Air quality prediction management method and system for air pollution prevention and control

Abstract

This invention relates to an air quality forecasting and management method and system for air pollution prevention and control, belonging to the field of environmental monitoring technology. The method mainly includes: calculating air quality forecast data for the target city and neighboring cities based on industrial emission data, traffic emission data, and real-time pollutant concentration monitoring data of the target city and neighboring cities respectively; updating the air quality forecast data of the target city to obtain the final air quality forecast data of the target city through meteorological forecast data, geographic information data, and air quality forecast data of neighboring cities; determining whether the final air quality forecast data of the target city is greater than a preset anomaly value; if it is greater than the preset anomaly value, managing and controlling industrial emissions and traffic emissions of the target city or its neighboring cities according to the cause of the final air quality forecast data.

Description

Technical Field

[0001] This invention relates to the field of environmental monitoring technology, and in particular to an air quality prediction and management method and system for air pollution prevention and control. Background Technology

[0002] Air pollution refers to the phenomenon where the concentration of certain substances in the atmosphere reaches a certain level, harming the natural and social environment and human health. Human activities such as industrial production, transportation, urban dust, and fuel combustion all contribute to air pollution. With increasing industrialization and urbanization, air pollution has become a challenge faced by all countries.

[0003] In recent years, air pollution has occurred frequently, and people are paying increasing attention to air quality. However, users can only find relevant air quality information on websites, which is difficult and inaccurate. Therefore, there is a need to develop an air quality monitoring system that can monitor air in real time, is simple and practical, and has a high cost-performance ratio. Summary of the Invention

[0004] The present invention aims to provide an air quality prediction and management method and system for atmospheric pollution prevention and control, so as to overcome the shortcomings of the existing technology. The technical problem to be solved by the present invention is achieved through the following technical solution.

[0005] This invention provides an air quality prediction and management method for air pollution prevention and control, the method comprising:

[0006] Acquire meteorological forecast data, industrial emission data, traffic emission data, geographic information data, and real-time pollutant concentration monitoring data for the target city and its neighboring cities;

[0007] Based on the industrial emission data, traffic emission data, and real-time pollutant concentration monitoring data of the target city and the adjacent cities, calculate the predicted air quality data for the target city and the adjacent cities respectively;

[0008] The final air quality forecast data for the target city is obtained by updating the air quality forecast data of the target city using meteorological forecast data and geographic information data of the target city and its neighboring cities, as well as air quality forecast data of the neighboring cities.

[0009] Determine whether the final air quality forecast data for the target city is greater than a preset anomaly value;

[0010] If the value is greater than the preset abnormal value, the cause of the final air quality prediction data is determined.

[0011] Based on the reasons for the generation of the final air quality forecast data, the management and control of industrial and traffic emissions in the target city or its neighboring cities shall be implemented.

[0012] In an optional embodiment, the step of calculating the predicted air quality data for the target city and the adjacent cities based on industrial emission data, traffic emission data, and real-time pollutant concentration monitoring data of the target city and the adjacent cities includes:

[0013] The industrial emission data, the traffic emission data, and the real-time pollutant concentration monitoring data are preprocessed, including anomaly filtering and data normalization.

[0014] The preprocessed industrial emission data, traffic emission data, and real-time pollutant concentration monitoring data are converted into a feature data matrix with a time dimension.

[0015] The feature data matrices corresponding to the target city and the adjacent cities are input into the air quality prediction model to obtain the air quality prediction data corresponding to the target city and the adjacent cities.

[0016] In an optional embodiment, updating the air quality forecast data of the target city using meteorological forecast data and geographic information data of the target city and its neighboring cities, as well as air quality forecast data of the neighboring cities, to obtain the final air quality forecast data of the target city includes:

[0017] The static attenuation value is calculated using geographic information data of the target city and its neighboring cities; the wind direction correction value and wind speed correction value are calculated using meteorological forecast data of the target city and its neighboring cities.

[0018] The city pollution weight values ​​between the target city and its neighboring cities are calculated based on the static attenuation value, the wind direction correction value, and the wind speed correction value.

[0019] The final air quality forecast data for the target city is obtained by updating the air quality forecast data of the target city using the city pollution weight value, the air quality forecast data of the target city and its neighboring cities, and meteorological forecast data.

[0020] In an optional embodiment, calculating the urban pollution weight value between the target city and its neighboring cities based on the static attenuation value, the wind direction correction value, and the wind speed correction value includes:

[0021] The product of the static attenuation value, the wind direction correction value, and the wind speed correction value is determined as the urban pollution weight value between the target city and its neighboring cities.

[0022] In an optional embodiment, the final air quality forecast data for the target city is obtained by updating the air quality forecast data of the target city using the city pollution weight value, air quality forecast data of the target city and its neighboring cities, and meteorological forecast data, including:

[0023] The city pollution weight value, the air quality prediction data of the target city and its neighboring cities, and the meteorological prediction data are converted into a feature vector matrix.

[0024] The feature vector matrix is ​​input into the temporal-space convolution model to obtain the final air quality prediction data for the target city.

[0025] In an optional embodiment, before calculating the wind direction correction value and wind speed correction value using meteorological forecast data from the target city and its neighboring cities, the method further includes:

[0026] Obtain historical meteorological data for the target city, and extract the monsoon cycle from the historical meteorological data using wavelet transform;

[0027] Determine the prevailing monsoon wind direction and average monsoon wind speed for each monsoon cycle;

[0028] Wind direction correction values ​​and wind speed correction values ​​are calculated using the prevailing monsoon wind direction, the average monsoon wind speed, and meteorological forecast data of the target city and its neighboring cities.

[0029] In an optional embodiment, the step of calculating wind direction correction values ​​and wind speed correction values ​​using the prevailing monsoon wind direction, the average monsoon wind speed, and meteorological forecast data of the target city and its neighboring cities includes:

[0030] The prevailing monsoon direction and average monsoon speed are calculated and updated by weighting for monsoon cycles of all lengths.

[0031] Wind direction corrections and wind speed corrections are calculated using updated monsoon prevailing wind direction, monsoon average wind speed, and meteorological forecast data for the target city and its neighboring cities.

[0032] In an optional embodiment, after inputting the feature vector matrix into a temporal-space convolutional model to obtain the final air quality prediction data for the target city, the method further includes:

[0033] Obtain the day-night diffusion index of the target city;

[0034] The final air quality forecast data for the target city is updated based on the diurnal diffusion index of the target city.

[0035] In an optional embodiment, obtaining the diurnal diffusion index of the target city includes:

[0036] Obtain the daytime and nighttime diffusion coefficients of the target city;

[0037] The initial diurnal diffusion index is obtained by calculating the ratio of the daytime diffusion coefficient to the nighttime diffusion coefficient of the target city.

[0038] The day-night diffusion index of the target city is calculated based on the green coverage rate of the target city and the initial day-night diffusion index of the target city.

[0039] This invention provides an air quality forecasting and management system for air pollution prevention and control, the system comprising:

[0040] The acquisition module is used to acquire meteorological forecast data, industrial emission data, traffic emission data, geographic information data, and real-time pollutant concentration monitoring data of the target city and its neighboring cities.

[0041] The calculation module is used to calculate the predicted air quality data for the target city and the adjacent cities based on the industrial emission data, traffic emission data, and real-time pollutant concentration monitoring data of the target city and the adjacent cities, respectively.

[0042] The update module is used to update the air quality forecast data of the target city using meteorological forecast data and geographic information data of the target city and its neighboring cities, as well as the air quality forecast data of the neighboring cities, to obtain the final air quality forecast data of the target city.

[0043] The determination module is used to determine whether the final air quality prediction data of the target city is greater than a preset anomaly value;

[0044] The determining module is further configured to determine the cause of the final air quality prediction data if the value is greater than a preset abnormal value.

[0045] The management module is used to manage and control industrial and traffic emissions in the target city or its neighboring cities based on the reasons for the generation of the final air quality prediction data.

[0046] The embodiments of the present invention have the following advantages:

[0047] This invention provides an air quality forecasting and management method and system for air pollution control. First, it acquires meteorological forecast data, industrial emission data, traffic emission data, geographic information data, and real-time pollutant concentration monitoring data for a target city and its neighboring cities. Then, based on these data, it calculates air quality forecast data for both the target city and its neighboring cities. Next, it updates the target city's air quality forecast data using the meteorological forecast data, geographic information data, and air quality forecast data from the neighboring cities to obtain the final air quality forecast data for the target city. It then determines whether the final air quality forecast data for the target city exceeds a preset anomaly value. If it does, it determines the cause of the increased air quality forecast data. Finally, based on the cause of the increased air quality forecast data, it manages and controls industrial and traffic emissions in the target city or its neighboring cities. Compared to existing technologies that only allow users to query air quality information online, this application calculates air quality forecast data based on various data points between cities and then updates the target city's air quality forecast data based on the air quality forecast data from neighboring cities. This improves the accuracy of air quality forecasting for the target city and thus enhances the effectiveness of air quality management in the target city. Attached Figure Description

[0048] Figure 1 This is a flowchart of an air quality prediction and management method for atmospheric pollution prevention and control provided by an embodiment of the present invention;

[0049] Figure 2 This is a flowchart illustrating the process of determining the final air quality prediction data for a target city, as provided in an embodiment of the present invention.

[0050] Figure 3 This is a schematic diagram of the structure of an air quality prediction and management system for atmospheric pollution prevention and control provided in an embodiment of the present invention. Detailed Implementation

[0051] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in this application can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0052] Please see Figure 1 This invention provides an air quality prediction and management method for air pollution prevention and control, specifically comprising steps S101-S106:

[0053] S101 acquires meteorological forecast data, industrial emission data, traffic emission data, geographic information data, and real-time pollutant concentration monitoring data for the target city and its neighboring cities.

[0054] The meteorological forecast data may include: temperature, humidity, wind speed, wind direction, air pressure, boundary layer height, etc.; industrial emission data include industrial emissions, dust, straw burning, etc.; traffic emission data is the exhaust emissions of vehicles in the city; geographic information data includes: city latitude and longitude, DEM elevation, land use type, vegetation cover index, etc.; real-time pollutant concentration monitoring data is the concentration data of pollutants such as PM2.5, SO2, and O3 detected by national / provincial monitoring stations.

[0055] It should be noted that real-time pollutant concentration monitoring data can be obtained from multiple monitoring points set up in the city, and then the average value of the obtained real-time pollutant concentration monitoring data is calculated to obtain the final real-time pollutant concentration monitoring data.

[0056] S102, calculate the predicted air quality data for the target city and neighboring cities based on industrial emission data, traffic emission data, and real-time pollutant concentration monitoring data for the target city and neighboring cities respectively.

[0057] In one optional embodiment provided in this application, the step of calculating the predicted air quality data for the target city and the adjacent cities based on industrial emission data, traffic emission data, and real-time pollutant concentration monitoring data of the target city and the adjacent cities respectively includes:

[0058] S1021, preprocess the industrial emission data, the traffic emission data, and the real-time pollutant concentration monitoring data, the preprocessing including abnormal data filtering and data normalization.

[0059] S1022, The preprocessed industrial emission data, traffic emission data, and real-time pollutant concentration monitoring data are converted into a feature data matrix with time dimension.

[0060] Specifically, the feature data matrix includes feature data of a certain time length, such as in days. This time length can be 7 days. That is, each row of data in the feature data matrix represents one day's industrial emission data, traffic emission data, and pollution concentration detection data.

[0061] S1023, input the feature data matrices corresponding to the target city and the adjacent cities respectively into the air quality prediction model to obtain the air quality prediction data corresponding to the target city and the adjacent cities.

[0062] In this embodiment, the air quality prediction model is a pre-trained time prediction model. After obtaining the feature data matrices corresponding to the target city and its neighboring cities, the feature data matrix of the target city is input into the air quality prediction model to obtain the control air quality prediction data of the target city; the feature data matrices of the neighboring cities are input into the air quality prediction model to obtain the control air quality prediction data of the neighboring cities.

[0063] S103 updates the air quality forecast data of the target city by using meteorological forecast data and geographic information data of the target city and its neighboring cities, as well as the air quality forecast data of the neighboring cities, to obtain the final air quality forecast data of the target city.

[0064] It should be noted that the air quality prediction data of the target city and neighboring cities calculated in step S102 are only used to represent the air quality affected by the local area itself. However, the control quality will be affected by other cities. For example, if a sandstorm occurs in a certain city, the neighboring cities will be affected by the sandstorm due to the wind direction. In view of the above reasons, this embodiment provides a method for predicting the final air quality data of the target city to solve the above problems.

[0065] Specifically, such as Figure 2 The diagram shown is a flowchart for determining the final air quality forecast data of a target city in this embodiment. The process for determining the final air quality forecast data of the target city includes:

[0066] S1031, calculate the static attenuation value using geographic information data of the target city and its neighboring cities; calculate the wind direction correction value and wind speed correction value using meteorological forecast data of the target city and its neighboring cities.

[0067] Specifically, it can be done through formulas Calculate the static attenuation value, where, This is the static attenuation value. Let be the geographical distance between target city i and its neighboring city j. This is the attenuation radius (usually taken as 200-500km, adjusted according to the regional scale). For example: the distance from city A to B is approximately 110km, if d0 = 200k, then d0 = 200k. base ≈0.58.

[0068] It should be noted that before calculating the wind direction correction value and wind speed correction value using the meteorological forecast data of the target city and its neighboring cities, the method further includes: acquiring historical meteorological data of the target city and performing wavelet transform on the historical meteorological data to extract the monsoon cycle; determining the dominant monsoon wind direction and average monsoon wind speed in each monsoon cycle; and calculating the wind direction correction value and wind speed correction value using the dominant monsoon wind direction, the average monsoon wind speed, and the meteorological forecast data of the target city and its neighboring cities.

[0069] Morlet wavelet transform was applied to the annual wind speed data to extract the annual cycle (12 months), seasonal cycle (3 months), and small-scale (e.g., hourly) data. For example, wavelet analysis of 10 years of wind speed data for a city revealed that the power spectra of the 12-month cycle (annual cycle) and the 3-month cycle (seasonal cycle) exceeded the 95% confidence level, thus identifying them as significant cycles. The dominant wind direction components (e.g., northwesterly winds in winter and southeasterly winds in summer) were reconstructed. Daubechies wavelet was applied to the hourly temperature data to extract the diurnal pattern of boundary layer height changes. For example, after reconstructing the winter monsoon cycle in City A, the dominant northwesterly wind period was from December 1st to February 28th, with an average wind speed of 4.2 m / s. The summer diurnal cycle in City C showed that the boundary layer height dropped below 200 meters at night, making PM2.5 more likely to accumulate.

[0070] Specifically, the step of calculating wind direction correction values ​​and wind speed correction values ​​using the prevailing monsoon wind direction, the average monsoon wind speed, and meteorological forecast data of the target city and its neighboring cities includes: performing weighted calculations on the prevailing monsoon wind direction and the average monsoon wind speed for all monsoon cycles of various lengths to update the prevailing monsoon wind direction and the average monsoon wind speed; and calculating the wind direction correction values ​​and wind speed correction values ​​using the updated prevailing monsoon wind direction, the average monsoon wind speed, and meteorological forecast data of the target city and its neighboring cities.

[0071] In this embodiment, through the formula The prevailing monsoon wind direction and average monsoon wind speed are updated using a weighted calculation. The weighting coefficients α, β, and γ can be determined through regression analysis of historical data. Then, wind direction and wind speed correction values ​​are calculated using the updated prevailing monsoon wind direction, average monsoon wind speed, and meteorological forecast data for the target city and its neighboring cities.

[0072] Specifically, this embodiment can be achieved through the formula Calculate the wind direction correction value using the formula. Calculate the wind speed correction value. Among them, This is a wind direction correction factor (0.2-0.5, fitted using historical data). The angle between the direction of the line connecting the target city i to the adjacent city j and the prevailing monsoon wind direction (0° represents a completely tailwind, and 180° represents a headwind). The average wind speed of the monsoon (in m / s). The reference wind speed is 3 m / s.

[0073] For example, the prevailing monsoon wind direction: such as the northwest wind in winter (315° azimuth), the direction of the city line connection: the azimuth from city A to city C is southeast (120°), the included angle is 𝜃 𝑖𝑗 =|315°−120°|=195°, but the minimum included angle of 165° is actually taken (i.e., 360°-195°). cos𝜃 𝑖𝑗 =cos165°≈−0.966, when cos𝜃 𝑖𝑗 >0: Tailwind, increases transmission weight; when cos 𝜃 𝑖𝑗 <0: Headwind, suppressing transmission weight. If 𝛼=0.3, then the correction factor is 1+0.3×(−0.966)≈0.711+0.3×(−0.966)≈0.71. If the average wind speed of the northwest wind in winter... =4.5k / k, base wind speed =3k / k, then the correction factor is: 4.5 / 3 = 1.5.

[0074] S1032, calculate the urban pollution weight value between the target city and its neighboring cities based on the static attenuation value, the wind direction correction value, and the wind speed correction value.

[0075] In this embodiment, the relationships between cities can be converted into a graph structure, where nodes represent cities (e.g., cities A, C, E, etc.), and each node contains a feature vector (e.g., PM2.5 concentration, industrial emissions, meteorological data, etc.). Edges in the graph structure represent pollution transmission relationships between cities (represented by city pollution weight values). Specifically, calculating the city pollution weight value between the target city and its neighboring cities based on the static attenuation value, the wind direction correction value, and the wind speed correction value includes: determining the product of the static attenuation value, the wind direction correction value, and the wind speed correction value as the city pollution weight value between the target city and its neighboring cities.

[0076] The urban pollution weight value is calculated using the following formula:

[0077]

[0078] For example, during the winter northwest wind season, the calculation process for the urban pollution weight value from city A to city B is as follows:

[0079] Distance: 110km → Base weight 𝑒 −110 / 200 ≈0.58.

[0080] Wind direction angle: 0° (completely downwind) → wind direction = 1 + 0.3 × 1 = 1.3.

[0081] Wind speed correction: 4.5m / s → wind speed = 4.5 / 3 = 1.5.

[0082] Urban pollution weight value: 0.58×1.3×1.5≈1.13.

[0083] S1033 updates the air quality forecast data of the target city by using the city pollution weight value, the air quality forecast data of the target city and its neighboring cities, and meteorological forecast data to obtain the final air quality forecast data of the target city.

[0084] In this embodiment, the air quality prediction data of the target city is updated using the city pollution weight value, the air quality prediction data of the target city and its neighboring cities, and the meteorological prediction data to obtain the final air quality prediction data of the target city. This includes: converting the city pollution weight value, the air quality prediction data of the target city and its neighboring cities, and the meteorological prediction data into a feature vector matrix; and inputting the feature vector matrix into a temporal-space convolution model to obtain the final air quality prediction data of the target city.

[0085] In this embodiment, the temporal-space convolutional model handles the transmission relationships between cities, and the propagation formula for each layer is as follows:

[0086]

[0087] in, For activation function, The adjacency matrix is ​​determined based on the city's pollution weight values ​​and is updated hourly. This is the feature vector matrix transformed from air quality forecast data and meteorological forecast data of the target city and its neighboring cities. This is a trainable weight matrix.

[0088] It should be noted that the loss function of the temporal-space convolution model in this embodiment consists of the prediction error loss. and physical constraint loss Composition, i.e., loss function When the weighted sum of the two loss values ​​is less than a predetermined value, the training of the temporal-space convolutional model is complete, where the prediction error loss... The calculation formula is:

[0089]

[0090] in, The model predicts the concentration of city A at time R+1. This represents the actual monitored concentration of PM2.5 in city A at time t+1. For example, if the model predicts that the PM2.5 concentration in city A is 85 KJ / k. 3 The actual value is 90 oz / oz 3 Therefore, the single-point error is (85−90)² = 25. After averaging over the 13 cities, =15.6.

[0091] The formula for physical constraint loss is:

[0092]

[0093] in, The concentration change represents the accumulation or dissipation of pollutants over time. Transport volume is the net input / output between cities, expressed by the formula... The calculation yielded the following result. The total amount of pollutants transported from upstream city A to city B. Let i be the pollution weight value for cities i to j. The concentration gradient drives the transport of pollutants from a high concentration zone (𝑖) to a low concentration zone (𝑗).

[0094] For example, the model predicts PM2.5 concentrations in 13 cities over the next 6 hours and calculates pollution transport pathways. Actual concentrations: [90, 120, 85, ...] 𝑇 (Unit: 𝜇𝑔 / 𝑚3); Predicted concentrations: [85, 115, 80, ...] 𝑇

[0095]

[0096] Transmission volume calculation: Based on the dynamic adjacency matrix, the net total input is +25kJ / k. 3 •h. Concentration change: P M 2 . 5 t + 1 − P M 2 . 5 t = [ + 5 , − 1 0 , + 3 , … ] T

[0097]

[0098] like =0.5, then

[0099] It should be noted that in this embodiment... Used to control the contribution weight of physical constraint terms to the total loss. If the size is too large, the model may over-satisfy physical constraints, potentially sacrificing prediction accuracy. If the value is too small, the constraints of physical laws are insufficient, and the model may output unreasonable results (such as pollutants appearing out of thin air). Therefore, this embodiment selects the appropriate value through cross-validation. ,generally Between 0.1 and 1.0.

[0100] In an optional embodiment provided in this application, after inputting the feature vector matrix into a temporal-space convolution model to obtain the final air quality prediction data of the target city, the method further includes: obtaining the diurnal diffusion index of the target city; and updating the final air quality prediction data of the target city based on the diurnal diffusion index of the target city.

[0101] The step of obtaining the diurnal diffusion index of the target city includes: obtaining the daytime diffusion coefficient and the nighttime diffusion coefficient of the target city; calculating the ratio of the daytime diffusion coefficient and the nighttime diffusion coefficient of the target city to obtain the initial diurnal diffusion index; and calculating the diurnal diffusion index of the target city based on the green coverage rate of the target city and the initial diurnal diffusion index of the target city.

[0102] In this embodiment, the formula for calculating the day-night diffusion index is:

[0103]

[0104]

[0105]

[0106]

[0107] in, The initial diurnal diffusion index, This is the daytime diffusion coefficient, which is positively correlated with boundary layer height and wind speed. This is the nighttime diffusion coefficient, which is affected by the thickness and stability of the inversion layer. This represents the daytime boundary layer height (which can be estimated using meteorological sounding data or models). The average wind speed during the day. This is the adjustment coefficient (usually taken as 0.1-0.3). Normalized Difference Vegetation Index (reflecting green coverage). This is the nighttime boundary layer height (typically 10%-30% of the daytime height). The average wind speed at night. This refers to the temperature difference between the Earth's surface and the bottom of the inversion layer. The greater the temperature difference, the weaker the diffusion capacity. This is the attenuation coefficient (empirical value approximately 0.05-0.1). The diurnal diffusion index of the target city. This is the greening adjustment coefficient (which needs to be locally calibrated). The target city's green coverage rate.

[0108] In this embodiment, the final air quality prediction data for the target city can be updated based on the diurnal diffusion index of the target city using the following formula:

[0109]

[0110] in, For updated air quality forecast data, This is the final air quality prediction data obtained through the temporal-space convolution model. This is the historical average DNI. For the current real-time calculation , The sensitivity coefficient is typically set to 0.3-0.6 and is obtained by fitting local data.

[0111] It should be noted that when > When: diffusion capacity is better than normal, predictable concentration is lowered; when < Time: Diffusion capacity deteriorates, leading to an upward revision of the predicted concentration.

[0112] Traditional models often underestimate the accumulation of pollution caused by nighttime temperature inversions. This embodiment can correct air quality forecast data. For example, if the predicted nighttime PM2.5 level for city D is 80 μg / m³, =8( =15), and the concentration can be increased by correcting it using the correction formula in this embodiment, that is The actual observed value was 108 μg / m³, so the error can be reduced from 37.5% to 1.9% using the correction formula in this embodiment.

[0113] However, with the increase in urban green space ratio, the DNI (Diffusion-Induced National Influence) rises due to enhanced nighttime diffusion. For example, if the green space ratio increases from 15% to 20%. The concentration increased from 12 to 13.5; the concentration can be decreased by correcting it using the correction formula in this embodiment. .

[0114] S104, determine whether the final air quality forecast data for the target city is greater than the preset anomaly value.

[0115] The preset abnormal values ​​can be set according to actual needs.

[0116] S105, if it is greater than the preset abnormal value, then determine the cause of the final air quality prediction data.

[0117] In this embodiment, the process of determining the cause of the final air quality prediction data can be as follows: calculate the pollution contribution rate, and then determine the cause of the final air quality prediction data based on the pollution contribution rate.

[0118] The formula for calculating the pollution contribution rate is: ;

[0119] in, The total amount of pollutants transported from upstream city A to city B. Let i be the pollution weight value for cities i to j. The concentration gradient drives the transport of pollutants from a high concentration zone (𝑖) to a low concentration zone (𝑗). This represents the total pollutant concentration of city A at time A+1, typically a monitoring value or model prediction value for a single spatiotemporal node.

[0120] S106, Based on the causes of the final air quality forecast data, manage and control industrial and traffic emissions in the target city or its neighboring cities.

[0121] For example, the PM2.5 concentration in city B at time 1+1. ,

[0122] If the sum of the transmission throughput from all upstream cities to city B , .

[0123] Based on this contribution rate, it can be determined that 60% of PM2.5 pollution in City B comes from external cities (such as City A, City D, etc.), while 40% originates from local emissions.

[0124] For industrial and traffic emissions within the city due to internal factors, the information in Table 1 can be used to manage these emissions. In the table, DNI is calculated using the formula... Calculated.

[0125] Table 1

[0126] DNI range Control Level Example of measures DNI > 15 Yellow Alert Strengthen nighttime road watering and restrict the use of construction waste trucks 5 ≤ DNI ≤ 15 Orange alert Shutting down 30% of high-emission enterprises and suspending outdoor classes in primary and secondary schools. DAYS < 5 Red Alert The entire region will implement odd-even license plate restrictions, and heavy industry production will be limited to 50%.

[0127] This embodiment provides an air quality forecasting and management method for air pollution prevention and control. First, it acquires meteorological forecast data, industrial emission data, traffic emission data, geographic information data, and real-time pollutant concentration monitoring data for a target city and its neighboring cities. Then, based on the industrial emission data, traffic emission data, and real-time pollutant concentration monitoring data of the target city and its neighboring cities, it calculates air quality forecast data for both the target city and its neighboring cities. Next, it updates the target city's air quality forecast data using the meteorological forecast data, geographic information data, and air quality forecast data of the neighboring cities to obtain the final air quality forecast data for the target city. It then determines whether the final air quality forecast data for the target city exceeds a preset anomaly value. If it does, it determines the cause of the increased final air quality forecast data. Finally, based on the cause of the increased final air quality forecast data, it manages and controls industrial and traffic emissions in the target city or its neighboring cities. Compared to existing technologies that only allow users to query relevant air quality information on websites, this application calculates air quality forecast data based on various data between cities and then updates the target city's air quality forecast data based on the air quality forecast data of neighboring cities. This improves the accuracy of air quality forecasting for the target city and thus enhances the effectiveness of air quality management in the target city.

[0128] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0129] In one embodiment, an air quality prediction and management system for air pollution prevention and control is provided. For example... Figure 3 As shown, the functional modules of this air quality forecasting and management system for air pollution prevention and control are described in detail below:

[0130] The acquisition module 31 is used to acquire meteorological forecast data, industrial emission data, traffic emission data, geographic information data, and real-time pollutant concentration monitoring data of the target city and its neighboring cities.

[0131] Calculation module 32 is used to calculate the air quality prediction data of the target city and the adjacent cities based on the industrial emission data, traffic emission data and real-time pollutant concentration monitoring data of the target city and the adjacent cities respectively;

[0132] The update module 33 is used to update the air quality forecast data of the target city using meteorological forecast data and geographic information data of the target city and its neighboring cities, as well as the air quality forecast data of the neighboring cities, to obtain the final air quality forecast data of the target city.

[0133] The determination module 34 is used to determine whether the final air quality prediction data of the target city is greater than a preset anomaly value;

[0134] The determining module 34 is also used to determine the cause of the final air quality prediction data if it is greater than a preset abnormal value.

[0135] Management module 35 is used to manage and control industrial and traffic emissions of the target city or its neighboring cities based on the reasons for the generation of the final air quality prediction data.

[0136] In an optional embodiment, the calculation module 32 is specifically used for:

[0137] The industrial emission data, the traffic emission data, and the real-time pollutant concentration monitoring data are preprocessed, including anomaly filtering and data normalization.

[0138] The preprocessed industrial emission data, traffic emission data, and real-time pollutant concentration monitoring data are converted into a feature data matrix with a time dimension.

[0139] The feature data matrices corresponding to the target city and the adjacent cities are input into the air quality prediction model to obtain the air quality prediction data corresponding to the target city and the adjacent cities.

[0140] In an optional embodiment, update module 33 is specifically used for:

[0141] The static attenuation value is calculated using geographic information data of the target city and its neighboring cities; the wind direction correction value and wind speed correction value are calculated using meteorological forecast data of the target city and its neighboring cities.

[0142] The city pollution weight values ​​between the target city and its neighboring cities are calculated based on the static attenuation value, the wind direction correction value, and the wind speed correction value.

[0143] The final air quality forecast data for the target city is obtained by updating the air quality forecast data of the target city using the city pollution weight value, the air quality forecast data of the target city and its neighboring cities, and meteorological forecast data.

[0144] In an optional embodiment, update module 33 is specifically used for:

[0145] The product of the static attenuation value, the wind direction correction value, and the wind speed correction value is determined as the urban pollution weight value between the target city and its neighboring cities.

[0146] In an optional embodiment, update module 33 is specifically used for:

[0147] The city pollution weight value, the air quality prediction data of the target city and its neighboring cities, and the meteorological prediction data are converted into a feature vector matrix.

[0148] The feature vector matrix is ​​input into the temporal-space convolution model to obtain the final air quality prediction data for the target city.

[0149] In an optional embodiment, the acquisition module 31 is further configured to acquire historical meteorological data of the target city and perform wavelet transform on the historical meteorological data to extract the monsoon cycle.

[0150] The determination module 34 is also used to determine the prevailing monsoon wind direction and the average monsoon wind speed in each monsoon cycle;

[0151] The calculation module 32 is also used to calculate wind direction correction values ​​and wind speed correction values ​​using the prevailing monsoon wind direction, the average monsoon wind speed, and meteorological forecast data of the target city and its neighboring cities.

[0152] In an optional embodiment, the calculation module 32 is specifically used for:

[0153] The prevailing monsoon direction and average monsoon speed are calculated and updated by weighting for monsoon cycles of all lengths.

[0154] Wind direction corrections and wind speed corrections are calculated using updated monsoon prevailing wind direction, monsoon average wind speed, and meteorological forecast data for the target city and its neighboring cities.

[0155] In an optional embodiment, the acquisition module 31 is further configured to acquire the day-night diffusion index of the target city;

[0156] The update module 33 is also used to update the final air quality prediction data of the target city based on the diurnal diffusion index of the target city.

[0157] In an optional embodiment, the acquisition module 31 is specifically used for:

[0158] Obtain the daytime and nighttime diffusion coefficients of the target city;

[0159] The initial diurnal diffusion index is obtained by calculating the ratio of the daytime diffusion coefficient to the nighttime diffusion coefficient of the target city.

[0160] The day-night diffusion index of the target city is calculated based on the green coverage rate of the target city and the initial day-night diffusion index of the target city.

[0161] It should be noted that the above detailed descriptions are exemplary and intended to provide further explanation of this application. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0162] For specific limitations regarding an air quality forecasting and management system for air pollution control, please refer to the limitations of an air quality forecasting and management method for air pollution control described above, which will not be repeated here. Each module in the aforementioned device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0163] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above.

[0164] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. An air quality prediction and management method for atmospheric pollution prevention and control, characterized in that the method... include: S101, acquire meteorological forecast data, industrial emission data, traffic emission data, geographic information data, and real-time pollutant concentration monitoring data of the target city and its neighboring cities; S102, calculate the predicted air quality data for the target city and neighboring cities based on industrial emission data, traffic emission data, and real-time pollutant concentration monitoring data for the target city and neighboring cities respectively; S103, by updating the air quality forecast data of the target city using meteorological forecast data and geographic information data of the target city and its neighboring cities, as well as the air quality forecast data of the neighboring cities, the final air quality forecast data of the target city is obtained. S104, determine whether the final air quality forecast data for the target city is greater than the preset anomaly value; S105, if it is greater than the preset abnormal value, then determine the cause of the final air quality forecast data; S106, Based on the causes of the final air quality forecast data, manage and control industrial and traffic emissions in the target city or its neighboring cities; S103 includes: S1031, calculate the static attenuation value using geographic information data of the target city and its neighboring cities; calculate the wind direction correction value and wind speed correction value using meteorological forecast data of the target city and its neighboring cities. Before calculating the wind direction and wind speed correction values ​​using meteorological forecast data of the target city and its neighboring cities, the method further includes: acquiring historical meteorological data of the target city and extracting monsoon cycles by performing wavelet transform on the historical meteorological data; determining the prevailing monsoon wind direction and average monsoon wind speed in each monsoon cycle; performing weighted calculations on the prevailing monsoon wind direction and average monsoon wind speed in monsoon cycles of all lengths to update the prevailing monsoon wind direction and average monsoon wind speed; and calculating the wind direction and wind speed correction values ​​using the updated prevailing monsoon wind direction, average monsoon wind speed, and meteorological forecast data of the target city and its neighboring cities. S1032, calculate the urban pollution weight value between the target city and its neighboring cities based on the static attenuation value, wind direction correction value, and wind speed correction value; wherein, the product of the static attenuation value, wind direction correction value, and wind speed correction value is determined as the urban pollution weight value between the target city and its neighboring cities. S1033, by updating the air quality forecast data of the target city using the city pollution weight value, the air quality forecast data of the target city and its neighboring cities, and the meteorological forecast data, the final air quality forecast data of the target city is obtained; this includes: converting the city pollution weight value, the air quality forecast data of the target city and its neighboring cities, and the meteorological forecast data into a feature vector matrix; and inputting the feature vector matrix into a temporal-space convolution model to obtain the final air quality forecast data of the target city. The temporal-space convolutional model handles inter-city transport relationships. The loss function of the temporal-space convolutional model consists of the prediction error loss. and physical constraint loss Composition; the temporal-space convolutional model handles inter-city transmission relationships, with the propagation formula for each layer as follows: in, For activation function, The adjacency matrix is ​​determined based on the city's pollution weight values ​​and is updated hourly. This is the feature vector matrix transformed from air quality forecast data and meteorological forecast data of the target city and its neighboring cities. The weight matrix is ​​trainable. Loss Function When the weighted sum of the two loss values ​​is less than a predetermined value, the training of the temporal-space convolutional model is complete, where the prediction error loss... The calculation formula is: ； ； in, The model predicts the concentration of city A at time R+1. The actual monitored concentration of city A at time R+1; The concentration change represents the accumulation or dissipation of pollutants over time. Transport volume is the net input / output between cities, expressed by the formula... Calculated; After inputting the feature vector matrix into the temporal-space convolutional model to obtain the final air quality prediction data for the target city, the method also includes: obtaining the day-night diffusion index of the target city; and updating the final air quality prediction data of the target city based on the day-night diffusion index. The process of obtaining the day-night diffusion index of the target city includes: obtaining the daytime diffusion coefficient and the nighttime diffusion coefficient of the target city; calculating the ratio of the daytime diffusion coefficient and the nighttime diffusion coefficient of the target city to obtain the initial day-night diffusion index; and calculating the day-night diffusion index of the target city based on the green coverage rate of the target city and the initial day-night diffusion index of the target city.

2. The method according to claim 1, characterized in that, Based on industrial emission data, traffic emission data, and real-time pollutant concentration monitoring data of the target city and neighboring cities, air quality forecast data for the target city and neighboring cities are calculated separately, including: Preprocessing is performed on industrial emission data, traffic emission data, and real-time pollutant concentration monitoring data. Preprocessing includes anomaly filtering and data normalization. Preprocessed industrial emission data, traffic emission data, and real-time pollutant concentration monitoring data are converted into a feature data matrix with a time dimension. The feature data matrices corresponding to the target city and neighboring cities are input into the air quality prediction model to obtain the air quality prediction data for the target city and neighboring cities.

3. An air quality forecasting and management system for atmospheric pollution prevention and control, wherein the system implements the method as described in claim 1, characterized in that the system... include: The acquisition module is used to acquire meteorological forecast data, industrial emission data, traffic emission data, geographic information data, and real-time pollutant concentration monitoring data of the target city and its neighboring cities. The calculation module is used to calculate the air quality prediction data for the target city and neighboring cities based on industrial emission data, traffic emission data, and real-time pollutant concentration monitoring data of the target city and neighboring cities, respectively. The update module is used to update the air quality forecast data of the target city by using meteorological forecast data and geographic information data of the target city and its neighboring cities, as well as the air quality forecast data of the neighboring cities, to obtain the final air quality forecast data of the target city. The determination module is used to determine whether the final air quality forecast data for the target city is greater than the preset anomaly value; The determination module is also used to determine the cause of the final air quality forecast data if it exceeds a preset outlier value. The management module is used to manage and control industrial and traffic emissions in the target city or its neighboring cities based on the causes of the final air quality forecast data.

Images