A method, equipment, medium, and product for predicting the transition corridor boundary of a tiltrotor aircraft.

Abstract

This application discloses a method, device, medium, and product for predicting the transition corridor boundary of a tiltrotor aircraft, relating to the field of tiltrotor aircraft design and aerodynamic simulation technology. The method includes: determining the design space of the tiltrotor aircraft; sampling points within the design space; performing dynamic transition simulation based on the sampling points to obtain transition corridor boundary information corresponding to each sampling point; forming sample pairs based on the sampling points and their corresponding transition corridor boundary information to generate a training set and a test set; constructing a network model based on a machine learning network; training and testing the network model using the training and test sets until the trained network model meets set conditions, thus obtaining a transition corridor boundary prediction model; and acquiring the design parameters of the tiltrotor aircraft to be predicted in real time, and using the transition corridor boundary prediction model to obtain the transition corridor boundary prediction result. This application enables rapid and accurate prediction of the transition corridor of a new configuration tiltrotor aircraft.

Description

Technical Field

[0001] This application relates to the field of tiltrotor aircraft design and aerodynamic simulation technology, and in particular to a method, equipment, medium and product for predicting the transition corridor boundary of a tiltrotor aircraft. Background Technology

[0002] A tiltrotor aircraft is an advanced vertical take-off and landing (VTOL) tiltrotor aircraft that combines the vertical take-off and landing capabilities of a helicopter with the high-speed cruise capabilities of a fixed-wing aircraft. Its core feature is that the rotor system can tilt between vertical take-off and landing modes and horizontal forward flight modes, thus achieving a change in flight mode. This transition occurs within a specific speed-rotor tilt angle flight envelope known as the conversion corridor (or transition corridor).

[0003] The transition corridor is one of the most critical performance envelopes for tiltrotor aircraft. It defines the set of all feasible flight state points from hover to cruise (or decelerate and land in reverse) that allow the tiltrotor to safely and stably accelerate from hover to cruise (or decelerate and land). The boundary of this corridor is influenced by multiple complex physical factors, including aerodynamic disturbances, dynamic stability, handling margin, and structural loads. Its accurate determination is crucial for the safe operation and overall design of the tiltrotor aircraft.

[0004] Currently, the mainstream method for determining the transition corridor boundary of a novel tiltrotor aircraft relies on high-fidelity flight dynamics simulation. This method typically requires establishing a complex simulation system that includes unsteady aerodynamic models of the rotor / wing, multibody dynamics models, and control system models. It also involves point-by-point dynamic simulation of a large number of candidate flight state points to determine whether they are within a stable and controllable transition region. This process is computationally intensive and extremely time-consuming, often requiring several days or even weeks of high-performance computing resources to complete a full envelope scan of a single configuration.

[0005] In the early conceptual design or iteration phases of tiltrotor aircraft, engineers typically need to quickly assess the impact of multiple different parameter configurations (such as weight, size, and power) on the transition corridor. However, the traditional simulation methods mentioned above, due to their high time and computational costs, are insufficient to support such rapid, multi-round design exploration requirements, severely limiting design efficiency. Although some studies have attempted to estimate using simplified models or empirical formulas, these methods generally suffer from insufficient accuracy and poor generalization ability, failing to accurately reflect the complex nonlinear relationships between different configurations. Summary of the Invention

[0006] To address the aforementioned problems, this application provides a method, device, medium, and product for predicting the transition corridor boundary of a tiltrotor aircraft, enabling rapid and accurate prediction of the transition corridor for a new configuration tiltrotor aircraft.

[0007] To achieve the above objectives, this application provides the following solution: In a first aspect, this application provides a method for predicting the transition corridor boundary of a tiltrotor aircraft, including: Determine the design space for the tiltrotor aircraft; the design space includes multiple design parameters. Sampling is performed in the design space to obtain sampling points; Dynamic transition simulation is performed based on the sampling points to obtain the transition corridor boundary information corresponding to each sampling point; Sample pairs are formed based on the sampling points and the corresponding transition corridor boundary information to generate training and test sets; Building network models based on machine learning networks; The network model is trained and tested using the training set and the test set until the trained network model meets the set conditions, and then the trained network model is obtained; the trained network model is used as the transition corridor boundary prediction model. The design parameters of the tiltrotor aircraft to be predicted are obtained in real time, and the prediction results of the transition corridor boundary are obtained by using the transition corridor boundary prediction model.

[0008] Secondly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the tiltrotor transition corridor boundary prediction method provided above.

[0009] Thirdly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the tiltrotor transition corridor boundary prediction method described above.

[0010] Fourthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the tiltrotor transition corridor boundary prediction method described above.

[0011] According to the specific embodiments provided in this application, this application has the following technical effects: This application provides a method, device, medium, and product for predicting the transition corridor boundary of a tiltrotor aircraft. By sampling within the design space, data redundancy and gaps can be avoided. Through dynamic transition simulation based on sampling points, transition corridor boundary information corresponding to each sampling point is obtained, achieving high-fidelity numerical simulation. Furthermore, by utilizing sample pairs formed from sampling points and corresponding transition corridor boundary information, along with a machine learning network, an efficient and accurate intelligent prediction model (i.e., a transition corridor boundary prediction model) is constructed. This enables rapid evaluation of the transition corridor of a new configuration tiltrotor aircraft, significantly reducing R&D costs and accelerating the design process. It addresses the problems of high computational costs and excessive time consumption in traditional high-fidelity flight dynamics simulations when evaluating transition corridors, providing efficient technical support for the overall design, performance evaluation, and scheme iteration of tiltrotor aircraft. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in 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.

[0013] Figure 1 A flowchart illustrating a method for predicting the transition corridor boundary of a tiltrotor aircraft, provided in an embodiment of this application; Figure 2 A schematic diagram of the implementation architecture of a method for predicting the transition corridor boundary of a tiltrotor aircraft, provided in an embodiment of this application; Figure 3 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0014] 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.

[0015] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0016] In recent years, machine learning technology has demonstrated its powerful ability to handle high-dimensional and nonlinear mapping relationships and has been applied in the aerospace field for tasks such as aerodynamic performance prediction and fault diagnosis. However, a mature and effective solution has yet to be found for systematically applying machine learning methods to the specific and complex dynamic envelope prediction problem of the transition corridor boundary of tiltrotor aircraft. To overcome this deficiency, this application provides a machine learning-based method, device, medium, and product for predicting the transition corridor boundary of tiltrotor aircraft.

[0017] In one exemplary embodiment, this application provides a method for predicting the transition corridor boundary of a tiltrotor aircraft. This method is executed by a computer device, specifically a terminal or server, or both. In this embodiment, the method is described using a server as an example. Figure 1 As shown, the method includes: Step 100: Determine the design space of the tiltrotor aircraft; the design space includes multiple design parameters, such as ① rotor / wing geometry parameters: rotor radius, blade chord length, twist angle distribution, airfoil, wing aspect ratio, installation angle, etc.; ② tiltrotor overall parameters: takeoff weight, center of gravity position, thrust-to-weight ratio, etc.; ③ control strategy parameters: tilt rate, collective pitch rate of change, etc.

[0018] Step 101: Sample in the design space to obtain sampling points; Step 102: Perform dynamic transition simulation based on sampling points to obtain the transition corridor boundary information corresponding to each sampling point; Step 103: Form sample pairs based on the sampling points and the corresponding transition corridor boundary information to generate training and test sets; Step 104: Construct a network model based on machine learning networks; Step 105: Train and test the network model using the training set and the test set until the trained network model meets the set conditions to obtain the trained network model; use the trained network model as the transition corridor boundary prediction model. Step 106: Obtain the design parameters of the tiltrotor aircraft to be predicted in real time, and use the transition corridor boundary prediction model to obtain the transition corridor boundary prediction results.

[0019] By implementing steps 100-106 above, this application can learn from existing transition corridor models of tiltrotor helicopters with different parameters using machine learning methods, obtain the influence law of different parameters of tiltrotor aircraft on the transition corridor boundary, and then predict the transition corridor boundary of the newly designed tiltrotor helicopter. Based on this, accurate calculation can be performed to quickly determine the transition corridor boundary of the new model, saving a lot of time and computing resources.

[0020] In an exemplary embodiment of this application, to ensure that the sample points uniformly cover the entire design space, the implementation of step 101 above can be described as follows: sampling is performed in the design space using a Latin hypercube sampling method or a modified Latin hypercube sampling method to obtain sample points. Wherein: (I) Latin Hypercube Sampling (LHS) is a statistical sampling method used to generate sample points in a multidimensional space, commonly applied in computational experimental design and other fields. Compared to simple random sampling, it can more uniformly cover the value space of input variables, thus obtaining information about the input variable space more efficiently with the same number of samples. Based on this, the process of sampling in the design space using the Latin Hypercube Sampling method to obtain sampling points includes: Step 1: Divide the space into multiple equally probable intervals; Step 2, Random Sampling: Randomly select a sample value within each equally probable interval; for example, for a variable, randomly select a value from each of its n intervals to obtain n sample values; Step 3, Pairing and Combining: Randomly pair the sample values ​​to form multiple (e.g., n) two-dimensional sample points; during the pairing process, each interval of each sample value can only be used once.

[0021] Two-dimensional sample points are used as sampling points.

[0022] (ii) When using the LHS method, its inherent randomness often leads to uneven sample distribution in high-dimensional parameter spaces. Therefore, this application can also perform sampling based on the maximum-minimum distance LHS. A specific algorithm is used to adjust the position of the sampling points so that the minimum distance between any two sample points in the entire design space is as large as possible, thereby ensuring a more uniform distribution of sampling points in the design space. Based on this, in order to significantly improve space filling and avoid sample clustering, and to be applicable to building high-precision surrogate models (such as Kriging, neural networks), this application uses an improved Latin hypercube sampling method (such as maximum-minimum distance Latin hypercube sampling) to sample in the design space. The process of obtaining sampling points includes: Using Monte Carlo methods, multiple sample sets are randomly generated within the design space; Determine the distance between samples in each sample set, and use the minimum distance between samples as the distance between corresponding sample sets; The sample set with the largest spacing is retained, and the samples in the sample set are used as sampling points.

[0023] Based on the above description, the core idea of ​​Maximin LHS (Latin Hypercube Sampling) is to select the set of samples that maximizes the minimum Euclidean distance between any two points from all sample sets satisfying the LHS constraints. Its mathematical expression is: x i For the i-th sampling point, x j Let j be the j-th sampling point.

[0024] In an exemplary embodiment of this application, in order to perform high-fidelity numerical simulation, step 102 above can employ a high-fidelity CFD calculation method. This involves performing dynamic transition simulation based on sampling points to obtain the transition corridor boundary corresponding to each sampling point; that is, finding the matching relationship between the rotor and wing aerodynamic forces under different rotor tilt angles and forward flight speeds. The transition corridor boundary is then quantized to obtain transition corridor boundary information. Specifically, quantizing the transition corridor boundary means representing it as a discrete point set, envelope function, or rasterized image.

[0025] The core objective of quantifying information at the boundary of the transition corridor is to transform the continuous, two-dimensional forward velocity... V Rotor tilt angle The feasible region boundary is transformed into a set of discrete numerical sequences that can be processed by machine learning models. Based on this, the process of quantifying the information of the transition corridor boundary includes: Step 1: Define the state space and constraints.

[0026] (1) The defined state space includes: forward velocity V (Unit: m / s or km / h) and rotor nacelle tilt angle (Unit: degree (°), usually defined as:) =90° is the vertical takeoff and landing mode. =0° is the level cruise mode.

[0027] (2) Physical / control constraints (determining boundary locations) include: 1) Lower boundary of the low-speed range: limited by wing stall or rotor vortex ring state (VRS). If the speed is too low and the rotor nacelle tilt angle... It has tilted forward, and insufficient lift is causing it to drop in altitude; 2) Upper boundary of the high-speed segment: limited by the engine's available power or the compressibility of the advancing blade / stall of the retreating blade. If the speed is too high and the rotor nacelle tilt angle... Insufficient forward tilt, excessive power, or aerodynamic instability.

[0028] In practical engineering, more attention is usually paid to the lower boundary (minimum permissible speed) because it is directly related to the transition safety.

[0029] Step 2: Discretize the tilt angle dimension.

[0030] Rotor nacelle tilt angle N tilt angle nodes are uniformly selected within the effective range.

[0031] The number of nodes N is set according to the required precision (typical value: 20–50). The denser the nodes, the more detailed the boundary description, but the higher the computational cost.

[0032] Step 3: Adjust the tilt angle of each rotor nacelle Solve for the corresponding boundary velocities.

[0033] For each fixed rotor nacelle tilt angle Perform the following operations through high-speed CFD simulation calculations: Step (1): Trim Analysis: Under fixed conditions... Under these conditions, adjust collective pitch, cyclic pitch, elevator and other control variables to find all possible steady-state level flight solutions (i.e., flight state with zero acceleration and stable attitude).

[0034] Step (2): Determine the feasible speed range: Scan speed, record all speed values ​​that meet the following conditions: 1) The aircraft is trimmable (a stable solution exists); 2) All state variables (angle of attack, pitch angle, power, etc.) are within safe limits; 3) The deflection of the control surface has not reached saturation.

[0035] Step (3): Extract boundary values: The lower boundary is represented as: .

[0036] The upper boundary is represented as: .

[0037] Step 4: Construct the standardized output vector Y.

[0038] Arranging the above boundary velocity sequence in order of tilt angle to form an N-dimensional output vector Y, we have: In the formula, T This indicates the matrix transpose.

[0039] This output vector Y is the digital representation of the lower boundary of the transition corridor corresponding to this tiltrotor aircraft configuration, and can be directly used as a label for supervised learning. The transition corridor boundary of a tiltrotor aircraft is the flight envelope formed by the forward airspeed and the rotor tilt angle as the XY axes.

[0040] Based on the above description, a large dataset is ultimately formed. Each data point in the dataset contains a set of design parameters (which can be encoded as a vector) and its corresponding transition corridor boundary (i.e., a sample pair). The dataset is divided according to a set ratio (e.g., 7:3), resulting in the training set and test set from step 103 above.

[0041] In an exemplary embodiment of this application, to establish a mapping model from design parameters to the transition corridor boundary, different machine learning networks can be selected based on the quantification results of the transition corridor boundary information. For example, if the transition corridor boundary information is a discrete set of points or parameters of the envelope function (the parameters of the envelope function are forward velocity or tilt angle), fully connected deep neural networks (DNNs) or graph neural networks (GNNs) can be used. DNNs have a simple structure but may struggle to capture the overall topological structure of the envelope. If the transition corridor boundary information is a rasterized image, a conditional image generation model can be constructed, using a variant of U-Net that performs well on image-to-image translation tasks to guide the network in generating a conditional envelope image, or a Conditional Generative Adversarial Network (cGAN), which excels at generating realistic images based on conditions, can be used. This approach may generate sharper, more physically intuitive boundaries.

[0042] In one exemplary embodiment of this application, in practical applications, before the design parameters are input into the model, the input design parameters can be normalized and subjected to simple data augmentation.

[0043] In an exemplary embodiment of this application, during the training process of the model in step 105 above, mean squared error (MSE) or mean absolute error (MAE) can be used to test the network model for point set or function class outputs; for image class outputs, pixel-level binary cross-entropy loss or standard adversarial loss plus L1 / L2 reconstruction loss can be used to test the network model. Furthermore, adaptive optimizers such as Adam can be used to train and optimize the network model.

[0044] Based on this, when evaluating the accuracy of the network model on the test set, metrics such as IoU (Intersection over Union) and Dice coefficient can be calculated for the image output to measure the overlap between the predicted and true envelopes. Several representative new configurations that were not involved in training are selected, and the envelopes predicted by the network model are compared with the true envelopes obtained from high-fidelity simulations for visualization analysis.

[0045] In one exemplary embodiment of this application, when designing a new tiltrotor aircraft, only the initial design parameters need to be input into a trained network model (i.e., the transition corridor boundary prediction model), and the predicted transition corridor envelope can be obtained within seconds. Using this rapid prediction result, the feasibility of the design scheme can be quickly determined, and subsequent expensive high-fidelity simulation resources can be focused on key areas near the predicted envelope boundary, rather than blindly searching across the entire velocity-altitude domain. This significantly saves computation time and cost, enabling rapid multi-scheme comparison and iterative optimization.

[0046] In one exemplary embodiment of this application, techniques such as Bayesian neural networks or Monte Carlo Dropout can also be introduced, allowing the network model to provide not only the prediction envelope but also the uncertainty of the prediction. This is crucial for engineering decision-making, helping transition corridor boundary prediction models predict unreliable design areas.

[0047] In an exemplary embodiment of this application, the specific implementation process of the tiltrotor transition corridor boundary prediction method provided above is described using a fully connected deep neural network as a machine learning network.

[0048] 1. Determine the set of input feature parameters: Key parameters characterizing the overall configuration of the tiltrotor aircraft are selected as the input vector X of the network model. These parameters include: maximum takeoff weight, rotor diameter, rotor disk load, wingspan and area, maximum engine power, thrust-to-weight ratio, center of gravity position, and rotor-wing relative layout.

[0049] 2. Define the output labels for the transition corridor: For each tiltrotor aircraft with a known configuration, its forward speed is determined through high-fidelity flight dynamics simulation. V Rotor nacelle tilt angle The stable transition feasible region on the two-dimensional plane is called the transition corridor.

[0050] To facilitate machine learning modeling, the corridor boundary is quantified as follows: within a preset tilt angle range (e.g., [0°, 90°], where 0° represents horizontal cruise and 90° represents vertical takeoff and landing, N tilt angle nodes are uniformly selected within this range. , ,..., , i=1,2,...,N.

[0051] For each The minimum / maximum permissible forward speed is determined through simulation. and (Usually focus on the lower boundary) (i.e., the boundary where speed does not decrease). The output label Y is defined as a velocity boundary sequence, such as Y=[ , ,..., ] T .

[0052] 3. Construct the sample dataset: Collect or generate M sets of sample pairs to form a training set D. Data sources can include historical simulations, publicly available literature, or a self-built high-fidelity transient process simulation platform.

[0053] 4. Data preprocessing: The input vector X is normalized; the output label Y can also be normalized, and the inverse transform parameters are recorded for subsequent prediction result reconstruction.

[0054] 5. Construct the network model architecture: A DNN is used as the network model, with input vector X and output N-dimensional label Y, corresponding to the minimum permissible forward speed at each tilt angle.

[0055] 6. Construct the loss function: A loss function is constructed using mean squared error (MSE) to measure the deviation between the predicted transition corridor boundary and the actual simulation boundary, as follows: .

[0056] In the formula, The value of the loss function. For the first j The lower boundary prediction results corresponding to each tilt angle node. For the first j The lower boundary corresponding to each tilt angle node.

[0057] 7. Conduct training and verification: The network is divided into training, validation, and test sets, and standard training strategies such as Adam optimizer, L2 regularization, and early stopping mechanism are adopted to ensure the generalization ability of the network model.

[0058] 8. Save the model: Save the weights and structure of the optimal network model to support subsequent inference deployment.

[0059] Based on the above description, the predictive analysis process for the transition corridor boundary of the novel tiltrotor aircraft mainly includes: 1. Input the design parameters for the new tiltrotor aircraft: The overall design parameters of the novel tiltrotor aircraft are extracted to form a new dataset, which is then preprocessed in the same way as the data used during training.

[0060] 2. Perform model inference: Input a new dataset into a trained DNN model (i.e., a transition corridor boundary prediction model), and output the predicted transition corridor boundary sequence.

[0061] 3. Results Visualization and Application: Will Point pairs drawn on On the plane, a predicted lower boundary curve of the transition corridor is formed, based on which it can be determined that at a certain tilt angle, the tiltrotor aircraft must maintain a speed no lower than the predicted speed to avoid stall or loss of control. This prediction result can also provide constraint boundaries for feasible trajectories in the automatic transition control law and facilitate rapid comparison of different configurations. The system offers greater flexibility and safety during in-plane transitions. Furthermore, by conducting high-fidelity simulations only near the transition corridor boundaries to verify stability margins, computational resources are significantly reduced. Specifically, when the tiltrotor's flight state is outside the transition corridor flight envelope (formed by the transition corridor boundaries), it will be unable to fly stably. Therefore, the flight control system must adjust the automatic transition control law to ensure that the tiltrotor's flight state remains within the transition corridor flight envelope.

[0062] Based on the above description, in this embodiment, the implementation architecture for predicting the transition corridor boundary of a tiltrotor aircraft based on a DNN model is as follows: Figure 2 As shown.

[0063] In summary, compared with the prior art, this application has at least the following advantages: 1. Significantly Improved Design Efficiency: This application constructs a machine learning-based network model that can predict the transition corridor boundary within seconds using only the key design parameters of the tiltrotor aircraft. This completely eliminates the reliance on high-fidelity dynamic simulations that take days or even weeks, as is traditional methods. This allows for rapid performance evaluation of numerous different configurations during the conceptual design and iteration phases of the tiltrotor aircraft, greatly shortening the development cycle.

[0064] 2. Significantly saves computing resources: This application transforms expensive high-fidelity simulation from full-envelope scanning to key area verification. In practical applications, preliminary transition corridor prediction results can be obtained first using the constructed transition corridor boundary prediction model. Then, only a small number of accurate simulations are performed near the prediction boundary for verification and fine-tuning, which can reduce computing resource consumption by more than 90% and effectively reduce R&D costs.

[0065] 3. Deep Mining of Data Value: This application effectively utilizes historically accumulated data (whether from simulations or flight tests) on different tiltrotor configurations and their corresponding transition corridor boundaries. By learning from this data through machine learning networks, the inherent, nonlinear mapping law between tiltrotor parameters and transition corridor boundaries is extracted, transforming scattered, isolated data points into reusable, systematic knowledge assets.

[0066] 4. Empowering Early Design Decisions: In the early stages of tiltrotor aircraft design, when detailed aerodynamic models are not yet complete, this application provides a reliable means of predicting transition corridor performance. This allows the transition corridor, a key performance indicator, to be incorporated into the early multidisciplinary optimization (MDO) framework, supporting more scientific and comprehensive design trade-offs and decisions, and avoiding major design rework due to the discovery of substandard transition performance later on.

[0067] 5. The model is highly versatile and easy to deploy: This application focuses on the overall design parameters of tiltrotor aircraft, without involving complex geometric details in modeling. Therefore, the constructed network model has good versatility and portability. The trained network model can be easily integrated into existing design software or evaluation platforms, thus providing a plug-and-play intelligent analysis tool.

[0068] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 3 As shown, the computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs in the non-volatile storage media to run. The database stores tiltrotor aircraft transition corridor boundary prediction data. The I / O interfaces are used for information exchange between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a tiltrotor aircraft transition corridor boundary prediction method.

[0069] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer equipment to which the present application is applied. Specific computer equipment may include, for example, [the following is a list of possible additional structures]. Figure 3 The diagram shows more or fewer components, or combinations of certain components, or different component arrangements.

[0070] In one exemplary embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0071] In one exemplary embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0072] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0073] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0074] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (RRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0075] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0076] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0077] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for predicting the boundary of a transition corridor for a tiltrotor aircraft, characterized in that, include: Determine the design space for the tiltrotor aircraft; The design space includes multiple design parameters; Sampling is performed in the design space to obtain sampling points; Dynamic transition simulation is performed based on the sampling points to obtain the transition corridor boundary information corresponding to each sampling point; Sample pairs are formed based on the sampling points and the corresponding transition corridor boundary information to generate training and test sets; Building network models based on machine learning networks; The network model is trained and tested using the training set and the test set until the trained network model meets the set conditions, and then the trained network model is obtained; the trained network model is used as the transition corridor boundary prediction model. The design parameters of the tiltrotor aircraft to be predicted are obtained in real time, and the prediction results of the transition corridor boundary are obtained by using the transition corridor boundary prediction model.

2. The method for predicting the transition corridor boundary of a tiltrotor aircraft according to claim 1, characterized in that, Sampling is performed in the design space to obtain sampling points, including: The sampling points are obtained by sampling in the design space using the Latin hypercube sampling method or an improved Latin hypercube sampling method.

3. The method for predicting the transition corridor boundary of a tiltrotor aircraft according to claim 2, characterized in that, The Latin hypercube sampling method is used to sample the design space to obtain the sampling points, including: The design space is divided into multiple equally probable intervals, and a sample value is randomly selected in each equally probable interval. The sample values ​​are randomly paired to form multiple two-dimensional sample points; The two-dimensional sample points are used as the sampling points.

4. The method for predicting the transition corridor boundary of a tiltrotor aircraft according to claim 2, characterized in that, An improved Latin hypercube sampling method is used to sample the design space to obtain the sampling points, including: Multiple sample sets are randomly generated within the design space using Monte Carlo methods. Determine the distance between samples in each sample set, and use the minimum distance between samples as the distance between corresponding sample sets; The sample set with the largest spacing is retained, and the samples in the sample set are used as the sampling points.

5. The method for predicting the transition corridor boundary of a tiltrotor aircraft according to claim 1, characterized in that, Dynamic transition simulation is performed based on the sampling points to obtain transition corridor boundary information corresponding to each sampling point, including: Using CFD calculation methods, dynamic transition simulation is performed based on the sampling points to obtain the transition corridor boundary corresponding to each sampling point; The information of the transition corridor boundary is obtained by performing information quantization processing on the boundary of the transition corridor.

6. The method for predicting the transition corridor boundary of a tiltrotor aircraft according to claim 5, characterized in that, Information quantization processing of the transition corridor boundary refers to representing the transition corridor boundary as a discrete point set, envelope function, or rasterized image.

7. The method for predicting the transition corridor boundary of a tiltrotor aircraft according to claim 1, characterized in that, The machine learning network can be any one of a fully connected deep neural network, a graph neural network, a U-Net network, or a conditional generative adversarial network.

8. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and capable of running on the processor, characterized in that the processor executes the computer program to implement the tiltrotor transition corridor boundary prediction method according to any one of claims 1-7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the tiltrotor transition corridor boundary prediction method as described in any one of claims 1-7.

10. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the tiltrotor transition corridor boundary prediction method as described in any one of claims 1-7.

Images