion parameters at the coordinates of each enterprise.

[0075] S223. Determine the source strength of each enterprise based on its dynamic emission inventory.

[0076] Understandably, source strength refers to the amount of pollutants emitted per unit of time.

[0077] S224. Substitute the source strength, wind speed data at the coordinates of each enterprise, lateral diffusion parameters at the coordinates of each enterprise, and vertical diffusion parameters at the coordinates of each enterprise into the Gaussian formula to obtain simulated data on the concentration distribution of various pollutants of each enterprise in the chemical industrial park and surrounding areas.

[0078] S23. Obtain real-time pollutant concentration data based on real-time monitoring of atmospheric composition data in and around the chemical industrial park. Using Euclidean distance, compare the simulated data of pollutant concentration distribution of various enterprises in and around the chemical industrial park with the real-time pollutant concentration data. Based on the comparison results, select enterprises whose simulated data of pollutant concentration distribution of various enterprises in and around the chemical industrial park has a high degree of fit with the real-time pollutant concentration data, and list them as potential pollution sources. At least one secondary pollution source includes potential pollution sources.

[0079] For example, S2 can be implemented using the following code.

[0080]

[0081] S3. In response to a pollution alarm event, acquire pollution alarm data, determine the location of at least one pollution source based on the pollution alarm data and the autoregressive moving average model, and determine at least one third pollution source based on the location of at least one pollution source.

[0082] In S3, pollution alarm data consists of real-time pollutant concentration distribution data and meteorological data within the first preset time period before and after the alarm time.

[0083] In some embodiments, such as Figure 7 As shown, the implementation methods of S3 can include S31 to S33.

[0084] S31. In response to the concentration of volatile organic compounds meeting the first preset condition, a pollution alarm event is determined to have occurred.

[0085] In S31, the first preset condition includes: within three consecutive preset monitoring periods, the concentration of a certain pollutant in the real-time pollutant concentration data is greater than a preset concentration threshold and exceeds the limit value by N times. N is a positive integer (for example, N is 2).

[0086] For example, ground-based fixed stations in a multi-level monitoring network can use the Modbus TCP protocol to transmit data (1Hz sampling), and mobile monitoring vehicles can transmit GPS positioning (±3m accuracy) and sensor data back via 5G CPE to facilitate the determination of whether a pollution alarm event has occurred.

[0087] For example, the first preset condition can be: the VOCs concentration exceeds the baseline value by 200% for three consecutive cycles (15s) and the absolute concentration is >500 μg / m³. 3 .

[0088] S32. In response to a pollution alarm event, acquire pollution alarm data.

[0089] For example, the first preset time period is 30 minutes, and the pollution alarm data obtained is the real-time pollutant concentration data distribution data and meteorological data within 30 minutes before and after the alarm time.

[0090] S33. Based on the autoregressive moving average model, model the pollution alarm data, analyze the potential causes before the pollutant concentration rises sharply, determine the location of at least one pollution source, and determine at least one third pollution source based on the location of at least one pollution source.

[0091] Understandably, the data before and after the alarm period is modeled based on the Autoregressive Integrated Moving Average (ARIMA) model. By identifying the trend, seasonality, and periodicity of the data sequence, the potential causes of the sudden increase in pollutant concentration are analyzed. For example, if the model finds that the concentration of a certain pollutant rises rapidly after a specific wind direction change, combined with the location of potential pollution sources identified during the emission source tracing phase, the suspected pollution sources can be further identified.

[0092] In some embodiments, such as Figure 8 As shown, the implementation method of S33 may include S331 to S334.

[0093] S331. Preprocess the pollution alarm data.

[0094] In S331, the preprocessing method includes: performing a stationarity test on the real-time pollutant concentration data distribution data within a first preset time period before and after the alarm time; and performing first-order difference to eliminate seasonal trends based on the wind speed sequence in the meteorological data within the first preset time period before and after the alarm time.

[0095] For example, a time series stationarity test (Augmented Dickey-Fuller, ADF) can be performed on the data 30 minutes before and after the alarm (p<0.05 rejects non-stationarity); the wind speed series can be first-differenced (d=1) to eliminate seasonal trends.

[0096] S332. Using the Akaike information content criterion, the optimal model parameters are determined by traversing the candidate model set through the grid search method.

[0097] In S332, the optimal model parameters are the model parameters of the candidate model set that have the smallest Akaike information content criterion value.

[0098] For example, the optimal (p, d, q) combination (e.g., (2, 1, 1)) model equation can be determined by grid search using the Akaike information criterion (AIC):

[0099] S333. Construct an autoregressive moving average model based on the optimal model parameters, fit the real-time pollutant concentration data distribution data within the first preset time period before and after the alarm time based on the autoregressive moving average model, calculate the fitted residual sequence, and identify at least one potential outlier based on the fitted residual sequence.

[0100] In S333, potential outliers are time points in the residual sequence that exceed G times the standard deviation. G is a positive integer, such as G = 3.

[0101] S334. While identifying any potential anomaly, check the wind direction data near the timestamp corresponding to the potential anomaly. In response to a change in wind direction near the timestamp corresponding to the potential anomaly that is greater than the first preset angle, output the azimuth angle of the suspected source based on the dominant wind direction angle recorded when the wind direction changes.

[0102] In S334, the location of at least one pollution source is determined based on the azimuth of the suspected source.

[0103] For example, when a 3σ abrupt change occurs in the residual sequence and matches the wind direction switching timestamp (e.g., 30°→120° is completed within 120s), the azimuth angle of the suspected source can be output: θ_suspect = the prevailing wind direction at the abrupt change time point ±15° tolerance.

[0104] S4. Obtain pollutant concentration data from each monitoring station based on atmospheric composition data of the chemical industrial park and surrounding areas. Determine areas of high concentration variation based on pollutant concentration data from each monitoring station, and identify at least one fourth pollution source within these areas.

[0105] In some embodiments, such as Figure 9 As shown, the implementation methods of S4 may include S41 to S42.

[0106] S41. Obtain pollutant concentration data for each monitoring station based on atmospheric composition data of the chemical industrial park and surrounding areas; use cluster analysis algorithm to perform cluster analysis on each monitoring station according to the similarity of pollutant concentration changes, divide each monitoring station into multiple clusters, and identify local high concentration change areas in the chemical industrial park.

[0107] In S41, at least one suspected emission source is identified based on enterprises located in areas of localized high concentration variation within chemical industrial parks.

[0108] For example, clustering analysis algorithms can be used to cluster monitoring stations within the park based on the similarity of pollutant concentration changes. Using the K-means clustering algorithm, the pollutant concentration vectors of each station at different time periods are used as samples to divide the stations into multiple clusters. Stations within the same cluster exhibit similar concentration change trends, indicating they are influenced by similar pollution sources. This helps to identify localized high-concentration change areas and accurately delineate the source tracing range.

[0109] For example, S41 can be implemented according to the following code.

[0110]

[0111] S42. Based on the Lagrange particle diffusion model, simulate the transmission trajectory of pollutants from each of the at least one suspected emission sources to each monitoring station, obtain the simulated data of each monitoring station, fit the simulated data of each monitoring station with the actual data of each monitoring station, and determine at least one fourth pollution source from the at least one suspected emission source based on the fitting results.

[0112] Understandably, based on the Lagrange particle diffusion model (LPDM), wind field data can be incorporated to drive particle motion simulation, tracking the transport trajectory of pollutants from potential emission sources to various monitoring stations. The model updates particle positions and concentrations in real time according to meteorological conditions, and combines this with measured concentration changes at monitoring stations to inversely calculate the location and emission intensity of pollution sources.

[0113] In some embodiments, such as Figure 10 As shown, the implementation method of S42 may include S421 to S424.

[0114] S421. Using the Lagrange particle diffusion model, the motion process of pollutant particles is simulated one by one from each of the at least one suspected emission sources, transmitted through the wind field to each monitoring station, to generate a simulated concentration time series.

[0115] S422. Compare the spatiotemporal variation characteristics of the simulated concentration time series with the actual data of each monitoring station, and construct the objective function.

[0116] S423. Use Bayesian optimization to optimize the parameters of the Lagrange particle diffusion model so that the objective function converges to a preset threshold, and obtain the optimized parameters of the Lagrange particle diffusion model.

[0117] Understandably, during the calculation process, Bayesian optimization algorithms are used to optimize key parameters in LPDM (such as particle diffusion coefficient and sedimentation velocity) to improve model accuracy and make the simulated concentration changes fit the measured values ​​best.

[0118] S424. Based on the optimized parameters of the Lagrange particle diffusion model, determine the location of at least one pollution source with a confidence level higher than a preset confidence threshold from at least one suspected emission source.

[0119] In S424, at least one fourth pollution source includes at least one pollution source location with a confidence level higher than a preset confidence threshold.

[0120] Understandably, S421 can be re-executed based on the optimized parameters of the Lagrange particle diffusion model to obtain the concentration change simulation and trend graph simulated by the Lagrange particle diffusion model, as shown below. Figure 11 As shown.

[0121] For example, S42 can be implemented according to the following code.

[0122]

[0123] For example, the reverse tracing based on the Lagrange ion diffusion model is shown in Table 2.

[0124] Table 2 Examples of Reverse Source Tracing

[0125]

[0126] S5. Determine the pollution source tracing results based on at least one first pollution source, at least one second pollution source, at least one third pollution source, at least one fourth pollution source, and the enterprise-pollutant-raw material correlation matrix within the chemical industrial park. The pollution source tracing results include at least one traceable enterprise, at least one pollutant, and at least one traceable raw material.

[0127] In some embodiments, such as Figure 12 As shown, the implementation methods of S5 may include S51 to S53.

[0128] S51. Take the first M results from at least one first pollution source, at least one second pollution source, at least one third pollution source, and at least one fourth pollution source respectively to obtain M first pollution sources, M second pollution sources, M third pollution sources, and M fourth pollution sources.

[0129] S52. Perform an intersection operation on M first pollution sources, M second pollution sources, M third pollution sources and M fourth pollution sources, and include the result of the intersection operation on at least one source tracing enterprise.

[0130] S53. Determine at least one traceable raw material based on the enterprise identifier of at least one traceable enterprise, at least one pollutant, and the enterprise-pollutant-raw material association matrix.

[0131] Understandably, the M first pollution sources, M second pollution sources, M third pollution sources, and M fourth pollution sources were obtained using different methods. If all the pollution sources obtained using different methods include company A, then company A is very likely the source of this pollution. The raw materials used by company A are the traceable raw materials. M is a positive integer and can be set according to actual needs; for example, M can be 5.

[0132] Some embodiments of the present invention provide a method for tracing the source of air pollution in chemical industrial parks, the advantages of which include:

[0133] 1. Multi-level data acquisition network: Innovatively integrating ground-based fixed monitoring stations, mobile monitoring vehicles, and drones equipped with monitoring equipment, a comprehensive and three-dimensional data acquisition system is constructed. This system can not only acquire information on the concentration of conventional air pollutants (sulfur dioxide, nitrogen oxides, etc.) and characteristic pollutants (volatile organic compounds and heavy metal compounds produced by specific chemical processes) in and around the park, but also capture meteorological data (wind speed, wind direction, temperature, humidity, and air pressure) in real time. At the same time, it connects with the production data system of enterprises in the park to obtain detailed operating information such as raw material usage and production processes, providing a massive and diverse data foundation for accurate source tracing.

[0134] 2. Multi-algorithm fusion tracing model:

[0135] Factor source tracing: Principal component analysis (PCA) algorithm is used to reduce the dimensionality of complex pollutant concentration data. The eigenvalues ​​and eigenvectors are accurately calculated through the covariance matrix. The top principal components with a cumulative contribution rate of over 85% are selected as key pollution factors, effectively focusing on the factors related to the core pollution source and significantly reducing the complexity of subsequent source tracing calculations.

[0136] Emission source tracing: Combining material balance and emission coefficient methods, based on the environmental impact assessment reports, pollutant discharge permit data, and real-time production conditions of enterprises in the park, the theoretical emission intensity of pollutants of each enterprise is dynamically calculated to construct an accurate emission inventory of enterprises; at the same time, using the Gaussian diffusion model, based on the Pasquill-Gifford diffusion curve method combined with meteorological conditions and the geographical location of enterprises, the concentration distribution of pollutants in and around the park is simulated to achieve preliminary accurate location of pollution sources.

[0137] Alarm source tracing: Based on the autoregressive moving average model (ARIMA), data before and after the alarm period are modeled to deeply analyze the potential causes before the sudden increase in pollutant concentration. By identifying the trend, seasonality and periodicity of the data sequence, and combining the emission source tracing results, suspected pollution sources can be quickly identified, enhancing the timeliness of source tracing.

[0138] Concentration change source tracing: The K-means clustering algorithm is used to divide the monitoring stations into clusters based on the pollutant concentration vectors at different time periods, accurately identifying local high concentration change areas; the Lagrange particle diffusion model (LPDM) is used to introduce wind field data to drive particle motion simulation, reversely inferring the location and emission intensity of pollution sources, and Bayesian optimization algorithm is used to optimize key parameters of LPDM to ensure that the simulated concentration changes are highly fitted to the measured values, achieving high-precision source tracing and location.

[0139] 3. Real-time and efficient system operation mechanism: A dedicated data center for the chemical industrial park is established, equipped with high-performance servers, to ensure that data from all monitoring devices can be transmitted and aggregated in real time via wired or wireless (4G / 5G) methods. In terms of software workflow, after the system starts, each module runs automatically in sequence, updating the source tracing results every 15 minutes. The visualization module instantly presents the source tracing information in an intuitive form, assisting environmental regulators in making rapid decisions and greatly improving the efficiency of pollution source tracing and response.

[0140] Some embodiments of the present invention provide a source tracing system for air pollution in chemical industrial parks, such as... Figure 13 As shown, the air pollution tracing system 100 in the chemical industrial park includes a first tracing module 101, a second tracing module 102, a third tracing module 103, a fourth tracing module 104, and a tracing result determination module 105.

[0141] The first source tracing module 101 is configured to: monitor atmospheric composition and meteorological data of the chemical industrial park and its surrounding areas in real time; identify at least one pollutant based on the atmospheric composition data of the chemical industrial park and its surrounding areas; and identify at least one primary pollution source based on the enterprise-pollutant-raw material correlation matrix within the chemical industrial park and the at least one pollutant. The enterprise-pollutant-raw material correlation matrix is ​​pre-generated and includes the correspondence between each enterprise in the chemical industrial park, the raw materials used in the production process of each enterprise, and the pollutants produced.

[0142] In some embodiments, the first source tracing module 101 is configured to: utilize a multi-level monitoring network to monitor atmospheric composition data and meteorological data of the chemical industrial park and its surrounding areas in real time. It extracts raw material information, chemical reaction formulas, and production process parameters from the production data systems of various enterprises within the chemical industrial park, combines this with a chemical engineering knowledge base, analyzes potential emission factors at each production stage, and constructs an enterprise-pollutant-raw material correlation matrix. In the event of a pollution incident occurring in the first area, it identifies at least one pollutant and, based on the at least one pollutant and the enterprise-pollutant-raw material correlation matrix, identifies at least one raw material. It then identifies at least one first pollution source based on the at least one raw material and the enterprise-pollutant-raw material correlation matrix. The multi-level monitoring network includes ground-based fixed stations, mobile monitoring vehicles, and equipment mounted on unmanned aerial vehicles (UAVs). The first area is a region within the chemical industrial park.

[0143] In some embodiments, the first source tracing module 101 is configured to: in the event of a pollution event in the first region, obtain the concentration data of pollutants of all enterprises in the first region based on the atmospheric composition data of the first region, and use principal component analysis to perform dimensionality reduction processing on the concentration data of pollutants of all enterprises in the first region to determine at least one pollutant.

[0144] The second source tracing module 102 is configured to: acquire real-time pollutant concentration data based on real-time monitoring of atmospheric composition data of the chemical industrial park and its surrounding areas; and, based on the real-time pollutant concentration data and Gaussian diffusion model simulation results, screen out enterprises with a high degree of fit to the real-time pollutant concentration data to identify at least one secondary pollution source.

[0145] In some embodiments, the second source tracing module 102 is configured to: calculate the theoretical emission intensity of various pollutants from each enterprise based on the environmental impact assessment reports and discharge permits of each enterprise in the chemical industrial park, combined with the real-time production conditions of each enterprise in the chemical industrial park, using material balance and emission coefficient methods, to obtain a dynamic emission inventory of each enterprise. Based on the dynamic emission inventory of each enterprise, simulated data of pollutant concentration distribution of various pollutants in the chemical industrial park and surrounding areas are obtained using Gaussian diffusion models and meteorological data. Real-time pollutant concentration data are obtained based on real-time monitored atmospheric composition data of the chemical industrial park and surrounding areas. Using Euclidean distance, the simulated data of pollutant concentration distribution of various pollutants in the chemical industrial park and surrounding areas are compared with the real-time pollutant concentration data. Based on the comparison results, enterprises with a high degree of fit between the simulated data and the real-time pollutant concentration data of various pollutants in the chemical industrial park and surrounding areas are selected and listed as potential pollution sources. At least one secondary pollution source includes a potential pollution source.

[0146] In some embodiments, the second source tracing module 102 is configured to: determine the coordinates of each enterprise and determine the wind speed data at the coordinates of each enterprise; calculate the lateral diffusion parameters and vertical diffusion parameters at the coordinates of each enterprise; determine the source strength of each enterprise based on the dynamic emission inventory of each enterprise; and substitute the source strength, wind speed data, lateral diffusion parameters, and vertical diffusion parameters of each enterprise into the Gaussian formula to obtain simulated data on the concentration distribution of various pollutants from each enterprise in the chemical industrial park and surrounding areas.

[0147] The third source tracing module 103 is configured to: in response to a pollution alarm event, acquire pollution alarm data, determine the location of at least one pollution source based on the pollution alarm data and an autoregressive moving average model, and determine at least one third pollution source based on the location of the at least one pollution source. The pollution alarm data consists of real-time pollutant concentration distribution data and meteorological data within a first preset time period before and after the alarm time.

[0148] In some embodiments, the third source tracing module 103 is configured to: determine that a pollution alarm event has occurred in response to the concentration of volatile organic compounds meeting a first preset condition; acquire pollution alarm data in response to the occurrence of a pollution alarm event; model the pollution alarm data based on an autoregressive moving average model, analyze the potential causes before the sudden increase in pollutant concentration, determine the location of at least one pollution source, and determine at least one third pollution source based on the location of the at least one pollution source. The first preset condition includes: within three consecutive preset monitoring periods, the concentration of a certain pollutant in the real-time pollutant concentration data is greater than a preset concentration threshold and exceeds the limit value by N times. N is a positive integer.

[0149] In some embodiments, the third source tracing module 103 is configured to: preprocess the pollution alarm data. This includes: performing a stationarity test on the real-time pollutant concentration data distribution data within a first preset time period before and after the alarm time; performing first-order difference to eliminate seasonal trends based on the wind speed sequence in the meteorological data within the first preset time period before and after the alarm time; using the Akaike information criterion and a grid search method to traverse the candidate model set to determine the optimal model parameters. The optimal model parameters are the model parameters of the model with the smallest Akaike information criterion value in the candidate model set; constructing an autoregressive moving average model based on the optimal model parameters; fitting the real-time pollutant concentration data distribution data within the first preset time period before and after the alarm time to the autoregressive moving average model; calculating the fitted residual sequence; and identifying at least one potential anomaly based on the fitted residual sequence. A potential anomaly is a time point in the residual sequence that exceeds G times the standard deviation. Simultaneously with identifying any potential anomaly, checking the wind direction data near the timestamp corresponding to the potential anomaly; and outputting the azimuth angle of the suspected source based on the dominant wind direction angle recorded at the time of the wind direction change when the wind direction near the timestamp corresponding to the potential anomaly changes by a greater than a first preset angle. The location of at least one pollution source is determined based on the azimuth angle of the suspected source. G is a positive integer.

[0150] The fourth source tracing module 104 is configured to: obtain pollutant concentration data from each monitoring station based on atmospheric composition data of the chemical industrial park and surrounding areas; determine high concentration variation areas based on pollutant concentration data from each monitoring station; and identify at least one fourth pollution source in the high concentration variation areas.

[0151] In some embodiments, the fourth source tracing module 104 is configured to: acquire pollutant concentration data of each monitoring station based on atmospheric composition data of the chemical industrial park and its surrounding areas; use a clustering analysis algorithm to perform clustering analysis on each monitoring station according to the similarity of pollutant concentration changes based on the pollutant concentration data of each monitoring station in the atmospheric composition data of the chemical industrial park and its surrounding areas, divide each monitoring station into multiple clusters, and identify local high-concentration change areas in the chemical industrial park; identify at least one suspected emission source based on enterprises located in local high-concentration change areas in the chemical industrial park; simulate the transmission trajectory of pollutants from each of the at least one suspected emission source to each monitoring station according to the Lagrange particle diffusion model, obtain simulated data for each monitoring station, fit the simulated data of each monitoring station with the actual data of each monitoring station, and determine at least one fourth pollution source from the at least one suspected emission source based on the fitting results.

[0152] In some embodiments, the fourth source tracing module 104 is configured to: employ a Lagrange particle diffusion model to simulate the movement of pollutant particles from each of the at least one suspected emission sources, via the wind field, to each monitoring station, generating a simulated concentration time series. The simulated concentration time series is compared with the spatiotemporal variation characteristics of the actual data from each monitoring station to construct an objective function. The parameters of the Lagrange particle diffusion model are optimized using a Bayesian optimization method to make the objective function converge to a preset threshold, and the optimized parameters of the Lagrange particle diffusion model are obtained. Based on the optimized parameters of the Lagrange particle diffusion model, at least one pollution source location with a confidence level higher than a preset confidence threshold is determined from the at least one suspected emission source. The at least one fourth pollution source includes at least one pollution source location with a confidence level higher than the preset confidence threshold.

[0153] The source tracing result determination module 105 is set to determine the pollution source tracing result based on at least one first pollution source, at least one second pollution source, at least one third pollution source, at least one fourth pollution source and the enterprise-pollutant-raw material association matrix within the chemical industrial park. The pollution source tracing result includes at least one source enterprise, at least one pollutant and at least one source raw material.

[0154] In some embodiments, the source tracing result determination module 105 is configured to: take the first M results from at least one first pollution source, at least one second pollution source, at least one third pollution source, and at least one fourth pollution source respectively to obtain M first pollution sources, M second pollution sources, M third pollution sources, and M fourth pollution sources. An intersection operation is performed on the M first pollution sources, M second pollution sources, M third pollution sources, and M fourth pollution sources, and at least one traceable enterprise is included in the result of the intersection operation. At least one traceable raw material is determined based on the enterprise identifier of at least one traceable enterprise, at least one pollutant, and the enterprise-pollutant-raw material association matrix.

[0155] The specific scheme and beneficial effects of the source tracing system for air pollution in chemical industrial parks provided by the embodiments of the present invention can be referred to the relevant description of the source tracing method for air pollution in chemical industrial parks provided by the embodiments of the present invention, which will not be repeated here.

[0156] Some embodiments of the present invention provide a computer device, such as Figure 14 As shown, the computer device 1400 includes a memory 1401 and a processor 1402. The memory 1401 stores a computer program. When the processor 1402 runs the computer program stored in the memory 1401, the processor 1402 executes the above-mentioned method for tracing the source of air pollution in the chemical industrial park.

[0157] For details on the specific solutions and beneficial effects of a computer device provided by some embodiments of the present invention, please refer to the relevant description of a cloud platform for tracing the source of air pollution in a chemical industrial park provided by some embodiments of the present invention, which will not be repeated here.

[0158] Some embodiments of the present invention provide a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the processor executes the above-described method for tracing the source of air pollution in a chemical industrial park.

[0159] For details on the specific solutions and beneficial effects of a computer-readable storage medium provided by some embodiments of the present invention, please refer to the relevant description of a cloud platform for tracing the source of air pollution in a chemical industrial park provided by some embodiments of the present invention, which will not be repeated here.

[0160] It is understood that the above embodiments are merely exemplary implementations used to illustrate the principles of the present invention, and the present invention is not limited thereto. For those skilled in the art, various modifications and improvements can be made without departing from the spirit and essence of the present invention, and these modifications and improvements are also considered to be within the scope of protection of the present invention.

Claims

1. A method for tracing the source of air pollution in a chemical industrial park, characterized in that, include: S1. Real-time monitoring of atmospheric composition data and meteorological data in and around the chemical industrial park; identification of at least one pollutant based on the atmospheric composition data of the chemical industrial park and its surrounding areas; and identification of at least one primary pollution source based on the enterprise-pollutant-raw material correlation matrix within the chemical industrial park and the at least one pollutant; the enterprise-pollutant-raw material correlation matrix is ​​pre-generated and includes the correspondence between the identifiers of each enterprise in the chemical industrial park, the raw materials used in the production process of each enterprise, and the pollutants produced. S2. Obtain real-time pollutant concentration data based on real-time monitoring of atmospheric composition data of the chemical industrial park and surrounding areas; screen out enterprises with high fitting degree to the real-time pollutant concentration data based on the real-time pollutant concentration data and Gaussian diffusion model simulation results, and determine at least one second pollution source based on the screening results. S3. In response to a pollution alarm event, acquire pollution alarm data, and determine the location of at least one pollution source based on the pollution alarm data and an autoregressive moving average model, and determine at least one third pollution source based on the location of the at least one pollution source; the pollution alarm data is real-time pollutant concentration data distribution data and meteorological data within a first preset time period before and after the alarm time; S4. Obtain pollutant concentration data for each monitoring station based on the atmospheric composition data of the chemical industrial park and its surrounding areas, determine high concentration variation areas based on the pollutant concentration data of each monitoring station, and identify at least one fourth pollution source in the high concentration variation areas. S5. Determine the pollution source tracing result based on the at least one first pollution source, the at least one second pollution source, the at least one third pollution source, the at least one fourth pollution source and the enterprise-pollutant-raw material correlation matrix within the chemical industrial park. The pollution source tracing result includes at least one traceable enterprise, the at least one pollutant and at least one traceable raw material.

2. The method for tracing the source of air pollution in chemical industrial parks according to claim 1, characterized in that, S1 includes: S11. Real-time monitoring of atmospheric composition and meteorological data in and around the chemical industrial park using a multi-level monitoring network; the multi-level monitoring network includes ground fixed stations, mobile monitoring vehicles, and equipment carried by unmanned aerial vehicles; S12. Extract raw material information, chemical reaction formulas, and production process parameters from the production data systems of various enterprises in the chemical industrial park. Combine this with a chemical professional knowledge base to analyze the potential emission factors of each production link and construct an enterprise-pollutant-raw material correlation matrix. S13. In the event of a pollution incident in the first region, at least one pollutant shall be identified based on the atmospheric composition data of the first region. The first area is one area within a chemical industrial park; S14. Determine at least one first source of pollution based on at least one pollutant and the enterprise-pollutant-raw material correlation matrix.

3. The method for tracing the source of air pollution in chemical industrial parks according to claim 2, characterized in that, S13 includes: In the event of a pollution incident in the first region, pollutant concentration data of all enterprises in the first region are obtained based on atmospheric composition data of the first region. Principal component analysis is then used to perform dimensionality reduction on the pollutant concentration data of all enterprises in the first region to identify at least one pollutant.

4. The method for tracing the source of air pollution in chemical industrial parks according to claim 1, characterized in that, S2 includes: S21. Based on the environmental impact assessment reports and pollutant discharge permits of each enterprise in the chemical industrial park, and combined with the real-time production conditions of each enterprise in the chemical industrial park, the theoretical emission intensity of various pollutants of each enterprise is calculated using the material balance and emission coefficient method, and the dynamic emission inventory of each enterprise is obtained. S22. Based on the dynamic emission inventories of each enterprise, simulated data on the concentration distribution of various pollutants of each enterprise in the chemical industrial park and surrounding areas are obtained using the Gaussian diffusion model and the meteorological data. S23. Obtain real-time pollutant concentration data based on real-time monitoring of atmospheric composition data of the chemical industrial park and surrounding areas. Using Euclidean distance, compare the simulated data of pollutant concentration distribution of various pollutants of each enterprise in the chemical industrial park and surrounding areas with the real-time pollutant concentration data. Based on the comparison results, select enterprises whose simulated data of pollutant concentration distribution of various pollutants of each enterprise in the chemical industrial park and surrounding areas has a high degree of fit with the real-time pollutant concentration data, and list them as potential pollution sources. The at least one second pollution source includes the potential pollution source.

5. The method for tracing the source of air pollution in chemical industrial parks according to claim 4, characterized in that, S22 includes: S221. Determine the coordinates of each enterprise and the wind speed data at each enterprise's coordinates; S222. Calculate the lateral diffusion parameters and vertical diffusion parameters at the coordinates of each enterprise. S223. Determine the source strength of each enterprise based on the dynamic emission inventory of each enterprise; S224. Substitute the source strength, wind speed data at the coordinates of each enterprise, lateral diffusion parameters at the coordinates of each enterprise, and vertical diffusion parameters at the coordinates of each enterprise into the Gaussian formula to obtain simulated data on the concentration distribution of various pollutants of each enterprise in the chemical industrial park and surrounding areas.

6. The method for tracing the source of air pollution in chemical industrial parks according to claim 1, characterized in that, S3 includes: S31. In response to the concentration of volatile organic compounds in the chemical industrial park and surrounding areas meeting a first preset condition, a pollution alarm event is determined to have occurred; the first preset condition includes: within three consecutive preset monitoring cycles, the concentration of a certain pollutant in the real-time pollutant concentration data is greater than a preset concentration threshold and exceeds the limit value by N times; N is a positive integer; S32. Obtain the pollution alarm data; S33. Model the pollution alarm data based on the autoregressive moving average model, analyze the potential causes before the pollutant concentration suddenly increases, determine the location of at least one pollution source, and determine at least one third pollution source based on the location of at least one pollution source.

7. The method for tracing the source of air pollution in chemical industrial parks according to claim 6, characterized in that, S33 includes: S331. Preprocess the pollution alarm data, including: performing a stationarity test on the real-time pollutant concentration data distribution data within a first preset time period before and after the alarm time; and performing first-order difference to eliminate seasonal trends based on the wind speed sequence in the meteorological data within the first preset time period before and after the alarm time. S332. Using the Akaike information content criterion, the optimal model parameters are determined by traversing the candidate model set through a grid search method; the optimal model parameters are the model parameters of the model with the smallest Akaike information content criterion value in the candidate model set. S333. Construct an autoregressive moving average model based on the optimal model parameters, fit the real-time pollutant concentration data distribution data within a first preset time period before and after the alarm time using the autoregressive moving average model, calculate the fitted residual sequence, and identify at least one potential outlier based on the fitted residual sequence; the potential outlier is a time point in the residual sequence that exceeds G times the standard deviation; G is a positive integer. S334. While identifying any potential anomaly, check the wind direction data near the timestamp corresponding to the potential anomaly. In response to a change in wind direction near the timestamp corresponding to the potential anomaly that is greater than a first preset angle, output the azimuth angle of the suspected source based on the dominant wind direction angle recorded at the time of the wind direction change. The location of the at least one pollution source is determined based on the azimuth angle of the suspected source.

8. The method for tracing the source of air pollution in chemical industrial parks according to claim 1, characterized in that, S4 includes: S41. Obtain pollutant concentration data for each monitoring station based on atmospheric composition data of the chemical industrial park and surrounding areas; use cluster analysis algorithm to perform cluster analysis on each monitoring station according to the similarity of pollutant concentration changes based on the pollutant concentration data of each monitoring station, divide each monitoring station into multiple clusters, and identify local high concentration change areas in the chemical industrial park; identify at least one suspected emission source based on enterprises located in local high concentration change areas in the chemical industrial park. S42. Based on the Lagrange particle diffusion model, simulate the transmission trajectory of pollutants from each of the at least one suspected emission sources to each monitoring station to obtain the simulated data of each monitoring station. Fit the simulated data of each monitoring station with the actual data of each monitoring station, and determine at least one fourth pollution source from the at least one suspected emission source based on the fitting results.

9. The method for tracing the source of air pollution in chemical industrial parks according to claim 8, characterized in that, S42 includes: S421. Using the Lagrange particle diffusion model, the motion process of pollutant particles is simulated one by one from each of the at least one suspected emission sources, transmitted through the wind field to each monitoring station, to generate a simulated concentration time series. S422. Compare the spatiotemporal variation characteristics of the simulated concentration time series with the actual data from each monitoring station, and construct the objective function; S423. The parameters of the Lagrange particle diffusion model are optimized using the Bayesian optimization method so that the objective function converges to a preset threshold, and the optimized parameters of the Lagrange particle diffusion model are obtained. S424. Based on the optimized parameters of the Lagrange particle diffusion model, determine at least one pollution source location with a confidence level higher than a preset confidence threshold from the at least one suspected emission source; the at least one fourth pollution source includes the at least one pollution source location with a confidence level higher than the preset confidence threshold.

10. The method for tracing the source of air pollution in chemical industrial parks according to claim 1, characterized in that, S5 includes: S51. Take the first M results from the at least one first pollution source, the at least one second pollution source, the at least one third pollution source, and the at least one fourth pollution source respectively to obtain M first pollution sources, M second pollution sources, M third pollution sources, and M fourth pollution sources; S52. Perform an intersection operation on the M first pollution sources, M second pollution sources, M third pollution sources and M fourth pollution sources, and at least one source tracing enterprise includes the result of the intersection operation; S53. Determine at least one traceable raw material based on the enterprise identifier of the at least one traceable enterprise, the at least one pollutant, and the enterprise-pollutant-raw material association matrix.

11. A source tracing system for air pollution in a chemical industrial park, characterized in that, include: The first source tracing module is configured to: monitor atmospheric composition data and meteorological data of the chemical industrial park and its surrounding areas in real time; determine at least one pollutant based on the atmospheric composition data of the chemical industrial park and its surrounding areas; and determine at least one first pollution source based on the enterprise-pollutant-raw material correlation matrix within the chemical industrial park and the at least one pollutant. The enterprise-pollutant-raw material correlation matrix is ​​pre-generated and includes the correspondence between the identifiers of each enterprise in the chemical industrial park, the raw materials used in the production process of each enterprise, and the pollutants produced. The second source tracing module is configured to: obtain real-time pollutant concentration data based on real-time monitoring of atmospheric composition data of the chemical industrial park and surrounding areas; screen out enterprises with a high degree of fit to the real-time pollutant concentration data based on the real-time pollutant concentration data and the simulation results of the Gaussian diffusion model; and determine at least one second pollution source based on the screening results. The third source tracing module is configured to: in response to a pollution alarm event, acquire pollution alarm data, determine the location of at least one pollution source based on the pollution alarm data and an autoregressive moving average model, and determine at least one third pollution source based on the location of the at least one pollution source; the pollution alarm data is real-time pollutant concentration data distribution data and meteorological data within a first preset time period before and after the alarm time; The fourth source tracing module is configured to: obtain pollutant concentration data of each monitoring station based on the atmospheric composition data of the chemical industrial park and its surrounding areas; determine high concentration variation areas based on the pollutant concentration data of each monitoring station; and determine at least one fourth pollution source in the high concentration variation areas. The pollution source tracing result determination module is configured to determine the pollution source tracing result based on the at least one first pollution source, the at least one second pollution source, the at least one third pollution source, the at least one fourth pollution source and the enterprise-pollutant-raw material association matrix within the chemical industrial park. The pollution source tracing result includes at least one traceable enterprise, the at least one pollutant and the at least one traceable raw material.

12. A computer device, characterized in that, It includes a memory and a processor, wherein the memory stores a computer program, and when the processor runs the computer program stored in the memory, the processor executes the source tracing method for air pollution in chemical industrial parks according to any one of claims 1 to 10.

13. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, the processor executes the source tracing method for air pollution in chemical industrial parks according to any one of claims 1 to 10.

Images