Urban atmospheric pollution transmission path and source contribution evaluation system and method based on machine learning and numerical model

Abstract

The invention belongs to the field of air pollution prevention and control, and particularly discloses an urban air pollution transmission path and source contribution evaluation system and method based on machine learning and a numerical mode, and the evaluation system comprises a meteorological simulation module, a pollution emission list module, a numerical traceability module and a machine learning transmission path evaluation module. According to the method, machine learning and a numerical mode are combined, the automation degree is high, automatic downloading, automatic running and online data analysis can be achieved without manual data statistics and analysis, operation is convenient, data accuracy is high, the response speed is high, and important support is provided for work such as atmospheric pollutant cause tracing and atmospheric pollution control scheme making.

Description

Technical Field

[0001] The present invention belongs to the technical field of air pollution control, and in particular relates to an evaluation system and method for urban air pollution transmission paths and source contributions based on machine learning and numerical models. Background Art

[0002] Studies have shown that the cross-regional transmission of atmospheric pollution outside the region can affect the local atmospheric pollutant concentration by 30%-80%, and the degree of transmission impact and path characteristics are often closely related to the meteorological circulation conditions. In addition, since the atmospheric pollution situation has regional characteristics, it is an effective way to improve the air quality environment to clarify the transmission path and source contribution of atmospheric pollution in specific cities so as to carry out targeted atmospheric pollution control based on the pollution source and transmission path. However, the transmission of atmospheric pollutants has the characteristics of significant geographical and meteorological influences, a wide range of influence, and rapid changes. It is more difficult to accurately explore the transmission path of atmospheric pollutants between regions and quantify their source contributions.

[0003] Existing methods for assessing the transmission paths and source contributions of urban air pollution often cluster air masses based on a relatively rough air mass trajectory model, or set different contribution scenarios based on regional chemical transport models, and then conduct numerical simulations to obtain the mutual transport contributions of air pollutants between regions. There is a single path assessment point, and the simulation of the source contribution scenario takes a long time, which cannot achieve accurate assessment of the upstream and downstream paths and pollution source contributions of air pollutant transmission between cities. Patent CN117708617A discloses an air pollution source tracing method based on multi-source big data and spatiotemporal matching of pollution characteristics. Based on the pollution point information mined from the trend of mobile monitoring data, combined with urban multi-source big data, local pollution sources are quickly and accurately locked according to characteristic pollutants, and small-scale air pollution source tracing is achieved. Patent CN116011317A discloses a method for tracing the source of atmospheric pollution at a small scale based on the fusion of multiple methods. Based on a machine learning prediction model, mathematical methods are used to improve the accuracy of meteorological element predictions at monitoring sites, thereby using spatial interpolation methods to obtain a highly accurate and refined forecast meteorological field, and combining real-time pollutant concentration monitoring data, coupled backward trajectory analysis and potential source contribution factor analysis methods to quickly carry out source tracing analysis, and achieve real-time tracing of local small-scale potential pollution sources. Patent CN115758711A discloses a method for tracing the source of atmospheric pollution based on pollution transmission paths. By analyzing the inherent laws between signals from different monitoring sites, the similarities are found, and based on the similarities, a new concept of in-and-out degrees is proposed and hot spots that need to be focused on are identified. Source analysis is performed through cross-convergence mapping to identify pollution paths. However, the above methods are either applied on a small scale, or the applicable target points are too single, and a regional path cannot be formed, or only semi-quantitative effects can be achieved, and quantitative evaluation of regional source contributions cannot be achieved, making it difficult to support the precise quantification of the transmission paths and source contributions of regional scale atmospheric pollutants between different cities.

[0004] Based on the above technical background, the present invention provides a system and method for evaluating the transmission path and source contribution of urban air pollution based on machine learning and numerical models. Summary of the invention

[0005] In view of the problems and shortcomings in the prior art, the purpose of the present invention is to provide an evaluation system and method for the transmission path and source contribution of urban air pollution based on machine learning and numerical models.

[0006] To achieve the purpose of the invention, the technical solution adopted by the present invention is as follows:

[0007] The first aspect of the present invention provides an evaluation system for the transmission path and source contribution of urban air pollution based on machine learning and numerical models, including a meteorological simulation module, a pollution emission inventory module, a numerical model tracing module and a machine learning transmission path evaluation module; the meteorological simulation module includes an initial meteorological field download unit, a meteorological field simulation unit and a meteorological simulation data extraction and post-processing unit, which are used to simulate and process high-precision meteorological field data of the target area; the pollution source emission inventory module includes a man-made source emission inventory generation unit, a natural source emission inventory generation unit and an emission inventory online marking unit, which are used to generate and mark emission data of various types of pollution sources in the region; the numerical model tracing module is used to simulate the concentration of atmospheric pollutants according to meteorological field data and pollution source emission data, and evaluate the contribution of emission sources in different regions to the concentration of atmospheric pollutants in the target area; the machine learning transmission path evaluation module is connected to the meteorological simulation module and the numerical tracing module, including a meteorological path clustering unit, a tracing result analysis unit, a transmission path matching unit and an impact contribution evaluation unit, which is used to analyze and evaluate the transmission path of pollutants and their contribution to urban air pollution through a machine learning algorithm based on the meteorological field data provided by the meteorological simulation module and the tracing contribution results of the numerical model tracing module.

[0008] As a preferred technical solution of the present invention, the initial meteorological field download unit is configured with an automatic meteorological initial field data download program for regular automatic download of meteorological initial field basic data; the meteorological field simulation unit is configured with a WRF meteorological model for high-precision meteorological field simulation of the region; the meteorological simulation data extraction and post-processing unit is used to clip and extract the required meteorological data.

[0009] As a more preferred technical solution of the present invention, the meteorological simulation data extraction and post-processing unit is used to trim and extract the meteorological field boundary and key data elements to output the hourly meteorological field required by the numerical model tracing module.

[0010] As a preferred technical solution of the present invention, the anthropogenic source emission inventory generating unit is configured with the total anthropogenic source emissions in the target area and the distribution coefficients of the corresponding pollutants for generating the anthropogenic source emission inventory.

[0011] As a more preferred technical solution of the present invention, the anthropogenic source emission inventory generation unit splits and grids the total anthropogenic source emissions based on characteristic factor coefficients, and generates an anthropogenic source emission inventory based on a program algorithm.

[0012] As a preferred technical solution of the present invention, the natural source emission inventory generation unit is configured with a MEGAN model, basic geographic information data of the study area, and an automatic generation program for the natural source emission inventory, which is used to simulate a high-resolution natural source emission inventory.

[0013] As a preferred technical solution of the present invention, the emission inventory online marking unit has a built-in operating program for automatically integrating and renaming species data in the anthropogenic emission inventory and the natural source emission inventory, so as to generate an emission inventory with regional markings.

[0014] As a preferred technical solution of the present invention, the numerical model tracing module receives and integrates the data output by the meteorological simulation module and the pollution source emission inventory module, simulates the concentration of ambient atmospheric pollutants through the built-in tracing chemical transport module, and quantitatively evaluates the contribution of emission sources in different regions to the concentration of atmospheric pollutants.

[0015] As a preferred technical solution of the present invention, the meteorological path clustering unit is internally configured with one of the HYPLIST or METEOINFO models, and air mass trajectories are clustered based on high-precision meteorological data to obtain the meteorological trajectory of the target city; the tracing result analysis unit has a built-in matching screening algorithm for processing the contribution results of atmospheric pollutants in different cities and screening upstream and downstream cities in different regions.

[0016] As a preferred technical solution of the present invention, the transmission path matching unit has a built-in machine learning algorithm, which connects the meteorological trajectories of various cities in the region and the contribution data of pollutant sources through the machine learning algorithm to form the atmospheric pollutant transmission path in the region.

[0017] As a preferred technical solution of the present invention, the impact contribution assessment unit integrates the contribution results of the air pollutant sources in each city and the air pollutant transport paths in the region to quantitatively assess the contribution of the pollution transmission paths to the regional air pollution.

[0018] The second aspect of the present invention further provides an evaluation method for urban air pollution transmission paths and source contributions based on machine learning and numerical models, using the evaluation system described in the first aspect, comprising the following steps:

[0019] S1. Start the meteorological simulation module to obtain the required meteorological data through downloading initial meteorological data, simulating the meteorological field and extracting meteorological data;

[0020] S2. Start the pollution source emission inventory module, generate anthropogenic source emission inventory and a high-precision natural source emission inventory by allocating the total amount of anthropogenic emissions and the meteorological data obtained in S1; merge the anthropogenic source emission inventory and the high-precision natural source emission inventory using an embedded program, and mark the merged emission inventory online according to the city division;

[0021] S3, based on the pollutant emission inventory marked online in S2 and the meteorological field simulated in S1, the numerical model source tracing module simulates the atmospheric pollutant concentration through the built-in source tracing chemical transport simulation program to quantitatively evaluate the contribution of pollution sources in different regions;

[0022] S4, a transmission path assessment module based on machine learning, combines the meteorological data of S1 and the pollution source contribution data of S3 to assess the transmission path of atmospheric pollutants and their contribution to regional pollution.

[0023] As a preferred technical solution of the present invention, the specific steps of starting the meteorological simulation module in step S1, obtaining the required meteorological data through meteorological initial data download, meteorological field simulation and meteorological data extraction are: starting the meteorological simulation module, carrying out high-precision multi-layer nested WRF simulation in the meteorological field simulation unit, and comparing the simulation results with the actual meteorological data obtained by the initial meteorological field download unit, adjusting the parameterization scheme of the WRF model until the simulation results reach the optimal; based on the optimized simulation results, the meteorological simulation data extraction and post-processing unit prunes and extracts the meteorological field boundary and key data elements, and outputs the standardized hourly meteorological field required by the subsequent numerical traceability module;

[0024] As a preferred technical solution of the present invention, step S4 specifically includes the following steps:

[0025] S401, a transmission path evaluation module based on machine learning, combines the meteorological data of S1, uses the meteorological path clustering model and the built-in clustering algorithm to obtain the meteorological trajectory of the target city;

[0026] S402, combined with the pollution source contribution information in S3, taking cities as units, analyze the atmospheric pollutant contribution results of each city and screen the upstream and downstream cities of the path, select RNN, CNN, SHARP and other algorithms, train the meteorological trajectory and the pollutant contribution data of the upstream and downstream cities, and use cross-validation to evaluate each algorithm model to form an atmospheric pollutant transport path evaluation system in the region;

[0027] S403. Integrate the source contribution results of the numerical model tracing module and the assessment results of the atmospheric pollutant transport paths, conduct quantitative analysis of the source contribution of atmospheric pollutants at the city level, and evaluate the contribution of different pathways to regional pollution in combination with the pollutant transport paths.

[0028] Compared with the prior art, the present invention has the following beneficial effects:

[0029] The present invention realizes meteorological field simulation, regional emission inventory generation and labeling through meteorological simulation module and pollution emission inventory module, and then realizes concentration simulation of environmental atmospheric pollutants and impact assessment of different regional emissions on atmospheric pollutant concentration through built-in traceability chemical transport module based on numerical model traceability module, and finally realizes meteorological path clustering and urban upstream and downstream paths matching through machine learning transport path assessment module, forms regional atmospheric pollutant transport path and realizes impact contribution assessment of regional atmospheric pollution. The present invention combines machine learning with numerical model, has high automation, can automatically download, automatically run and analyze data online without manual statistical analysis, is easy to operate, has high data accuracy and fast response speed, and provides important support for tracing the causes of atmospheric pollutants and formulating atmospheric pollution control plans. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 A schematic diagram of the structure of the evaluation system for the transmission path and source of urban air pollution based on machine learning and numerical models of the present invention;

[0031] Figure 2 A flow chart of the method for evaluating the transmission paths and sources of urban air pollution based on machine learning and numerical models of the present invention;

[0032] Figure 3 This is the impact of the out-of-region transmission air mass on Zhengzhou, a typical city in the application example of the present invention;

[0033] Figure 4 The PM values ​​of typical months in four quarters of a typical year in 18 cities in Henan Province in the application example of the present invention are shown in Table 1. 2.5 The conveying path;

[0034] Figure 5 The PM in a typical month in Zhengzhou, a typical city in Henan Province, in the application example of the present invention is 2.5 Affected by the contribution of transmission from the upper and lower paths, the unit in the figure is tons / day. DETAILED DESCRIPTION

[0035] In order to enable those skilled in the art to more clearly understand the technical solution of the present invention, the technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0036] Example 1

[0037] This embodiment provides an evaluation system for the transmission path and source contribution of urban air pollution based on machine learning and numerical models, and its structural diagram is shown in FIG. Figure 1 As shown, it includes a meteorological simulation module, a pollution emission inventory module, a numerical model tracing module and a machine learning transportation path assessment module.

[0038] The meteorological simulation module includes an initial meteorological field download unit, a meteorological field simulation unit, and a meteorological simulation data extraction and post-processing unit. The initial meteorological field download unit downloads the initial meteorological data through regular access, the meteorological simulation unit is used to simulate the regional and urban meteorological fields with high precision, and the meteorological simulation data extraction and post-processing unit is used to trim and extract the boundaries and key data elements of the regional and urban meteorological fields, and output the hourly meteorological field data required by the subsequent numerical model traceability module.

[0039] The pollution emission inventory module includes an anthropogenic emission inventory generation unit, a natural source emission inventory generation unit and an emission inventory online marking unit. The anthropogenic emission inventory generation unit is used to split and grid the total anthropogenic emission based on the characteristic factor coefficient, and generate an anthropogenic emission inventory based on the program algorithm; the natural source emission inventory generation unit is used to receive the meteorological data file output by the meteorological simulation data extraction and post-processing unit, and generate a natural source emission inventory based on the program algorithm; the emission inventory online marking unit is used to receive the anthropogenic emission data and natural source emission data calculated by the anthropogenic emission inventory generation unit and the natural source emission inventory generation unit, and mark the emission data based on the regional / city classification, generate an emission inventory with regional markings, and transmit the generated emission inventory data with regional markings to the numerical model traceability module.

[0040] The numerical model traceability module is used to receive the data information integrated and output by the meteorological simulation module and the pollution emission inventory module. It simulates the concentration of ambient atmospheric pollutants through the built-in traceability chemical transport module and evaluates the impact of emissions from different regions on the concentration of atmospheric pollutants. The spatial resolution of the output data is 4km×4km.

[0041] The machine learning transport path assessment module includes a meteorological path clustering unit, a traceability result analysis unit, a transmission path matching unit, and an impact contribution assessment unit. The meteorological path clustering unit is based on the high-precision meteorological field simulation data output by the meteorological simulation module, and obtains the meteorological trajectory of the target city through a built-in clustering algorithm; the traceability result analysis unit is used to receive the contribution results of atmospheric pollutants in different cities with regional labels output by the numerical model traceability module, and screen the upstream and downstream cities of the path; the transmission path matching unit has a built-in machine learning algorithm, and establishes a transmission path matching unit through learning and training, and connects the meteorological trajectory of the target city, all cities in the region, and the results of upstream and downstream cities to form the transport path of atmospheric pollutants in the region; the impact contribution assessment unit will use the results of the numerical model traceability module to integrate the contribution results of atmospheric pollutants from different cities and the results of atmospheric pollutants transport paths in the region to assess the contribution of pollution transmission paths to regional atmospheric pollution.

[0042] The meteorological simulation module and the pollution emission inventory module of the present invention are both connected to the numerical model tracing module, wherein the WRF model in the meteorological field simulation unit of the meteorological simulation module adopts three-layer nesting, and the spatial resolutions are 36km×36km, 12km×12km, 4km×4km, and 17 layers in the vertical direction. The simulation variables include meteorological parameters such as wind temperature and humidity pressure. The meteorological simulation data extraction and post-processing unit has a built-in MCIP program, which cuts the NC data simulated by WRF into the standard format required by the numerical model tracing module according to the set grid information. The anthropogenic source emission inventory generation unit of the pollution emission inventory module has built-in gridded anthropogenic source emission inventories of different scales such as East Asia and China, and monthly, weekly and hourly emission allocation coefficients of different pollutants and different emission sources. The emission inventory online marking unit has built-in grid data based on the geographical location of the city, with spatial resolution. The rate is 4km×4km. After marking different types of emissions, a spatial resolution of 4km×4km with regional marking and a time resolution of 1 hour are obtained, which includes an emission inventory of species such as SO2, NO2, PM, VOC, EC, OC, etc., and is transmitted to the numerical model traceability module. The numerical model traceability module is configured with an air quality mode, where the air quality mode is one or more of WRF-CHEM, CMAQ, and CAMx. It is based on the standardized meteorological field data output by the meteorological simulation module and the emission inventory with regional markings output by the pollution source emission inventory module. The concentration simulation of environmental atmospheric pollutants and the impact of emissions from different regions on the concentration of atmospheric pollutants are realized through the built-in traceability chemical transport module. The spatial resolution of the output data is 4km×4km, which is the simulation result of atmospheric pollutant concentration with source contribution marking.

[0043] The meteorological simulation module and the numerical model tracing module are connected to the machine learning transmission path evaluation module. The machine learning transmission path evaluation module includes a meteorological path clustering unit, a tracing result analysis unit, a transmission path matching unit and an impact contribution evaluation unit. The meteorological path clustering unit is based on the high-precision meteorological field simulation data output by the meteorological simulation module, and obtains the meteorological trajectory of the target city through built-in clustering algorithms such as HYPLIST or METEOINFO. The tracing result analysis unit is used to receive the marked city-by-city atmospheric pollutant contribution results output by the numerical model tracing module and screen the upstream and downstream cities of the path. The transmission path matching unit has a built-in machine learning algorithm, and selects RNN, CNN, SHARP and other algorithms to establish a transmission path matching unit through learning and training to connect the meteorological trajectory of the target city, all cities in the region and the upstream and downstream city results to form the atmospheric pollutant transmission path in the region. The impact contribution evaluation unit and the numerical tracing module are used to evaluate the contribution of the pollution transmission path to the regional atmospheric pollution by integrating the contribution results of the atmospheric pollutant sources of the cities and the atmospheric pollutant transmission path results in the region.

[0044] The system for evaluating the transmission path and source contribution of urban air pollution based on machine learning and numerical models provided by the present invention realizes meteorological field simulation, regional emission inventory generation and labeling through a meteorological simulation module and a pollution emission inventory module, and then realizes the concentration simulation of environmental air pollutants and the impact assessment of different regional emissions on the concentration of air pollutants through a built-in traceability chemical transport module based on the numerical model traceability module, and finally realizes the clustering of meteorological paths and the matching of urban upstream and downstream paths through a machine learning transport path assessment module, so as to form a regional air pollutant transport path and realize the impact contribution assessment of regional air pollution. The present invention combines machine learning with numerical models, has a high degree of automation, can automatically download, automatically run and analyze data online without manual operation, is easy to operate, has high data accuracy, and has a fast response speed, providing important support for tracing the causes of air pollutants and formulating air pollution control plans.

[0045] Example 2

[0046] This embodiment provides an evaluation method for the transmission path and source contribution of urban air pollution based on machine learning and numerical models. For a flowchart, see Figure 2 , using the evaluation system described in Example 1, specifically comprising the following steps:

[0047] S1. Start the meteorological simulation module to obtain the required meteorological data through downloading initial meteorological data, simulating the meteorological field and extracting meteorological data;

[0048] Specifically, the meteorological simulation module is started, the initial meteorological data is downloaded in the initial meteorological field download unit, high-precision multi-layer nested WRF simulation is carried out in the meteorological simulation unit, the meteorological field boundary and key data elements are trimmed and extracted in the meteorological simulation data extraction and post-processing unit, and the standardized hourly meteorological field data required by the subsequent numerical model tracing module is output;

[0049] S2. Start the pollution source emission inventory module, generate anthropogenic source emission inventory and a high-precision natural source emission inventory by allocating the total amount of anthropogenic emissions and the meteorological data obtained in S1; merge the anthropogenic source emission inventory and the high-precision natural source emission inventory using an embedded program, and mark the merged emission inventory online according to the city division;

[0050] Specifically, the pollution emission inventory module is started, and the total amount of anthropogenic emissions and the generation of a high-precision gridded anthropogenic source emission inventory are carried out in the pollution emission inventory module. Based on the standardized meteorological data obtained in the above step S1, a high-precision natural source emission inventory is generated based on the MEGAN model simulation, and then the anthropogenic source emission inventory is merged with the natural source emission inventory and the pollution emission inventory based on urban zoning is online marked based on the embedded program;

[0051] S3, based on the pollutant emission inventory marked online in S2 and the meteorological field simulated in S1, the numerical model source tracing module simulates the atmospheric pollutant concentration through the built-in source tracing chemical transport simulation program to quantitatively evaluate the contribution of pollution sources in different regions;

[0052] Specifically, based on the online marked pollutant dynamic emission inventory generated in step S2, and combined with the high-precision meteorological field simulated in step S1 to drive the model, parameters are set in the chemical transport model, and atmospheric pollutant simulation is carried out using one or more of WRF-CHEM, CMAQ, and CAMx to quantitatively evaluate the pollution sources and contributing factors of atmospheric pollutants in the simulation unit;

[0053] S4, a transmission path assessment module based on machine learning, combines the meteorological data of S1 and the pollution source contribution data of S3 to assess the transmission path of atmospheric pollutants and their contribution to regional pollution.

[0054] Specifically, step S4 includes the following steps:

[0055] S401, a transmission path evaluation module based on machine learning, combines the meteorological data obtained in S1, and uses the meteorological path clustering model HYPLIST or METEOINFO through a built-in clustering algorithm to obtain the meteorological trajectory of the target city;

[0056] S402, combined with the pollution source contribution information in S3, taking cities as units, analyze the atmospheric pollutant contribution results of each city and screen the upstream and downstream cities of the path, select RNN, CNN, SHARP and other algorithms, train the urban meteorological trajectory and the pollutant contribution data of upstream and downstream cities, and use cross-validation to evaluate each algorithm model to form an atmospheric pollutant transport path evaluation system in the region;

[0057] S403. Integrate simulation results to quantify the contribution of air pollutant sources in different cities and evaluate the transport pathways of air pollutants at the city level within the region, and combine the transport pathways to achieve quantitative evaluation of the contribution of different pollution transmission pathways to regional air pollution.

[0058] Example 3

[0059] This embodiment provides an electronic device, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, any step in the method for evaluating the transmission path and source contribution of urban air pollution based on machine learning and numerical models as described in Example 2 is implemented.

[0060] Further, the evaluation method of the urban air pollution transmission path and source contribution based on machine learning and numerical model described in Example 2 can be implemented as a computer software program. For example, the present embodiment includes a computer program product, which includes a computer program carried on a computer readable medium, and the computer program contains a program code for executing the method. In such an embodiment, the computer program can be downloaded and installed from the network, and / or installed from a removable medium. When the computer program is executed by the processor, the above functions defined in the method of the present application are executed.

[0061] Example 4

[0062] A computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements any step in a method for evaluating the transmission path and source contribution of urban air pollution based on machine learning and numerical models as described in Example 2.

[0063] The computer-readable medium described in the present application may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to, an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which a computer-readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0064] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0065] Application Examples

[0066] This application example is based on the evaluation system of Example 1 or the evaluation method of Example 2 to conduct an example analysis, and evaluate the air pollution transmission path and source contribution in Henan Province. Typical cities are affected by air masses transmitted from outside the region, such as Figure 3 As shown in the figure, the PM of 18 cities in Henan Province after the merger 2.5 The transport path of concentration is as follows Figure 4 As shown in the figure, the results of the upstream and downstream contributions to the city are as follows Figure 5 shown.

[0067] Figure 3 This is the impact of the out-of-region air mass transmission from Zhengzhou, a typical city, on the transport trajectory from outside the region to Zhengzhou. The trajectory with the highest probability is the trajectory from the southeast of Zhengzhou to Zhengzhou, which comes from Xuchang, Zhoukou and Bozhou, with a probability of 22.2%; the second is the northeast direction along Anyang-Hebi-Xinxiang, with a probability of 15%; the transmission along the Zhengzhou-Kaifeng-Heze direction accounts for 14%.

[0068] Figure 4 PM in typical months of four quarters of a typical year in 18 cities in Henan Province 2.5Concentration transmission paths: In winter (January), there are four main transmission paths: from Anyang to Jiyuan to the southwest, from Jiaozuo to the southeast to Xinyang and Zhoukou respectively, and from Sanmenxia to the southeast to Xinyang. In spring, the main transmission paths are from Jiyuan to the northeast to Anyang, from Sanmenxia to the east to Kaifeng and Shangqiu respectively, and from Nanyang to the northeast to Shangqiu. In summer, the main transmission paths are from Sanmenxia to the northeast to Anyang, from Sanmenxia to the east to Shangqiu, from Nanyang to the northeast to Kaifeng, and from Xinyang to the northeast to Shangqiu. In autumn, the main transmission paths are from Anyang to the southwest to Jiyuan, from Jiaozuo to the southeast to Xinyang, from Sanmenxia to the southeast to Xinyang, and from Zhoukou to the southwest to Xinyang. In summary, PM2.5 in Henan Province is 2.447 million tons, 1.78 million tons of PM2.5 is 2.11 million tons, and 1.75 million tons of PM2.5 is 2.11 million tons, which is 1.3 times the PM2.5. 2.5 In autumn and winter, it is mostly transmitted from northeast to southwest and from northwest to southeast; in spring, it is mostly transmitted from west to east; and in summer, it is mostly transmitted from southwest to northeast.

[0069] Figure 5 PM in Zhengzhou City in a typical month 2.5 The contribution of transmission from upstream to downstream is as follows: PM from upstream cities Luoyang, Jiaozuo, and Xinxiang 2.5 The net inflows were 328 tons / day, 623 tons / day and 325 tons / day respectively. 2.5 The downstream cities are Pingdingshan, Xuchang and Kaifeng, with transport contributions of 61.2 tons / day, 459 tons / day and 194 tons / day respectively. 2.5 It stays in the local area and causes pollution.

[0070] The above embodiments are specific implementation methods of the present invention, but the implementation methods of the present invention are not limited to the above embodiments. Any other combination, change, modification, substitution and simplification that does not exceed the design concept of the present invention shall fall within the protection scope of the present invention.

Claims

1. An evaluation system for urban air pollution transmission paths and source contributions based on machine learning and numerical models, characterized in that: It includes a meteorological simulation module, a pollution emission inventory module, a numerical model tracing module and a machine learning transportation path evaluation module; the meteorological simulation module includes an initial meteorological field download unit, a meteorological field simulation unit and a meteorological simulation data extraction and post-processing unit, which are used to simulate and process high-precision meteorological field data of the target area; the pollution source emission inventory module includes a man-made source emission inventory generation unit, a natural source emission inventory generation unit and an emission inventory online marking unit, which are used to generate and mark emission data of various pollution sources in the region; the numerical model tracing module is used to simulate the concentration of atmospheric pollutants according to meteorological field data and pollution source emission data, and evaluate the contribution of emission sources in different regions to the concentration of atmospheric pollutants in the target area; the machine learning transportation path evaluation module is connected to the meteorological simulation module and the numerical tracing module, including a meteorological path clustering unit, a tracing result analysis unit, a transmission path matching unit and an impact contribution evaluation unit, which are used to analyze and evaluate the transportation path of pollutants and their contribution to urban air pollution through machine learning algorithms based on the meteorological field data provided by the meteorological simulation module and the tracing contribution results of the numerical model tracing module.

2. The system for evaluating the transmission paths and source contributions of urban air pollution based on machine learning and numerical models according to claim 1, characterized in that: The initial meteorological field download unit is configured with an automatic meteorological initial field data download program for regularly and automatically downloading the basic data of the meteorological initial field; the meteorological field simulation unit is configured with a WRF meteorological model for high-precision meteorological field simulation of the region; the meteorological simulation data extraction and post-processing unit is used to clip and extract the required meteorological data.

3. The system for evaluating the transmission paths and source contributions of urban air pollution based on machine learning and numerical models according to claim 1, characterized in that: The anthropogenic source emission inventory generation unit is configured with the total anthropogenic source emission amount in the target area and the distribution coefficient of the corresponding pollutant, and is used to generate the anthropogenic source emission inventory.

4. The system for evaluating the transmission paths and source contributions of urban air pollution based on machine learning and numerical models according to claim 1, characterized in that: The natural source emission inventory generation unit is configured with a MEGAN model, basic geographic information data of the research area, and an automatic generation program for the natural source emission inventory, which is used to simulate a high-resolution natural source emission inventory.

5. The system for evaluating the transmission paths and source contributions of urban air pollution based on machine learning and numerical models according to claim 1, characterized in that: The emission inventory online labeling unit has a built-in operating program for automatically integrating and renaming species data in the anthropogenic emission inventory and the natural source emission inventory, and is used to generate an emission inventory with regional labels.

6. The system for evaluating the transmission paths and source contributions of urban air pollution based on machine learning and numerical models according to claim 1, characterized in that: The numerical model tracing module receives and integrates the data output by the meteorological simulation module and the pollution source emission inventory module, simulates the concentration of ambient air pollutants through the built-in tracing chemical transport module, and quantitatively evaluates the contribution of emission sources in different regions to the concentration of air pollutants.

7. The system for evaluating the transmission paths and source contributions of urban air pollution based on machine learning and numerical models according to claim 1, characterized in that: The meteorological path clustering unit is internally configured with one of the HYPLIST or METEOINFO models, and performs air mass trajectory clustering based on high-precision meteorological data to obtain the meteorological trajectory of the target city; The traceability result analysis unit has a built-in matching screening algorithm for processing the atmospheric pollutant contribution results of each city and screening the upstream and downstream cities in different regions.

8. The system for evaluating the transmission paths and source contributions of urban air pollution based on machine learning and numerical models according to claim 1, characterized in that: The transmission path matching unit has a built-in machine learning algorithm, which connects the meteorological trajectories of cities in the region and the contribution data of pollutant sources to form the transmission path of atmospheric pollutants in the region.

9. The system for evaluating the transmission paths and source contributions of urban air pollution based on machine learning and numerical models according to claim 1, characterized in that: The impact contribution assessment unit integrates the contribution results of air pollutant sources in different cities and the results of air pollutant transport paths in the region, and quantitatively assesses the contribution of pollution transmission paths to regional air pollution.

10. A method for evaluating the transmission paths and source contributions of urban air pollution based on machine learning and numerical models, characterized in that: The application of the evaluation system described in any one of claims 1 to 9 comprises the following steps: S1. Start the meteorological simulation module to obtain the required meteorological data through downloading initial meteorological data, simulating the meteorological field and extracting meteorological data; S2. Start the pollution source emission inventory module, generate anthropogenic source emission inventory and a high-precision natural source emission inventory by allocating the total amount of anthropogenic emissions and the meteorological data obtained in S1; merge the anthropogenic source emission inventory and the high-precision natural source emission inventory using an embedded program, and mark the merged emission inventory online according to the city division; S3, based on the pollutant emission inventory marked online in S2 and the meteorological field simulated in S1, the numerical model source tracing module simulates the atmospheric pollutant concentration through the built-in source tracing chemical transport simulation program to quantitatively evaluate the contribution of pollution sources in different regions; S4, a transmission path assessment module based on machine learning, combines the meteorological data of S1 and the pollution source contribution data of S3 to assess the transmission path of atmospheric pollutants and their contribution to regional pollution.

Images