Real-time generation method for influence domain of air separation of a combination aircraft
By using quasi-steady-state numerical analysis and machine learning methods, the influence domain of aerial separation of combined aircraft is generated, which solves the problems of large computational load and long cycle during the separation of combined aircraft, realizes the rapid determination of the appropriate separation trajectory of individual aircraft, and supports real-time control.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- NANJING UNIV OF AERONAUTICS & ASTRONAUTICS
- Filing Date
- 2023-03-22
- Publication Date
- 2026-05-19
AI Technical Summary
Existing technologies involve large computational loads and long cycles in analyzing aerodynamic interference problems during the separation of combined aircraft, making it difficult to quickly determine the appropriate separation trajectory for a single aircraft.
Using quasi-steady-state numerical analysis and machine learning methods, the airborne separation influence domain of the combined aircraft is generated through numerical simulation and algorithms. Combining the sliced infinitesimal method and program algorithms, the aerodynamic influence domain of the individual aircraft is generated in real time, simplifying the model to determine a suitable separation corridor online.
It enables the rapid and efficient generation of separation trajectories for single-unit aircraft, solving the problems of large computational load and long cycle time, and providing real-time control support for the separation process of combined aircraft.
Smart Images
Figure CN116451604B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of aerodynamic analysis technology for aircraft, and in particular to a method for real-time generation of the airborne separation influence domain of a combined aircraft. Background Technology
[0002] The aerodynamic impact of separation from combined aircraft is one of the important issues that must be considered in the development of various aerospace combined aircraft. The main goal of the study on the aerodynamic influence domain of combined aircraft separation is to quantify and visualize the aerodynamic characteristics influence domain generated by the unsteady flow field during separation. At the same time, in order to ensure the attitude stability of the individual aircraft after separation, a suitable separation flight corridor must be found within the influence domain.
[0003] The aerodynamic interference problem during the separation of combined aircraft is caused by the relative motion and turbulent flow field during the separation transient. To obtain the aerodynamic influence region during the separation of a single aircraft, aerodynamic analysis of a large number of calibration points within a limited space is required. Currently, the main methods used to address the aerodynamic interference problem during the separation of combined aircraft include theoretical analysis and numerical simulation. Theoretical analysis qualitatively narrows the separation influence region of the combined aircraft and determines the implementation scheme and approach for subsequent numerical simulations. Numerical simulation primarily employs CFD simulation, performing force analysis on each mesh node to compare the differences in forces experienced by a single aircraft in steady and unsteady flow fields. However, the large amount of data at the test points leads to a long simulation analysis cycle.
[0004] In the aerospace field, the aerodynamic impact of separation from combined aircraft has always been a complex aerodynamic analysis problem involving coupled relative motion in a complex and irregular space flow field. Previous studies have often employed indirect prediction (IPM) methods to analyze the actual separation trajectory of a single aircraft under aerodynamic disturbances. This involves using correlation simulation to generate a database of disturbance aerodynamic forces of the separated body at different positions and attitudes within the carrier's disturbed flow field. These forces are then input into the six-degree-of-freedom equations of motion of the separated body for simulation. This method is also known as Dynamic Modeling and Simulation (DMS). The database is established using a grid measurement method: first, the static disturbance forces and moments of the separated body at different positions and attitude angles under the combined aircraft's disturbed flow field are measured, thus forming a static database. However, its drawbacks are obvious: the number of grid points required for calculation is enormous, with most data points being wasted and consuming a significant amount of unnecessary computational space. Summary of the Invention
[0005] The key problem addressed by this invention is to simplify the model of the aerodynamic influence domain in space, based on the aerodynamic force variation law in three axes in the spatial coordinate system obtained from limited numerical simulation. From the perspective of mathematical modeling, a real-time generation method for the airborne separation influence domain of combined aircraft is proposed. This method obtains the aerodynamic influence domain after the separation of individual aircraft and generates it online at different altitudes and speeds. Subsequently, by combining the slicing micro-element method and a certain program algorithm, a suitable separation corridor can be quickly determined.
[0006] The main objective of this invention is to solve the problem of online generation of the influence domain of the turbulent flow field on the separation path of individual aircraft during the separation of aerospace combined aircraft, based on a limited number of aerodynamic data obtained by numerical simulation and algorithmic methods, and to provide means to support the design of the optimal separation trajectory of individual aircraft in combined aircraft.
[0007] To achieve the above objectives, the present invention adopts the following technical solution:
[0008] A method for real-time generation of an aerial separation impact box for a combined aircraft includes the following steps:
[0009] Step 1: Using the quasi-steady-state numerical analysis method, assuming a small disturbance environment for the single aircraft, the sub-single aircraft is placed at different positions in the spatial flow field generated by the main aircraft. Numerical simulation is used to conduct numerical analysis on the aerodynamic force state in the three axes at each calibration point. The simulation will determine the influence range that is conducive to increasing the lift of the single aircraft and reducing aerodynamic drag, i.e., the aerodynamic influence range.
[0010] Step 2: Preliminary determination of the aerodynamic influence domain; Taking the parallel combined aircraft configuration as the research object, based on the aircraft wake effect model, analyze the trajectory trend of the single aircraft after separation. When it is in the upwash region of the wake, it is conducive to increasing the lift of the single aircraft. When it is in a certain range above and behind the main aircraft, it has the effect of drag reduction. Therefore, it can be preliminarily determined that the aerodynamic influence domain is roughly distributed in the space region above and behind the main aircraft on the right and in the symmetrical space region above and behind the left.
[0011] Step 3: Determine the boundaries of the influence domain; simplify the aerodynamic influence domain spatial model, using level flight as the initial sampling condition, and conduct numerical simulation analysis at a certain altitude and speed. The following boundaries are proposed to be constructed:
[0012] 1) Rear boundary in the longitudinal direction; First, analyze its axial boundary. Taking the separation point as the initial point, select several calibration points parallel to the rear of the main aircraft's longitudinal axis. The point where the rate of change of aerodynamic drag between two adjacent points approaches zero is defined as the boundary point. The plane containing this boundary point and perpendicular to the longitudinal axis is used as the rear boundary of the aerodynamic influence domain.
[0013] 2) Forward boundary in the longitudinal direction; the plane containing the transverse and vertical axes at the initial separation position of the single aircraft is taken as the forward boundary of the aerodynamic influence domain;
[0014] 3) Lower boundary in the vertical axis direction; the horizontal plane containing the canard chord of the single aircraft is taken as the lower boundary of the aerodynamic influence domain;
[0015] 4) Spanning inner boundary; the plane containing the wingtip section of the main aircraft is taken as the inner boundary of the aerodynamic influence box;
[0016] 5) Extending to the outer boundary; when the calibration point is on the lower boundary surface and parallel to the horizontal axis, the point where the lateral force on the single aircraft approaches zero is the lateral boundary point of the influence domain.
[0017] 6) Curved boundary in the vertical axis direction; Perform slice analysis parallel to the vertical axis direction at continuous positions in the vertical axis direction, select its calibration point, and approximate the boundary point where the rate of change of lift of adjacent points approaches zero;
[0018] The distances from the surface containing the boundary points in each of the above directions to the initial separation position constitute the parameter set of the current aerodynamic influence domain;
[0019] Step 4: Construct an influence domain parameter database; Based on Step 2 and Step 3 above, the aerodynamic influence domain size under the current flight state is initially enclosed by each boundary; Similarly, under several other flight conditions at different altitudes and speeds, Step 1 to Step 3 are repeated to obtain a finite number of influence domain parameter databases under different flight states.
[0020] Step 5: Machine learning model training and fitting; Based on the influence domain parameter database of the limited samples obtained in Step 4, firstly, clean the parameters in the database, discarding or supplementing some data. Then, segment the different classes of parameters of the influence bins in the database, select a deep learning method in machine learning, and fit the training set in the database. Use the remaining data to test and observe the parameter fitting effect. If the goodness of fit is less than 0.8, change the model and refit the training set until the goodness of fit is above 0.8.
[0021] Step Six: Algorithm Implementation; Based on the machine learning model trained in Step Five, a real-time generation algorithm is written using computer language to realize the real-time generation of the influence domain of the combined aircraft's air separation, providing input of environmental parameter variables for the air separation process control of the combined aircraft.
[0022] The aforementioned combined aircraft specifically refers to a parallel combined aircraft, and the aerodynamic characteristics such as the flight envelope of this type of combined aircraft have been determined.
[0023] Furthermore, in step three, a quasi-steady-state analysis method is adopted, meaning that the attitude of the individual aircraft remains unchanged during the separation process.
[0024] In step three, the sampling point spacing first increases and then gradually decreases, and finally, the boundary value is approximated using the bisection method.
[0025] In step three, the sampling points are passed as input parameters to the CFD solver. By solving the flow field under the disturbance of the combined aircraft, the drag, lateral force, and lift aerodynamic data of the single aircraft are obtained.
[0026] In step four, the parameters in the airborne separation influence domain database include not only the influence domain boundary parameters under different flight states, but also the flight angle of attack, flight altitude and speed, and environmental parameters under standard atmospheric conditions corresponding to each boundary parameter.
[0027] In step five, the fitting is not limited to using algorithms such as multi-layer neural networks in machine learning. Ultimately, the goodness of fit is guaranteed to be above 0.8.
[0028] In step six, the real-time generation algorithm program for the airborne separation influence domain is installed in the flight control system of the main aircraft computer to provide an environmental reference for the safe and stable separation of the individual aircraft.
[0029] The present invention adopts the above technical solution and has the following advantages compared with the prior art:
[0030] (1) This invention proposes the concept of separation influence domain of combined aircraft. The characterization method is realized by sensitivity analysis and model inverse problem solving. Through the proposed measurement and characterization method, the separation influence domain can be generated online. Subsequently, combined with the slice micro-element method and a certain program algorithm, a suitable separation corridor can be quickly determined.
[0031] (2) The method for real-time generation of the air separation influence domain of a combined aircraft proposed in this invention can effectively solve the problems of large computational load and long cycle in the research process of suitable separation trajectory design for single aircraft. Attached Figure Description
[0032] The present invention will be described by way of example and with reference to the accompanying drawings, wherein:
[0033] Figure 1 This is a flowchart of the online generation method for the separation impact box of the combined aircraft;
[0034] Figure 2 This is a conceptual diagram of the impact domain of the combined aircraft separation in the implementation case.
[0035] Figure 3 This is a graph showing the changing trend of lateral forces acting on individual aircraft during the separation process of the combined aircraft.
[0036] Figure 4This is a graph showing the trend of lift force changes experienced by individual aircraft during the separation process of the combined aircraft.
[0037] Figure 5 This is a graph showing the trend of drag changes experienced by individual aircraft during the separation process of the combined aircraft;
[0038] Figure 6 This is a demonstration diagram of the online generation algorithm for the separation influence domain of the combined aircraft. Specific Implementation
[0039] To enable those skilled in the art to better understand the technical solution of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.
[0040] For example, such as Figure 2 As shown, the online generation method for the separation influence box of the combined aircraft of the present invention is analyzed using the following configuration example:
[0041] Specifically, such as Figure 2 The combined aircraft shown is assumed to undergo symmetrical separation, with the sub-aircraft positioned at different locations within the wake field of the main space. A geometric model was exported using SOLIDWORDS, input into ICEM for mesh generation, and finally, FLUENT numerical analysis was used to obtain steady aerodynamic data for each different location.
[0042] The variations in lift, drag, and lateral force were observed when the single-unit aircraft was positioned above the main body, along the axial direction, and horizontally.
[0043] In this implementation case, the upper rear part of the main body, i.e. the suitable separation area, is taken as the research object. Under the cruise state with a flight altitude of H = 2000m and a flight speed of V = 66.2m / s, an example numerical simulation analysis is carried out, as shown in Tables 1a to 1c below.
[0044] First, it was observed that when the single aircraft was located at different lateral positions of the main aircraft, the magnitude of the lateral force it experienced gradually decreased, as shown in Table 1a. At the same time, the trend of change also tended to be gradual, thus indicating that the boundary of the influence domain had been reached.
[0045] Table 1a Lateral Force Analysis Table
[0046]
[0047] Similarly, regarding the range of the aerodynamic influence domain above the main aircraft, by translating the single aircraft directly above the initial separation position, the trend of lift variation was numerically simulated. Table 1b shows that the lift gradually decreases and tends to stabilize. Thus, the influence boundary of the directly above influence domain can be defined.
[0048] Table 1b Lift Force Analysis Table
[0049]
[0050] When studying the rear boundary of the aerodynamic influence domain, the single aircraft was placed at different positions directly behind the main aircraft. According to Table 1c, it can be seen that as the distance increases, the drag gradually increases and tends to stabilize. This indicates that the aircraft has escaped the influence boundary of the interfering airflow, that is, it has reached the rear boundary of the influence domain.
[0051] Table 1c Resistance Force Analysis Table
[0052]
[0053]
[0054] Combination Figures 3 to 5 As shown, based on the numerical simulation results above, it can be concluded that under the flight conditions of 2000m and 66.2m / s, the side boundary, upper boundary, and rear boundary of the separation aerodynamic influence box can be obtained at distances of 700mm, 4000mm, and 20000mm from the initial separation position, respectively.
[0055] Similarly, take the same steps, such as Figure 6 As shown in Table 2, the boundary parameters of the aerodynamic influence box at different altitudes and speeds can be obtained.
[0056] Table 2 Statistical Table of Influence Box Parameters
[0057] Height H / m Speed V / m / s Side boundary / mm upper boundary / mm Back boundary / mm 2000 66.2 700 4000 20000 4000 73.38 1100 4900 40000 6000 81.76 1400 5600 55000 8000 91.61 1600 6100 65000 10000 103.3 1700 6400 70000 11000 109.98 1800 6600 72000
[0058] After fitting curves to the above raw data, an online generation algorithm is implemented using the C++ programming language.
Claims
1. A method for real-time generation of the airborne separation influence domain of a combined aircraft, characterized in that, Includes the following steps: Step 1: Using the quasi-steady-state numerical analysis method, assuming a small disturbance environment for the single aircraft, the sub-single aircraft is placed at different positions in the spatial flow field generated by the main aircraft. Numerical simulation is used to conduct numerical analysis on the aerodynamic force state in the three axes at each calibration point. The simulation will determine the influence range that is conducive to increasing the lift of the single aircraft and reducing aerodynamic drag, i.e., the aerodynamic influence range. Step 2: Preliminary determination of the aerodynamic influence domain; Taking the parallel combined aircraft configuration as the research object, based on the aircraft wake effect model, analyze the trajectory trend of the single aircraft after separation. When it is in the upwash region of the wake, it is conducive to increasing the lift of the single aircraft. When it is in a certain range above and behind the main aircraft, it has the effect of drag reduction. Therefore, it can be preliminarily determined that the aerodynamic influence domain is roughly distributed in the space region above and behind the main aircraft on the right and in the symmetrical space region above and behind the left. Step 3: Determine the boundaries of the influence domain; simplify the aerodynamic influence domain spatial model, using level flight as the initial sampling condition, and conduct numerical simulation analysis at a certain altitude and speed. The following boundaries are proposed to be constructed: 1) Rear boundary in the longitudinal direction; First, analyze its axial boundary. Taking the separation point as the initial point, select several calibration points parallel to the rear of the main aircraft's longitudinal axis. The point where the rate of change of aerodynamic drag between two adjacent points approaches zero is defined as the boundary point. The plane containing this boundary point and perpendicular to the longitudinal axis is used as the rear boundary of the aerodynamic influence domain. 2) Forward boundary in the longitudinal direction; the plane containing the transverse and vertical axes at the initial separation position of the single aircraft is taken as the forward boundary of the aerodynamic influence domain; 3) Lower boundary in the vertical axis direction; the horizontal plane containing the separated canard chord of the single aircraft is taken as the lower boundary of the aerodynamic influence domain; 4) Spanning inner boundary; the plane containing the wingtip section of the main aircraft is taken as the inner boundary of the aerodynamic influence box; 5) Extending to the outer boundary; when the calibration point is on the lower boundary surface and parallel to the horizontal axis, the point where the lateral force on the single aircraft approaches zero is the lateral boundary point of the influence domain. 6) Curved boundary in the vertical axis direction; Perform slice analysis parallel to the vertical axis direction at continuous positions in the vertical axis direction, select its calibration point, and approximate the boundary point where the rate of change of lift of adjacent points approaches zero; The distances from the surface containing the boundary points in each of the above directions to the initial separation position constitute the parameter set of the current aerodynamic influence domain; Step 4: Construct an influence domain parameter database; Based on Step 2 and Step 3 above, the aerodynamic influence domain size under the current flight state is initially enclosed by each boundary; Similarly, under several other flight conditions at different altitudes and speeds, Step 1 to Step 3 are repeated to obtain a finite number of influence domain parameter databases under different flight states. Step 5: Machine learning model training and fitting; Based on the influence domain parameter database of the limited samples obtained in Step 4, firstly, clean the parameters in the database, discarding or supplementing some data. Then, segment the different classes of parameters of the influence bins in the database, select a deep learning method in machine learning, and fit the training set in the database. Use the remaining data to test and observe the parameter fitting effect. If the goodness of fit is less than 0.8, change the model and refit the training set until the goodness of fit is above 0.
8. Step Six: Algorithm Implementation; Based on the machine learning model trained in Step Five, a real-time generation algorithm is written using computer language to realize the real-time generation of the influence domain of the combined aircraft's air separation, providing input of environmental parameter variables for the air separation process control of the combined aircraft.
2. The method for real-time generation of the airborne separation influence domain of a combined aircraft according to claim 1, characterized in that, The aforementioned combined aircraft specifically refers to a parallel combined aircraft, and the aerodynamic characteristics such as the flight envelope of this type of combined aircraft have been determined.
3. The method for real-time generation of the airborne separation influence domain of a combined aircraft according to claim 1, characterized in that, In step three, a quasi-steady-state analysis method is adopted, that is, the attitude of the single aircraft remains unchanged during the separation process, and the single aircraft is translated to different positions in the spatial flow field generated by the main body for numerical simulation analysis.
4. The method for real-time generation of the airborne separation influence domain of a combined aircraft according to claim 1, characterized in that, In step three, the sampling point spacing first increases and then gradually decreases, and finally, the boundary value is approximated using the bisection method.
5. The method for real-time generation of the airborne separation influence domain of a combined aircraft according to claim 1, characterized in that, In step three, the sampling points are passed as input parameters to the CFD solver. By solving the flow field under the disturbance of the combined aircraft, the drag, lateral force, and lift aerodynamic data of the single aircraft are obtained.
6. The method for real-time generation of the airborne separation influence domain of a combined aircraft according to claim 1, characterized in that, In step four, the parameters in the airborne separation influence domain database include not only the influence domain boundary parameters under different flight states, but also the flight angle of attack, flight altitude and speed, and environmental parameters under standard atmospheric conditions corresponding to each boundary parameter.
7. The method for real-time generation of the airborne separation influence domain of a combined aircraft according to claim 1, characterized in that, In step five, the fitting is not limited to using algorithms such as multi-layer neural networks in machine learning. Ultimately, the goodness of fit is guaranteed to be above 0.
8.
8. The method for real-time generation of the airborne separation influence domain of a combined aircraft according to claim 1, characterized in that, In step six, the real-time generation algorithm program for the airborne separation influence domain is installed in the flight control system of the main aircraft computer to provide an environmental reference for the safe and stable separation of the individual aircraft.