Aircraft tail floor mopping risk assessment method

Abstract

The embodiment of the invention provides an aircraft tail floor mopping risk assessment method, and relates to the technical field of aviation safety. The method comprises the steps of obtaining flight data of a target flight; extracting risk assessment features from the flight data; the risk assessment features comprise flight state features, operation behavior features and time sequence analysis features; the time sequence analysis feature is used for representing a change trend or a dynamic sequence of one or more flight parameters in a preset flight stage; inputting the risk assessment features into a risk assessment model to obtain a predicted grounding pitch angle output by the risk assessment model; the risk assessment model can represent a linear relation and a non-linear relation between the risk assessment characteristics and the grounding pitch angle; and determining tail floor mopping risk information of the target flight based on the predicted grounding pitch angle. The method and the device can be used in flight quality monitoring and pilot training evaluation processes, and are used for solving the problems of dependence on artificial experience and low evaluation accuracy in the prior art.

Description

Technical Field

[0001] This application relates to the field of aviation safety technology, and in particular to a method for assessing the risk of aircraft tail scraping the ground. Background Technology

[0002] In aviation, tail scrambling during landing refers to the phenomenon where an aircraft's tail contacts the runway surface during landing due to excessive pitch angle, landing gear compression, or other factors. This is a common aviation safety hazard that can lead to structural damage to the aircraft, increased maintenance costs, and even more serious flight accidents. To mitigate such risks, existing technologies primarily rely on pilots' operational experience, simulation training, and post-analysis based on limited flight data, resulting in low levels of automation and inaccurate assessment results. Summary of the Invention

[0003] The purpose of this application is to provide a method for assessing the risk of aircraft tail-to-ground contact, aiming to improve the accuracy of such assessments.

[0004] Firstly, a method for assessing the risk of aircraft tail-to-ground contact is provided. The method includes: acquiring flight data of a target flight; extracting risk assessment features from the flight data; the risk assessment features include flight state features, operational behavior features, and time-series analysis features; the time-series analysis features are used to characterize the changing trend or dynamic sequence of one or more flight parameters within a preset flight phase; inputting the risk assessment features into a risk assessment model to obtain the predicted ground contact pitch angle output by the risk assessment model; the risk assessment model can characterize the linear and nonlinear relationships between the risk assessment features and the ground contact pitch angle; and determining the tail-to-ground contact risk information of the target flight based on the predicted ground contact pitch angle.

[0005] The tail-to-ground strike risk assessment method provided in this application first acquires flight data of the target flight and accurately extracts risk assessment features covering flight state characteristics, operational behavior characteristics, and time-series analysis characteristics. The time-series analysis features effectively characterize the changing trends or dynamic sequences of flight parameters within a preset flight phase, providing comprehensive and dynamic data support for risk assessment. Next, the risk assessment features are input into a risk assessment model that can simultaneously characterize both linear and nonlinear relationships between the features and the ground-to-ground pitch angle. This ensures the model accurately captures the correlation between multiple features and the ground-to-ground pitch angle, thereby outputting an accurate predicted ground-to-ground pitch angle. Finally, based on the predicted ground-to-ground pitch angle, the tail-to-ground strike risk information of the target flight is determined, achieving accurate assessment of the tail-to-ground strike risk. This provides a reliable basis for flight operation safety warnings and operational optimization, effectively reducing the probability of tail-to-ground strike accidents and improving flight safety levels.

[0006] In conjunction with the first aspect mentioned above, in one possible implementation, the risk assessment features are input into a risk assessment model to obtain a predicted ground elevation angle, including: inputting the risk assessment features into a first risk assessment sub-model to obtain a predicted first ground elevation angle; the first risk assessment sub-model is used to characterize the linear relationship between the risk assessment features and the ground elevation angle; inputting the risk assessment features into a second risk assessment sub-model to obtain a predicted second ground elevation angle; the second risk assessment sub-model is used to characterize the nonlinear relationship between the risk assessment features and the ground elevation angle; and determining the predicted ground elevation angle based on the predicted first ground elevation angle and the predicted second ground elevation angle.

[0007] In conjunction with the first aspect mentioned above, in one possible implementation, the risk assessment features are input into the first risk assessment sub-model to obtain the predicted value of the first ground elevation angle, including: converting the continuous numerical features in the risk assessment features to a preset numerical range to obtain the first risk assessment sub-feature; encoding the discrete state features in the risk assessment features to obtain the second risk assessment sub-feature; and inputting the first risk assessment sub-feature and the second risk assessment sub-feature into the first risk assessment sub-model to obtain the predicted value of the first ground elevation angle output by the first risk assessment sub-model.

[0008] In conjunction with the first aspect mentioned above, in one possible implementation, the first risk assessment sub-model is constructed in the following way: from multiple risk assessment features, a subset of features whose contribution to the predicted grounding pitch angle reaches a first preset threshold is selected; based on the feature subset, the weight coefficient corresponding to each feature is determined by linear regression to generate the first risk assessment sub-model.

[0009] In conjunction with the first aspect mentioned above, in one possible implementation, the second risk assessment sub-model is a prediction model based on an integrated decision tree; inputting risk assessment features into the second risk assessment sub-model to obtain the predicted value of the second ground elevation angle includes: inputting risk assessment features into the prediction model based on an integrated decision tree; judging the input features through multiple decision trees respectively, summarizing the output results of each decision tree, and obtaining the predicted value of the second ground elevation angle.

[0010] In conjunction with the first aspect mentioned above, in one possible implementation, the second risk assessment sub-model is trained as follows: multiple training subsets are extracted from historical flight data in chronological order; for each training subset, a primary prediction model is trained based on the training subset, and the performance of the primary prediction model is verified using the next training subset following the training subset; the second risk assessment sub-model is determined based on the performance evaluation results of all primary prediction models.

[0011] In conjunction with the first aspect above, in one possible implementation, after obtaining the predicted second ground pitch angle, the method further includes the following steps: identifying key risk assessment features that have a high degree of influence on the predicted second ground pitch angle from the risk assessment features; identifying a combination of risk assessment features that have a synergistic effect on the predicted second ground pitch angle from the risk assessment features; and determining the tail-scratching risk information of the target flight, including generating a risk report containing at least one of the predicted ground pitch angle, key risk assessment features, and combinations of risk assessment features.

[0012] In conjunction with the first aspect above, in one possible implementation, key risk assessment features are determined from the risk assessment features, including: performing multiple simulation predictions for the second risk assessment sub-model, masking one or more risk assessment features in each simulation prediction; quantifying the contribution of each risk assessment feature to the predicted value of the second ground pitch angle based on the fluctuations in the results of the multiple simulation predictions; and identifying risk assessment features with a contribution higher than a first preset threshold as key risk assessment features.

[0013] In conjunction with the first aspect mentioned above, in one possible implementation, the risk assessment feature combination is determined from the risk assessment features, including: quantifying the synergistic influence of any two risk assessment features on the predicted value of the second ground pitch angle through simulation prediction; and determining the feature pairs with a synergistic influence higher than a second preset threshold as the risk assessment feature combination that produces the synergistic influence.

[0014] In conjunction with the first aspect mentioned above, in one possible implementation, the tail-to-ground risk information of the target flight is determined based on the predicted ground pitch angle, including: determining the corresponding maximum permissible pitch angle and safety margin according to the aircraft type of the target flight; calculating the safety margin between the maximum permissible pitch angle and the predicted ground pitch angle; if the safety margin is greater than the preset safety margin, it is determined to be low risk; if the safety margin is less than or equal to the safety margin but greater than zero, it is determined to be medium risk; if the safety margin is less than or equal to zero, it is determined to be high risk.

[0015] Secondly, this application provides an aircraft tail-to-ground friction risk assessment device, comprising: a data acquisition module, a feature extraction module, a risk assessment module, and an output module; the data acquisition module is used to acquire flight data of a target flight; the feature extraction module is used to extract risk assessment features from the flight data; the risk assessment features include flight state features, operational behavior features, and time-series analysis features; the time-series analysis features are used to characterize the changing trend or dynamic sequence of one or more flight parameters within a preset flight phase; the risk assessment module is used to input the risk assessment features into a risk assessment model to obtain the predicted ground pitch angle output by the risk assessment model; the risk assessment model can characterize the linear and nonlinear relationships between the risk assessment features and the ground pitch angle; the output module is used to determine the tail-to-ground friction risk information of the target flight based on the predicted ground pitch angle.

[0016] Thirdly, this application provides an electronic device comprising: a processor and a memory; the memory storing processor-executable instructions; when the processor is configured to execute the instructions, the electronic device implements the method of the first aspect described above.

[0017] Fourthly, this application provides a computer-readable storage medium comprising: computer software instructions; which, when executed in an electronic device, cause the electronic device to implement the method described in the first aspect.

[0018] Fifthly, this application provides a computer program product that, when run on a computer, causes the computer to perform the steps of the relevant method described in the first aspect above, so as to implement the method of the first aspect above.

[0019] The beneficial effects of the second to fifth aspects mentioned above can be referred to the corresponding description of the first aspect, and will not be repeated here. Attached Figure Description

[0020] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 A flowchart illustrating a method for assessing the risk of aircraft tail scraping the ground, provided in an embodiment of this application; Figure 2 A flowchart illustrating a method for calculating and predicting ground elevation angle provided in an embodiment of this application; Figure 3A flowchart illustrating another method for assessing the risk of aircraft tail-end ground scratching provided in this application embodiment; Figure 4 A schematic diagram of the composition of an aircraft tail-side ground-scraping risk assessment device provided in this application embodiment; Figure 5 This is a schematic diagram of the structure of an aircraft tail-side ground-scraping risk assessment device provided in an embodiment of this application. Detailed Implementation

[0022] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0023] It should be noted that in the embodiments of this application, the words "exemplarily" or "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design scheme described as "exemplarily" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design schemes. Specifically, the use of the words "exemplarily" or "for example" is intended to present the relevant concepts in a specific manner.

[0024] In the embodiments of this application, the terms "first," "second," "third," "fourth," "fifth," and "sixth" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined with "first," "second," "third," "fourth," "fifth," and "sixth" may explicitly or implicitly include one or more of that feature.

[0025] In embodiments of this application, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0026] "A and / or B" includes the following three combinations: A only, B only, and a combination of A and B.

[0027] As described in the background section, in the aviation field, tail scrambling during landing refers to the phenomenon where an aircraft's tail contacts the runway surface during landing due to excessive pitch angle, landing gear compression, or other factors. This is a common aviation safety hazard that can lead to structural damage to the aircraft, increased maintenance costs, and even more serious flight accidents. To mitigate such risks, existing technologies primarily rely on pilots' operational experience, simulation training, and post-analysis based on limited flight data, resulting in low levels of automation and inaccurate assessment results.

[0028] Based on this, this application provides a method for assessing the risk of aircraft tail strike. First, flight data of the target flight is acquired, and risk assessment features covering flight state characteristics, operational behavior characteristics, and time-series analysis characteristics are accurately extracted. The time-series analysis features effectively characterize the changing trends or dynamic sequences of flight parameters within a preset flight phase, providing comprehensive and dynamic data support for risk assessment. Next, the risk assessment features are input into a risk assessment model that can simultaneously characterize both linear and nonlinear relationships between the features and the ground pitch angle, ensuring that the model accurately captures the correlation between multiple features and the ground pitch angle, thereby outputting an accurate predicted ground pitch angle. Finally, the tail strike risk information of the target flight is determined based on the predicted ground pitch angle, achieving accurate assessment of the tail strike risk. This provides a reliable basis for flight operation safety warnings and operational optimization, effectively reducing the probability of aircraft tail strike accidents and improving flight safety levels.

[0029] The aircraft tail strike risk assessment method provided in this application embodiment can be applied to an aviation safety assessment system with computing and data storage capabilities. This application embodiment does not impose any restrictions on the specific implementation form of the aviation safety assessment system.

[0030] For example, an aviation safety assessment system can be deployed on a cloud server in the aviation industry. This cloud server can be a single centralized server, or a server cluster consisting of multiple sub-servers. In some embodiments, the server cluster can also be a distributed cluster where edge clouds and central clouds collaborate, enabling localized collection, processing, and centralized aggregation and analysis of flight data. This application does not impose any limitations on this approach.

[0031] For example, the aviation safety assessment system can also be directly deployed on airborne terminal devices. These airborne terminal devices can be referred to as: airborne flight data processing terminals, flight parameter monitoring devices, airborne safety assessment systems, etc. Specifically, they can be flight data recorder processing modules with real-time data processing and storage capabilities, airborne electronic flight instrument system expansion units, airborne domain controllers, etc. Furthermore, the aviation safety assessment system can also be deployed on ground-based operational monitoring terminals, such as monitoring terminals in airline operations control centers, airport security terminals, and other terminal equipment capable of establishing communication connections with aircraft or aviation data networks. This application does not limit this specific deployment.

[0032] The following detailed description of a method for assessing the risk of aircraft tail scraping the ground, provided by this application, is based on specific embodiments and accompanying drawings.

[0033] Figure 1 This is a flowchart illustrating a method for assessing the risk of aircraft tail-end ground scratches, provided as an embodiment of this application. Specifically, as shown... Figure 1 As shown, it includes the following: S101. Obtain flight data of the target flight.

[0034] In the embodiments of this application, flight data refers to various parameter data generated and recorded by the aircraft during flight, covering information such as aircraft attitude, flight performance, operation input, environmental conditions and equipment status, which can comprehensively reflect the real-time operation status and flight phase characteristics of the aircraft.

[0035] In one implementation, the aviation safety assessment system (hereinafter referred to as the system) acquires flight data of a target flight through a quick access recorder (QAR) mounted on the aircraft. The QAR collects and stores flight data of the target flight in real time at a preset sampling frequency. The sampling frequency can be set to 1Hz or higher according to actual needs to ensure the timeliness and integrity of the data. The collected flight data is organized and stored in a time-series format, facilitating subsequent extraction and analysis of the flight data according to flight phases.

[0036] In one possible implementation, when acquiring flight data for a target flight, the system can simultaneously connect to multiple data storage nodes. In addition to directly acquiring raw flight data from the aircraft's fast access recorder, backup flight data for the target flight can also be synchronously acquired from the airline's ground data center or cloud server. Through multi-source data comparison and verification, abnormal data caused by equipment failure or transmission interference is eliminated, ensuring the accuracy of the acquired flight data. This multi-source data acquisition method further enhances data reliability and provides a high-quality data foundation for subsequent risk assessment.

[0037] In one possible implementation, after acquiring the flight data of the target flight, the system can also perform preliminary data filtering. Based on preset flight phase division rules, it filters out flight data corresponding to key flight phases relevant to tail-scratching risk from the acquired complete time-series flight data, such as takeoff and landing data. This reduces the interference of non-critical data with subsequent processing and improves the efficiency of risk assessment. This method, by pre-filtering data, can focus on core analytical objects and shorten the time spent on subsequent feature extraction and model calculation.

[0038] As can be seen from S101 above, by acquiring the flight data of the target flight, it is possible to ensure that the acquired flight data is complete, timely and accurate, providing a high-quality data foundation for the subsequent extraction of risk assessment features, thereby ensuring the reliability of the risk assessment results of the aircraft tail scratching the ground, and providing effective data support for aviation safety prevention and control.

[0039] S102. Extract risk assessment features from flight data.

[0040] Among them, risk assessment features include flight status features, operational behavior features, and time series analysis features. Time series analysis features are used to characterize the changing trend or dynamic sequence of one or more flight parameters within a preset flight phase.

[0041] In this embodiment, the risk assessment features are a set of key indicators extracted from flight data that reflect the risk of the aircraft's tail scraping the ground. Among them, flight state features characterize the aircraft's real-time operating attitude and performance, operational behavior features reflect the pilot's control actions, and time series analysis features capture the dynamic change patterns of parameters within a preset flight phase.

[0042] In one implementation, when the system extracts risk assessment features from flight data, it can extract flight status features, operational behavior features, and time-series analysis features based on preset feature extraction rules, categorizing different parameters in the flight data. Flight status features are extracted directly from the raw flight data, including parameters reflecting the real-time operating status of the aircraft such as indicated airspeed, vertical speed, radio altitude, and pitch angle. Operational behavior features are extracted based on crew operation-related data, including parameters reflecting operational actions such as throttle position, control stick input, and throttle release timing. Time-series analysis features are extracted by analyzing flight parameters in preset flight phases (such as the 500-foot to touchdown phase during landing), extracting parameter change trends, dynamic sequences, and energy management-related indicators, comprehensively covering static and dynamic factors affecting tail-to-ground strike risk.

[0043] In one possible implementation, the effectiveness of time-series analysis features can be enhanced by combining sliding window techniques and mathematical calculation methods. For time-series flight data of a preset flight phase, a fixed-duration sliding window is used to segment the parameters. Linear fitting is used to calculate the slope of changes in parameters such as airspeed and glide slope deviation, characterizing the dynamic trends of these parameters. Simultaneously, based on energy management logic, indicators such as energy altitude and the ratio of sink rate to airspeed are calculated and incorporated into the time-series analysis features, achieving a precise characterization of the aircraft's energy state and parameter variation patterns. This approach comprehensively captures the dynamic interaction relationships of parameters, improving the features' ability to represent risks.

[0044] In one possible implementation, after extracting risk assessment features, feature optimization can be performed. Correlation analysis is conducted on the extracted features to eliminate redundant or overly correlated features, avoiding information overlap between features. Simultaneously, the features are standardized, converting features with different dimensions to a unified numerical range to ensure the stability and accuracy of subsequent model training. This approach, by optimizing feature quality, reduces interference from invalid information, improves the training efficiency and prediction accuracy of the risk assessment model, and thus enhances the reliability of tail-scratching risk assessment.

[0045] As can be seen from S102 above, by extracting risk assessment features, it is possible to comprehensively and accurately capture the key factors affecting the risk of aircraft tails touching the ground. This includes both static flight status and operational behavior information, as well as dynamic temporal change patterns, laying a data foundation for subsequent accurate prediction of the grounding pitch angle and risk assessment.

[0046] S103. Input the risk assessment features into the risk assessment model to obtain the predicted ground elevation angle output by the risk assessment model.

[0047] Among them, the risk assessment model can characterize the linear and nonlinear relationships between risk assessment characteristics and ground elevation angle.

[0048] In this embodiment, the risk assessment model refers to a predictive model used to establish the correlation between risk assessment characteristics and the grounding pitch angle. This model can simultaneously capture both the linear correlation and nonlinear interaction effects between the characteristics and the grounding pitch angle, achieving accurate prediction of the grounding pitch angle. The grounding pitch angle refers to the fuselage pitch angle when the aircraft is about to touch the ground, and is a core indicator for judging the risk of tail-to-ground contact.

[0049] In one implementation, when the system inputs risk assessment features into the risk assessment model to obtain the predicted ground pitch angle, it must first perform feature adaptation processing to ensure that the format and dimensions of the input features are consistent with the input requirements of the risk assessment model. Subsequently, the processed flight status features, operational behavior features, and time-series analysis features are input into the risk assessment model as a whole. The model can simultaneously analyze the linear and nonlinear relationships between the features and the ground pitch angle through its built-in algorithm logic, automatically calculate and output the corresponding predicted ground pitch angle, achieving efficient conversion from feature input to result output.

[0050] In one possible implementation, a feature priority ranking step can be added before inputting risk assessment features into the risk assessment model. Based on the validation results of historical flight data, priority weights are assigned to various risk assessment features. High-priority features are input into the model first for core calculations, while low-priority features are used as auxiliary supplements, improving the model's computational efficiency while ensuring prediction accuracy. This approach optimizes the model's computational process and adapts to the real-time requirements of different scenarios.

[0051] In one possible implementation, the system can incorporate a model input anomaly monitoring mechanism. This mechanism monitors in real time for issues such as missing data or numerical anomalies in the input risk assessment features. If anomalies are detected, it automatically triggers completion or correction strategies, such as using interpolation to fill in missing data or removing outliers outside the reasonable range, ensuring the quality of the feature data input to the model. This anomaly monitoring mechanism can prevent abnormal data from affecting the model's prediction results, enhancing the stability and reliability of the risk assessment model.

[0052] As can be seen from S103 above, by inputting risk assessment features into the risk assessment model, and by leveraging the model's comprehensive ability to represent linear and nonlinear relationships, the predicted ground pitch angle can be accurately output, providing a basis for accurately judging the risk of the aircraft's tail scraping the ground, thus improving the scientificity and accuracy of the risk assessment.

[0053] S104. Based on the predicted ground pitch angle, determine the tail-to-ground risk information of the target flight.

[0054] In this embodiment of the application, the tail-to-ground contact risk information refers to the level of risk of the aircraft tail contacting the ground, which is determined based on the predicted ground pitch angle and the relevant judgment criteria. It is mainly used to intuitively present the safety status of the target flight during the critical flight phase.

[0055] In one implementation, when determining the tail-scratching risk information of a target flight based on the predicted ground pitch angle, the system needs to consider the aircraft type parameters of the target flight. First, based on the specific aircraft type of the target flight, the system queries and determines the maximum permissible pitch angle and the preset safety margin corresponding to that aircraft type. Next, the difference between the maximum permissible pitch angle and the predicted ground pitch angle is calculated to obtain the safety margin. Finally, the risk level is determined based on the range of the safety margin: if the safety margin is greater than the preset safety margin, it is determined to be low risk; if the safety margin is less than or equal to the safety margin but greater than zero, it is determined to be medium risk; if the safety margin is less than or equal to zero, it is determined to be high risk, thus completing the determination of the tail-scratching risk information.

[0056] In one possible implementation, after determining the tail-end grazing risk information, a risk description document can be generated simultaneously. In addition to clearly defining the risk level, the document can also include specific values ​​for the safety margin, corresponding judgment criteria, and preliminary prevention and control suggestions. For example, in cases of high risk, it can advise pilots to adjust their operating posture, providing clear decision-making references for ground control personnel and pilots. This approach enriches the practicality of risk information and facilitates the rapid implementation of targeted prevention and control measures.

[0057] In one possible implementation, safety margin parameters can be dynamically adjusted for different airport scenarios. By considering environmental factors such as runway gradient, altitude, and climate conditions, the preset safety margin can be adaptively modified. For example, the safety margin threshold could be appropriately increased at high-altitude airports, making risk assessment more aligned with actual operational scenarios. This approach improves the accuracy of risk information and better adapts to the safety assessment needs in complex environments.

[0058] As can be seen from S104 above, by determining the tail-to-ground-scratch risk information based on the predicted ground pitch angle, combined with aircraft characteristics and quantitative judgment standards, the risk level can be objectively and accurately classified, so that the tail-to-ground-scratch risk can be clearly and intuitively presented, providing a reliable basis for safety warning and decision-making in aviation operations, and reducing the probability of tail-to-ground-scratch accidents.

[0059] In this embodiment of the application, the risk assessment model includes at least a first risk assessment sub-model and a second risk assessment model. Furthermore, the risk assessment model can obtain the predicted grounding elevation angle based on the prediction results of the first risk assessment sub-model and the prediction results of the second risk assessment model. Figure 2 As shown, the risk assessment features are input into the risk assessment model to obtain the predicted ground elevation angle, including the following steps S201-S203: S201. Input the risk assessment features into the first risk assessment sub-model to obtain the predicted value of the first grounding elevation angle.

[0060] The first risk assessment sub-model is used to characterize the linear relationship between risk assessment features and grounding pitch angle.

[0061] In the embodiments of this application, the first risk assessment sub-model is a prediction model used to capture the linear correlation between risk assessment features and ground elevation angle. It can output the ground elevation angle prediction result based on linear law by establishing a linear functional relationship between features and target values.

[0062] In one implementation, to input risk assessment features into a first risk assessment sub-model to obtain a predicted first ground pitch angle, feature preprocessing and adaptation must first be performed. The process is as follows: First, the continuous numerical features in the risk assessment features are converted to a preset numerical range to obtain the first risk assessment sub-feature. Then, the discrete state features in the risk assessment features are encoded to obtain the second risk assessment sub-feature. Finally, the first and second risk assessment sub-features are input into the first risk assessment sub-model to obtain the predicted first ground pitch angle output by the first risk assessment sub-model. For example, the system first standardizes the continuous numerical features in the risk assessment features, converting them to a preset numerical range of [-1,1]. For example, pitch angle data with an original range of -5° to 15° is converted into the first risk assessment sub-feature within the range of [-1,1] using a standardization formula. Then, the discrete state features are one-hot encoded. For example, the "retracted" and "deployed" states of the discrete feature "landing gear status" are encoded into vector forms of [1,0] and [0,1] respectively to obtain the second risk assessment sub-feature. Finally, the preprocessed first risk assessment sub-features and second risk assessment sub-features are input together into the first risk assessment sub-model, and the model outputs the predicted value of the first ground elevation angle through linear calculation.

[0063] In one implementation, the first risk assessment sub-model is constructed through feature selection and linear fitting. First, feature selection is performed from the initial risk assessment features using a multiple stepwise linear regression method. Specifically: First, a subset of features whose contribution to the predicted grounding pitch angle reaches a first preset threshold is selected from multiple risk assessment features. Then, based on the feature subset, the weight coefficient corresponding to each feature is determined through linear regression, generating the first risk assessment sub-model. For example, the first preset threshold is set to 0.8 (i.e., the feature contribution must reach 80%), and a subset of features such as pitch angle change rate, energy altitude, and the ratio of sinking rate to airspeed that contribute to the predicted grounding pitch angle are selected. Next, based on this feature subset, the weight coefficient corresponding to each feature is calculated using a linear regression algorithm. For example, the weight coefficient for "pitch angle change rate" is determined to be 0.3, and the weight coefficient for "energy altitude" is determined to be 0.25. Finally, the first risk assessment sub-model is generated. For example, based on the determined weight coefficients, the model is defined as: First ground pitch angle prediction value = β + 0.3 × pitch angle change rate + 0.25 × energy height + ... + ε (where β is a constant term and ε is an error term).

[0064] In one possible implementation, when transforming continuous numerical features, the processing method can be flexibly selected according to the feature type. For continuously distributed features (such as airspeed), standardization is used to transform them to a preset range; for irregularly distributed features (such as vertical velocity), normalization is used to ensure data stability and further improve the reliability of the first risk assessment sub-feature. This flexible processing method can adapt to the distribution characteristics of different features, providing high-quality input data for the model.

[0065] In one possible implementation, when constructing the first risk assessment sub-model, a variance inflation factor can be introduced to diagnose multicollinearity. If multicollinearity is detected in the selected feature subset (e.g., variance inflation factor greater than 10), the impact of multicollinearity can be reduced by removing redundant features or performing principal component analysis before calculating the weight coefficients. This approach avoids model bias caused by multicollinearity and enhances the stability and prediction accuracy of the first risk assessment sub-model.

[0066] As can be seen from S201 above, by constructing the first risk assessment sub-model and the feature input method, we can ensure the standardization of the input data through precise feature preprocessing, focus on the core influencing factors through feature screening, and achieve interpretable prediction of the grounding pitch angle by leveraging the advantages of the linear model. This provides basic data support that is both accurate and readable for subsequent comprehensive risk assessment.

[0067] S202. Input the risk assessment features into the second risk assessment sub-model to obtain the predicted value of the second ground elevation angle.

[0068] The second risk assessment sub-model is used to characterize the nonlinear relationship between risk assessment features and ground elevation angle.

[0069] In this embodiment, the second risk assessment sub-model refers to a predictive model specifically designed to capture the nonlinear correlation between risk assessment features and ground pitch angle. It is constructed using an integrated decision tree algorithm and has the advantages of processing high-dimensional data and mining complex interactive effects of features, and can accurately adapt to the nonlinear relationship between flight parameters.

[0070] In one implementation, the second risk assessment sub-model is constructed using a random forest (a model based on ensemble decision trees). The process of inputting risk assessment features into the model to obtain the predicted value of the second ground pitch angle is as follows: First, the risk assessment features are input into the prediction model based on ensemble decision trees. Then, multiple decision trees judge the input features separately, and the output results of each decision tree are summarized to obtain the predicted value of the second ground pitch angle. For example, the system first inputs the preprocessed flight state features, operational behavior features, and time series analysis features (such as high descent rate, late throttle cut-off, pitch angle change rate, etc.) into the random forest model. The decision trees within the model independently judge the input features based on different feature subsets. For example, tree 1 focuses on the features of "descent rate to airspeed ratio" and "pitch angle change rate" and outputs a predicted value of 8.2°, tree 2 focuses on the features of "energy altitude" and "throttle cut-off timing" and outputs a predicted value of 8.5°, and the remaining decision trees complete the judgment in sequence. Finally, the output results of all decision trees are summarized by a weighted average method to obtain the final predicted value of the second ground pitch angle of 8.3°.

[0071] In one implementation, the second risk assessment sub-model is constructed through iterative training using time-series training subsets. The specific steps are as follows: First, multiple training subsets are extracted from historical flight data in chronological order. Then, for each training subset, a primary prediction model is trained based on that subset, and the performance of this primary prediction model is validated using the next training subset. Finally, the second risk assessment sub-model is determined based on the performance evaluation results of all primary prediction models. For example, firstly, one training subset is extracted from historical flight data from the past three years, chronologically, every three months, resulting in 12 training subsets (denoted as S1 to S12). Then, a primary model M1 is trained based on training subset S1, and its performance is validated using subsequent subset S2 (e.g., calculating prediction error). A primary model M2 is trained based on S2 and validated using S3, and so on, to complete the training and validation of all primary models. Finally, the performance evaluation results of all primary models were statistically analyzed, and eight models with prediction errors below 0.5° (such as M3 to M10) were selected for integration. The final second risk assessment sub-model was generated by adjusting the weight coefficients of each model to ensure that the model is adapted to the latest flight data patterns.

[0072] In one possible implementation, the second risk assessment sub-model based on the ensemble decision tree can introduce a feature importance assessment mechanism. After training, by calculating the split contribution of features in all decision trees, the key features with the greatest impact on nonlinear prediction (such as "pitch angle change rate" and "sinking rate") are identified, and these core features are preferentially retained in subsequent model iterations, while redundant features with a contribution of less than 5% are removed, further improving the computational efficiency of the model.

[0073] In one possible implementation, data augmentation can be performed on each training subset when training the second risk assessment sub-model. To address the issue of limited data samples during the landing phase, the sample size of the training subset can be increased through time series interpolation, adding appropriate noise, etc. For example, the sample size of S1 can be increased from 1,000 to 3,000 samples to avoid overfitting due to insufficient samples and to enhance the model's generalization ability in complex flight scenarios.

[0074] As shown in S202 above, through model construction based on ensemble decision tree, feature input method, and time-series training method, the second risk assessment sub-model can accurately capture the nonlinear correlation and complex interaction effects between risk assessment features and ground pitch angle. At the same time, it adapts to the time-series variation pattern of flight data, and the output second ground pitch angle prediction value has high accuracy and strong adaptability, which can provide data support for improving the reliability of risk assessment by combining the results of the two sub-models.

[0075] S203. Determine the predicted ground elevation angle based on the predicted values ​​of the first and second ground elevation angles.

[0076] In this embodiment of the application, the predicted ground pitch angle refers to the core pitch angle index determined by combining the linear prediction results of the first risk assessment sub-model and the nonlinear prediction results of the second risk assessment sub-model. It is used to judge the risk of tail rubbing. It integrates the advantages of the two models to improve the prediction accuracy.

[0077] In one possible implementation, the step of determining the final predicted ground elevation angle based on the two predicted values ​​can be performed using a fixed-weighted fusion strategy. For example, the weight of the first predicted ground elevation angle is first set to 0.4 (to accommodate the interpretability advantage of the linear model), and the weight of the second predicted ground elevation angle is set to 0.6 (to accommodate the accuracy advantage of the nonlinear model). Assuming the first sub-model outputs a predicted value of 8.1° and the second sub-model outputs a predicted value of 8.3°, the final predicted ground elevation angle can be calculated using the formula "final predicted value = first predicted value × 0.4 + second predicted value × 0.6" as 8.1° × 0.4 + 8.3° × 0.6 = 8.22°, thus completing the fusion calculation.

[0078] In one possible implementation, the step of determining the final predicted ground pitch angle based on two predicted values ​​can also employ a dynamic weight adjustment strategy. For example, a weight adjustment rule can be constructed based on historical flight data: when the difference between the two predicted values ​​is less than 0.3°, both are assigned the same weight (0.5 each); when the difference is between 0.3° and 0.8°, the weights are adjusted according to the model's recent prediction accuracy; for example, if the second sub-model has a lower prediction error in the past month, it is assigned a weight of 0.7, and the first sub-model a weight of 0.3; if the difference is greater than 0.8°, a manual review mechanism is triggered. Assuming the first predicted value is 7.9°, the second predicted value is 8.5° (difference 0.6°), and the second sub-model has higher recent accuracy, then the final predicted value = 7.9° × 0.3 + 8.5° × 0.7 = 8.32°.

[0079] In one possible implementation, a reasonableness verification step for the predicted ground elevation angle can be added before determining the final predicted ground elevation angle. A reasonable range for the ground elevation angle is set (e.g., 2° or 12°, depending on the parameters of mainstream models). If the predicted value of any sub-model exceeds this range, its weight is reduced to 0.2, while the weight of another reasonable predicted value is increased to 0.8. If both exceed the range, it is marked as an anomaly and data reprocessing is triggered. This method avoids abnormal predicted values ​​affecting the final result and enhances the reliability of the prediction.

[0080] In one possible implementation, the fusion strategy can be dynamically adjusted based on the flight scenario. For example, a fixed-weight fusion strategy can be used in conventional airport landing scenarios. In complex scenarios such as high altitudes or strong winds, a dynamic weight strategy can be switched to, and additional scenario features (such as wind speed and runway slope) can be introduced to assist in correcting the weights, making the final prediction value more closely match the actual flight conditions. This scenario-based adaptation method can further improve prediction accuracy.

[0081] As can be seen from S203 above, by integrating the predicted values ​​of the first and second ground pitch angles, the advantages of the linear model in terms of stability and interpretability are fully utilized, while the advantages of the nonlinear model in terms of accuracy in capturing complex relationships are also utilized. Through a reasonable weighting strategy, the optimal integration of the two results is achieved, and the final output predicted ground pitch angle has high accuracy and reliability, providing effective data support for the subsequent tail rubbing risk assessment.

[0082] In this embodiment of the application, after obtaining the predicted value of the second ground pitch angle in step S202, the risk assessment features that can affect the predicted value of the second ground pitch angle can be determined from the risk assessment features through the second risk assessment sub-model. Specifically, this includes: determining key risk assessment features with a high degree of influence on the predicted value of the second ground pitch angle from the risk assessment features; and determining a combination of risk assessment features that have a synergistic effect on the predicted value of the second ground pitch angle from the risk assessment features. Furthermore, when determining the tail-side ground scraping risk information of the target flight, this includes: generating a risk report containing at least one of the predicted ground pitch angle, key risk assessment features, and combinations of risk assessment features.

[0083] In this application embodiment, the key risk assessment feature refers to a single feature that has a significant impact on the predicted value of the second ground pitch angle, and the risk assessment feature combination with synergistic effect refers to a feature pair that has a significant superimposed or correlated impact on the predicted value when the two features work together. Both are used to deepen the risk cause analysis and enrich the content of the risk report.

[0084] In one possible implementation, the process of determining key risk assessment features employs a feature masking simulation method, as follows: First, for the second risk assessment sub-model, multiple simulation predictions are performed, masking one or more risk assessment features in each simulation prediction. Then, based on the fluctuations in the results of multiple simulation predictions, the contribution of each risk assessment feature to the predicted value of the second grounding pitch angle is quantified. Finally, risk assessment features with a contribution exceeding a first preset threshold are identified as key risk assessment features. For example, first, for the second risk assessment sub-model (random forest model), risk assessment features such as "pitch angle change rate," "sinking rate," and "energy altitude" are selected, and simulation predictions are performed, masking one feature in each simulation. For instance, after masking "pitch angle change rate," the predicted value changes from 8.3° to 7.5°, a fluctuation of 0.8°; after masking "energy altitude," the predicted value fluctuates by 0.2°. Based on the fluctuation amplitude, the contribution of each feature is quantified, and a first preset threshold of 0.5° is set. Then, "pitch angle change rate" and "sinking rate," whose contributions meet the threshold, are identified as key risk assessment features.

[0085] In one possible implementation, determining the risk assessment feature combination with synergistic impact requires pairwise feature synergistic simulation, as follows: First, through simulation prediction, the synergistic impact of any two risk assessment features on the predicted value of the second grounding pitch angle is quantified. Then, feature pairs with synergistic impact values ​​higher than a second preset threshold are identified as risk assessment feature combinations that generate synergistic impact. For example, all risk assessment features are paired to generate multiple feature pairs, and the synergistic impact of each feature pair is quantified through simulation prediction. For instance, when simulating the combination of "high subsidence rate + late throttle reduction," the predicted value increases by 0.6° compared to considering the two features individually, resulting in a synergistic impact of 0.6°. When simulating the combination of "pitch angle change rate + wind speed," the synergistic impact is 0.2°. Setting the second preset threshold to 0.4°, "high subsidence rate + late throttle reduction" is identified as a risk assessment feature combination that generates synergistic impact.

[0086] In one possible implementation, the risk report should integrate core assessment information. The report should first clearly state the final predicted grounding pitch angle (e.g., 8.22°) and its corresponding risk level (e.g., medium risk); secondly, it should list key risk assessment characteristics and indicate their contribution (e.g., "pitch angle change rate: contribution 0.8°"); finally, it should explain the combination of synergistic impact characteristics and their effects (e.g., "high subsidence rate + late throttle reduction: synergistic increase in risk 0.6°"), while also supplementing targeted prevention and control recommendations, such as "it is recommended to optimize the lever operation rhythm to avoid late throttle reduction under high subsidence rates," thus forming a complete risk report.

[0087] In one possible implementation, a time dimension analysis can be introduced when determining key risk assessment features. Feature masking simulations are performed for different time periods of the landing phase (e.g., 500 feet to 200 feet, 200 feet to touchdown) to identify key features for each time period. For example, "energy altitude" is a key feature for the 500 feet to 200 feet phase, and "pitch angle change rate" is a key feature for the 200 feet to touchdown phase, making the risk causation analysis more closely aligned with the characteristics of the flight phase.

[0088] In one possible implementation, a tiered reporting mechanism can be added after a risk report is generated. Reports are sent to the appropriate recipients based on their risk level: low-risk reports are sent only to the airline's safety management department for filing; medium-risk reports are sent to the corresponding flight crew and safety department; and high-risk reports are sent in real-time to the flight crew, ground control center, and safety management department, along with emergency response guidelines, thus improving the timeliness of risk prevention and control.

[0089] In some embodiments, by identifying key risk assessment features, analyzing collaborative feature combinations, and generating risk reports, it is possible not only to accurately pinpoint the core causes of tail-end rubbing risks, but also to provide clear risk tracing and control basis for the crew and ground control personnel through rich report content, effectively improving the practicality of risk assessment and helping to optimize operating procedures in a targeted manner and reduce the probability of accidents.

[0090] In summary, in one specific embodiment, Figure 3 This is a flowchart illustrating another method for assessing the risk of aircraft tail-end skidding, provided in an embodiment of this application. Figure 3 As shown, the aircraft tail-to-ground contact risk assessment method of this application can be implemented through the following complete process: First, raw QAR data is acquired. This data is a collection of various flight parameters recorded in real time during the flight, covering multiple dimensions of information such as aircraft attitude, performance, and operation. Next, based on aviation industry experience, QAR feature variables related to tail-to-ground contact risk are extracted from the raw QAR data, such as pitch angle and vertical speed for flight status, throttle cut-off timing and pitch angle change rate for operational behavior, and parameter change trend sequences for time series analysis. Subsequently, the extracted QAR feature variables are used to predict the ground contact pitch angle. In this process, linear models and random forest models are used to capture the linear and nonlinear relationships between the features and the ground contact pitch angle, respectively, to obtain the corresponding prediction results. Then, the model prediction results are normalized and converted into a risk score of 0-100, with a higher risk score indicating a higher tail-to-ground contact risk. Finally, the flight's risk score is statistically added to the corresponding operator's risk score to comprehensively evaluate the operator's flight operation risk level, and then targeted safety improvement measures are formulated based on this.

[0091] The tail-to-ground strike risk assessment method provided in this application first acquires flight data of the target flight and accurately extracts risk assessment features covering flight state characteristics, operational behavior characteristics, and time-series analysis characteristics. The time-series analysis features effectively characterize the changing trends or dynamic sequences of flight parameters within a preset flight phase, providing comprehensive and dynamic data support for risk assessment. Next, the risk assessment features are input into a risk assessment model that simultaneously characterizes both linear and nonlinear relationships between the features and the ground pitch angle. This ensures the model accurately captures the correlation between multiple features and the ground pitch angle, thereby outputting an accurate predicted ground pitch angle. Finally, based on the predicted ground pitch angle, the tail-to-ground strike risk information of the target flight is determined, achieving accurate assessment of the tail-to-ground strike risk. This provides a reliable basis for flight operation safety warnings and operational optimization, effectively reducing the probability of tail-to-ground strike accidents and improving flight safety levels.

[0092] In an exemplary embodiment, Figure 4This is a schematic diagram illustrating the composition of an aircraft tail-side ground-scraping risk assessment device provided in an embodiment of this application. Figure 4 As shown, the aircraft tail-to-ground-strike risk assessment device includes: a data acquisition module 401, a feature extraction module 402, a risk assessment module 403, and an output module 404. The data acquisition module 401 acquires flight data of the target flight. The feature extraction module 402 extracts risk assessment features from the flight data. These features include flight state features, operational behavior features, and time-series analysis features. The time-series analysis features characterize the changing trends or dynamic sequences of one or more flight parameters within a preset flight phase. The risk assessment module 403 inputs the risk assessment features into a risk assessment model to obtain the predicted ground-to-ground pitch angle output by the model. The risk assessment model can characterize the linear and nonlinear relationships between the risk assessment features and the ground-to-ground pitch angle. The output module 404 determines the tail-to-ground-strike risk information of the target flight based on the predicted ground-to-ground pitch angle.

[0093] In an exemplary embodiment, this application also provides an electronic device, which may be the aircraft tail-side ground-scraping risk assessment device in the above method embodiments. Figure 5 This is a schematic diagram of a risk assessment device for aircraft tail-end ground scratching provided in an embodiment of this application. Figure 5 As shown, the aircraft tail-side ground-scraping risk assessment device may include: a processor 501 and a memory 502; the memory 502 stores instructions executable by the processor 501; when the processor 501 is configured to execute instructions, it causes electronic devices or network devices or managers to perform the system functions described in the foregoing method embodiments.

[0094] Through the above description of the implementation methods, those skilled in the art can clearly understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0095] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0096] The units described as separate components may or may not be physically separate. A component shown as a unit can be one or more physical units; that is, it can be located in one place or distributed in multiple different locations. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0097] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0098] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a readable storage medium. Based on this understanding, the technical solutions of the embodiments of this application, essentially, or the parts that contribute to the prior art, or all or part of the technical solutions, can be embodied in the form of a software product. This software product is stored in a storage medium and includes several instructions to cause a device (which may be a microcontroller, chip, etc.) or processor to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, ROM, RAM, magnetic disks, or optical disks.

[0099] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for assessing the risk of aircraft tail scraping the ground, characterized in that, The method includes: Obtain flight data for the target flight; Risk assessment features are extracted from the flight data; the risk assessment features include flight status features, operational behavior features, and time series analysis features; the time series analysis features are used to characterize the changing trend or dynamic sequence of one or more flight parameters within a preset flight phase. The risk assessment features are input into the risk assessment model to obtain the predicted ground elevation angle output by the risk assessment model; the risk assessment model can characterize the linear and nonlinear relationships between the risk assessment features and the ground elevation angle. Based on the predicted ground pitch angle, the tail-side ground scraping risk information of the target flight is determined.

2. The method according to claim 1, characterized in that, The step of inputting the risk assessment features into the risk assessment model to obtain the predicted ground elevation angle includes: The risk assessment features are input into the first risk assessment sub-model to obtain the predicted value of the first ground elevation angle; the first risk assessment sub-model is used to characterize the linear relationship between the risk assessment features and the ground elevation angle; The risk assessment features are input into the second risk assessment sub-model to obtain the predicted value of the second ground elevation angle; the second risk assessment sub-model is used to characterize the nonlinear relationship between the risk assessment features and the ground elevation angle. The predicted ground elevation angle is determined based on the first predicted ground elevation angle and the second predicted ground elevation angle.

3. The method according to claim 2, characterized in that, The step of inputting the risk assessment features into the first risk assessment sub-model to obtain the predicted value of the first ground elevation angle includes: The continuous numerical features in the risk assessment features are converted to a preset numerical range to obtain the first risk assessment sub-feature; The discrete state features in the risk assessment features are encoded to obtain the second risk assessment sub-features; The first risk assessment sub-feature and the second risk assessment sub-feature are input into the first risk assessment sub-model to obtain the first ground elevation angle prediction value output by the first risk assessment sub-model.

4. The method according to claim 3, characterized in that, The first risk assessment sub-model is constructed in the following way: From the multiple risk assessment features, a subset of features whose contribution to the predicted grounding pitch angle reaches a first preset threshold is selected; Based on the feature subset, the weight coefficients corresponding to each feature are determined by linear regression to generate the first risk assessment sub-model.

5. The method according to claim 2, characterized in that, The second risk assessment sub-model is a prediction model based on an ensemble decision tree; the step of inputting the risk assessment features into the second risk assessment sub-model to obtain the predicted value of the second ground elevation angle includes: The risk assessment features are input into a prediction model based on an ensemble decision tree; The input features are evaluated by multiple decision trees, and the output results of each decision tree are summarized to obtain the second ground elevation angle prediction value.

6. The method according to claim 5, characterized in that, The second risk assessment sub-model is trained in the following way: Multiple training subsets were extracted from historical flight data in chronological order. For each training subset, a primary prediction model is trained based on the training subset, and the performance of the primary prediction model is verified using the next training subset after the training subset. Based on the performance evaluation results of all primary prediction models, the second risk assessment sub-model is determined.

7. The method according to claim 2, characterized in that, After obtaining the second predicted ground elevation angle, the method further includes the following steps: From the aforementioned risk assessment features, identify the key risk assessment features that have a high degree of influence on the predicted value of the second ground pitch angle; From the risk assessment features, determine the combination of risk assessment features that have a synergistic effect on the predicted value of the second ground pitch angle; The determination of the tail-scratching risk information of the target flight includes: Generate a risk report that includes at least one of the predicted ground elevation angle, the key risk assessment features, and combinations of the risk assessment features.

8. The method according to claim 7, characterized in that, The step of identifying key risk assessment features that have a high impact on the predicted value of the second ground elevation angle from the risk assessment features includes: For the second risk assessment sub-model, perform multiple simulation predictions, and mask one or more of the risk assessment features in each simulation prediction; Based on the fluctuations in the results of the multiple simulations and predictions, the contribution of each risk assessment feature to the predicted value of the second ground elevation angle is quantified; Risk assessment features whose contribution exceeds a first preset threshold are identified as key risk assessment features.

9. The method according to claim 7, characterized in that, The step of determining the combination of risk assessment features that have a synergistic impact on the predicted value of the second ground pitch angle from the risk assessment features includes: Through simulation and prediction, the synergistic influence of any two risk assessment features on the predicted value of the second grounding pitch angle is quantified. Feature pairs with a synergistic impact level higher than a second preset threshold are identified as the risk assessment feature combination that generates synergistic impact.

10. The method according to claim 1, characterized in that, The determination of the tail-scratching risk information of the target flight based on the predicted ground pitch angle includes: Based on the aircraft type of the target flight, determine the corresponding maximum permissible pitch angle and safety margin; Calculate the safety margin between the maximum permissible pitch angle and the predicted ground pitch angle; If the safety margin is greater than the preset safety threshold, it is determined to be low risk; If the safety margin is less than or equal to the safety tolerance and greater than zero, it is determined to be of medium risk; If the safety margin is less than or equal to zero, it is considered high risk.

Images