Longitudinal aerodynamic parameter identification and calculation method in small angle-of-attack range in closed-loop state
By optimizing the longitudinal aerodynamic model parameters using a genetic algorithm, the problem of aerodynamic characteristic identification within a small angle of attack range under UAV closed-loop control was solved, achieving efficient and accurate aerodynamic parameter identification, applicable to various UAV types and complex flight conditions.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-31
- Publication Date
- 2026-04-10
AI Technical Summary
In the closed-loop control state of UAVs, it is difficult to identify the longitudinal aerodynamic characteristics within a small angle of attack range, and existing technologies cannot accurately obtain aerodynamic parameters.
A combinatorial optimization method based on genetic algorithms is adopted. By acquiring the main parameters of the longitudinal aerodynamic model and flight test data, a flight simulation model is established to simulate the flight state changes of the UAV in a closed loop. The parameters of the longitudinal aerodynamic model are iteratively optimized using genetic algorithms. The optimal aerodynamic model parameters are identified by combining the sum of squared errors between the flight simulation data and the actual test data.
It improves the accuracy and efficiency of aerodynamic parameter identification, is applicable to different types of UAVs, and enhances the model's applicability under complex flight conditions.
Smart Images
Figure CN121835200A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The application belongs to the technical field of unmanned aerial vehicle design, and particularly relates to a longitudinal aerodynamic parameter identification calculation method in a closed loop state and in a small angle of attack range. BACKGROUND
[0002] Aircraft aerodynamic characteristic parameters are key factors in modern aircraft design, performance analysis and flight simulation, and are of great significance to the research and optimization of aircrafts. In actual engineering applications, the main means of obtaining aircraft aerodynamic characteristics include numerical simulation calculation, wind tunnel test and flight test verification.
[0003] Among them, the numerical simulation method is widely used due to its efficiency and economy, but the results are often limited by factors such as calculation model accuracy and grid division; while the wind tunnel test can provide more accurate data, but due to the limitation of ground facility conditions, it is difficult to completely reproduce the complex flow state in the real flight environment. In contrast, the aerodynamic parameters obtained through actual flight test have higher reference value and application value because they directly reflect the working characteristics of the aircraft in the real atmospheric environment.
[0004] For the identification of longitudinal aerodynamic characteristics of unmanned aerial vehicles in a closed loop control state and in a small angle of attack range, an optimization calculation method based on flight data is proposed. SUMMARY
[0005] In order to solve the above problems, the application provides a longitudinal aerodynamic parameter identification calculation method in a closed loop state and in a small angle of attack range, to solve the problem that the longitudinal aerodynamic characteristics of unmanned aerial vehicles in a closed loop control state and in a small angle of attack range are difficult to identify in the prior art.
[0006] The technical scheme of the application is: a longitudinal aerodynamic parameter identification calculation method in a closed loop state and in a small angle of attack range,
[0007] Preferably, it comprises:
[0008] Obtaining the main parameters of the longitudinal aerodynamic model and the key parameters in the flight test data;
[0009] Based on the main parameters of the longitudinal aerodynamic model and the key parameters in the flight test data, a longitudinal aerodynamic model and a flight simulation model are established, the flight state change history of the unmanned aerial vehicle in a closed loop state and in a small angle of attack range is simulated based on the flight simulation model, and the dynamic characteristics of the aircraft under different longitudinal aerodynamic conditions are obtained;
[0010] The dynamic characteristics of the aircraft under different longitudinal aerodynamic conditions are calculated to obtain flight simulation data, and further to obtain the error sum of squares between the flight simulation data and the actual flight test data;
[0011] The longitudinal aerodynamic model parameter combination set is iteratively optimized by a genetic algorithm, and the optimal longitudinal aerodynamic model parameter combination is obtained by combining the error sum of squares between flight simulation data and actual flight test data;
[0012] The optimal longitudinal aerodynamic model parameters are simulated and analyzed, and when the preset accuracy requirement is met, the verification is completed.
[0013] Preferably, the longitudinal aerodynamic model includes a lift coefficient model, a drag coefficient model, and a pitch moment coefficient model.
[0014] Preferably, the flight simulation model is modeled based on a six-degree-of-freedom kinematics equation or a simplified dynamics model.
[0015] Preferably, in the flight simulation model, the operation instructions of the pilot in actual flight or the control input generated by the automatic pilot system are collected and input into the flight simulation model, different flight control strategies are set for flight simulation, and dynamic response data of different flight control strategies are obtained.
[0016] Preferably, the method for parameter identification calculation is:
[0017] According to the design experience and existing data of the aircraft, an initial longitudinal aerodynamic model parameter combination set containing different longitudinal aerodynamic model parameters is generated;
[0018] Based on the initial longitudinal aerodynamic model parameter combination set, a plurality of longitudinal aerodynamic model parameter combinations are randomly generated.
[0019] For each longitudinal aerodynamic model parameter combination, the flight simulation model is run to simulate the flight trajectory and dynamic response of the aircraft under the action of different control laws, and the error sum of squares between the flight simulation data and the actual flight test data is calculated.
[0020] Preferably, the error sum of squares between the flight simulation data and the actual flight test data is calculated by using a mean square error function.
[0021] Preferably, when simulating the flight trajectory and dynamic response of the aircraft under the action of different control laws, the elevator deflection angle and throttle change data under different flight states are simultaneously input.
[0022] Preferably, the specific method for iteratively optimizing the longitudinal aerodynamic model parameter combination set by the genetic algorithm is:
[0023] The longitudinal aerodynamic model parameter combination set is taken as the initial population of the genetic algorithm;
[0024] Fitness evaluation: calculate the error squares of all longitudinal aerodynamic model parameter combinations and calculate the fitness values of each individual;
[0025] Genetic operation: the individuals with higher fitness are reserved by selection operation, and new longitudinal aerodynamic model parameter combinations are generated by crossover operation to obtain iteratively convergent individuals;
[0026] The fitness evaluation and genetic operation process is repeated until the maximum iteration number is reached or the fitness value converges, and the individual with the highest fitness is selected as the optimal longitudinal aerodynamic model parameter.
[0027] Preferably, when generating new longitudinal aerodynamic model parameter combinations by crossover operation, random disturbance is introduced by mutation operation.
[0028] Preferably, the initial state of the simulation calculation is set to be consistent with the initial state of the actual flight test data, the flight simulation results are compared with the actual test data, and the error sum of squares of the flight simulation data and the actual flight test data and the fitness of the current longitudinal aerodynamic model parameter are calculated.
[0029] Preferably, the method for obtaining the actual flight test data is to perform a complete longitudinal step response test under the specified configuration and height conditions, and to collect real flight data.
[0030] The longitudinal aerodynamic parameter identification calculation method in the closed loop state under small angle of attack range of the present application has the following advantages:
[0031] The parameter identification is performed by using a combination optimization method, and the optimization method is a genetic algorithm, which has low requirements for the initial values of the longitudinal aerodynamic coefficients in the aerodynamic model module;
[0032] The range of the angle of attack in the flight simulation data and the flight test data should not be too large to ensure that the longitudinal aerodynamic coefficients in the small angle of attack range change approximately linearly with the angle of attack;
[0033] The influence of the control law on the flight test during the flight process is considered, and the longitudinal aerodynamic coefficient parameter identification under the closed loop state is more accurate;
[0034] It is universal and suitable for longitudinal aerodynamic coefficient parameter identification of different types of unmanned aerial vehicles, improves the efficiency of parameter identification, and enhances the applicability of the model under complex flight conditions. BRIEF DESCRIPTION OF DRAWINGS
[0035] Figure 1 The algorithm flowchart for solving and calculating the longitudinal aerodynamic model parameters of the present application by using a combination optimization method;
[0036] Figure 2 The graph of the change of the aircraft state parameters over time of the present application. DETAILED DESCRIPTION
[0037] For the purpose, technical solutions and advantages of the embodiments of the present application to be clearer, the technical solutions in the embodiments of the present application will be described in more detail below with reference to the drawings in the embodiments of the present application. In the drawings, the same or similar notations represent the same or similar elements or elements with the same or similar functions throughout. The described embodiments are part of the embodiments of the present application, not all of the embodiments. The embodiments described below by reference to the drawings are exemplary and are intended to explain the present application, and cannot be understood as limiting the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of protection of the present application. The embodiments of the present application will be described in detail below with reference to the drawings.
[0038] The first aspect of the present application provides a longitudinal aerodynamic parameter identification calculation method in a closed loop state and a small angle of attack range. First, the main parameters of the longitudinal aerodynamic model and the parameters that should be included in the flight test data are determined, and then the corresponding flight simulation model is established. It should be noted that the effect of the control law must be fully considered in the modeling process; The method of combined optimization is used for longitudinal aerodynamic model parameter identification calculation. First, the flight simulation data under the given longitudinal aerodynamic model is calculated (such as Figure 1 ), then the error sum of squares of the flight simulation data and the flight test data is calculated, and finally the genetic algorithm is used to optimize the calculation according to the error sum of squares under different longitudinal aerodynamic model parameters, as shown in Figure 2 , the final longitudinal aerodynamic model parameters are obtained.
[0039] As shown in Figures 1-2 , the method comprises the following steps:
[0040] Step 1, obtain the main parameters of the longitudinal aerodynamic model and the key parameters in the flight test data.
[0041] These parameters usually include lift coefficient, drag coefficient, pitch moment coefficient and other aerodynamic parameters. At the same time, the necessary information that should be included in the flight test data is determined, such as flight altitude, speed, attitude angle and its rate of change and other dynamic parameters. These parameters and data will provide the basis for subsequent modeling and optimization.
[0042] The longitudinal aerodynamic model includes the lift coefficient model, the drag coefficient model and the pitch moment coefficient model.
[0043] Step 2, based on the main parameters of the longitudinal aerodynamic model and the key parameters in the flight test data, establish the longitudinal aerodynamic model and the flight simulation model, simulate the flight state change history of the unmanned aerial vehicle under the condition of closed loop state and small angle of attack range based on the flight simulation model, and obtain the dynamic characteristics of the aircraft under different longitudinal aerodynamic conditions.
[0044] Flight simulation models are essentially a representation of Newton's second law, and flight simulation data refers to the flight state change history calculated using the current longitudinal aerodynamic model and the flight simulation model. The longitudinal aerodynamic model refers to the law governing how the aircraft's aerodynamic characteristics change with flight conditions.
[0045] Preferably, the flight simulation model is modeled based on six-degree-of-freedom kinematic equations or simplified dynamic models (such as point mass models).
[0046] During flight simulation, the pilot's operating commands or the control inputs generated by the autopilot system during actual flight are collected and input into the flight simulation model. Different flight control strategies are set to conduct flight simulation and obtain dynamic response data for different flight control strategies, so as to ensure that the model can truly reflect the dynamic response of the aircraft under different control strategies.
[0047] Step 3: Perform parameter identification and calculation on the dynamic characteristics of the aircraft under different longitudinal aerodynamic conditions to obtain flight simulation data, and further obtain the sum of squared errors between the flight simulation data and the actual flight test data.
[0048] Preferably, the method for parameter identification and calculation is as follows:
[0049] Step 3.1: Generate the initial parameter set. Based on the aircraft design experience and existing data, generate a set of initial longitudinal aerodynamic model parameter combinations containing different longitudinal aerodynamic model parameters.
[0050] Step 3.2: Simulation and Evaluation. For each combination of longitudinal aerodynamic model parameters, run the flight simulation model to simulate the flight trajectory and dynamic response of the aircraft under different control laws. It is particularly important to input elevator deflection angle and throttle variation data under different flight conditions simultaneously when simulating the flight trajectory and dynamic response of the aircraft under different control laws to ensure the comprehensiveness and accuracy of the simulation results.
[0051] Step 3.3: Calculate the sum of squared errors. By defining an appropriate evaluation function (such as mean squared error), the difference between the simulation data and the experimental data is quantified to obtain the sum of squared errors for each candidate parameter set.
[0052] Step 4: Iteratively optimize the set of longitudinal aerodynamic model parameters using a genetic algorithm, and obtain the optimal set of longitudinal aerodynamic model parameters by combining the sum of squared errors between flight simulation data and actual flight test data.
[0053] To achieve efficient searching for the optimal longitudinal aerodynamic model parameters, this invention introduces a genetic algorithm. The specific steps are as follows:
[0054] Step 4.1: Population Initialization. Use the initial parameter set as the initial population for the genetic algorithm.
[0055] Step 4.2: Fitness Assessment. Calculate the fitness value for each individual sample (i.e., each set of parameters) based on the sum of squared errors (or the reciprocal of the sum of squared errors).
[0056] Step 4.3: Selection, Crossover, and Mutation. Individuals with high fitness are retained through selection, new combinations of longitudinal aerodynamic model parameters are generated through crossover, and random perturbations are introduced through mutation to enhance population diversity.
[0057] Step 4.4: Iterative Optimization. Repeat the fitness evaluation and genetic operations until the preset termination conditions are met (such as reaching the maximum number of iterations or fitness value convergence). Finally, a set of optimal parameters is obtained, enabling the simulation results to achieve the best match with the experimental data under the control law.
[0058] Step 5: Perform simulation verification and error analysis on the optimal longitudinal aerodynamic model parameters. Once the preset accuracy requirements are met, the verification is complete.
[0059] Preferably, during simulation verification, the current longitudinal aerodynamic model parameters are input into the flight simulation model, the initial state of the simulation calculation is set to be consistent with the initial state of the actual flight test data, the flight simulation results are compared with the actual test data, and the sum of squared errors between the flight simulation data and the actual flight test data and the fitness of the current longitudinal aerodynamic model parameters are calculated.
[0060] The method for obtaining actual flight test data is as follows: a complete longitudinal step response test is conducted under specified configuration and altitude conditions to collect real flight data.
[0061] If the model performs poorly under certain conditions, it may be necessary to readjust the parameters or optimize the strategy until the accuracy requirements are met. Ultimately, through this systematic identification and optimization process, an aerodynamic model that accurately reflects the longitudinal motion characteristics of the aircraft can be obtained, providing a reliable basis for subsequent flight control design and performance analysis.
[0062] In summary, this application has the following advantages:
[0063] Parameter identification is performed using a combinatorial optimization approach, with the optimization method being a genetic algorithm. This approach has relatively low requirements for the initial values of the longitudinal aerodynamic coefficients in the aerodynamic model module.
[0064] The range of angle of attack variation in flight simulation data and flight test data should not be too large, so as to ensure that the longitudinal aerodynamic coefficient changes approximately linearly with the angle of attack within a small angle of attack range;
[0065] The influence of the control law on the flight test during flight was taken into account, and the longitudinal aerodynamic coefficient parameter identification under the corresponding closed-loop state was more accurate.
[0066] It is versatile and suitable for identifying longitudinal aerodynamic coefficient parameters of different types of UAVs, improving the efficiency of parameter identification and enhancing the applicability of the model under complex flight conditions.
[0067] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art 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 identifying and calculating longitudinal aerodynamic parameters in a small angle-of-attack range under closed-loop conditions, characterized in that, include: Obtain the main parameters of the longitudinal aerodynamic model and the key parameters from the flight test data; Based on the main parameters of the longitudinal aerodynamic model and the key parameters in the flight test data, a longitudinal aerodynamic model and a flight simulation model are established. Based on the flight simulation model, the flight state change process of the UAV under closed-loop conditions with a small angle of attack is simulated, and the dynamic characteristics of the aircraft under different longitudinal aerodynamic conditions are obtained. The dynamic characteristics of the aircraft under different longitudinal aerodynamic conditions are parameterized and calculated to obtain flight simulation data, and the sum of squared errors between the flight simulation data and the actual flight test data is further obtained. The optimal combination of longitudinal aerodynamic model parameters is obtained by iteratively optimizing the set of parameters using a genetic algorithm and combining the sum of squared errors between flight simulation data and actual flight test data. The optimal longitudinal aerodynamic model parameters are simulated and verified, and error analysis is performed. The verification is completed once the preset accuracy requirements are met.
2. The method for identifying and calculating longitudinal aerodynamic parameters in a small angle-of-attack range under closed-loop conditions as described in claim 1, characterized in that, The longitudinal aerodynamic model includes the lift coefficient model, the drag coefficient model, and the pitching moment coefficient model.
3. The method for identifying and calculating longitudinal aerodynamic parameters in a small angle-of-attack range under closed-loop conditions as described in claim 1, characterized in that, The flight simulation model is modeled based on the six-degree-of-freedom kinematic equations or simplified dynamic models.
4. The method for identifying and calculating longitudinal aerodynamic parameters in a small angle-of-attack range under closed-loop conditions as described in claim 1, characterized in that, In the flight simulation model, the pilot's operating commands or the control inputs generated by the autopilot system during actual flight are collected and input into the flight simulation model. Different flight control strategies are set to conduct flight simulation and obtain dynamic response data of different flight control strategies.
5. The method for identifying and calculating longitudinal aerodynamic parameters in a small angle-of-attack range under closed-loop conditions as described in claim 1, characterized in that, The method for parameter identification and calculation is as follows: Based on aircraft design experience and existing data, an initial set of longitudinal aerodynamic model parameter combinations is generated, containing different longitudinal aerodynamic model parameters. Multiple longitudinal aerodynamic model parameter combinations are randomly generated based on the initial set of longitudinal aerodynamic model parameter combinations; For each combination of longitudinal aerodynamic model parameters, a flight simulation model is run to simulate the flight trajectory and dynamic response of the aircraft under different control laws, and the sum of squared errors between the flight simulation data and the actual flight test data is calculated.
6. The method for identifying and calculating longitudinal aerodynamic parameters in a small angle-of-attack range under closed-loop conditions as described in claim 5, characterized in that, The mean square error function is used to calculate the sum of squared errors between flight simulation data and actual flight test data.
7. The method for identifying and calculating longitudinal aerodynamic parameters in a small angle-of-attack range under closed-loop conditions as described in claim 5, characterized in that, When simulating the flight trajectory and dynamic response of an aircraft under different control laws, the elevator deflection angle and throttle change data under different flight states are simultaneously input.
8. The method for identifying and calculating longitudinal aerodynamic parameters in a small angle-of-attack range under closed-loop conditions as described in claim 7, characterized in that, The specific method for iteratively optimizing the parameter combination set of the longitudinal aerodynamic model using a genetic algorithm is as follows: The set of parameters for the longitudinal aerodynamic model is used as the initial population for the genetic algorithm. Perform fitness assessment: Calculate the squared error of all longitudinal aerodynamic model parameter combinations and calculate the fitness value for each individual; Genetic operation: By selecting individuals with high fitness, a new combination of longitudinal aerodynamic model parameters is generated using a crossover operation, resulting in individuals that converge iteratively. Repeat the fitness evaluation and genetic operation process until the maximum number of iterations is reached or the fitness value converges, and select the individual with the highest fitness as the optimal longitudinal aerodynamic model parameter.
9. The method for identifying and calculating longitudinal aerodynamic parameters in a small angle-of-attack range under closed-loop conditions as described in claim 8, characterized in that, When generating new combinations of longitudinal aerodynamic model parameters using cross operations, random perturbations are introduced through mutation operations.
10. The method for identifying and calculating longitudinal aerodynamic parameters in a small angle-of-attack range under closed-loop conditions as described in claim 8, characterized in that, During simulation verification, the current longitudinal aerodynamic model parameters are input into the flight simulation model. The initial state of the simulation calculation is set to be consistent with the initial state of the actual flight test data. The flight simulation results are compared with the actual test data, and the sum of squared errors between the flight simulation data and the actual flight test data, as well as the fitness of the current longitudinal aerodynamic model parameters, are calculated.
11. The method for identifying and calculating longitudinal aerodynamic parameters in a small angle-of-attack range under closed-loop conditions as described in claim 1, characterized in that, The method for obtaining actual flight test data is as follows: a complete longitudinal step response test is conducted under specified configuration and altitude conditions to collect real flight data.