Airport emission list construction method and system

Abstract

The invention discloses an airport emission list construction method and system, and relates to the technical field of atmospheric emission, and the method comprises the steps: building an airport traffic operation simulation model of an airport scene and airspace, carrying out a simulation experiment facing airport emission, and obtaining a multivariate data set related to airport emission parameters; a k-nearest neighbor mutual information feature selection method is adopted to select related variable subsets of airport atmospheric emission image factors from the multivariate data set; determining an attribution value of each target feature in the correlation variable quantum set based on an SHAP model interpretation method, and determining a correction coefficient of airport localized emission according to the attribution value of each target feature; and correcting the airport emission output by the airport traffic operation simulation model according to the correction coefficient to obtain a corrected airport emission list.

Description

Technical Field

[0001] The present application relates to the field of atmospheric emission technology, and in particular to a method and system for constructing an airport emission inventory. Background Art

[0002] Airports are one of the important sources of atmospheric pollutants and carbon dioxide emissions. Compiling a high-resolution airport emission inventory and finding out the emission and changing patterns of atmospheric pollutants and carbon dioxide can provide scientific and theoretical support for the coordinated and efficient work of pollution reduction and carbon reduction in related fields. At present, the construction of the airport emission inventory does not take into account the correlation between various influencing factors. Therefore, the accuracy of the construction of the airport emission inventory needs to be improved. Summary of the invention

[0003] The purpose of this application is to provide a method and system for constructing an airport emission inventory, which can improve the accuracy of constructing an airport emission inventory.

[0004] To achieve the above objectives, this application provides the following solutions:

[0005] In a first aspect, the present application provides a method for constructing an airport emission inventory, comprising:

[0006] Build an airport traffic operation simulation model for the airport surface and airspace, conduct simulation experiments for airport emissions, and obtain a multivariate data set related to airport emission parameters;

[0007] A k-nearest neighbor mutual information feature selection method is used to select a subset of relevant variables of airport atmospheric emission image factors from the multivariate data set;

[0008] Determine the attribution value of each target feature in the relevant variable subset based on the SHAP model interpretation method, and determine the correction coefficient of airport localized emissions according to the attribution value of each target feature;

[0009] The airport emissions output by the airport traffic operation simulation model are corrected according to the correction coefficient to obtain a corrected airport emissions inventory.

[0010] Optionally, each sample in the multivariate data set includes characteristic variables and target variables, the characteristic variables including atmospheric temperature, wind speed and direction, horizontal speed, control interval and weather phenomena, and the target variables including the operating time of each stage of each flight landing and take-off cycle, fuel flow rate and fuel consumption.

[0011] Optionally, a k-nearest neighbor mutual information feature selection method is used to select a subset of relevant variables of the airport atmospheric emission image factors from the multivariate data set, specifically including:

[0012] Initialize the relevant variable subset to be empty, each feature variable constitutes a feature variable set, and each target variable constitutes a target variable set;

[0013] Calculate the mutual information between each feature variable and the target variable set, and store the feature variable corresponding to the maximum value in each mutual information as the target feature in the relevant variable subset;

[0014] Sorting the mutual information in descending order to obtain a mutual information ranking;

[0015] Traverse the mutual information sorting to calculate the high-dimensional mutual information, and output the relevant variable subset: when traversing to the current traversal mutual information, calculate the mutual information of the current relevant variable subset, the characteristic variables corresponding to the current traversal mutual information and the target variable set, and record it as the current first mutual information; calculate the mutual information of the current relevant variable subset and the target variable set, and record it as the current second mutual information; if the current first mutual information is greater than the current second mutual information, store the characteristic variables corresponding to the current traversal mutual information as the target feature in the relevant variable subset.

[0016] Optionally, the formula for calculating the mutual information between each feature variable and the target variable set is:

[0017]

[0018] Among them, x i is the i-th feature variable in the feature variable set, Y is the target variable set, MI(x i , Y) is x i The mutual information with Y, n is the total number of samples in the multivariate data set, ψ() is the double Γ function, α xi is the data set of the feature variable set and x i Distance less than ρ i / 2 the number of data points, α yi is the data set of the feature variable set and y i Distance less than ρ i / 2 the number of data points, ρ i For x i The distance to the kth nearest neighbor, k is a constant, 2≤k≤6.

[0019] Optionally, the formula for calculating the mutual information between the current subset of relevant variables, the feature variables corresponding to the current traversal mutual information, and the target variable set, and calculating the mutual information between the current subset of relevant variables and the target variable set is:

[0020]

[0021] Among them, MI(X1, X2, ..., X m) represents variable X1, variable X2, ..., variable X m The mutual information between xim Represents variable X m The value in the i-th sample, α xi1 represents the value of variable X1 in the i-th sample, α xi2 Represents the value of variable X2 in the i-th sample.

[0022] Optionally, the formula for determining the correction factor of airport localized emissions based on the attribution value of each target characteristic is:

[0023]

[0024] in, is the correction coefficient, θ is the air traffic control condition, ω is the meteorological condition, is the average attribution value of the lth target feature, f_vc a，j，l It is the fuel consumption output value of aircraft a performing flight phase j in the benchmark simulation model corresponding to the lth target feature.

[0025] Optionally, the airport emissions output by the airport traffic operation simulation model are corrected according to the correction coefficient to obtain a corrected airport emissions inventory, specifically including:

[0026]

[0027] Among them, E′ a，j，b is the corrected emission of pollutant b when aircraft a performs flight phase j, EI a，j，b is the emission of pollutant b when aircraft a performs flight phase j output by the airport traffic operation simulation model, t_vc a，j is the simulation running time corresponding to the flight phase j of aircraft a, n a Indicates the number of engines of aircraft a, FF a，j represents the fuel flow rate of aircraft a performing flight phase j.

[0028] Optionally, an airport traffic operation simulation model of the airport surface and airspace is built to conduct simulation experiments for airport emissions and obtain a multivariate data set related to airport emission parameters, including:

[0029] The whole airspace and airport fast-time simulation tools are used to build airport traffic operation simulation models of the airport scene and airspace, and simulation experiments for airport emissions are carried out to obtain a multivariate data set related to airport emission parameters.

[0030] Optionally, according to the correction coefficient, the input data of the airport traffic operation simulation model includes airport surface physical data, airport surface operation status, flight plan, airport approach control area static traffic capacity and airport approach control area dynamic operation status.

[0031] In a second aspect, the present application provides an airport emission inventory construction system, comprising:

[0032] The data set determination module is used to build an airport traffic operation simulation model for the airport surface and airspace, conduct simulation experiments for airport emissions, and obtain a multivariate data set related to airport emission parameters;

[0033] A relevant variable subset determination module, used for selecting a relevant variable subset of the airport atmospheric emission image factor from the multivariate data set by using a k-nearest neighbor mutual information feature selection method;

[0034] A correction coefficient determination module, used to determine the attribution value of each target feature in the relevant variable subset based on the SHAP model interpretation method, and determine the correction coefficient of the airport localized emissions according to the attribution value of each target feature;

[0035] The airport emission inventory correction module is used to correct the airport emissions output by the airport traffic operation simulation model according to the correction coefficient to obtain a corrected airport emission inventory.

[0036] According to the specific embodiments provided in this application, this application discloses the following technical effects:

[0037] The present application provides a method and system for constructing an airport emission inventory, which adopts a k-nearest neighbors mutual information feature selection method to select a subset of relevant variables of airport atmospheric emission image factors from a multivariate data set, identify key factors (target features) of image airport emissions, determine the attribution value of each target feature in the relevant variable subset based on the SHapley Additive Explanations (SHAP) model interpretation method, and determine the correction coefficient of airport localized emissions according to the attribution value of each target feature, thereby realizing the correction of the airport emission inventory by utilizing the impact of key factors affecting airport traffic emissions on airport emissions, and improving the accuracy of constructing the airport emission inventory. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0039] Figure 1 A flowchart of a method for constructing an airport emission inventory provided in one embodiment of the present application;

[0040] Figure 2 A schematic diagram of a simulation modeling process provided by an embodiment of the present application;

[0041] Figure 3 A flow chart of a k-nearest neighbor mutual information algorithm provided in one embodiment of the present application;

[0042] Figure 4 A schematic diagram of the construction principle of the airport emission localization correction model based on SHAP provided in one embodiment of the present application;

[0043] Figure 5 A detailed flowchart of a method for constructing an airport emission inventory provided in one embodiment of the present application;

[0044] Figure 6 A schematic diagram of a gridded HC emission inventory for an airport provided in an embodiment of the present application;

[0045] Figure 7 A schematic diagram of a gridded CO emission inventory for an airport provided in an embodiment of the present application;

[0046] Figure 8 An airport NO provided in an embodiment of the present application X Schematic diagram of gridded emissions inventory;

[0047] Fig. 9 An airport SO provided in an embodiment of the present application X Schematic diagram of gridded emissions inventory;

[0048] Fig.10 A schematic diagram of a PM grid emission inventory for an airport provided in an embodiment of the present application;

[0049] Fig.11 A schematic diagram of a gridded CO2 emission inventory for an airport provided in an embodiment of the present application;

[0050] Fig.12 A schematic diagram of the functional modules of an airport emissions inventory construction system provided in one embodiment of the present application. DETAILED DESCRIPTION

[0051] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0052] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0053] This application provides a method for constructing an airport emission inventory, such as Figure 1 As shown, a method for constructing an airport emission inventory includes:

[0054] Step 101: Build an airport traffic operation simulation model of the airport surface and airspace, conduct a simulation experiment for airport emissions, and obtain a multivariate data set related to airport emission parameters.

[0055] Step 102: Select a subset of relevant variables of airport atmospheric emission image factors from the multivariate data set using a k-nearest neighbor mutual information feature selection method.

[0056] Step 103: Determine the attribution value of each target feature in the relevant variable subset based on the SHAP model interpretation method, and determine the correction coefficient of the airport localized emissions according to the attribution value of each target feature.

[0057] Step 104: Correcting the airport emissions output by the airport traffic operation simulation model according to the correction coefficient to obtain a corrected airport emissions inventory.

[0058] This application uses the k-nearest neighbor mutual information feature selection method to select a subset of relevant variables of airport atmospheric emission image factors from a multivariate data set, identifies the key factors of image airport emissions, determines the attribution value of each target feature in the relevant variable subset based on the SHAP model interpretation method, and determines the correction coefficient of airport localized emissions according to the attribution value of each target feature. It realizes the use of the key factors affecting airport traffic emissions to correct the airport emission inventory, improves the accuracy of constructing the airport emission inventory, and realizes the construction of a refined airport emission inventory. Accurate quantification of atmospheric pollutant emissions generated by airport traffic operations is the primary task and basic prerequisite for carrying out airport air pollution prevention and control, which can promote various airports to improve the airport's joint working mechanism for air pollution prevention and control and carry out in-depth air pollution prevention and control.

[0059] Among them, step 101 specifically includes: using the total airspace and airport fast time simulation tool to build an airport traffic operation simulation model of the airport scene and airspace, conducting simulation experiments for airport emissions, and obtaining a multivariate data set related to airport emission parameters. More specifically, conducting a "multi-scenario-multi-factor" simulation experiment for airport emissions to output a multivariate data set of emission parameters such as operating time, fuel flow rate, and fuel consumption. The total airspace and airport fast time simulation tool includes the total airspace and airport model (Total Airspace and Airport Modelle, TAAM).

[0060] Each sample in the multivariate data set includes characteristic variables and target variables. The characteristic variables include atmospheric temperature, wind speed and direction, horizontal speed of spacecraft, control interval and weather phenomena. The target variables include the running time of each stage of the landing and take-off cycle of each flight, fuel flow rate and fuel consumption. Weather phenomena refer to four optional weathers set in advance.

[0061] In this application, the process of building an airport traffic operation simulation model of the airport scene and airspace using the full airspace and airport fast time simulation tools mainly includes three parts: static model construction, dynamic model calibration and operation data output, such as Figure 2 shown.

[0062] (1) The static model specifically refers to the entire airspace and airport model. The construction of the static model mainly involves the input of relevant data for model establishment, including the airport runways and taxiways, waypoints and navigation stations, standard approach and departure procedures, sector structure, flight plans, etc.

[0063] Based on the previous survey of the airport runway system layout, waypoint and navigation station configuration, route structure, approach and departure procedure structure and historical flight plan data of flights, the static model was built. The input data types of the airport simulation model (airport traffic operation simulation model) mainly include: airport surface physical data, airport surface operation status, flight plan, airport approach control area static traffic capacity and airport approach control area dynamic operation status. The specific input data, data description and source in the process of developing and building the airport traffic operation simulation model are shown in Table 1.

[0064] Table 1 Simulation model input data

[0065]

[0066]

[0067] (2) Dynamic model calibration: After the static model of the airport is built, the simulation operation of the TAAM model is started. The TAAM model is calibrated and dynamically corrected according to the real-time operation status. Problems in operation are discovered, and rules and restrictions are added to make timely adjustments, so that the resulting airport benchmark simulation model is closer to the actual operation status.

[0068] A typical busy day at the airport was selected to establish a benchmark simulation operation scenario for a certain airport, and the output results of the benchmark simulation operation scenario were compared with the actual situation to verify the accuracy and objectivity of the airport benchmark simulation model. The key reference indicators calibrated by the airport benchmark simulation model (airport traffic operation simulation model) were used: the simulation output airport take-off and landing flight schedule distribution (flights / hour), traffic distribution in each corridor direction (flights), and average delay time (minutes) in each runway direction were compared with the actual operation of the airport to ensure the reliability of the output data of the airport benchmark simulation model.

[0069] (3) Operational data output: After completing static modeling and ensuring that the airport benchmark simulation model is close to the actual situation and reliable, run the airport benchmark simulation model and screen the relevant output data required for the project. On the basis of clarifying the research needs and objectives, collect data such as the current airspace, control operation, flight schedule, specific flight status, and flight fuel consumption of the airport simulated by the airport benchmark simulation model. For the calculation of airport air pollution emissions, as shown in Table 2, run the airport benchmark simulation model and output the required airspace, runway and surface related data.

[0070] Table 2 Simulation model output data list

[0071]

[0072]

[0073] Based on the current conditions, facilities and operation status of the airport, this application mainly extracts the influencing factors of atmospheric pollutant emissions from airport control instructions and meteorological conditions, and sets up multi-factor and multi-scenario simulation experiments of airport aircraft flow control, horizontal speed regulation, weather phenomena, atmospheric temperature, and (near-ground) wind speed and direction. It also outputs data such as the operating time, fuel flow rate and fuel consumption of each stage of the landing and take-off (LTO) cycle of each flight in each simulation experiment scenario, so as to prepare the necessary training data set, i.e., multivariate data set, for the subsequent in-depth exploration of the correlation characteristics and correlation analysis of the influencing factors of airport traffic emissions.

[0074] The multi-factor and multi-scenario simulation experiment of airport air pollution emissions conducted in this application specifically includes: by setting multiple sets of input data, conducting sensitivity tests, obtaining multiple sets of input-output data combinations, and then identifying the main factors affecting the output results and their influencing mechanisms; the input data combination is multiple factors in the scenes such as control actions and meteorological conditions, and the output data is aircraft fuel flow and pollutant emissions. Each set of input-output data combination constitutes a sample in the multivariate data set.

[0075] This application obtains the target features through key factor correlation analysis based on k-nearest neighbor mutual information.

[0076] Mutual information measures the amount of shared information between two or more random variables. This measurement criterion is not affected by whether the variables are linearly related. Specifically, given a set of feature variables X(x1, x2, …, x n ) and a target variable set Y(y1,y2,…,y m ), for each feature x i , perform k-nearest neighbor matching with the target variable Y, that is, find the k samples most relevant to the feature. In other words, count the number of co-occurrences of each feature and the target variable, that is, the number of times the feature and the target variable appear at the same time in the k nearest neighbor samples.

[0077] The mutual information MI(X,Y) between two sets of variables can be defined by formula (1).

[0078]

[0079] In the formula, X represents the set of feature variables, Y represents the set of target variables, p(x) and p(y) are the marginal probability distributions of X and Y respectively, and p(x, y) is the joint probability distribution between two random variables. The larger the mutual information MI(X, Y), the greater the correlation between X and Y; conversely, the smaller the mutual information MI(X, Y), the smaller the correlation between X and Y.

[0080] In order to avoid the problem of "dimensionality curse" where the target variable dimension is too large and the probability distribution estimation accuracy decreases, a method for directly estimating the variable entropy and mutual information is proposed. The basic idea is: in the sample space Φ(X,Y) composed of X and Y, take ρ i / 2 is a point The distance to its k nearest neighbors, ρ xi and ρ yi Points The distances mapped to the X-axis and Y-axis are used to calculate the mutual information between each feature variable and the target variable to obtain the importance ranking of the feature variables, thereby selecting the most relevant features for subsequent research and analysis.

[0081] The estimated entropy under the k-nearest neighbor method can be directly calculated by formula (2).

[0082]

[0083] Where H is the estimated entropy, ρ i For x i The distance between the kth nearest neighbor point; c m is the volume of the unit sphere; Ψ(x) is a double Γ function and satisfies Ψ(1) = -0.5773 and Ψ(x+1) = Ψ(x)+x -1 , m represents the total number of variables, n represents the total number of samples. The value of the ln function can be calculated by formula (3).

[0084] ln[p(x)]=ψ(k)-ψ(n)-m, E(lnρ)-ln(c m ) (3)

[0085] According to formula (4), the mutual information between the feature and target variable set is calculated.

[0086]

[0087] Among them, x i is the i-th feature variable in the feature variable set, Y is the target variable set, MI(x i , Y) is x i The mutual information with Y, n is the total number of samples in the multivariate data set, ψ() is the double Γ function, α xi is the data set of the feature variable set and x i Distance less than ρ i / 2 the number of data points, α yi is the data set of the feature variable set and y i Distance less than ρ i / 2 the number of data points, ρ i For x i and the kth nearest neighbor (i.e. x i The distance between the sample and the kth nearest neighbor in the multivariate data set, k is a constant, 2≤k≤6, ψ(x) is a double Γ function, and satisfies ψ(1)≈-0.5773 and ψ(x+1)=ψ(x)+x -1 .

[0088] Multiple variables (X1, X2, ..., X m ) can be expanded from formula (4) to obtain formula (5).

[0089]

[0090] Among them, MI(X1, X2, ..., X m) represents variable X1, variable X2, ..., variable X m The mutual information between xim Represents variable X m The value in the i-th sample, α xi1 represents the value of variable X1 in the i-th sample, α xi2 represents the value of variable X2 in the i-th sample, k is variable X1, variable X2, ..., variable X m The total number of possible values.

[0091] To improve the accuracy and interpretability of airport emission calculations, Figure 3 As shown, the present application is based on the k-nearest neighbor mutual information algorithm, which evaluates the relevance and importance of features by calculating the mutual information between feature variables and target variables, preliminarily determines the composition characteristics of the sample feature space, and then identifies and determines the strongest relevant subset of factors affecting airport air pollution emissions.

[0092] Wherein, step 102 specifically includes:

[0093] Initialize the relevant variable subset S' to be empty, each feature variable constitutes a feature variable set X, and each target variable constitutes a target variable set Y.

[0094] According to formula (4), each feature variable x is calculated i The mutual information with the target variable set Y is calculated, and the feature variable corresponding to the maximum value in each mutual information is stored as the target feature in the relevant variable subset.

[0095] The mutual information is sorted in descending order to obtain a mutual information sorting.

[0096] The mutual information is traversed and sorted to calculate the high-dimensional mutual information, and a subset of related variables is outputted: when the current traversal mutual information is traversed, the mutual information of the current subset of related variables, the characteristic variables corresponding to the current traversal mutual information and the target variable set is calculated according to formula (5), which is recorded as the current first mutual information; the mutual information of the current subset of related variables and the target variable set is calculated according to formula (5), which is recorded as the current second mutual information; if the current first mutual information is greater than the current second mutual information, the characteristic variables corresponding to the current traversal mutual information are stored as target features in the subset of related variables.

[0097] The mathematical description of SHAP is an additive feature attribution method that interprets the results of the airport traffic operation simulation model as the sum of the attribution values ​​(SHAP values) of each input feature: Assume that the i-th feature is x i , the output value of the model for the i-th sample is y i (x i ), as shown in formula (6). The definition of SHAP value is shown in formula (7).

[0098]

[0099] Among them, y i (x i ) is the output result of the airport traffic operation simulation model, that is, the model to be explained; n is the number of input features; μ0 is the mean of the output results of all sample features; μ i Represents a single target feature x i The attributed value of {x1, x2, ..., x n} is the overall feature set of the input; S means removing a single feature x i The remaining feature subset after z i It is a 0-1 variable introduced for text features, representing feature x i The marginal contribution value of each feature is measured by the above formula, and the magnitude and direction of influence are judged according to the size and positivity of the SHAP value. When the SHAP value of the target feature is greater than zero, it means that the target feature has a promoting effect on the model output result. When the SHAP value of the target feature is less than zero, it means that the target feature has an inhibitory effect on the model output. The larger the absolute value of the SHAP value, the stronger the impact on the model output.

[0100] In order to achieve more accurate calculation of airport air pollutant emissions, this application analyzes the correlation characteristics of airport emission influencing factors, identifies the key influencing factors of airport emissions, and proposes a SHAP-based airport emission localization correction model, such as Figure 4 shown.

[0101] According to the experimental results of each scenario output by the simulation, the flight time of each flight under different scenario conditions, the corresponding operating time of each LTO stage, the fuel flow rate, and the fuel consumption are extracted. The SHAP value of each feature is obtained using the SHAP interpretation model, that is, the marginal contribution value of the key emission factors to the fuel consumption of Fuzhou Airport. The ratio of the average SHAP value of each characteristic variable to the output value of the fuel consumption result of the baseline simulation model is used as the correction coefficient Δ(θ,ω) of the key emission factors to the localized emissions of Fuzhou Airport. Its essential meaning is the marginal contribution rate of the key emission factors to the fuel consumption of Fuzhou Airport, thereby revealing the complex influence relationship between the key factors on aircraft fuel consumption and emissions. The SHAP value of each feature includes the characteristic value SHAP of atmospheric temperature. T1 SHAP T2 , …, SHAP T5 ; Wind speed and direction characteristic value SHAP Wind1 SHAP Wind2 , …, SHAP Wind61 ; Horizontal velocity eigenvalue SHAP V1 SHAPV2 , …, SHAP V9 ; Control interval characteristic value SHAP S1 SHAP S2 , …, SHAP S6 ; Weather phenomenon characteristic value SHAP Weather1 SHAP Weather2 , …, SHAP Weather4 .

[0102] According to the above model design, the localized correction coefficient based on the International Civil Aviation Organization (ICAO) benchmark emission calculation model is determined, and the standard time of each LTO stage of ICAO is replaced by the simulation running time. Thus, formula (8) can be obtained as the pollutant correction emission of the aircraft in different flight stages.

[0103] Among them, in step 103, the formula for determining the correction coefficient of airport localized emissions according to the attribution value of each target feature is:

[0104]

[0105] in, is the correction coefficient, i.e., the marginal contribution rate of the influencing factors to airport fuel consumption. In this study, it mainly depends on the air traffic control conditions θ and meteorological conditions ω. is the average attribution value of the lth target feature, that is, the specific contribution value of the corresponding influencing factor to the airport fuel consumption, f_vc a，j，l It is the fuel consumption output value of aircraft a performing flight phase j in the benchmark simulation model corresponding to the lth target feature. The airport traffic operation simulation model is the same experimental platform for parameter adjustment, and each target feature corresponds to a correction coefficient.

[0106] Wherein, step 104 specifically includes:

[0107]

[0108] Among them, E′ a，j，b is the corrected emission of pollutant b when aircraft a performs flight phase j, EI a，j，b is the emission index of pollutant b when aircraft a performs flight phase j output by the airport traffic operation simulation model, t_vc a，j is the simulation running time corresponding to the flight phase j of aircraft a, n a Indicates the number of engines of aircraft a, FF a，j represents the fuel flow rate (kg / s) of aircraft a performing flight phase j.

[0109] The detailed application process is as follows Figure 5 As shown in the figure, the whole airspace and airport fast-time simulation tools are used to build airport traffic operation simulation models for airport scenes and airspace, and a "multi-scenario-multi-factor" simulation experiment for airport emissions is carried out to output a multivariate data set of emission parameters such as operating time and fuel flow rate; the k-nearest neighbor mutual information is used to mine the correlation characteristics of factors affecting airport traffic emissions, and an attribution analysis is performed based on the SHAP model interpretation method. From the local to the global, the different contributions of various influencing factors to the fuel consumption and atmospheric pollutant emissions in the airport LTO cycle stage are explored in depth, and an airport emission localization correction coefficient model is constructed; on this basis, a refined airport emission inventory is compiled to explore the main atmospheric pollutants HC, CO, NO x The spatiotemporal distribution characteristics of SO2, PM and CO2 emissions, such as the gridded HC emission inventory of an airport Figure 6 As shown in Figure 2, the gridded CO emission inventory is as follows: Figure 7 shown, NO x Gridded emission inventories such as Figure 8 As shown in the figure, the gridded SO2 emission inventory is as follows: Fig. 9 The gridded PM emission inventory is shown in Fig.10 As shown in the figure, the gridded CO2 emission inventory is as follows: Fig.11 As shown, Figures 6 to 11 The medium spatial resolution is ≤10m; the temporal resolution is ≤1h, which enables the refined construction of the airport emission inventory.

[0110] This application aims at the particularity of the spatiotemporal aggregation of airport pollution emissions, combines the advantages of repeatable experiments in traffic simulation, and uses machine learning methods to identify the key factors of airport traffic emissions based on multi-scenario traffic operation simulation experiments for airport emissions, revealing the complex influence relationship between key factors and aircraft fuel consumption and emissions, and then constructs a refined airport emission inventory method model based on traffic simulation. This improves the accuracy of building airport emission inventories.

[0111] Based on the same inventive concept, the embodiment of the present application also provides an airport emission inventory construction system for implementing the airport emission inventory construction method involved above. The implementation solution provided by the system to solve the problem is similar to the implementation solution recorded in the above method, so the specific limitations in one or more airport emission inventory construction system embodiments provided below can refer to the limitations of the airport emission inventory construction method above, and will not be repeated here.

[0112] In an exemplary embodiment, Fig.12 As shown, a system for constructing an airport emission inventory is provided, comprising:

[0113] The data set determination module 201 is used to build an airport traffic operation simulation model of the airport surface and airspace, conduct simulation experiments for airport emissions, and obtain a multivariate data set related to airport emission parameters.

[0114] The relevant variable subset determination module 202 is used to select the relevant variable subset of the airport atmospheric emission image factor from the multivariate data set by using the k-nearest neighbor mutual information feature selection method.

[0115] The correction coefficient determination module 203 is used to determine the attribution value of each target feature in the relevant variable subset based on the SHAP model interpretation method, and determine the correction coefficient of the airport localized emissions according to the attribution value of each target feature.

[0116] The airport emission inventory correction module 204 is used to correct the airport emissions output by the airport traffic operation simulation model according to the correction coefficient to obtain a corrected airport emission inventory.

[0117] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0118] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A method for constructing an airport emission inventory, characterized in that: The airport emission inventory construction method comprises: Build an airport traffic operation simulation model for the airport surface and airspace, conduct simulation experiments for airport emissions, and obtain a multivariate data set related to airport emission parameters; A k-nearest neighbor mutual information feature selection method is used to select a subset of relevant variables of airport atmospheric emission image factors from the multivariate data set; Determine the attribution value of each target feature in the relevant variable subset based on the SHAP model interpretation method, and determine the correction coefficient of airport localized emissions according to the attribution value of each target feature; The airport emissions output by the airport traffic operation simulation model are corrected according to the correction coefficient to obtain a corrected airport emissions inventory.

2. The method for constructing an airport emission inventory according to claim 1, characterized in that: Each sample in the multivariate data set includes characteristic variables and target variables. The characteristic variables include atmospheric temperature, wind speed and direction, horizontal speed, control interval and weather phenomena. The target variables include the operating time of each stage of the landing and take-off cycle of each flight, fuel flow rate and fuel consumption.

3. The method for constructing an airport emission inventory according to claim 2, characterized in that: The k-nearest neighbor mutual information feature selection method is used to select a subset of relevant variables of airport atmospheric emission image factors from the multivariate data set, specifically including: Initialize the relevant variable subset to be empty, each feature variable constitutes a feature variable set, and each target variable constitutes a target variable set; Calculate the mutual information between each feature variable and the target variable set, and store the feature variable corresponding to the maximum value in each mutual information as the target feature in the relevant variable subset; Sorting the mutual information in descending order to obtain a mutual information ranking; Traverse the mutual information sorting to calculate the high-dimensional mutual information, and output the relevant variable subset: when traversing to the current traversal mutual information, calculate the mutual information of the current relevant variable subset, the characteristic variables corresponding to the current traversal mutual information and the target variable set, and record it as the current first mutual information; calculate the mutual information of the current relevant variable subset and the target variable set, and record it as the current second mutual information; if the current first mutual information is greater than the current second mutual information, store the characteristic variables corresponding to the current traversal mutual information as the target feature in the relevant variable subset.

4. The method for constructing an airport emission inventory according to claim 2, characterized in that: The formula for calculating the mutual information between each feature variable and the target variable set is: Among them, x i is the i-th feature variable in the feature variable set, Y is the target variable set, MI(x i ,Y) is x i The mutual information with Y, n is the total number of samples in the multivariate data set, ψ( ) is the double Γ function, α xi is the data set of the feature variable set and x i Distance less than ρ i / 2 the number of data points, α yi is the data set of the feature variable set and y i Distance less than ρ i / 2 the number of data points, ρ i For x i The distance to the kth nearest neighbor, k is a constant, 2≤k≤6.

5. The method for constructing an airport emission inventory according to claim 4, characterized in that: The formula for calculating the mutual information between the current subset of relevant variables, the feature variables corresponding to the current traversal mutual information and the target variable set, and the mutual information between the current subset of relevant variables and the target variable set is: Among them, MI(X1,X2,…,X m ) represents variable X1, variable X2, ..., variable X m The mutual information between xim Represents variable X m The value in the i-th sample, α xi1 represents the value of variable X1 in the i-th sample, α xi2 Represents the value of variable X2 in the i-th sample.

6. The method for constructing an airport emission inventory according to claim 1, characterized in that: The formula for determining the correction factor for airport localized emissions based on the attribution value of each target characteristic is: in, is the correction coefficient, θ is the air traffic control condition, ω is the meteorological condition, is the average attribution value of the lth target feature, f_vc a,j,l It is the fuel consumption output value of aircraft a performing flight phase j in the benchmark simulation model corresponding to the lth target feature.

7. The method for constructing an airport emission inventory according to claim 6, characterized in that: According to the correction coefficient, the airport emissions output by the airport traffic operation simulation model are corrected to obtain a corrected airport emissions inventory, which specifically includes: Among them, E' a,j,b is the corrected emission of pollutant b when aircraft a performs flight phase j, EI a,j,b is the emission of pollutant b when aircraft a performs flight phase j output by the airport traffic operation simulation model, t_vc a,j is the simulation running time corresponding to the flight phase j of aircraft a, n a Indicates the number of engines of aircraft a, FF a,j represents the fuel flow rate of aircraft a performing flight phase j.

8. The method for constructing an airport emission inventory according to claim 1, characterized in that: Build an airport traffic operation simulation model for the airport surface and airspace, conduct simulation experiments for airport emissions, and obtain a multivariate data set related to airport emission parameters, including: The whole airspace and airport fast-time simulation tools are used to build airport traffic operation simulation models of the airport scene and airspace, and simulation experiments for airport emissions are carried out to obtain a multivariate data set related to airport emission parameters.

9. The method for constructing an airport emission inventory according to claim 1, characterized in that: According to the correction coefficient, the input data of the airport traffic operation simulation model includes airport surface physical data, airport surface operation status, flight plan, airport approach control area static traffic capacity and airport approach control area dynamic operation status.

10. An airport emission inventory construction system, characterized in that: The airport emission inventory construction system comprises: The data set determination module is used to build an airport traffic operation simulation model for the airport surface and airspace, conduct simulation experiments for airport emissions, and obtain a multivariate data set related to airport emission parameters; A relevant variable subset determination module, used for selecting a relevant variable subset of the airport atmospheric emission image factor from the multivariate data set by using a k-nearest neighbor mutual information feature selection method; A correction coefficient determination module, used to determine the attribution value of each target feature in the relevant variable subset based on the SHAP model interpretation method, and determine the correction coefficient of the airport localized emissions according to the attribution value of each target feature; The airport emission inventory correction module is used to correct the airport emissions output by the airport traffic operation simulation model according to the correction coefficient to obtain a corrected airport emission inventory.

Images