Multi-scene simulation and data evaluation method for flight performance of unmanned aerial vehicle

Abstract

The invention relates to the technical field of unmanned aerial vehicles, in particular to an unmanned aerial vehicle flight performance multi-scene simulation and data evaluation method, which comprises the following steps of: firstly, establishing a three-dimensional physical model of an unmanned aerial vehicle by using computer aided design, and performing aerodynamics and structural mechanics simulation through computational fluid mechanics and finite element analysis; secondly, fusing real-time meteorological data and topographic information to simulate the flight performance of the unmanned aerial vehicle in a complex flight environment; a simulation process is accelerated by using high-performance calculation, and nonlinear parameters in a simulation model are optimized by applying a machine learning algorithm; finally, actual flight data of the unmanned aerial vehicle are used for calibration, sensor data are processed through Kalman filtering and other technologies, a lift coefficient and a resistance coefficient are inversely calculated, and a closed-loop optimization system is formed; the method can effectively improve simulation precision and multi-scene adaptability, reduces actual flight test times, reduces research and development cost and risks, and is suitable for performance evaluation and optimization of various unmanned aerial vehicles.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicles, and in particular to a method for multi-scenario simulation and data evaluation of unmanned aerial vehicle flight performance. Background Art

[0002] With the rapid development of drone technology, drones have been widely used in many fields such as military, logistics, agriculture, and environmental monitoring. However, in the actual application of drones, how to accurately evaluate their flight performance has become an important technical problem. Traditional flight performance evaluation methods mostly rely on actual flight tests, but this method is costly and risky, and it is difficult to fully cover various flight conditions in complex flight environments. Therefore, flight performance evaluation methods based on computer simulation have gradually become the focus of research. Through simulation technology, different flight scenarios can be simulated in a virtual environment, and the flight performance of drones can be estimated, greatly reducing the cost and risk of actual flight tests.

[0003] However, existing UAV simulation methods have several shortcomings. First, the simulation accuracy is not high enough, especially in aerodynamic modeling and structural mechanics modeling, simplified physical models are often used, ignoring important factors such as turbulence effects, complex terrain influences, and meteorological changes, resulting in a large difference between the simulation results and the actual flight performance. Secondly, the existing simulation technology lacks sufficient multi-scenario adaptability, especially when it comes to extreme flight conditions or complex environments (such as mountains, urban buildings, etc.), it is difficult to simulate real flight dynamics. Finally, the existing technology has a weak data verification and feedback mechanism, and cannot fully utilize actual flight data to calibrate and optimize the simulation model, resulting in insufficient credibility of the simulation results.

[0004] In order to solve the above problems, the present invention proposes a multi-scenario simulation and data evaluation method for UAV flight performance that combines high-precision aerodynamic simulation, structural mechanics simulation, real-time meteorological data fusion and high-performance computing. By using advanced simulation technologies such as computational fluid dynamics (CFD) and finite element analysis (FEM), combined with machine learning optimization algorithms, the present invention can accurately predict the flight performance of UAVs in a variety of complex environments, and calibrate through actual flight data to form a closed-loop optimization system, thereby improving the accuracy and adaptability of the simulation. This method provides strong technical support for the design, performance optimization and safety evaluation of UAVs. Summary of the invention

[0005] This paper proposes a multi-scenario simulation and data evaluation method for UAV flight performance, aiming to solve the problems of insufficient simulation accuracy, lack of data verification, and limited multi-scenario application capabilities in the prior art. By comprehensively using high-precision modeling, environmental data fusion, high-performance computing, and machine learning, this method can accurately simulate the performance of UAVs in complex flight environments and improve the reliability and practicality of simulation results.

[0006] Furthermore, this method starts with building a high-precision three-dimensional physical model of the drone. Using computer-aided design (CAD) software, detailed models of the drone's fuselage, wings, tail, propellers and other components are constructed to ensure that the size and shape of each component are consistent with the actual drone. For example, key parameters such as the wingspan, fuselage length, and propeller diameter are set according to actual design data. This refined model lays the foundation for subsequent simulation analysis.

[0007] Furthermore, in aerodynamic simulation, computational fluid dynamics (CFD) software is used to simulate the airflow around the drone. By setting boundary conditions such as flight speed, angle of attack, and air density, the aerodynamic parameters such as lift and drag of the drone under different flight conditions are solved. In order to more accurately simulate the turbulence effect in the actual flight environment, the standard k-ε turbulence model is used. This process can help us understand the aerodynamic performance of the drone under various flight conditions and provide data support for design optimization.

[0008] Furthermore, structural mechanics simulation is performed. Finite element analysis (FEM) software is used to apply the aerodynamic loads obtained from the aerodynamic simulation to the structural model of the UAV to analyze the stress and deformation of the UAV under flight loads. By selecting appropriate material properties (such as the elastic modulus and Poisson's ratio of carbon fiber composite materials), the strength and safety of the UAV structure can be evaluated, possible stress concentration areas can be identified, and the reliability of the structural design can be ensured.

[0009] Furthermore, this method integrates multi-source environmental data. By accessing real-time meteorological data such as wind speed, wind direction, temperature and humidity, the flight performance of the UAV under real meteorological conditions can be simulated. At the same time, the high-precision terrain model of the flight area is obtained using the Geographic Information System (GIS), and the terrain data is imported into the simulation environment to evaluate the impact of terrain undulations on airflow and UAV flight performance. This fusion of environmental data makes the simulation results closer to the actual situation.

[0010] Furthermore, in order to solve the huge amount of calculation required for high-precision simulation, this method introduces high-performance computing technology. Through GPU parallel computing, complex simulation tasks are divided into multiple subtasks and assigned to different computing cores for simultaneous calculation, which greatly shortens the simulation time. For example, after using GPU acceleration, the simulation calculation time can be shortened from dozens of hours to a few hours, improving work efficiency.

[0011] Furthermore, this method also applies machine learning algorithms to optimize the simulation model. By collecting historical flight data of the UAV under different flight conditions, the deep neural network model is trained to optimize key parameters in the simulation model, such as lift coefficient and drag coefficient. The machine learning model can better capture nonlinear and complex flight dynamics, and improve the adaptability and prediction accuracy of the simulation model.

[0012] Furthermore, data verification and model calibration are performed. Various sensors, such as inertial measurement unit (IMU), GPS and pitot tube, are installed on the drone to collect the flight data of the drone in real time. By processing and analyzing these data, the actual flight data is compared with the simulation results, the key parameters in the simulation model are back-calculated, and the model is calibrated. This process forms a closed-loop optimization system, which continuously improves the accuracy of the simulation model.

[0013] Beneficial effects of the present invention:

[0014] First, through high-precision aerodynamic simulation and structural mechanics simulation, combined with real-time meteorological data and complex terrain modeling, the present invention can accurately predict the flight performance of drones in a variety of flight environments. Compared with traditional simulation methods, the present invention significantly improves the simulation accuracy and can more realistically reflect the flight performance of drones in complex environments. In addition, the use of high-performance computing technology to accelerate the simulation process has greatly shortened the time required for simulation and improved the efficiency of simulation.

[0015] Secondly, the present invention introduces a machine learning algorithm and calibrates the simulation model in combination with actual flight data to establish a closed-loop optimization system. The system can continuously optimize the simulation model in multiple iterations to ensure that the error between the simulation results and the actual flight data is minimized, thereby improving the reliability of the simulation results. This solution not only reduces the number and cost of actual flight tests, but also enhances the adaptability of drones in complex scenarios, providing strong technical support for the design optimization and performance evaluation of drones. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0017] Figure 1 A schematic diagram of a closed-loop process of multi-scenario flight simulation and result calibration according to an embodiment of the present invention;

[0018] Figure 2 A schematic diagram of the overall system architecture of an embodiment of the present invention;

[0019] Figure 3 A schematic diagram of the architecture of an actual flight data collection system according to an embodiment of the present invention;

[0020] Figure 4 Schematic diagram of environmental data fusion according to an embodiment of the present invention. DETAILED DESCRIPTION

[0021] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. At the same time, it is explained here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments, and those skilled in the art may also adopt other alternatives to implement some known technologies; and the accompanying drawings are only for more specific description of the embodiments, and are not intended to specifically limit the present invention.

[0022] It should be noted that the references to "one embodiment", "embodiment", "exemplary embodiments", "some embodiments" and the like in the specification indicate that the embodiments described may include specific features, structures or characteristics, but not every embodiment may include the specific features, structures or characteristics. In addition, when a specific feature, structure or characteristic is described in conjunction with an embodiment, it should be within the knowledge of a person skilled in the art to implement such feature, structure or characteristic in conjunction with other embodiments (whether or not explicitly described).

[0023] In general, a term can be understood, at least in part, from its use in context. For example, depending, at least in part, on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in the singular sense, or can be used to describe a combination of features, structures, or characteristics in the plural sense. Additionally, the term "based on" can be understood as not necessarily intended to convey an exclusive set of factors, but can instead, depending, at least in part, on the context, allow for the presence of other factors that are not necessarily explicitly described.

[0024] In order to better illustrate the multi-scenario simulation and data evaluation method of the UAV flight performance of the present invention, the present invention is described in detail below in conjunction with specific embodiments. This embodiment is intended to help those skilled in the art to better understand and implement the present invention, but does not limit the scope of protection of the present invention.

[0025] The present invention provides a multi-scenario simulation and data evaluation method for UAV flight performance, aiming to improve the accuracy and reliability of UAV flight performance evaluation. The method realizes the performance evaluation of UAV in complex flight environment by establishing a high-precision simulation model, integrating multi-source environmental data, applying high-performance computing and machine learning algorithms, and strengthening data verification and model calibration. The method of the present invention will be described in detail below.

[0026] like Figures 1 to 4 Shown

[0027] 1. Establish a high-precision simulation model

[0028] 1. Construction of 3D physical model

[0029] 1.1 Establishment of geometric model

[0030] Use computer-aided design (CAD) software (such as SolidWorks) to build a 3D physical model of the fixed-wing UAV. The model should accurately reflect the actual structure and size of the UAV, including the following components:

[0031] Wing: The wingspan is set at 2 meters, with an elliptical airfoil, and the chord length gradually decreases from 0.5 meters at the root to 0.2 meters at the wingtip.

[0032] Fuselage: The length is set at 1.5 meters, the cross-section is elliptical, the maximum diameter is 0.3 meters, and the streamlined design is used to reduce resistance.

[0033] Tail: includes horizontal tail and vertical tail, the horizontal tail has a wingspan of 0.6 meters and the vertical tail has a height of 0.3 meters.

[0034] Propeller: Diameter is set to 30 cm, number of blades is 2, pitch is 10 inches.

[0035] 1.2 Grid division

[0036] After building the geometric model, use meshing software (such as ANSYSMeshing) to mesh the model. The specific steps are as follows:

[0037] Meshing type: Unstructured tetrahedral mesh is used to accommodate complex geometries.

[0038] Grid density: In key areas (such as the leading edge of the wing, the trailing edge, and near the propeller), a dense grid is used, with a minimum grid size of 1 mm; in areas far away from the fuselage, a coarser grid is used, with a size of up to 50 mm.

[0039] Total number of grids: controlled between 5 million and 10 million cells to balance simulation accuracy and computing resources.

[0040] 2. Aerodynamics Simulation

[0041] 1. Simulation settings

[0042] 2.1 Software Selection

[0043] Aerodynamic simulation was performed using computational fluid dynamics (CFD) software ANSYS Fluent.

[0044] 2.2 Physical model settings

[0045] Fluid Model: Select a three-dimensional, steady-state, incompressible flow model.

[0046] Turbulence model: The standard k-ε turbulence model is used, which is suitable for external flows with high Reynolds numbers.

[0047] Air Properties: Density

[0048] Dynamic viscosity Pa s

[0049] 2. Boundary condition setting

[0050] 2.3 Inlet Boundary

[0051] Speed ​​entry: set the flight speed m / s

[0052] Turbulence Intensity: Set to 5%

[0053] Turbulent Viscosity Ratio: Set to 10

[0054] 2.4 Export Boundary

[0055] Pressure outlet: set to ambient static pressure Pa

[0056] 2.5 Wall conditions

[0057] Drone surface: set to no-slip wall (speed is zero), hot wall temperature is set to ambient temperature K

[0058] Ground: If simulating low altitude flight, the ground is set as a moving wall with the same speed as the free stream to simulate ground effect.

[0059]

[0060] 2.6 Numerical methods

[0061] Pressure-velocity coupling: using SIMPLE algorithm

[0062] Discrete format: Second-order upwind format to improve computational accuracy

[0063] Convergence criteria: residual less than

[0064] 4. Calculation and result analysis

[0065] 2.7 Calculation process

[0066] Initialization: The initial velocity of the whole field is 30m / s, and the direction is consistent with the entrance velocity.

[0067] Iterations: Generally, 1000 to 5000 iterations are required, depending on the convergence

[0068] 2.8 Result Extraction

[0069] Lift coefficient and drag coefficient Calculation:

[0070]

[0071] in:

[0072] : The lift force on the drone is perpendicular to the direction of the incoming flow, and the software automatically calculates it.

[0073] The resistance force on the drone is the force along the direction of the incoming flow.

[0074] Pressure distribution and velocity field: Extract the pressure coefficient distribution on the drone surface and the velocity vector of the flow field to analyze the aerodynamic characteristics

[0075] 3. Structural Mechanics Simulation

[0076] 1. Simulation settings

[0077] 3.1 Software Selection

[0078] Structural mechanics simulation using finite element analysis (FEM) software ANSYS Mechanical

[0079] 3.2 Material properties

[0080] Wing and fuselage materials: Carbon fiber composites

[0081] Elastic modulus GPa

[0082] Poisson's ratio

[0083] density

[0084] 2. Loads and constraints

[0085] 3.3 Load application

[0086] Aerodynamic loads: Import the pressure distribution obtained from CFD simulation into the structural model as surface loads

[0087] Inertial load: Considering the weight of the drone, in the gravity field, the acceleration of gravity m / .

[0088] 3.4 Constraints

[0089] Fixed support: Since the drone is free and unconstrained during flight, there is no need to add

[0090] However, in order to prevent rigid body motion, a displacement constraint can be applied to a certain point on the fuselage.

[0091] 3. Solution and result analysis

[0092] 3.5 Solver Settings

[0093] Analysis type: Static analysis

[0094] Solution Control: Default Settings

[0095] 3.6 Result Extraction

[0096] Stress analysis: Extract the stress distribution of the wing and fuselage, focusing on the maximum principal stress and equivalent stress (vonMist

[0097] Deformation analysis: observe the deformation of the structure and ensure that the deformation is within the allowable range

[0098] 3.7 Safety Assessment

[0099] Strength check: Ensure that the maximum stress is less than the allowable stress of the material and calculate the safety factor:

[0100]

[0101] in,

[0102] 4. Fusion of multi-source environmental data

[0103] 1. Real-time weather data introduction

[0104] 4.1 Data Acquisition

[0105] Get real-time weather data for the flight area from a weather data API (such as OpenWeatherMap), including

[0106] Wind speed: 10m / s

[0107] Wind direction: Southwest (225°)

[0108] Temperature: 20°C

[0109] Humidity: 70%

[0110] 4.2 Simulation Adjustment

[0111] Wind speed and direction: In CFD simulation, the inlet velocity is adjusted to the synthetic incoming flow velocity, taking into account the synthesis of flight speed and wind speed.

[0112] Temperature and humidity: Adjust the density and viscosity of the air based on the standard atmospheric model.

[0113] 2. Terrain data integration

[0114] 4.3 Terrain model acquisition

[0115] Use GIS software (such as ArcGIS) to obtain digital elevation model (DEM) data of the flight area and generate a three-dimensional terrain model.

[0116] 4.4 The influence of terrain on airflow

[0117] Terrain import: Import the terrain model into the CFD simulation environment as the bottom boundary of the calculation domain.

[0118] Meshing: Encrypt the mesh on the terrain surface to capture the impact of terrain details on airflow.

[0119] Simulation run: Re-run the CFD simulation to observe the changes in airflow caused by terrain undulations, such as vortices, updrafts, etc.

[0120] 5. Applying high performance computing and machine learning to optimize models

[0121] High Performance Computing Acceleration

[0122] 5.1 GPU Parallel Computing

[0123] Hardware configuration: Use a high-performance computing server equipped with NVIDIA Tesla V100 GPU, each GPU has 5120 CUDA cores.

[0124] Parallel strategy: Utilize CUDA parallel computing to divide the computational domain into multiple subdomains, with each subdomain being computed by a GPU core.

[0125] Acceleration effect: The simulation time is shortened from the traditional 30 hours to 6 hours, improving computing efficiency.

[0126] Machine Learning Model Optimization

[0127] 5.2 Dataset Preparation

[0128] Training data: Collect historical data of the drone under different flight conditions, including flight speed, angle of attack, wind speed, and lift coefficient. and drag coefficient . Data volume: Contains at least 1,000 sets of valid data.

[0129] 5.3 Model construction

[0130] Algorithm selection: Use deep neural network (DNN) model.

[0131] Network structure:

[0132] Input layer: flight speed, angle of attack, wind speed; hidden layer: 3 layers, 128 neurons per layer, activation function is ReLU; output layer: lift coefficient and drag coefficient ,The activation function is a linear function.

[0133] 5.4 Model training; loss function:

[0134]

[0135] Optimization algorithm: Adam optimizer is used, and the learning rate is set to 0.001.

[0136] Training rounds: Iterate training for 1000 rounds and monitor the convergence of the loss function.

[0137] 5.5 Model Validation

[0138] Validation set: 20% of the dataset is divided as the validation set.

[0139] Evaluation indicators: Calculate the mean square error (MSE) and the coefficient of determination (

[0140] Coefficient of determination:

[0141]

[0142]

[0143] in, is the true value, is the predicted value, is the average of the true values.

[0144] 6. Data Verification and Model Calibration Actual Flight Data Collection 6.1 Sensor Configuration

[0145] IMU (Inertial Measurement Unit): measures the acceleration and angular velocity of the drone, with a sampling frequency of 100Hz. GPS module: records the location information of the drone, with an accuracy of less than 1 meter. Pitot tube: measures the relative air speed of the drone, with an accuracy of 0.1m / s. 6.2 Data Recording

[0146] Flight conditions: Test flights are conducted at different speeds, angles of attack and environmental conditions to collect multiple sets of data.

[0147] Data Storage: Use an onboard data logger to store data in real time or transmit wirelessly to a ground station.

[0148] Data processing: 6.3 Kalman filter

[0149] Filtering purpose: remove sensor noise and improve data accuracy.

[0150] Filter model: state vector: ,in For speed, is the angle of attack, is the pitch angle. State transfer equation:

[0151]

[0152] Observation equation:

[0153]

[0154] Noise assumptions: and is Gaussian white noise.

[0155] Filtering steps:

[0156] predict:

[0157]

[0158]

[0159]

[0160]

[0161]

[0162] 3. Simulation model calibration

[0163] 6.4 Parameter Inversion

[0164] Calculation of lift coefficient:

[0165]

[0166] in, , is the vertical acceleration, which comes from the IMU data. Calculation of the drag coefficient:

[0167]

[0168] in, , is the horizontal acceleration. 6.5 Model Adjustment

[0169] Comparative analysis: The actual calculated and Compare with the predicted value of the simulation model. Parameter adjustment: If the error exceeds the predetermined threshold (such as 5%), adjust the aerodynamic parameters in the simulation model and re-run the CFD simulation.

[0170] 1. Closed-loop optimization process

[0171] 7.1 Iteration Process

[0172] step:

[0173] 2. Test flight data collection: obtain actual flight data. 3. Data processing and calibration: calculate the actual and , adjust the model. 4. Model update: update simulation model parameters. 5. Re-simulation: perform new simulation calculations.

[0174] Iteration times: Usually 3-5 iterations are required, depending on the convergence of the model. 2. Multi-scenario simulation and evaluation 7.2 Flight scenario setting

[0175] Scenario 1: High wind speed, wind speed 20m / s, wind direction changes randomly. Scenario 2: High angle of attack flight, angle of attack reaches 15°, simulated climb and dive. Scenario 3: Complex terrain, such as mountainous areas and canyons, with dramatic terrain fluctuations. 7.3 Simulation and Analysis

[0176] Performance indicators: lift, drag, pitch moment, sideslip, etc. Stability evaluation: Analyze the flight stability of the UAV in different scenarios, such as whether there are stalls, sideslips, and other adverse phenomena. Maneuverability evaluation: Evaluate the response speed and control accuracy of the UAV. 7.4 Application of results

[0177] Design optimization: Based on the simulation results, the structure and control system of the UAV are optimized, such as adjusting the wing airfoil and increasing the stabilizer area.

[0178] Flight strategy formulation: formulate corresponding flight plans and countermeasures based on the simulation results of different scenarios. 8. Visualization and report generation 1. Visualization of simulation results 8.1 Software tools

[0179] 8.2 Graphical Display

[0180] Flow field visualization: Use ANSYS CFD-Post or Tecplot software. Structural result visualization: Use the post-processing tools provided by ANSYS Mechanical. Isobars and isobaric surfaces: Display the pressure distribution on the surface of the drone. Streamlines and vector diagrams: Display the direction and speed of airflow. Stress cloud diagram: Display the stress distribution of structural parts and identify high stress areas. Deformation diagram: Display the deformation of the structure and magnify the deformation for observation.

[0181] 2. Report Generation

[0182] 8.3 Report Content

[0183] Simulation settings: Detailed description of the simulation boundary conditions, model parameters, and calculation methods.

[0184] Result analysis: including numerical results and visual graphics to evaluate the performance of the UAV.

[0185] Optimization suggestions: Based on simulation and actual data, make suggestions for improvements in design and operation.

[0186] 8.4 Format and Output

[0187] Document format: PDF or Word format, including table of contents, chapters and appendices.

[0188] Chart clarity: Ensure all charts are of high resolution for easy reading and presentation.

[0189] IX. Precautions

[0190] 9.1 Data Accuracy

[0191] Sensor calibration: Before data collection, all sensors are calibrated to ensure that the data is accurate and reliable.

[0192] Data backup: Back up collected data in real time to prevent data loss.

[0193] 9.2 Computing Resources

[0194] Hardware requirements: High performance computing requires sufficient hardware support, and computing resources should be planned in advance.

[0195] Software licensing: Make sure all software used has a valid license.

[0196] 9.3 Security

[0197] Flight test safety: During actual flight tests, comply with flight regulations to ensure the safety of personnel and equipment.

[0198] Data security: Protect the confidentiality of data and prevent unauthorized access.

[0199] The present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention. In order to make the public have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, but those skilled in the art can fully understand the present invention without the description of these details. In addition, in order to avoid unnecessary confusion about the essence of the present invention, well-known methods, processes, procedures, components and circuits are not described in detail.

[0200] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A multi-scenario simulation and data evaluation method for UAV flight performance, characterized in that: The following steps are involved: Step 1: 3D physical model construction: Use computer-aided design (CAD) software to build a 3D geometric model of the drone, including the drone's fuselage, wings, tail, and propeller components, to ensure that the model can reflect the actual geometric dimensions of the drone; according to the drone design parameters, the wingspan is set to 2 meters, the fuselage length is set to 1.5 meters, and the propeller diameter is set to 30 centimeters; this geometric model is used for aerodynamic and structural mechanics analysis in subsequent simulations; Step 2: Aerodynamic simulation Use computational fluid dynamics (CFD) simulation software to input flight speed, angle of attack and air density parameters, where the flight speed is set to 30m / s, the angle of attack range is set to -5°C to 15°, and the air density is 1.225kg / m'; use the standard k-ε model in the turbulence model to simulate the turbulent airflow around the drone; solve the Navier-Stokes equations, calculate the airflow distribution on the surface of the drone, and output the lift coefficient and drag coefficient; obtain the lift coefficient of the drone at different angles of attack through CFD simulation and drag coefficient , used for flight performance evaluation; Step 3: Structural mechanics simulation: Use finite element analysis (FEM) software to perform structural mechanics simulation, apply the aerodynamic loads obtained by CFD simulation to the three-dimensional geometric model of the UAV; select material parameters, and input flight loads and boundary conditions; calculate the structural stress and strain of the UAV under flight loads, output stress concentration areas and deformation conditions, and evaluate the safety of the structural design; Step 4: Integrate real-time environmental data: Obtain real-time meteorological data from external meteorological data sources, including wind speed of 10m / s, wind direction of northeast wind, temperature of 20°C, and humidity of 60%, and input the data into the CFD simulation software; use the geographic information system GIS to obtain a high-precision terrain model of the flight area, use the digital elevation model DEM to generate a three-dimensional terrain map of the flight area, and import the terrain model into the CFD simulation environment to analyze the impact of terrain undulations on airflow and UAV flight performance; Step 4: High-performance computing and machine learning optimization: GPU parallel computing is used to accelerate CFD and FEM simulation tasks. The simulation tasks are divided into multiple subtasks and assigned to GPU cores for parallel computing to improve the simulation calculation speed. The deep neural network DNN algorithm in machine learning is used to optimize the aerodynamic parameters of the UAV simulation model. The neural network parameters are optimized by minimizing the following loss function: Step 6: Data verification and calibration: The simulation model is more accurate in nonlinear flight conditions by training the model with back propagation algorithm and actual flight data. An inertial measurement unit (IMU), GPS module and pitot tube are installed on the UAV to collect the position, speed, acceleration and air velocity data of the UAV in flight. The flight data is processed by Kalman filter to remove noise and outliers in the data. The update equation is: in, is the estimated value of the state, is the measured value, is the Kalman gain; the lift coefficient is calculated by inverse calculation using actual flight data , calculated by the following lift formula: Among them, lift , For drone quality, is the acceleration due to gravity, is the flight speed, is the air density, is the wing area; Step 7: Closed-loop optimization simulation model: The simulation model that has been verified and calibrated with actual data is reused for simulation, and the model parameters are iteratively optimized until the error between the simulation results and the actual flight data is less than a predetermined threshold; the simulation and data calibration process is repeated until a high-precision UAV flight performance evaluation model is obtained.

2. The multi-scenario simulation and data evaluation method for UAV flight performance according to claim 1 is characterized in that: In establishing the three-dimensional physical model, the wingspan is set to 2 meters, the fuselage length is 1.5 meters, and the propeller diameter is 30 centimeters. The specific parameters can be adjusted according to the actual design file of the UAV.

3. The multi-scenario simulation and data evaluation method for UAV flight performance according to claim 1 is characterized in that: The standard k-ε turbulence model is used in the CFD simulation process to simulate the turbulence effect of the airflow, so as to more accurately simulate the aerodynamic performance of the UAV at different angles of attack and wind speeds.

4. The multi-scenario simulation and data evaluation method for UAV flight performance according to claim 1 is characterized in that: The high-performance computing uses GPU parallel computing technology to divide the simulation task into 64 parallel computing units, and performs calculations on multiple GPU cores simultaneously, thereby reducing the simulation calculation time from 30 hours to 6 hours.

5. The multi-scenario simulation and data evaluation method for UAV flight performance according to claim 1 is characterized in that: The terrain model is constructed by using GIS data to generate a digital elevation model (DEM), and importing it into a simulation environment to analyze the impact of terrain undulations in the flight area on the flight performance of the UAV.

6. The multi-scenario simulation and data evaluation method for UAV flight performance according to claim 1 is characterized in that: The machine learning optimization process trains the lift coefficient and drag coefficient in historical flight data through a deep neural network (DNN) algorithm, and uses a back-propagation algorithm to adjust the neural network parameters to improve the adaptability and accuracy of the simulation model.

7. The multi-scenario simulation and data evaluation method for UAV flight performance according to claim 1 is characterized in that: During the data verification and calibration process, data is collected through actual flight and processed using a Kalman filter, aerodynamic parameters in the simulation model are back-calculated, and the model is iteratively calibrated multiple times to ensure that the error between the simulation result and the actual data is less than 1.1%.

8. The multi-scenario simulation and data evaluation method for UAV flight performance according to claim 1 is characterized in that: By inputting a variety of meteorological parameters and terrain models, different simulation scenarios are generated. The flight performance of the UAV under different airflow and terrain conditions is simulated when the wind speed ranges from 5m / s to 20m / s and the angle of attack ranges from -10° to 20°.

9. The multi-scenario simulation and data evaluation method for UAV flight performance according to claim 1 is characterized in that: The closed-loop optimization simulation model generates the flight performance curve of the UAV through multiple data feedback and model iterations, and provides flight strategy recommendations for different scenarios after reaching the set error threshold.

Images