Key point location identification method, electronic equipment, storage medium and program product

Abstract

The invention belongs to the field of environmental engineering, and particularly discloses a key point location identification method, electronic equipment, a storage medium and a program product. The method comprises the following steps: constructing a multi-dimensional data array based on various data output by an air quality model; performing dimension reduction on the multi-dimensional data array based on a principal component analysis method to obtain a first principal component; and setting a load threshold value and selecting a key point location from the first principal component based on the load threshold value. The method has the advantages that a regional atmosphere monitoring network can be optimized, and the monitoring and analysis capability of a monitoring station on a key transmission channel when regional pollution occurs is improved.

Description

A method for identifying key locations, and an electronic device, storage medium, and program product. Technical Field

[0001] This application relates to the field of environmental engineering, and in particular to a method for identifying key locations, as well as electronic devices, storage media, and program products. Background Technology

[0002] As my country's efforts to control air pollution deepen, pollution control strategies are shifting from extensive, generalized approaches to precise, scientific, and source-tracing methods. Against this backdrop, how to scientifically and rationally deploy environmental monitoring stations to capture the regional spatial distribution characteristics and formation mechanisms of pollution at the lowest possible cost has become a crucial technical issue in environmental management.

[0003] Currently, methods for selecting key locations largely rely on statistical analysis of historical observation data or simple spatial interpolation and clustering of pollutant concentration fields simulated by air quality models. However, these traditional methods have significant limitations. First, they over-rely on the surface spatiotemporal distribution of pollutant concentrations, completely ignoring the complex physical transport and chemical transformation processes driving the concentration field. For example, a high-concentration location may be caused by local emissions or regional transport, and its representativeness of the overall regional pollution pattern is fundamentally different. Second, faced with the massive, multi-dimensional output data generated by air quality models, existing methods lack effective data mining and feature extraction capabilities, making it difficult to identify, at the mechanistic level, the key spatial locations that play a decisive role in the regional pollution distribution pattern. Summary of the Invention

[0004] To address the aforementioned problems in the prior art, this invention provides a key point identification method, as well as an electronic device, storage medium, and program product.

[0005] First, this invention provides a key point identification method, comprising the following steps:

[0006] A multidimensional data array is constructed based on various data outputs from the air quality model;

[0007] The multidimensional data array is reduced in dimensionality using principal component analysis to obtain the first principal component.

[0008] Set a load threshold and select key points from the first principal component based on the load threshold.

[0009] The above method, in constructing the air quality model, includes the following steps:

[0010] Determine the target space and target time;

[0011] Meteorological field data is used to process emission inventory data to match the resolution of the target space and the target time.

[0012] Set environmental variables to activate the process analysis module to analyze the physicochemical process contribution and chemical reaction pathways of pollutants;

[0013] Run and verify the reliability of the air quality model.

[0014] In the above method, the process analysis module includes an integrated process rate module and an integrated reaction rate module, wherein:

[0015] The integrated process rate module directly outputs the contribution rate of each physicochemical process term to the concentration change of the pollutant by integrating and separating the various physicochemical process terms in the model equation online.

[0016] The integrated reaction rate module tracks and outputs the cumulative rate of specific reactions in the chemical mechanism to quantify the contribution of a single chemical reaction to the generation or consumption of the pollutant, thereby obtaining the chemical reaction pathway.

[0017] In the above method, constructing a multidimensional data array based on various data output from the air quality model includes:

[0018] Extract pollutant concentration field data, process contribution field data of integrated process rates, and reaction path contribution field data of integrated reaction rates within the period of interest;

[0019] The pollutant concentration field data, process contribution field data, and reaction pathway contribution field data are integrated into a total number of variables to construct a three-dimensional data array with dimensions n×i×j, where n is the total number of variables, i is the number of longitude grids, and j is the number of latitude grids.

[0020] In the above method, the integrated process rate includes at least: horizontal advection process, vertical convection process, vertical diffusion process, dry deposition process, emission process, gas phase chemical process, aerosol process, cloud process, and wet deposition process.

[0021] In the above method, the reaction pathway for the integrated reaction rate includes at least: the chemical generation and consumption of ozone, NO... x Key cyclic reactions, generation and consumption of OH radicals, HO X The generation and consumption of VOCs.

[0022] In the above method, the step of reducing the dimensionality of the multidimensional data array based on principal component analysis to obtain the first principal component includes:

[0023] Data preprocessing involves reconstructing the three-dimensional data array into a two-dimensional matrix X, where the dimension of the two-dimensional matrix X is n×(i×j).

[0024] Principal component extraction: Principal component analysis is performed on the two-dimensional matrix X, and the principal component with the highest variance contribution rate is selected as the first principal component.

[0025] Spatial feature field reconstruction involves mapping the first principal component back to the geospatial grid as a vector to obtain the spatial distribution feature field of the first principal component of i×j.

[0026] In the above method, setting a load threshold and selecting key points from the first principal component based on the load threshold includes:

[0027] The two-dimensional matrix X is orthogonally rotated to obtain the first principal component load matrix, and then Varimax principal component analysis is performed on it.

[0028] Based on a preset load threshold, the spatial grid that contributes the most to the first principal component is selected from the first principal component load matrix as key points.

[0029] The present invention also provides an electronic device, comprising:

[0030] At least one memory for storing computer instructions;

[0031] At least one processor implements the above method when the at least one processor executes the computer instructions.

[0032] The present invention also provides a computer-readable storage medium storing at least one computer instruction, which is loaded and executed by a processor to implement the above-described method.

[0033] The present invention also provides a computer program product, the computer program product including computer instructions, which, when executed, implement the above-described method.

[0034] Compared with existing technologies, this application employs a process analysis approach when establishing an air quality model. After obtaining multidimensional data through model operation, principal component analysis is applied to reduce the dimensionality and analyze the acquired data. This simplifies the massive multidimensional data and descriptively reveals the causal relationships between different data by exposing the covariant structure among them, thereby identifying key data points, i.e., critical locations. This process not only reflects the spatial distribution hotspots of pollutant concentrations but also reveals the spatial hotspots of the core physicochemical processes driving this distribution (such as transport, chemical formation, and deposition). The locations selected by this method thus possess deeper mechanistic representativeness and higher practical application value, enabling the optimization of regional atmospheric monitoring networks and improving the monitoring and analysis capabilities of monitoring stations on key transport channels during regional pollution events. Attached Figure Description

[0035] Figure 1 shows a flowchart of a key point identification method according to some embodiments of this application;

[0036] Figure 2 shows a sub-flowchart of S1 in the flowchart shown in Figure 1;

[0037] Figure 3 shows the sub-flowchart of S2 in the flowchart shown in Figure 1;

[0038] Figure 4 shows the sub-flowchart of S3 in the flowchart shown in Figure 1;

[0039] Figure 5 shows the sub-flowchart of S4 in the flowchart shown in Figure 1;

[0040] Figure 6 shows the point distribution map of a certain region, in which the loading of the first principal component is marked;

[0041] Figure 7 is a hardware structure block diagram of the electronic device of the present invention. Detailed Implementation

[0042] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0043] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used in this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items. Words such as “comprising” or “including” mean that the element or object preceding “comprising” or “including” covers the element or object listed following “comprising” or “including” and its equivalents, and does not exclude other elements or objects. Words such as “connected” or “linked” are not limited to physical or mechanical connections and can include electrical connections, whether direct or indirect.

[0044] It should be understood that although the terms first, second, third, etc., may be used in this application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0045] The embodiments of this application will now be described in further detail with reference to the accompanying drawings.

[0046] Figure 1 shows a flowchart of a key point identification method according to some embodiments of this application. As shown in the figure, the method proposed in the embodiments of this application mainly includes four steps:

[0047] S1, Construct an air quality model;

[0048] S2, construct a multidimensional data array;

[0049] S3, perform dimensionality reduction on the multidimensional data array to obtain the first principal component;

[0050] S4, select key points from the first principal component.

[0051] The specific steps are explained below with reference to Figures 2-5.

[0052] First, referring to Figure 2, constructing the air quality model (step S1) includes the following steps:

[0053] S11 allows you to build a basic air quality model by setting the simulation range and simulation time in an air quality model (such as a CMAQ or CAMx model) to determine the target space and the target time period (segment) to be studied. For example, the simulation range can be configured with parameters such as grid resolution, vertical stratification, and map projection; the simulation time can be set with a "spin-up" period to eliminate the influence of initial conditions. The meteorological field of the air quality model is driven by WRF (Weather Research and Forecasting Model). WRF, CMAQ, and CAMx models are all open-source computer numerical models used for weather and air quality forecasting. It is not an instrument, but an extremely complex set of mathematical programs and physical schemes running on supercomputers to simulate and predict atmospheric behavior.

[0054] S12 utilizes the grid and time information set in S11 to process high-precision emission inventory data (such as MEIC), performs spatial and temporal allocation to match the temporal and spatial resolution of the air quality model, and prepares initial and boundary condition files.

[0055] S13. Configure the air quality model before compilation. Set environmental variables to initiate the Process Analysis (PA) module, which analyzes the physicochemical contributions and chemical reaction pathways of pollutants. The PA module typically quantifies the contribution of the following processes to changes in the concentration of specific pollutants (such as O3 and PM2.5):

[0056] • Advection transport: Transport carried by horizontal wind fields.

[0057] • Turbulent diffusion: a vertical mixing process.

[0058] • Dry deposition: The deposition and removal of pollutants to the ground surface.

[0059] • Gas phase chemistry: The chemical reactions (formation or consumption) of pollutants in the air.

[0060] • Aerosol processes include nucleation, condensation, evaporation, and gas-particle partitioning.

[0061] • Emissions: Local and regional sources of pollutant emissions.

[0062] • Cloud processes and liquid phase chemistry (if the model includes them).

[0063] Its typical output is the contribution rate (μg / m³ / h or ppb / h) for each grid, each time point, each pollutant, and each process. By integration, the net contribution of each process in the entire pollution event can be obtained.

[0064] In the operation control file, explicitly set the output process analysis data and specify the pollutant species to be analyzed (e.g., select O3 or PM). 2.5 ).

[0065] The IPR (Integrated Process Rate) module directly outputs the contribution rate of each process (such as horizontal advection, vertical diffusion, gas-phase chemistry, aerosol processes, etc.) to species concentration changes by integrating and separating the various physicochemical process terms in the model equations online. The IRR (Integrated Reaction Rate) module tracks and outputs the cumulative rate of specific reactions in the chemical mechanism, used to quantify the contribution of a single chemical reaction to the generation or consumption of a certain species, and accurately analyze complex chemical reaction pathways.

[0066] S14. Run and verify the reliability of the air quality model. After completing the steps in S11-13 above, the air quality model is ready to run. Run it several times to ensure that the error between the simulated concentration values ​​and the observed values ​​is within an acceptable range, thus guaranteeing the reliability of the analysis results in subsequent processes.

[0067] The second step is to construct a multidimensional data array (step S2), which includes extracting the following three types of data within the time period t of interest based on the results of the air quality model in S1, thereby constructing a standardized multidimensional data array.

[0068] The three types of data include:

[0069] 1. Extract pollutant concentration field data for the period of interest. Specifically, this can be done by obtaining the target pollutant (O3 or PM2.5). 2.5 The hourly average data of grid point concentrations.

[0070] 2. Extract process contribution field data for integrated process rates. Specifically, the integrated process rates include at least: horizontal advection processes, vertical convection processes, vertical diffusion processes, dry deposition processes, emission processes, gas phase chemical processes, aerosol processes, and cloud processes and wet deposition processes.

[0071] 3. Extract reaction pathway contribution field data for integrated reaction rates. Specifically, the reaction pathways for integrated reaction rates should include at least: the chemical generation and consumption of ozone, NO... x Key cyclic reactions, generation and consumption of OH radicals, HO X The generation and consumption of VOCs.

[0072] After preparing the above three types of data, perform the following steps, referring to Figure 3:

[0073] S21, the above-mentioned multiple pollutant concentration field data, process contribution field data, and reaction pathway contribution field data are integrated into a single variable count, constructing a three-dimensional data array with dimensions n×i×j, where n is the total number of variables, i is the number of longitude grids, and j is the number of latitude grids. For example, in one embodiment, the above-mentioned multiple pollutant concentration field data, process contribution field data, and reaction pathway contribution field data include: ozone concentration, chemical generation, cloud physics, dry deposition, horizontal advection, horizontal diffusion, vertical diffusion, vertical advection, HOx generation, ozone depletion, ozone generation, OH generation, and VOCs depletion, a total of 13 variables; the spatial longitude and latitude grids for collecting the above data are all 112, thus constructing a three-dimensional data array with dimensions of 13×112×112. The data in this three-dimensional data array is constructed by dividing the area of ​​interest in this embodiment into 112×112 grids, and collecting 13 types of data in each grid.

[0074] The third step involves dimensionality reduction of the multidimensional data array obtained in the second step using principal component analysis (S3) to obtain the first principal component. The first principal component is the most important comprehensive direction in the multidimensional data. It uses a new variable that maximally summarizes the overall variation structure of the original multidimensional data, representing the best one-dimensional representation of the "main signal" hidden in the original variables. It can compress multidimensional data into the core low-dimensional space with minimal information loss, comprehensively representing the spatial distribution characteristics of chemical processes; it can be considered a weighted algorithm.

[0075] Referring to Figure 4, the specific steps include:

[0076] S31. Data preprocessing: Reconstruct the three-dimensional data array into a two-dimensional matrix X with dimensions n×(i×j), where the value of each variable on all spatial grids constitutes an eigenvector. Corresponding to the embodiment in step S21, the dimension of the two-dimensional matrix X is 13×12544 (i.e., 112×112).

[0077] S32. Perform PCA analysis to calculate the variance contribution rate of each principal component and select the first principal component. PCA (Principal Component Analysis) is performed on the two-dimensional matrix X. PCA transforms a set of potentially correlated variables into a new set of linearly uncorrelated variables (principal components) through linear transformation, and sorts them by variance. The principal component with the highest variance contribution rate can be selected as the first principal component. By analyzing the spatial distribution characteristics of the first principal component and the loadings of each variable (component), the importance of various pollutant processes can be analyzed, and the correlation patterns between variables (components) can be indirectly observed to assist in analyzing the driving mechanisms of pollutant physical transport and chemical transformation processes.

[0078] S33. Spatial Feature Field Reconstruction: The first principal component is used as a vector and mapped back to the geospatial grid to obtain an i×j dimensional spatial distribution feature field of the first principal component. For example, the dimension 12544 of the two-dimensional matrix X in step S31 is mapped back to 112×112. This distribution field effectively condenses the comprehensive characteristics of regional pollution formation and evolution, and can effectively provide meaningful data for pollution analysis.

[0079] The fourth step is to select key points from the first principal component (step S4). A load threshold can be set, and key points can be selected from the first principal component based on the load threshold. Specifically, this includes:

[0080] S41. Perform orthogonal PCA analysis on matrix X. The two-dimensional matrix X can be orthogonally rotated to obtain the first principal component load matrix, and then subjected to Varimesis Principal Component Analysis (PCA).

[0081] S42. Based on a preset load threshold, select the spatial grids that contribute the most to the first principal component from the first principal component load matrix as key points. Typically, the absolute value of the load threshold is preferably greater than or equal to 0.9. The selected high-load grid points represent the intensity and variation of various physicochemical processes (IPR) and reaction pathways (IRR) at that location, and are highly synchronized and positively correlated with the core pollution spatial distribution mode of the entire region (i.e., the first principal component distribution characteristic field in step three). Therefore, they can be identified as key points, maximally representing the overall distribution characteristics and formation mechanism of the region, providing a scientific basis for observation network optimization and precise source tracing.

[0082] In the above embodiments, combining IRR, IPR, and pollutant concentration data allows for the reflection of the role of each location in the region from three levels: pollutant distribution characteristics, macroscopic physicochemical reaction characteristics, and microscopic chemical reaction mechanisms. Rotational PCA applies a mathematical transformation to the principal components extracted by PCA, aiming to optimize the factor loading structure. This allows each principal component to be more clearly strongly correlated with several specific original variables, thus obtaining clearer and more easily interpretable physical or practical meaning. Applying rotational PCA to the multidimensional process analysis of IPR and IRR, which include pollution formation mechanisms, from air quality model outputs can produce significant synergistic effects. This deep cross-disciplinary integration ensures that the ultimately selected key locations not only reflect spatial distribution hotspots of pollutant concentrations but also reveal spatial hotspots of the core physicochemical processes driving this distribution (such as transport, chemical formation, and deposition). Therefore, the locations selected by this method possess deeper mechanistic representativeness and higher practical application value, successfully solving long-standing technical problems and achieving good technical results.

[0083] Figure 6 shows the data point distribution map of a certain region, with the loading of the first principal component indicated by the shades of color. Interpreting Figure 6 reveals that regions A, B, and C, which are located around the bay, have high data loadings, reflecting the most significant pollutant concentrations, IRR, or IPR characteristics in these areas, making them key grids requiring close monitoring.

[0084] Further analysis of the weights of the 13 variables (ozone concentration, chemical formation, cloud physics, dry deposition, horizontal advection, horizontal diffusion, vertical diffusion, vertical advection, HOx formation, ozone depletion, ozone formation, OH formation, and VOCs depletion) projected onto the first principal component, based on Table 1 below, reveals that ozone depletion and ozone formation have the largest weights, followed by chemical processes. This indicates that in this case, the ozone formation and depletion process is a key factor influencing site selection.

[0085]

[0086] The embodiments provided by this invention can be implemented in various user devices, such as mobile terminals, computer terminals, or similar computing devices and systems (e.g., Electronic Control Units). Figure 7 shows a hardware structure block diagram of a user device 200.

[0087] As shown in Figure 7, the user equipment 200 may include one or more (only one is shown in the figure) processors 210 (processors 210 may include, but are not limited to, central processing units (CPUs), graphics processing units (GPUs), digital signal processors (DSPs), microprocessors (MCUs), or programmable logic devices (FPGAs), input / output interfaces 220 for user interaction, a memory 230 for data storage, a transmission device 240 for communication functions, and a bus 250. Those skilled in the art will understand that the structure shown in Figure 7 is merely illustrative and does not limit the structure of the aforementioned electronic device.

[0088] The memory 230 may include internal and external memory, or it may include volatile and non-volatile memory, or it may include magnetic storage devices, solid-state memory, flash memory, floppy disks, and hard disks, etc. The memory 230 can be used to store the operating system and the data generated by the operating system during operation, the software program of application software and the data generated by the application software during operation, and the software program of dedicated modules and the data generated by the program during operation. For example, in one embodiment of the present invention, the above-described key point identification method exists in the form of program instructions / modules, and the processor 210 in the user equipment 200 implements various functions / applications and data processing by running the program instructions stored in the memory 230. In some embodiments, the memory 230 may also be a remote memory (relative to a conventional memory) remotely located relative to the processor 210, and these remote memories can be connected to the processor 210 via a network connection.

[0089] The transmission device 240 is used to receive or send data via a network. In this invention, the transmission device 240 includes a programmable monitoring device 241, through which it communicates with other network devices. As an example, the transmission device 240 may be a device responsible for network communication, such as a switch chipset or a network interface card (NIC), and the monitoring device 241 may be programmable hardware, such as a programmable network chip, for example, an ASIC chip, a P4 chip, or a Tofino chip. These programmable hardware components are alternatively or additionally incorporated into the transmission device 240, and this invention does not limit this.

[0090] The embodiments of the present invention are not limited to the type of application; they can be system applications or software applications.

[0091] It should be noted that in the examples and description of this patent, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one" does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0092] Although this application has been illustrated and described with reference to certain preferred embodiments thereof, those skilled in the art should understand that various changes in form and detail may be made thereto without departing from the spirit and scope of this application.

Claims

1. A method for identifying key points, characterized in that, The process includes the following steps: constructing a multidimensional data array based on various data output from an air quality model; performing dimensionality reduction on the multidimensional data array using principal component analysis to obtain a first principal component; setting a load threshold and selecting key points from the first principal component based on the load threshold.

2. The method as described in claim 1, characterized in that, Constructing the air quality model includes the following steps: determining the target space and target time; and processing emission inventory data using meteorological field data to match the resolution of the target space and target time. Set environmental variables to start the process analysis module to analyze the physicochemical process contributions and chemical reaction pathways of pollutants; run and verify the reliability of the air quality model.

3. The method as described in claim 2, characterized in that, The process analysis module includes an integrated process rate module and an integrated reaction rate module. The integrated process rate module directly outputs the contribution rate of each physicochemical process term to the concentration change of the pollutant by integrating and separating the physicochemical process terms in the model equation online. The integrated reaction rate module tracks and outputs the cumulative rate of specific reactions in the chemical mechanism to quantify the contribution of a single chemical reaction to the generation or consumption of the pollutant and obtain the chemical reaction pathway.

4. The method as described in claim 1, characterized in that, The multidimensional data array is constructed based on various data output from the air quality model, including: extracting pollutant concentration field data, integrating process contribution field data of process rates, and integrating reaction path contribution field data of reaction rates within the period of interest; integrating the pollutant concentration field data, process contribution field data, and reaction path contribution field data into a total number of variables to construct a three-dimensional data array with dimensions n×i×j, where n is the total number of variables, i is the number of longitude grids, and j is the number of latitude grids.

5. The method as described in claim 4, characterized in that, The integrated process rate includes at least the following processes: horizontal advection, vertical convection, vertical diffusion, dry deposition, emission, gas phase chemical processes, aerosol processes, cloud processes, and wet deposition processes.

6. The method as described in claim 4, characterized in that, The reaction pathway of the integrated reaction rate includes at least: the chemical generation and consumption of ozone, NO x Key cyclic reactions, generation and consumption of OH radicals, HO X The generation and consumption of VOCs.

7. The method as described in claim 1, characterized in that, The step of reducing the dimensionality of the multidimensional data array based on principal component analysis to obtain the first principal component includes: data preprocessing, reconstructing the three-dimensional data array into a two-dimensional matrix X, wherein the dimension of the two-dimensional matrix X is n×(i×j); principal component extraction, performing principal component analysis on the two-dimensional matrix X, and selecting the principal component with the highest variance contribution rate as the first principal component; and spatial feature field reconstruction, mapping the first principal component as a vector back to a geospatial grid to obtain the spatial distribution feature field of the i×j first principal component.

8. The method as described in claim 7, characterized in that, The step of setting a load threshold and selecting key points from the first principal component based on the load threshold includes: performing an orthogonal rotation on the two-dimensional matrix X to obtain the first principal component load matrix, and performing Varimax principal component analysis on it; and selecting the spatial grid that contributes the most to the first principal component from the first principal component load matrix as key points according to the preset load threshold.

9. An electronic device, characterized in that, include: At least one memory for storing computer instructions; at least one processor for implementing the method as described in any one of claims 1-8 when the at least one processor executes the computer instructions.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores at least one computer instruction, which is loaded and executed by a processor to implement the method as described in any one of claims 1-8.

11. A computer program product, characterized in that, The computer program product includes computer instructions that, when executed, implement the method as described in any one of claims 1-8.

Images