Flight carbon emission prediction and evaluation method, system and device based on multi-source data and medium

Abstract

The application discloses a flight carbon emission prediction and evaluation method and system based on multi-source data, equipment and a medium, and the method comprises the following steps: constructing a multi-source factor sample data set; constructing a carbon emission prediction model comprising a spatial hierarchical clustering module and a carbon emission training and prediction module; the spatial hierarchical clustering module performs spatial unit clustering and importance screening of carbon emission factors in the spatial unit based on a flight path as a spatial attribute, carbon emission factors as variable attributes and a carbon emission rate as a target result, and forms a variable factor set in the spatial unit to input the carbon emission training and prediction module; the carbon emission training and prediction module performs carbon emission rate prediction training in the spatial unit based on the multi-source factor sample data set; and the carbon emission prediction model predicts the carbon emission rate based on research flight data. The application realizes full-process, multi-dimensional and high-precision carbon emission prediction and accounting of flights, and improves the carbon emission monitoring capability and emission reduction management level of the industry.

Description

Technical Field

[0001] This invention relates to the field of flight carbon emission assessment, and in particular to a method, system, equipment and medium for predicting and assessing flight carbon emissions based on multi-source data. Background Technology

[0002] In the current global context of advocating for low-carbon and environmentally friendly practices, the aviation transportation industry, as a significant source of carbon emissions, urgently needs to establish an efficient and accurate carbon emission assessment mechanism. Accurately assessing flight carbon emissions and effectively guiding flight emission reduction is a crucial issue that must be studied and prioritized. It is a vital foundation for assessing and diagnosing the emission reduction potential of civil aviation, achieving refined flight management, and promoting green flight. Traditional flight carbon emission assessments primarily rely on manual collection of flight data, item-by-item calculation, and aggregation. This is not only inefficient but also prone to data bias, failing to meet the demands of the ever-increasing number of flights and refined management requirements. Currently, internationally used methods for calculating flight carbon emissions are mainly divided into the Phase Takeoff and Landing (LTO) and Cruise Phase (CCD) phases. The LTO phase often uses the ICAO-published Aircraft Engine Emissions Database (EEDB) to estimate emissions based on fuel consumption and emission factors under standard operating conditions. However, in actual flight, due to factors such as pilot operation, airport layout, and traffic control, the engine thrust conditions during the LTO phase often deviate from the standard values, resulting in errors in carbon emission calculations. While the cruise phase accounts for over 80% of a flight's fuel consumption, the changes in temperature, humidity, pressure, and wind at high altitudes have a significant impact on engine performance and fuel consumption. Existing databases are largely based on ground-based standard environmental test results, which fail to reflect real-world flight conditions and carbon emissions. Furthermore, current assessment methods are mostly limited to single airports or single flights, lacking a unified assessment system for multi-airport, multi-flight operation scenarios. This makes horizontal comparison of carbon emission data difficult, hindering the improvement of the industry's overall carbon emission management level.

[0003] Chinese patent (application number 202110244033.0) discloses a method and system for rapidly calculating aviation carbon dioxide emissions. The method includes: obtaining aircraft type parameters corresponding to a specific flight from an airport in a specific city, based on real-time acquired flight information; calculating the total carbon dioxide emissions per flight based on these parameters; allocating the total carbon dioxide emissions per flight to the departure and arrival airports based on spatial allocation principles; and summarizing the total carbon dioxide emissions from departure and arrival airports for all flights within the city's airport to obtain the total aviation carbon dioxide emissions within that airport. Currently, methods for calculating aviation carbon dioxide emissions primarily rely on top-down calculations based on publicly available macroeconomic statistical data. For example, calculations are made based on statistics on oil consumption from statistical yearbooks, energy statistical yearbooks, and civil aviation statistical bulletins, combined with carbon dioxide emission factors per unit of oil consumption. The current methods for calculating aviation carbon emissions suffer from several drawbacks. First, they lack timeliness, fail to assess emissions from different aircraft types at a micro-level, and struggle to evaluate emissions differences across different operational phases. The aforementioned patent proposes a rapid, bottom-up method for calculating aviation carbon dioxide emissions, addressing issues such as poor timeliness in urban air transport emissions assessment, emissions differences between aircraft types, and spatial allocation of emissions across regional air transport. While this patent can objectively contribute to the effective assessment of flight carbon emissions, it still relies on a single-trip carbon dioxide emission factor to estimate emissions and allocates total flight emissions to departure and arrival airports using a calculation formula. This approach significantly deviates from actual flight emissions and requires extensive data collection on flight and aircraft emissions before calculation. Therefore, it fails to improve the precision of flight carbon emission assessment or achieve intelligent evaluation.

[0004] In summary, existing technologies do not provide sufficient explanation on how to achieve refined and intelligent assessment of flight carbon emissions. Due to the significant differences in flight operating conditions and the exceptionally complex changes in fuel flow rates of aircraft engines throughout the flight, simply estimating flight carbon emissions using emission factors under four thrust conditions from the engine emission database provided by the International Civil Aviation Organization (ICAO) is no longer sufficient to meet the requirements for civil aviation carbon emission verification and reduction. A more intelligent and refined method and system for assessing flight carbon emissions must be established. Summary of the Invention

[0005] The purpose of this invention is to provide a method, system, device and medium for predicting and assessing flight carbon emissions based on multi-source data, to construct a carbon emission prediction model, and to use multi-source factor sample datasets to perform spatial unit clustering and select carbon emission factors within units to form a variable factor set, and then to perform carbon emission rate prediction training within spatial units, thereby realizing intelligent accounting and dynamic assessment of flight carbon emissions in the whole process, in multiple dimensions and with high precision.

[0006] The objective of this invention is achieved through the following technical solution: A method for predicting and assessing flight carbon emissions based on multi-source data, the method comprising: S1. Construct a multi-source sample dataset for flights. Identify and extract carbon emission factors from the multi-source sample dataset for flights to construct a multi-source factor sample dataset. The carbon emission factor types in the multi-source factor sample dataset include aircraft type factors, engine performance and fuel factors, flight status factors, and meteorological environment factors. The multi-source factor sample dataset contains flight tracks and carbon emission rate label data. S2. Construct a carbon emission prediction model. The carbon emission prediction model includes a spatial hierarchical clustering module and a carbon emission training and prediction module. The spatial hierarchical clustering module uses flight tracks as spatial attributes, carbon emission factors as variable attributes, and carbon emission rate as the target result to cluster spatial units and screen the importance of carbon emission factors within spatial units to form a set of variable factors within the spatial unit, which is then input into the carbon emission training and prediction module. The carbon emission training and prediction module uses a multi-source factor sample dataset to train carbon emission rate prediction within the spatial unit. S3. Obtain multi-source flight data for the study flight and extract multi-source factor data from the multi-source flight data, inputting it into the carbon emission prediction model. The carbon emission prediction model extracts the spatial unit to which the flight track belongs and predicts the carbon emission rate within each spatial unit.

[0007] To better realize the present invention, the carbon emission rate is the ratio of the carbon emissions generated by a certain segment of the flight path of the aircraft to the fuel consumption; in method S3, the fuel consumption corresponding to the flight path is extracted from the multi-source data of the flight, and the carbon emissions corresponding to each flight path are calculated. Carbon emissions = fuel consumption × carbon emission rate. The carbon emissions of all flight paths in the study are added together to obtain the total carbon emissions of the flight.

[0008] Preferably, the flight multi-source sample dataset includes flight information, QAR time-series data, and meteorological environmental data. The aircraft type factor, engine performance and fuel factor, and flight status factor are identified and extracted from the flight information and QAR time-series data, while the meteorological environmental factor is identified and extracted from the meteorological environmental data. The aircraft type factor is the aircraft type information corresponding to the flight. The engine performance and fuel factor includes fuel flow, engine speed, combustion chamber pressure, and engine thrust. The flight status factor includes flight stage, flight altitude, latitude and longitude, flight attitude, flight speed, horizontal speed, vertical speed, and control surface deflection. The meteorological environmental factor includes wind direction, wind speed, temperature, and humidity. The carbon emission rate label data is obtained by measuring and calculating the carbon emission rate of the flight. The carbon emission factors in the multi-source factor sample dataset undergo spatiotemporal alignment processing.

[0009] Preferably, the carbon emission prediction model visualizes the carbon emission rate and carbon emission amount corresponding to each flight track in the GIS map according to the flight track of the study flight; the flight stages are divided by combining all flight tracks of the study flight with QAR time series data, and the flight stages include takeoff, initial climb, climb, cruise, initial descent, descent, approach and landing. The carbon emission amount of the study flight is accumulated according to the flight stage to obtain the carbon emission amount of each flight stage.

[0010] Preferably, the carbon emission prediction model defines a study area, which includes the airspace of take-off and landing airports. All flights and their flight tracks within the study area are obtained during the study period. The carbon emissions corresponding to all flight tracks within the study area during the study period are summed to obtain the total regional carbon emissions of the study area during the study period.

[0011] Preferably, the method for obtaining the set of variable factors within the spatial unit is as follows: using carbon emission factors as variable factors, calculating the Pearson correlation coefficient for all variable factors to obtain a correlation coefficient matrix, converting the correlation in the correlation coefficient matrix into distance for clustering and constructing a distance matrix, using the Ward hierarchical clustering algorithm and employing the Ward minimum variance method to perform hierarchical clustering processing to generate a clustering tree and form several clusters, and performing carbon emission factor removal processing, using a random forest model to sort the variable factors according to the importance of the carbon emission rate, and selecting the set of variable factors within the spatial unit according to preset parameters.

[0012] Preferably, the carbon emission training and prediction module obtains the spatial units and variable factor sets within the spatial units input by the spatial hierarchical clustering module, and the carbon emission training and prediction module extracts multi-source factor sample datasets and performs carbon emission rate prediction training on sample data within the spatial units.

[0013] A flight carbon emission prediction and assessment system based on multi-source data includes a sample database processing module, a data acquisition module, and a carbon emission prediction model. The sample database processing module includes a multi-source factor sample dataset and a multi-source factor sample dataset. The module identifies and extracts carbon emission factors from the flight multi-source sample dataset to construct the multi-source factor sample dataset. The carbon emission factor types in the multi-source factor sample dataset include aircraft type factors, engine performance and fuel factors, flight status factors, and meteorological environment factors. The multi-source factor sample dataset contains flight tracks and carbon emission rate label data. The carbon emission prediction model may include a GIS map and includes spatial... The system comprises a hierarchical clustering module and a carbon emission training and prediction module. The spatial hierarchical clustering module uses flight tracks as spatial attributes, carbon emission factors as variable attributes, and carbon emission rate as the target result to cluster spatial units and screen the importance of carbon emission factors within spatial units to form a set of variable factors within the spatial units, which is then input into the carbon emission training and prediction module. The carbon emission training and prediction module uses a multi-source factor sample dataset within the spatial units to train carbon emission rate prediction. The data acquisition module acquires multi-source flight data of the research flights and extracts multi-source factor data from the flight multi-source data, which is then input into the carbon emission prediction model. The carbon emission prediction model extracts the spatial units to which the flight tracks of the research flights belong and predicts the carbon emission rate within each belonging spatial unit.

[0014] An electronic device includes at least one processor, at least one memory, and a data bus; wherein: the processor and the memory communicate with each other via the data bus; the memory stores program instructions that are executed by the processor, and the processor calls the program instructions to perform the steps of implementing the flight carbon emission prediction and assessment method of the present invention.

[0015] A storage medium includes a memory and a processor, the memory storing an executable program, and the processor executing the executable program to implement the steps of the flight carbon emission prediction and assessment method of the present invention.

[0016] Compared with the prior art, the present invention has the following advantages and beneficial effects: (1) This invention identifies and extracts carbon emission factors from multi-source sample datasets of flights to construct a multi-source factor sample dataset, and constructs a carbon emission prediction model. The carbon emission prediction model uses the multi-source factor sample dataset to perform spatial unit clustering and select carbon emission factors within the unit to form a variable factor set, and then performs carbon emission rate prediction training within the spatial unit. This realizes intelligent accounting and dynamic assessment of carbon emissions throughout the entire process of flights or airports, multi-dimensional and high-precision, improves the industry's ability to monitor flight carbon emissions and the level of flight emission reduction management, and solves the problem that existing methods are difficult to adapt to complex flight conditions and changes in the high-altitude environment.

[0017] (2) This invention enables refined calculation of carbon emissions throughout the entire flight process, significantly improving the accuracy of carbon emission assessment and accurately reflecting the carbon emission situation at each stage of the flight. Through the multi-airport and multi-flight data integration assessment mechanism, the carbon emission situation of different airports can be uniformly calculated and compared, improving the systematicness and consistency of industry carbon emission monitoring. Through carbon emission statistical analysis and visualization processing, users can intuitively understand the total carbon emissions and efficiency of each airport, thereby reducing manual intervention and data processing costs, improving assessment efficiency and data availability, and thus improving the level of industry carbon emission monitoring and management. Attached Figure Description

[0018] Figure 1 This is a flowchart of the flight carbon emission prediction and assessment method of the present invention; Figure 2 This is a correlation diagram of a portion of the variable factors extracted from the correlation coefficient matrix in the example; Figure 3 This is a correlation network diagram of a portion of the variable factors extracted from the correlation coefficient matrix in the example; Figure 4 This is a ranking of the importance of some variable factors on carbon emission rates from high to low in the examples. Figure 5 The following is an example of the correlation diagram between some factors in the engine performance fuel factor and the carbon emission rate. Figure 6 This is a comparison chart showing the importance of the impact of selected variable factors within a certain cluster on carbon emission rates in the embodiments. Figure 7 This is a graph showing the relationship between engine thrust and fuel consumption at different altitudes in the embodiment. Figure 8 This is a graph showing the relationship between aircraft vacuum speed and fuel consumption at different altitudes in the embodiment. Figure 9 This is a software interface diagram of the flight carbon emission prediction and assessment system in the embodiment; Figure 10 This is a schematic diagram illustrating engine fuel consumption and carbon emission data for a certain aircraft model during the cruise phase, as shown in the example. Figure 11 This is a schematic diagram illustrating the carbon emissions of a flight at different altitudes and speeds, as shown in the example. Figure 12 This is a schematic diagram illustrating the comparison of carbon emission rates of some airports as an example in the embodiments. Detailed Implementation

[0019] The present invention will be further described in detail below with reference to embodiments: Example

[0020] like Figure 1As shown, a method for predicting and assessing flight carbon emissions based on multi-source data includes the following steps: S1. Construct a multi-source flight sample dataset. The multi-source data sources for this dataset include flight information, QAR time-series data, and meteorological environmental data. Establish automated data quality control rules to identify and remove outliers or jumps caused by sensor malfunctions or packet loss. For missing short-series data, use linear interpolation or spline interpolation based on flight status (e.g., climb, cruise, descent) to fill in the gaps. Extract carbon emission factors from the multi-source flight sample dataset to construct a multi-source factor sample dataset. The carbon emission factor types in this dataset include aircraft type factors, engine performance and fuel factors, flight status factors, and meteorological environmental factors. The multi-source factor sample dataset includes flight tracks and carbon emission rate label data. The carbon emission rate is the ratio of carbon emissions generated by a specific segment of the aircraft's flight track to fuel consumption. Aircraft type factors, engine performance and fuel factors, and flight status factors are identified and extracted from flight information and QAR time-series data. Meteorological environment factors are identified and extracted from meteorological environment data. The aircraft type factor is the aircraft type information corresponding to the flight. The engine performance and fuel factors include fuel flow, engine speed, combustion chamber pressure, engine thrust (i.e., Mach number), and engine exhaust temperature. The flight status factors include flight stage, flight altitude, latitude and longitude, flight attitude, flight speed, horizontal speed, vertical speed, and control surface deflection. The meteorological environment factors include wind direction, wind speed, temperature, and humidity. The carbon emission rate label data is obtained by measuring the carbon emissions of the flight and calculating the carbon emission rate. The carbon emission factors and carbon emission rate label data of the multi-source factor sample dataset are spatiotemporally aligned according to the flight trajectory. The spatial location and time of the data at the track point or track segment are spatiotemporally aligned at the second level to construct a high-precision multi-source factor sample dataset after spatiotemporal alignment. This dataset serves as the data foundation for subsequent carbon emission prediction model training, laying the data foundation for high-precision and high-timeliness carbon emission assessment.

[0021] S2. Construct a carbon emission prediction model. The carbon emission prediction model includes a spatial hierarchical clustering module and a carbon emission training and prediction module. The spatial hierarchical clustering module uses flight tracks as spatial attributes, carbon emission factors as variable attributes, and carbon emission rates as the target result to cluster spatial units and screen the importance of carbon emission factors within spatial units to form a set of variable factors within the spatial unit, which is then input into the carbon emission training and prediction module. The preferred method for obtaining the set of variable factors within spatial units in this invention is as follows: Carbon emission factors are used as variable factors (in the importance screening of carbon emission factors, carbon emission factors are called variable factors). Pearson correlation coefficients are calculated for all variable factors to obtain a correlation coefficient matrix. In this embodiment, the correlation relationships of a portion of the variable factors are extracted from the obtained correlation coefficient matrix as follows: Figure 2As shown, the importance of some variable factors on carbon emission rates is ranked as follows: Figure 4 As shown. Preferably, after constructing the correlation coefficient matrix, a correlation threshold can be set for correlation screening. This embodiment uses the correlation between some factors in engine performance fuel factors and carbon emission rates as an example to screen engine performance fuel factors whose correlation with carbon emission rates is higher than the correlation threshold, such as... Figure 5 As shown, the correlation coefficient matrix is ​​transformed into distances for clustering, and a distance matrix is ​​constructed. A partial subset of variable factors is extracted from the correlation coefficient matrix to construct the correlation distance matrix network, as shown below. Figure 3 As shown, the Ward hierarchical clustering algorithm and Ward minimum variance method are used to generate cluster trees and form several clusters. Carbon emission factors are then removed. Specifically, the Ward hierarchical clustering algorithm is used to perform collinearity analysis of carbon emission factors in the carbon emission rate, and acceptable carbon emission factors are selected (unacceptable carbon emission factors are removed). Taking engine performance fuel factor as an example, after intra-cluster screening, the importance of the remaining variable factors on the carbon emission rate is ranked as follows: Figure 6 As shown, a random forest model was used to rank variables according to their importance to carbon emission rates, and a set of variables within a spatial cell was selected based on preset parameters. Within this set of variables, each variable has both direct and indirect effects on carbon emission rates. For example, engine thrust affects fuel consumption, which in turn affects carbon emission rates. The relationship between engine thrust and fuel consumption at different altitudes is shown in the figure. Figure 7 As shown; aircraft vacuum speed affects fuel consumption and thus carbon emission rate, with the relationship between aircraft vacuum speed and fuel consumption at different altitudes as follows: Figure 8 As shown.

[0022] The carbon emission training and prediction module obtains the spatial cells and variable factor sets within the spatial cells from the spatial hierarchical clustering module. It then extracts multi-source factor sample datasets and uses these samples within the spatial cells to train carbon emission rate prediction. The carbon emission training and prediction module uses the multi-source factor sample dataset within the spatial cells to train carbon emission rate prediction.

[0023] S3. Acquire multi-source flight data for the research flights and extract multi-source factor data from the flight multi-source data, inputting it into the carbon emission prediction model. The carbon emission prediction model extracts the spatial unit to which the flight track belongs (extracting and identifying the spatial unit to which the flight track belongs) and predicts the carbon emission rate within each spatial unit. This invention significantly improves the accuracy and reliability of flight carbon emission data in the spatial dimension, enabling carbon emission analysis to be deepened from the flight level to the route segment level. In this embodiment, the multi-source factor sample dataset is divided into a training set, a validation set, and a test set. The training set is used for training the carbon emission prediction model, the validation set is used for model tuning, and the carbon emission prediction model is evaluated using the test set. Its coefficient of determination R² reaches over 0.85, and the prediction accuracy meets the needs of practical applications, exhibiting advantages such as strong robustness, strong generalization ability, and strong prediction ability.

[0024] In some embodiments, this invention extracts fuel consumption corresponding to flight tracks from multi-source flight data, calculates the carbon emissions corresponding to each track, where carbon emissions = fuel consumption × carbon emission rate, and sums the carbon emissions of all tracks in the study to obtain the total carbon emissions of the flight. This invention supports multi-dimensional filtering of flights by aircraft type, route, time period, etc., and generates box plots of total carbon emissions and carbon emission rates or performs ranking comparisons. For high-carbon-emission flights, it generates carbon emission diagnostic reports, pointing out possible causes. This invention transforms innovative assessment methods into stable, efficient, and easy-to-use systematic solutions, realizing multi-level, intelligent carbon emission assessment and management from single-flight analysis to fleet, airport, route, and even global route network.

[0025] In some embodiments, the carbon emission prediction model visualizes the carbon emission rate and amount corresponding to each flight path on a GIS map. It divides the flight into stages by combining all flight paths with QAR time-series data. Flight stages include takeoff, initial climb, climb, cruise, initial descent, descent, approach, and landing. The carbon emissions of the study flight are accumulated according to each flight stage to obtain the carbon emissions for each stage. For example, the flight altitude for the initial climb can be identified and divided according to the rule of "35 feet < altitude < 1500 feet, climb rate > 500 feet / minute". This embodiment can introduce a lightweight time-series classification model based on LSTM or CNN to accurately identify and divide the flight stages of takeoff, initial climb, climb, cruise, initial descent, descent, approach, and landing. This invention can also construct vertical profile indicators, time profile indicators, and horizontal comparative analysis indicators for carbon emission rates. Examples of vertical profile indicators for carbon emission rates include: "Carbon emission rate from 30,000 feet to landing," defined as: (Total carbon emissions during this phase) / (Horizontal flight distance from 30,000 feet to landing). Similar carbon emission rate indicators are defined for 20,000 feet and 10,000 feet to landing. These indicators can effectively assess the optimization level of arrival management (CDA) and descent profiles. Examples of time profile indicators for carbon emission rates include: "Carbon emission intensity in the last half hour of a flight," defined as (Total carbon emissions in the last half hour) / (Flight distance in the last half hour), used to assess terminal area operational efficiency. Examples of horizontal comparative analysis indicators are as follows: A green flight benchmark profile is established for the same aircraft type, route, and similar conditions. The percentage deviation of carbon emissions at each stage of the flight from the benchmark value (which can be the historical median) is calculated, ultimately yielding the horizontal and vertical carbon emission rate indicators for the flight at different stages. This decomposes the macroscopic total carbon emissions into micro-indicators directly related to specific operational actions, air traffic control instructions, and pilot operations, providing an actionable diagnostic tool for precise emission reduction. During evaluation, this invention automatically associates flight data with the characteristic profiles of takeoff and landing airports. When calculating efficiency indicators (such as departure carbon emission rate), it automatically filters out the influence of inherent airport characteristics, making the carbon emission performance of flights between different airports comparable. This embodiment uses the cruise phase of a certain aircraft type as an example. The engine fuel consumption and carbon emission data statistics for this aircraft type during the cruise phase are as follows: Figure 10 As shown. This invention can select several combinations of carbon emission factors from a multi-source factor sample dataset for flight carbon emission analysis. For example, this embodiment takes a flight at different altitudes and speeds as an example, where altitude and speed are combinations of carbon emission factors. The carbon emissions of this flight at different altitudes and speeds are as follows: Figure 11 As shown.

[0026] In some embodiments, a study area is defined in the carbon emission prediction model. This study area includes the airspace of takeoff and landing airports (attributed as airport area). All flights and their flight tracks within the study area are acquired during the study period. The carbon emissions corresponding to all flight tracks within the study area during the study period are summed to obtain the total regional carbon emissions for the study area during the study period. The carbon emission prediction model of this invention incorporates a GIS map, enabling comparative display of carbon emission assessments for different airports and dynamic display of flight route carbon emissions. This embodiment uses a selection of airports as examples and statistically compares the carbon emission rates of these example airports. Figure 12 As shown.

[0027] A flight carbon emission prediction and assessment system based on multi-source data includes a sample database processing module, a data acquisition module, and a carbon emission prediction model. The sample database processing module includes a multi-source factor sample dataset and a multi-source factor sample dataset. The module identifies and extracts carbon emission factors from the flight multi-source sample dataset to construct the multi-source factor sample dataset. The carbon emission factor types in the multi-source factor sample dataset include aircraft type factors, engine performance and fuel factors, flight status factors, and meteorological environment factors. The multi-source factor sample dataset contains flight tracks and carbon emission rate label data. The carbon emission prediction model includes a spatial hierarchical clustering module and a carbon emission training module. The carbon emission prediction module and the spatial hierarchical clustering module use flight tracks as spatial attributes, carbon emission factors as variable attributes, and carbon emission rates as the target result to cluster spatial units and screen the importance of carbon emission factors within spatial units to form a set of variable factors within the spatial unit, which is then input into the carbon emission training prediction module. The carbon emission training prediction module trains carbon emission rate prediction within the spatial units using a multi-source factor sample dataset. The data acquisition module acquires multi-source flight data for the research flights and extracts multi-source factor data from the flight multi-source data, which is then input into the carbon emission prediction model. The carbon emission prediction model extracts the spatial unit to which the flight track belongs and predicts the carbon emission rate within each belonging spatial unit. The software interface of the flight carbon emission prediction and assessment system (also known as the flight carbon emission intelligent assessment system) of this invention is as follows: Figure 9 As shown, it provides a graphical interface that visualizes carbon emission data in map form, supports comparative analysis of carbon emissions from arriving and departing flights at different airports worldwide, and offers emission reduction strategy suggestions. Users can interactively select specific airports, routes, or time ranges to view detailed carbon emission data and trend analysis.

[0028] An electronic device includes at least one processor, at least one memory, and a data bus; wherein: the processor and the memory communicate with each other via the data bus; the memory stores program instructions that are executed by the processor, and the processor calls the program instructions to execute the steps of the flight carbon emission prediction and assessment method of the present invention.

[0029] A storage medium includes a memory and a processor. The memory stores an executable program, and the processor executes the executable program to implement the steps of the flight carbon emission prediction and assessment method of the present invention.

[0030] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for predicting and assessing flight carbon emissions based on multi-source data, characterized in that: The methods include: S1. Construct a multi-source sample dataset for flights. Identify and extract carbon emission factors from the multi-source sample dataset for flights to construct a multi-source factor sample dataset. The carbon emission factor types in the multi-source factor sample dataset include aircraft type factors, engine performance and fuel factors, flight status factors, and meteorological environment factors. The multi-source factor sample dataset contains flight tracks and carbon emission rate label data. S2. Construct a carbon emission prediction model. The carbon emission prediction model includes a spatial hierarchical clustering module and a carbon emission training and prediction module. The spatial hierarchical clustering module uses flight tracks as spatial attributes, carbon emission factors as variable attributes, and carbon emission rate as the target result to cluster spatial units and screen the importance of carbon emission factors within spatial units to form a set of variable factors within the spatial unit, which is then input into the carbon emission training and prediction module. The carbon emission training and prediction module uses a multi-source factor sample dataset to train carbon emission rate prediction within the spatial unit. S3. Obtain multi-source flight data for the study flight and extract multi-source factor data from the multi-source flight data, inputting it into the carbon emission prediction model. The carbon emission prediction model extracts the spatial unit to which the flight track belongs and predicts the carbon emission rate within each spatial unit.

2. The flight carbon emission prediction and assessment method based on multi-source data according to claim 1, characterized in that: The carbon emission rate is the ratio of carbon emissions generated by a certain segment of the flight path to fuel consumption. In method S3, the fuel consumption corresponding to the flight path is extracted from the multi-source data of the flight, and the carbon emission corresponding to each flight path is calculated. Carbon emission = fuel consumption × carbon emission rate. The total carbon emission of the flight is obtained by adding up the carbon emission of all flight paths in the study.

3. The flight carbon emission prediction and assessment method based on multi-source data according to claim 1, characterized in that: The flight multi-source sample dataset comprises flight information, QAR time-series data, and meteorological environmental data. Aircraft type factors, engine performance and fuel factors, and flight status factors are extracted from flight information and QAR time-series data, while meteorological environmental factors are extracted from meteorological environmental data. Aircraft type factors are the aircraft type information corresponding to the flight. Engine performance and fuel factors include fuel flow, engine speed, combustion chamber pressure, and engine thrust. Flight status factors include flight stage, flight altitude, latitude and longitude, flight attitude, flight speed, horizontal speed, vertical speed, and control surface deflection. Meteorological environmental factors include wind direction, wind speed, temperature, and humidity. The carbon emission rate label data is obtained by measuring and calculating the carbon emission rate of the flight. The carbon emission factors in the multi-source factor sample dataset undergo spatiotemporal alignment processing.

4. The flight carbon emission prediction and assessment method based on multi-source data according to claim 2, characterized in that: The carbon emission prediction model visualizes the carbon emission rate and carbon emission amount corresponding to each flight track in the GIS map according to the flight track of the study flight. The flight stages are divided by combining all the flight tracks of the study flight with QAR time series data. The flight stages include takeoff, initial climb, climb, cruise, initial descent, descent, approach and landing. The carbon emission amount of the study flight is accumulated according to the flight stage to obtain the carbon emission amount of each flight stage.

5. The flight carbon emission prediction and assessment method based on multi-source data according to claim 2 or 3, characterized in that: The carbon emission prediction model defines a study area, which includes the airspace of take-off and landing airports. All flights and their flight tracks within the study area are acquired during the study period. The carbon emissions corresponding to all flight tracks within the study area during the study period are summed to obtain the total regional carbon emissions of the study area during the study period.

6. The flight carbon emission prediction and assessment method based on multi-source data according to claim 1, characterized in that: The method for obtaining the set of variable factors within the spatial unit is as follows: using carbon emission factors as variable factors, calculating the Pearson correlation coefficient for all variable factors to obtain a correlation coefficient matrix, converting the correlation in the correlation coefficient matrix into distance for clustering and constructing a distance matrix, using the Ward hierarchical clustering algorithm and the Ward minimum variance method to perform hierarchical clustering to generate a clustering tree and form several clusters, and performing carbon emission factor removal processing, using the random forest model to sort the variable factors according to the importance of the carbon emission rate, and selecting the set of variable factors within the spatial unit according to preset parameters.

7. The flight carbon emission prediction and assessment method based on multi-source data according to claim 1 or 6, characterized in that: The carbon emission training and prediction module obtains the spatial units and variable factor sets within the spatial units from the spatial hierarchical clustering module. The carbon emission training and prediction module extracts multi-source factor sample datasets and performs carbon emission rate prediction training on the sample data within the spatial units.

8. A flight carbon emission prediction and assessment system based on multi-source data, characterized in that: The system includes a sample database processing module, a data acquisition module, and a carbon emission prediction model. The sample database processing module comprises a multi-source factor sample dataset and a multi-source factor sample dataset. The sample database processing module identifies and extracts carbon emission factors from the flight multi-source sample dataset to construct the multi-source factor sample dataset. The carbon emission factor types in the multi-source factor sample dataset include aircraft type factors, engine performance and fuel factors, flight status factors, and meteorological environment factors. The multi-source factor sample dataset contains flight tracks and carbon emission rate label data. The carbon emission prediction model may include a GIS map and includes a spatial hierarchical clustering module and a carbon emission... The training and prediction module and the spatial hierarchical clustering module perform spatial unit clustering with flight tracks as spatial attributes, carbon emission factors as variable attributes, and carbon emission rate as the target result. They also screen the importance of carbon emission factors within spatial units to form a set of variable factors within the spatial units, which is then input into the carbon emission training and prediction module. The carbon emission training and prediction module uses a multi-source factor sample dataset to train carbon emission rate prediction within the spatial units. The data acquisition module acquires multi-source flight data of the research flights and extracts multi-source factor data from the flight multi-source data, which is then input into the carbon emission prediction model. The carbon emission prediction model extracts the spatial unit to which the flight track belongs and predicts the carbon emission rate within each belonging spatial unit.

9. An electronic device, characterized in that: It includes at least one processor, at least one memory, and a data bus; wherein: the processor and the memory communicate with each other through the data bus; the memory stores program instructions that are executed by the processor, and the processor calls the program instructions to perform the steps of implementing the flight carbon emission prediction and assessment method according to any one of claims 1 to 7.

10. A storage medium comprising a memory and a processor, wherein the memory stores an executable program, characterized in that... When the processor executes the executable program, it implements the steps of the flight carbon emission prediction and assessment method according to any one of claims 1 to 7.

Images