A method for constructing a regional air pollutant transmission network

Abstract

The present invention relates to a method for constructing a regional air pollutant transmission network, comprising the following steps: Step 1: obtaining a large amount of air pollutant monitoring data and geographic information data within a region and performing data preprocessing; Step 2: solving the single-source diffusion impact factor between any two effective air quality monitoring stations at a corresponding time, and mining the possible basic air pollutant transmission paths between any two effective air quality monitoring stations based on the single-source diffusion impact factor; Step 3: based on the possible basic air pollutant transmission paths between any two effective air quality monitoring stations obtained in Step 2, screening the real paths with a causal mechanism as a guide, thereby constructing a regional air pollutant transmission network. The present invention can effectively improve the efficiency of analyzing the interaction between regional air pollutants and the accuracy and real-time performance of constructing an inter-regional air pollutant transmission network.

Description

Technical Field

[0001] The present invention belongs to the technical field of air quality data mining and relates to a method for constructing a pollutant transmission network, in particular to a method for constructing a regional air pollutant transmission network. Background Art

[0002] Air pollution is a serious global environmental problem. The cross-regional transport of pollutants leads to environmental degradation and makes governance difficult. Air is ubiquitous and has no specific form. The transport of air pollutants is highly volatile, uncontrolled, and subject to the interaction of multiple factors.

[0003] The movement of air pollutants is affected by multi-dimensional factors, and their transmission trajectories are difficult to describe. Therefore, the process of constructing an air pollutant transmission path network is very complex.

[0004] Several methods for constructing air pollutant transmission path networks have been proposed in the prior art. For example, the VG method proposed by Lacasa et al., which converts time series into a visual graph of a network, can retain many of the attributes of the original series and provides a simple and easy-to-use approach to network modeling of air quality. In the VG method, time series are used as nodes, and a path exists when the correlation between nodes is within the visible range. Cabezas et al. used the VG method to convert the time series of tropospheric ozone into a transmission network and verified the feasibility of this method in the field of air quality by analyzing the dynamic changes of pollutants in urban and rural areas.

[0005] After extracting various characteristics of air pollution, Song et al. established a complex network model, conducted a comprehensive analysis of the existence of paths between node pairs in the network, and mined the key nodes and paths of pollutant transmission based on an improved PageRank algorithm. They found that they were basically consistent with terrain characteristics and atmospheric flow changes.

[0006] However, the above existing methods for constructing regional air pollutant transmission networks still have the following defects:

[0007] 1. The establishment of the air pollutant transmission path network relies on pollution source lists and remote sensing data. However, pollution source lists and remote sensing data are difficult to obtain, and the data labeling and processing processes are complicated. This leads to delayed data updates in the construction of the air pollutant transmission path network and high computational costs.

[0008] 2. Insufficient data accuracy and completeness. Air quality and meteorological data are monitored and recorded by two separate types of monitoring stations, with differences in monitoring frequency and location. Existing methods cannot effectively integrate multi-source data, and the data composition structure is incomplete. Furthermore, due to incomplete monitoring records, data may be missing in certain regions or time periods, resulting in inaccurate results in the construction of the air pollutant transmission network.

[0009] 3. The model structure constructed by the existing method is simple and fails to fully consider the impact of various influencing factors such as wind, temperature, and distance during the transmission process.

[0010] After searching, no public documents of the prior art that are identical or similar to the present invention were found. Summary of the Invention

[0011] The purpose of the present invention is to overcome the shortcomings of the existing technology and propose a method for constructing a regional air pollutant transmission network, which can effectively improve the analysis efficiency of the interaction relationship between regional air pollutants and the accuracy and real-time performance of the construction of the regional air pollutant transmission network.

[0012] The present invention solves the practical problem by adopting the following technical solutions:

[0013] A method for constructing a regional air pollutant transmission network comprises the following steps:

[0014] Step 1: Obtain a large amount of air pollutant monitoring data and geographic information data in the region, perform data preprocessing, obtain monitoring data from valid air quality monitoring stations, and then perform data filling, normalization, and data balancing.

[0015] Step 2: Based on the various factors that affect pollutant transmission, define the single-source diffusion impact factor between a certain effective air quality monitoring station and other effective air quality monitoring stations. Based on the preprocessed data in step 1, solve the single-source diffusion impact factor between any two effective air quality monitoring stations at different times. Based on this single-source diffusion impact factor, explore the possible basic paths of air pollutant transmission between any two effective air quality monitoring stations.

[0016] Step 3: Based on the possible basic paths of air pollutant transmission between any two effective air quality monitoring stations obtained in step 2, the real paths are screened with the causal mechanism as the guide, and then the regional air pollutant transmission network is constructed.

[0017] Moreover, the specific steps of step 1 include:

[0018] (1) Grid division of the area containing the air quality monitoring station: Based on the size of the area and the division accuracy, the area is divided into m*n squares separated by n horizontal lines and m vertical lines using distance or longitude and latitude as horizontal and vertical lines. Starting from the upper left, each square is numbered 1 to m*n.

[0019] (2) Based on the regional grid division results of step (1), effective air quality monitoring stations are selected: the number of air quality monitoring stations in each grid ng is counted, and the number of monitoring stations in the effective grid nsta is set. If the total number of monitoring stations ngi in grid i is less than nsta, grid i is regarded as an invalid grid; for effective grid j, if ngj = nsta, all monitoring stations in the grid are selected as effective; if ngj > nsta, nsta air quality monitoring stations are selected from it.

[0020] (3) For the effective air quality monitoring station selected in step (2), find the meteorological monitoring station closest to it, and connect the meteorological data and air quality monitoring data at the same time as the monitoring data of the effective air quality monitoring station.

[0021] (4) filling and normalizing the monitoring data of the effective air quality monitoring stations obtained in step (3);

[0022] (5) Finally, the data is balanced by randomly deleting the valid air quality monitoring stations and timestamp indexes with large data volumes, so that the number of data entries in each index is basically the same.

[0023] Moreover, the specific method of step (4) of step 1 is:

[0024] Compare the list of valid air quality monitoring stations and retain only the pollutant and meteorological information of the valid air quality monitoring stations in the list. Remove redundant information and calculate the mean of the missing values within the previous and next 24 hours and fill in the missing values.

[0025] Normalize the monitoring data of effective air quality monitoring stations and standardize each item in the monitoring data according to the following formula:

[0026]

[0027] In the above formula, Con std is the normalized value, x is the data value to be normalized, min is the minimum value of all data in this data project, and max is the maximum value of all data in this data project;

[0028] Moreover, the specific steps of step 2 include:

[0029] (1) Comprehensively considering different influencing factors, under the influence of wind, distance and concentration, the single-source diffusion influence factor Inf of effective air quality monitoring station i on effective air quality monitoring station j at time t is calculated as ij (t) is defined as:

[0030]

[0031] In the above formula, ΔcenDij (t) is the concentration difference between sites i and j at time t; ΔF ij (t) is the wind speed difference between sites i and j at time t; dis ij is the distance between sites i and j; Inf ij The larger the value of (t), the higher the possibility of the existence of the transmission path. Set the threshold a, when Inf ij When (t)>a, it is considered that there is a transmission basis between the two monitoring stations;

[0032] (2) According to the distance between effective air quality monitoring stations and the value of the single-source diffusion impact factor, the basic path of air pollutant transmission that may exist between any two effective air quality monitoring stations can be obtained.

[0033] Moreover, the specific steps of step 3 include:

[0034] (1) Set the time interval: In a given time interval, the pollutant concentration at monitoring station j must increase with that at monitoring station i. The time interval is defined as the interval (m, n) after the current moment, where the start and end times are defined as follows:

[0035]

[0036] In the above two formulas, m is the starting time point of the interval, and n is the ending time point of the interval. ij is the distance between monitoring stations i and j.

[0037] (2) The pollutant concentration change requirement is met: according to the time interval range of step (1), when the pollutant concentration at monitoring station j increases by more than 10% compared with the initial time within the specified time range, that is, when the following formula is established, the transmission path is considered to exist:

[0038] make

[0039] In the formula, m, n are time interval marks, con j (k) is the concentration of monitoring station j at time k, con j (start) is the pollutant concentration at monitoring station j when calculating the single-source impact factor, i.e., the initial moment;

[0040] (3) Accumulate the real paths at each moment and construct a regional air pollutant transmission network.

[0041] Advantages and beneficial effects of the present invention:

[0042] 1. This paper proposes a method for constructing a regional air pollutant transmission network. This method focuses on the network construction of air pollutant transmission paths. Starting from pollutant concentration data from monitoring stations, it comprehensively considers various factors and statistically accumulates and screens air pollutant transmission paths. This method constructs pollutant transmission paths between different areas within a region, generates an air quality network structure, and visualizes and analyzes the characteristics of pollutant transmission processes. This method effectively improves the efficiency of analyzing regional pollutant interactions, providing an effective reference for pollution prevention and control and air quality control.

[0043] 2. This invention comprehensively considers various factors influencing air pollutant transmission, screens air pollutant transmission paths based on causal mechanisms, and constructs an air pollutant transmission network structure through statistical accumulation, thereby enabling visualization and characteristic analysis of the air pollutant transmission process. This implementation thoroughly analyzes the air pollutant transmission mechanism and comprehensively considers the influencing factors. Furthermore, using only data from ground-based monitoring stations, data can be obtained in real time, improving the accuracy and real-time nature of the construction of inter-regional air pollutant transmission relationships.

[0044] 3. This invention breaks away from the limitations of traditional methods, such as pollution source inventories and remote sensing data. The air pollutant transmission network is constructed using data from ground-based monitoring stations as input. This data is published in real time on the Ministry of Ecology and Environment's official website and is easily accessible.

[0045] 4. The present invention comprehensively considers a variety of influencing factors, and the resulting air pollutant transmission path network is in line with reality and has high accuracy. The present invention proposes the concept of single-source diffusion influencing factors and constructs an air pollutant transmission path network based on a causal mechanism. Taking air quality monitoring stations as nodes and air pollutant transmission processes as edges, the air quality system is constructed as a graph structure model that is easy to analyze and calculate. The extraction method of spatiotemporal influencing factors affecting pollutant transmission is analyzed, and the concept of single-source diffusion influencing factors is proposed based on the dynamic diffusion mechanism of pollutants, and the pollutant transmission process is probabilistically characterized. The causal behavior of pollutant concentrations is extracted from historical data, and reasonable transmission paths are screened. The path screening process is reasonable, the factors are fully considered, the network structure is consistent with the actual situation, and the accuracy is high.

[0046] 5. The practical results achieved by this invention demonstrate that it conducts in-depth mining based on historical time series, corresponding to the actual geographical distribution and spatiotemporal transmission state. Based on this air pollutant transmission network structure, it is possible to analyze the transmission properties of air pollutants at different spatial scales from monitoring stations to regions. This has effectively promoted the study of pollutant dynamic behavior. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 A schematic diagram of the basic framework of a method for constructing a regional air pollutant transmission network according to the present invention;

[0048] Figure 2 This is an example diagram of the causal mechanism path screening process of the present invention;

[0049] Figure 3 This is a sample diagram of the network construction results of the present invention;

[0050] Figure 4 Schematic diagram of transmission paths corresponding to different single-source diffusion influence factors of the present invention (taking February 15, 2022 as an example). The smaller the threshold of the single-source diffusion factor, the more paths there are in the network;

[0051] Figure 5 This is the pollutant transmission path diagram of the present invention. DETAILED DESCRIPTION

[0052] The embodiments of the present invention are further described below in conjunction with the accompanying drawings:

[0053] The present invention is based on meteorological factors, geographical factors and historical data of pollutants in monitoring stations, and establishes a method for constructing a regional air pollutant transmission network based on the causal mechanism, such as Figure 1 As shown, the following steps are included:

[0054] Step 1: Obtain a large amount of air pollutant monitoring data and geographic information data in the region, perform data preprocessing, obtain monitoring data from valid air quality monitoring stations, and then perform data filling, normalization, and data balancing.

[0055] In this embodiment, the data resources include massive amounts of meteorological monitoring data, pollutant concentration detection data, and site geographic information data obtained from various monitoring stations. Valid monitoring stations need to be screened through a grid-based approach, and pollutant concentration monitoring information and meteorological information need to be connected. Directly using all monitoring information as basic data will cause data imbalance in location and time, which in turn will cause an imbalance in path distribution. Therefore, valid monitoring stations are first screened based on location distribution, and then meteorological and air quality monitoring stations are connected by time, one-to-one. Finally, all data is balanced and cleaned, normalized, and redundant data is deleted.

[0056] The specific steps of step 1 include:

[0057] (1) Grid division of the area containing the air quality monitoring station: Based on the size of the area and the division accuracy, the area is divided into m*n squares separated by n horizontal lines and m vertical lines using distance or longitude and latitude as horizontal and vertical lines. Starting from the upper left, each square is numbered 1 to m*n.

[0058] (2) Based on the regional grid division results of step (1), effective air quality monitoring stations are selected: the number of air quality monitoring stations in each grid ng is counted, and the number of monitoring stations in the effective grid nsta is set. If the total number of monitoring stations ngi in grid i is less than nsta, grid i is regarded as an invalid grid; for effective grid j, if ngj = nsta, all monitoring stations in the grid are selected as effective; if ngj > nsta, nsta air quality monitoring stations are selected from it.

[0059] (3) For the effective air quality monitoring station selected in step (2), find the meteorological monitoring station closest to it, and connect the meteorological data and air quality monitoring data at the same time as the monitoring data of the effective air quality monitoring station.

[0060] For the effective air quality monitoring stations selected in step (2), such stations cannot monitor meteorological information. That is, the air quality monitoring stations and meteorological monitoring stations are built in different locations, and the pollutant information does not correspond to the meteorological information. For a certain air quality monitoring station, it is necessary to determine its temperature, wind speed and other related meteorological information based on its neighboring meteorological monitoring stations. The connection method is as follows: for a certain air quality monitoring station, find the meteorological monitoring station closest to it, connect the meteorological data and air quality monitoring data at the same time, and use them as the monitoring data of the air quality monitoring station.

[0061] (4) filling and normalizing the monitoring data of the effective air quality monitoring stations obtained in step (3);

[0062] The specific method of step (4) is:

[0063] Compare the list of valid air quality monitoring stations and retain only the pollutant and meteorological information of the valid air quality monitoring stations in the list. Remove redundant information and calculate the mean of the missing values within the previous and next 24 hours and fill in the missing values.

[0064] Normalize the monitoring data of effective air quality monitoring stations and standardize each item in the monitoring data according to the following formula:

[0065]

[0066] In the above formula, Con std is the normalized value, x is the data value to be normalized, min is the minimum value of all data in this data project, and max is the maximum value of all data in this data project;

[0067] (5) Finally, the data is balanced by randomly deleting the data items of valid air quality monitoring stations and timestamp indexes with large data volumes, so that the number of data items contained in each index is basically the same, reducing the impact caused by data imbalance.

[0068] Step 2: Based on the various factors that affect the spread of pollutants, define the single-source diffusion impact factor between a certain effective air quality monitoring station and other effective air quality monitoring stations. According to the preprocessed data in step 1, solve the single-source diffusion impact factor between any two effective air quality monitoring stations at different times. Based on this single-source diffusion impact factor, explore the possible basic paths of air pollutant transmission between any two effective air quality monitoring stations.

[0069] The specific steps of step 2 include:

[0070] (1) Comprehensively considering different influencing factors, under the influence of wind, distance and concentration, the single-source diffusion influence factor Inf of effective air quality monitoring station i on effective air quality monitoring station j at time t is calculated as ij (t) is defined as:

[0071]

[0072] In the above formula, ΔcenD ij (t) is the concentration difference between sites i and j at time t; ΔF ij (t) is the wind speed difference between sites i and j at time t; dis ij is the distance between sites i and j; Inf ij The larger the value of (t), the higher the possibility of the existence of the transmission path. Set the threshold a, when Inf ij When (t)>a, it is considered that there is a transmission basis between the two monitoring stations;

[0073] (2) According to the distance between effective air quality monitoring stations and the value of the single-source diffusion impact factor, the basic path of air pollutant transmission that may exist between any two effective air quality monitoring stations can be obtained.

[0074] In this embodiment, various factors affecting the spread of air pollutants are considered separately. Taking into account different factors, the air pollutant transmission paths are different under the influence of wind, distance, and concentration:

[0075] 1) Distance. Distance affects the diffusion of pollutants between two locations. For source monitoring station A, the farther away from destination monitoring station B, the weaker its influence. Because the Earth is a sphere, the distance between two points is not a straight line on a two-dimensional plane, but rather an arc length on the spherical surface.

[0076] The location of two points is represented by the longitude and latitude in the geographic coordinate system. The specific method of determining the distance arc is as follows: let the center of the earth be O, the coordinates of station i be (Lon i ,lat i ), the coordinates of site j are (lon j ,latj ), the distance between points i and j is calculated as follows:

[0077] dis ij =R*arccos[sin(lat i )*sin(lat j )*cos(lon i -lon j )+cos(lat i )*cos(lat j )]

[0078] In the above formula, R is the radius of the Earth.

[0079] 2) Wind factor. Wind can promote the spread of pollutants. The greater the wind speed difference between two places, the lower the spread cost. Pollutants will spread to other areas along the direction of the wind. The greater the wind speed difference between different areas, the faster the spread speed. When i is the source diffusion area and j is the destination, the wind speed difference ΔF in the direction of the line connecting the two places is ij (t) is:

[0080] ΔF ij (t)=|F i (T)|*COS(β-α)+|F j (t)|*cos(β-γ)

[0081] In the above formula, F i is the vector wind at site i, and its wind direction angle is α. j is the vector wind at site j, the wind direction angle is γ, and β is the angle between the line from site i to j and the north direction.

[0082] 3) Pollutant concentration difference factor. The gas diffusion process is actually the process of molecules moving to equilibrium between high-concentration and low-concentration areas. Introducing the influence of concentration difference in the process of air pollutant propagation, we define the new concentration difference as:

[0083]

[0084] Among them, cen i (t) is the pollutant concentration at site i at time t, ΔcenD ij (t) is the concentration difference between sites i and j at time t.

[0085] 4) Comprehensively consider different influencing factors, under the influence of wind, distance and concentration, the single source diffusion influence factor Inf of effective air quality monitoring station i on effective air quality monitoring station j at time t is calculated as ij (t) is defined as:

[0086]

[0087] In the above formula, ΔcenD ij (t) is the concentration difference between sites i and j at time t. ΔF ij (t) is the wind force difference between sites i and j at time t. dis ij is the distance between sites i and j. ij The larger the value of (t), the higher the possibility of the existence of the transmission path. Set the threshold a, when Inf ij When (t)>a, it is considered that there is a transmission basis between the two monitoring stations.

[0088] This invention addresses the transmission of pollutants near the surface. Near-surface transmission is hindered by surface objects, and pollutants attenuate more with increasing distance. The maximum distance for near-surface transmission is set at 200 km. Furthermore, when the distance is too close, a large number of closed loops are easily formed between monitoring stations. Therefore, the minimum distance between transmission paths is set at 20 km. The basic transmission path between monitoring stations can be derived based on distance and the single-source diffusion impact factor.

[0089] Step 3: Based on the possible basic paths of air pollutant transmission between any two effective air quality monitoring stations obtained in step 2, the real paths are screened with the causal mechanism as the guide, and then the regional air pollutant transmission network is constructed.

[0090] At time t, assuming that the transmission path P ij If the path really exists, the spread of pollutants from place i to place j will inevitably lead to changes in pollutant concentrations. We believe that this path only exists when the pollutant concentration in place j increases in a certain period in the future.

[0091] The specific steps of step 3 include:

[0092] (1) Set the time interval: In a given time interval, the pollutant concentration at monitoring station j must increase with that at monitoring station i. The time interval is defined as the interval (m, n) after the current moment, where the start and end times are defined as follows:

[0093]

[0094] In the above two formulas, m is the starting time point of the interval, and n is the ending time point of the interval. ij is the distance between monitoring stations i and j.

[0095] (2) The pollutant concentration change requirement is met: according to the time interval range of step (1), when the pollutant concentration at monitoring station j increases by more than 10% compared with the initial time within the specified time range, that is, when the following formula is established, the transmission path is considered to exist:

[0096] make

[0097] In the formula, m, n are time interval marks, con j (k) is the concentration of monitoring station j at time k, con j (start) is the pollutant concentration at monitoring station j when calculating the single-source impact factor, i.e., the initial moment;

[0098] (3) Accumulate the real paths at each moment and construct a regional air pollutant transmission network.

[0099] In this embodiment, the working principle of step 3 is:

[0100] 1) Time interval setting. Because wind speeds fluctuate constantly, longer distances require longer transmission times, and the transmission process introduces more uncertainties. Therefore, the specified distance interval is divided into three segments. As distance increases, the distance and time intervals are gradually relaxed. The time interval is defined as (m, n). That is, if monitoring station i transmits pollutants to monitoring station j, the pollutant concentration at monitoring station j will increase within the next time period (m, n). The start and end times are defined as follows:

[0101]

[0102] In the above two formulas, m is the starting time point of the interval, and n is the ending time point of the interval. ij is the distance between monitoring stations i and j.

[0103] 2) Concentration change. Based on the above time interval, if the concentration of pollutants at monitoring station j increases by more than 10% compared to the initial time within the specified time range, that is, if the following formula holds, the transmission path is considered to exist:

[0104] make

[0105] In the formula, m, n are time interval marks, con j (k) is the concentration of monitoring station j at time k, con j (start) is the pollutant concentration at monitoring station j when calculating the single-source impact factor, i.e., the initial moment;

[0106] Figure 2 The figure below shows an example of path selection. For four monitoring stations at time t, three possible transmission paths exist. The time interval corresponding to the distance between each monitoring station is calculated, and the change in pollutant concentration within this interval is calculated. The two edges (1→2 and 3→4) whose pollutant concentration changes do not meet the requirements are deleted. After causal filtering, the corresponding transmission path between the monitoring stations at that time (2→3) is obtained.

[0107] 3) For steps 2) and 3), the solution is the propagation path at a single moment. This network can only describe the dynamics at a certain moment and cannot characterize the regional laws and characteristics hidden behind large-scale data. For the pollutant transmission path network within a period of time or a certain cycle, it is necessary to map the propagation path at each moment (t, t+1, t+2,…, t+n) to the entire valid site set, overlap the transmission paths corresponding to each timestamp in the time period, accumulate them in sequence, and use the number of occurrences of the path as the weight of the corresponding edge. The nodes in the network are air quality monitoring stations, the edges are pollutant transmission paths, and the direction of the edge is the direction of wind difference. For efficient storage in the computer, the paths are mapped one-to-one with the nodes and converted into a transmission matrix.

[0108] In this embodiment, combined with the single-source diffusion impact factor, based on the causal relationship of historical data of monitoring stations, the path is screened, the network structure at each moment is generated, and the network structure in different time periods can be obtained after statistical accumulation, and the transmission relationship is displayed in the form of a graph. Taking February 15, 2022 as an example, a transmission network between monitoring stations is constructed. The nodes in the network are air quality monitoring stations, and the edges represent the pollutant transmission relationship between nodes at that moment. The transmission path network of that day is as follows: Figure 3 As shown in the figure, the nodes are marked with the monitoring station numbers, and the node color indicates the node's degree in the network, with the deeper the color, the greater the degree. The size of the directed edge arrow represents the weight of the edge. The larger the weight, the more frequent the transmission path appears, indicating that the transmission of pollutants between the two locations is more frequent.

[0109] The working principle of the present invention is:

[0110] The present invention relates to a method for constructing a regional air pollutant transmission network. The specific process of the method is: first, a large amount of monitoring data and geographic information data are processed, including grid division, selection of effective monitoring stations, data normalization, connection and filling. Secondly, the single-source diffusion influence factor is defined and the basic path is solved. According to the various factors affecting the propagation of pollutants, the influence factor of a certain monitoring station on other monitoring stations is defined. The single-source diffusion influence factor between any two monitoring stations at the corresponding time is solved to obtain the basic transmission path between the monitoring stations. Finally, the basic transmission path is screened and the network is constructed through the causal mechanism. Combined with the single-source diffusion influence factor, the causal relationship of the historical data of the monitoring station is measured based on historical data, the path is screened, and the network structure at each moment is generated. The network structure at all moments in the specified time period is counted to generate the pollutant transmission path network in the area.

[0111] The innovation of the present invention lies in:

[0112] Step A: Process a large amount of air quality related data.

[0113] Step B: Based on the various factors that affect the spread of pollutants, define the single-source diffusion impact factor of a certain effective monitoring station on other effective monitoring stations, and solve the basic transmission path between monitoring stations based on this factor.

[0114] Step C: Based on the possible transmission basic probability paths between monitoring stations, the real path screening is carried out guided by the causal mechanism.

[0115] Step D: Accumulate the paths at each moment to generate a pollutant transmission network.

[0116] The data processing process of step A is as follows:

[0117] The area containing the air quality monitoring stations is divided into grids using longitude and latitude lines as dividing lines. An equal number of monitoring stations are selected in each grid as valid monitoring stations.

[0118] When connecting meteorological data and pollutant concentration data, for a certain air quality monitoring station, find the meteorological monitoring station closest to it, and connect the meteorological data and air quality monitoring data at the same time as the monitoring data of the air quality monitoring station.

[0119] Normalize the monitoring data of effective air quality monitoring stations and standardize each item in the monitoring data according to the following formula:

[0120]

[0121] In the above formula, Con std is the normalized value, x is the data value to be normalized, min is the minimum value of all data in this data project, and max is the maximum value of all data in this data project;

[0122] The calculation and solution process of the single-source diffusion impact factor in step B is as follows:

[0123] Comprehensively considering different influencing factors, under the effects of wind, distance and concentration, the single-source diffusion impact factor Inf of effective air quality monitoring station i on effective air quality monitoring station j at time t is calculated as ij (t) is defined as:

[0124]

[0125] In the above formula, ΔcenD ij (t) is the concentration difference between sites i and j at time t. ΔF ij (t) is the wind speed difference in the direction of the line connecting the two places, dis ij is the distance between sites i and j.

[0126] The path screening process in step C is as follows:

[0127] In a given time interval, the pollutant concentration at monitoring station j increases with time i. The time interval is defined as the interval (m, n) after the current time. The start and end times are defined as follows:

[0128]

[0129] In the above two formulas, m is the starting time point of the interval, and n is the ending time point of the interval. ij is the distance between monitoring stations i and j.

[0130] The requirement for pollutant concentration change is: based on the above range, when the pollutant concentration at monitoring station j increases by more than 10% compared to the initial time within the specified time range, that is, when the following formula is established, the transmission path is considered to exist:

[0131] make

[0132] In the formula, m, n are time interval marks, con j (k) is the concentration of monitoring station j at time k, con j (start) is the pollutant concentration at monitoring station j when calculating the single-source impact factor, i.e., the initial moment;

[0133] The present invention uses air quality monitoring stations as nodes and transmission paths as edges to design a method for constructing a regional air pollutant transmission network. First, multi-source data is processed to extract effective air pollutant transmission related information; then, a single-source diffusion influence factor is defined based on various factors affecting pollutant transmission to find possible paths between monitoring stations; secondly, a causal mechanism is used to screen paths; finally, the paths at each moment are accumulated to generate a network structure. The different values of the single-source diffusion influence factor proposed in the present invention represent different probabilities of path existence. The higher the single-source influence factor, the higher the probability of the path existing. Different single-source diffusion influence factors are selected to display paths. A low factor corresponds to more paths, while a high factor filters paths with low probability of occurrence and only displays fewer paths. Figure 4 Shown are the transmission paths corresponding to different single-source diffusion impact factors.

[0134] Pollutant propagation paths were calculated over a weekly period. Pollutant concentrations varied little from moment to moment, so data was collected every three hours to reduce algorithm runtime. Paths with high single-source impact factors were selected to generate daily paths. Figure 5 This is the path network for the four weeks of January. Since January has 31 days, the fourth time period contains a 10-day path collection. Figure 5The topographic structure of the region is also indicated, with the green portion representing the North China Plain and the yellow portion representing mountains and hills. The image shows that transport processes of varying distances are widely distributed across the region. This is particularly true in the eastern coastal areas of Cangzhou and Tianjin, situated on the plains. During periods of heavy pollution, transport pathways between these two regions are numerous and frequent. Shijiazhuang and Beijing also experience poor air quality, but the identified inter-regional transport pathways are less frequent. This is closely related to the two regions' geographical location, which is surrounded by mountains. This semi-basin terrain makes it difficult for pollutants to be transported outward. Comparing the four images reveals significant variations in the number of transport pathways across different time periods. This is because the transport process is directly influenced by pollutant concentration and wind speed. Uncertainties in wind speed, direction, and pollutant emissions cause transport pathways to vary dramatically. High pollutant concentrations and strong winds create a greater number of long-distance transport pathways.

[0135] It should be emphasized that the embodiments described in the present invention are illustrative rather than restrictive. Therefore, the present invention includes but is not limited to the embodiments described in the specific embodiments. Any other embodiments derived by those skilled in the art based on the technical solutions of the present invention also fall within the scope of protection of the present invention.

Claims

1. A method for constructing a regional air pollutant transmission network, characterized by: The following steps are involved: Step 1: Obtain a large amount of air pollutant monitoring data and geographic information data in the region, perform data preprocessing, obtain monitoring data from valid air quality monitoring stations, and then perform data filling, normalization, and data balancing. Step 2: Based on the various factors that affect pollutant transmission, define the single-source diffusion impact factor between a certain effective air quality monitoring station and other effective air quality monitoring stations. Based on the preprocessed data in step 1, solve the single-source diffusion impact factor between any two effective air quality monitoring stations at different times. Based on this single-source diffusion impact factor, explore the possible basic paths of air pollutant transmission between any two effective air quality monitoring stations. Step 3: Based on the possible basic air pollutant transmission paths between any two valid air quality monitoring stations obtained in Step 2, the real paths are screened with the causal mechanism as the guide, and then the regional air pollutant transmission network is constructed; The specific steps of step 2 include: (1) Comprehensively considering different influencing factors, under the influence of wind, distance and concentration, the single-source diffusion influence factor Inf of effective air quality monitoring station i on effective air quality monitoring station j at time t is calculated as ij (t) is defined as: In the above formula, ΔcenD ij (t) is the concentration difference between sites i and j at time t; ΔF ij (t) is the wind speed difference between sites i and j at time t; dis ij is the distance between sites i and j; Inf ij The larger the value of (t), the higher the possibility of the existence of the transmission path. Set the threshold a, when Inf ij When (t)>a, it is considered that there is a transmission basis between the two monitoring stations; (2) According to the distance between effective air quality monitoring stations and the value of the single-source diffusion impact factor, the possible basic path of air pollutant transmission between any two effective air quality monitoring stations is obtained; The specific steps of step 3 include: (1) Set the time interval: In a given time interval, the pollutant concentration at monitoring station j must increase with that at monitoring station i. The time interval is defined as the interval (m, n) after the current moment, where the start and end times are defined as follows: In the above two formulas, m is the starting time point of the interval, and n is the ending time point of the interval; dist ij is the distance between monitoring stations i and j; (2) The pollutant concentration change requirement is met: according to the time interval range of step (1), when the pollutant concentration at monitoring station j increases by more than 10% compared with the initial time within the specified time range, that is, when the following formula is established, the transmission path is considered to exist: make In the formula, m, n are time interval marks, con j (k) is the concentration of monitoring station j at time k, con j (start) is the pollutant concentration at monitoring station j when calculating the single-source impact factor, i.e., the initial moment; (3) Accumulate the real paths at each moment and construct a regional air pollutant transmission network.

2. The method for constructing a regional air pollutant transmission network according to claim 1, characterized in that: The specific steps of step 1 include: (1) Gridding the area containing the air quality monitoring station: Based on the size and accuracy of the area, use distance or longitude and latitude as horizontal and vertical lines to divide the entire area into m*n squares separated by n horizontal lines and m vertical lines. Starting from the upper left, number each square 1 to m*n; (2) Based on the regional grid division result of step (1), select effective air quality monitoring stations: count the number of air quality monitoring stations ng in each grid, set the number of monitoring stations in the effective grid nsta, if the total number of monitoring stations ngi in grid i is less than nsta, then grid i is considered an invalid grid; for effective grid j, if ngj = nsta, then all monitoring stations in the grid are selected as effective; if ngj > nsta, then nsta air quality monitoring stations are selected from it; (3) For the effective air quality monitoring station selected in step (2), find the meteorological monitoring station closest to it, and connect the meteorological data and air quality monitoring data at the same time as the monitoring data of the effective air quality monitoring station; (4) filling and normalizing the monitoring data of the effective air quality monitoring stations obtained in step (3); (5) Finally, the data is balanced by randomly deleting the valid air quality monitoring stations and timestamp indexes with large data volumes, so that the number of data entries in each index is basically the same.

3. The method for constructing a regional air pollutant transmission network according to claim 2, characterized in that: The specific method of step (4) of step 1 is: Compare the list of valid air quality monitoring stations and retain only the pollutant and meteorological information of the valid air quality monitoring stations in the list. Remove redundant information and calculate the mean of the missing values within the previous and next 24 hours and fill in the missing values. Normalize the monitoring data of effective air quality monitoring stations and standardize each item in the monitoring data according to the following formula: In the above formula, Con std is the normalized value, x is the data value that needs to be normalized, min is the minimum value of all the data in this data item, and max is the maximum value of all the data in this data item.

Images