CN-122018359-A - Method for constructing high-efficiency hybrid simulation system for unmanned aerial vehicle control design and evaluation

CN122018359ACN 122018359 ACN122018359 ACN 122018359ACN-122018359-A

Abstract

The invention discloses a method for constructing an efficient hybrid simulation system for unmanned aerial vehicle control design and evaluation, which can realize efficient and real turbulence simulation in a borderless space, support dynamic gathering and scattering of multiple unmanned aerial vehicles, obviously inhibit false waves caused by dynamic addition and deletion of grid blocks through self-adaptive blocks and Laplacian initialization, improve numerical stability, avoid boundary layer grid refinement by adopting ALM/ASM (ALM/ASM), couple tightly supporting cores to give consideration to precision and stability, enhance stability of convection opening boundaries under concurrent working conditions of inlet and outlet, be superior to the traditional Neumann/original convection conditions, simulate parameters with high efficiency, directly calibrate parameters through a gradient method, improve result reliability, and meet design iteration of a real-time or interactive level controller under the condition of capturing complex turbulence motion compared with the traditional CFD method.

Inventors

WANG JIWEI
LIU XIAOPEI
WANG YANG
SONG WENBIN

Assignees

上海科技大学

Dates

Publication Date: 20260512
Application Date: 20260112

Claims (10)

1. The method is characterized in that a simulation domain is decomposed into a far-field domain and a near-object domain, wherein the far-field domain adopts a GPU-optimized self-adaptive block lattice Boltzmann solver, the near-object domain adopts a parameterized algebraic empirical model, and the far-object field and the near-object field are subjected to force-dependent and fluid state exchange coupling so as to form an efficient and stable integrated FSI simulation system.
2. The method for constructing an efficient hybrid simulation system for unmanned aerial vehicle control design and evaluation according to claim 1, wherein for the far field domain, the infinite field is discretized into uniform cubes which do not overlap each other, each cube has a dimension of L, and each cube is disposed therein Is a uniform grid of (1), each cubic cell within a cubic block is of size And solving the self-adaptive block lattice Boltzmann solver in a cube, wherein the cubes which only allocate memory for simulation are called active blocks, and other cubes are called inactive blocks.
3. The method for constructing the efficient hybrid simulation system for unmanned aerial vehicle control design and evaluation according to claim 2, wherein collision is locally performed at each grid node, a cumulative relaxation model is used as a collision model, wherein high-order parameters are dynamically calculated based on flow field adaptation to reduce numerical dissipation and dispersion errors, and turbulence modeling under the condition of high Reynolds number is realized by introducing turbulence viscosity by adopting WALE sub-grid scale model on the basis.
4. The method for constructing the efficient hybrid simulation system for unmanned aerial vehicle control design and evaluation according to claim 3, wherein the moving blocks are dynamically determined according to flow conditions, two types of processing are carried out on block boundaries, namely, a conventional distribution function is carried out on adjacent moving block surfaces to flow across the blocks, adjacent non-moving block surfaces are used as open-domain boundaries, enhanced convection boundary conditions are adopted, inflow/outflow stability is met, the moving blocks are judged and selected by adopting a classification strategy based on turbulence characteristics such as local vorticity and the like, buffer bands with the width d are arranged on each moving block, inner and outer nodes of the buffer bands are classified according to strong/medium/weak turbulence levels, moving blocks are created, maintained or deleted according to the classification, a newly created block is used for initializing a macro field by solving a density and speed Laplace equation and applying a Neumann condition of Dirichlet and domain boundaries of the adjacent moving blocks, and the distribution function is reconstructed on the basis of the macro field and gradients, and false compression waves caused by constant value initialization are avoided.
5. The method of claim 1, wherein for the near object field, the algebraic empirical model used by the rotor of the unmanned aerial vehicle is an actuation line model ALM, and the algebraic empirical model used by the other components of the unmanned aerial vehicle than the rotor is an actuation surface model ASM.
6. The method for constructing an efficient hybrid simulation system for unmanned aircraft control design and evaluation according to claim 5, wherein in said actuation line model ALM, the blades are spread into segments, force points are defined at the center of each segment, and the force points are based on the relative incoming flow velocity Local density of Chord length Spanwise dimension Angle of attack Using lift/drag coefficients And And calculating segment force, and decomposing and accumulating the segment force along the lifting/resisting direction to obtain the total thrust and moment of the rotor wing.
7. The method for constructing an efficient hybrid simulation system for unmanned aircraft control design and evaluation as set forth in claim 6, wherein the rotor section has a coefficient of lift/drag And Pre-generated and interpolated by XFoil software, or calculated by CFD, the rise/drag coefficients of the body's surface elements And Using a simplified approximation, there are , 。
8. The method for constructing an efficient hybrid simulation system for unmanned aerial vehicle control design and evaluation according to claim 5, wherein in said active surface model ASM, the body surface is uniformly sampled, each sampling point is regarded as a small planar element, and the relative incoming flow velocity is based on Local density of Equivalent area of Angle of attack The rise/drag is calculated as an empirical function and superimposed.
9. The method for constructing the efficient hybrid simulation system for unmanned aerial vehicle control design and evaluation according to claim 1, wherein probe points are arranged in the near object field, local flow velocity and density are interpolated from the far field blocks, a grid scale is arranged in front of the rotor section probe along the rotation direction, and the body sampling probe is moved outwards by one grid scale along the normal direction, so that near wall mutual interference is reduced, and stability and reliability of reference incoming flow are improved.
10. The method for constructing the efficient hybrid simulation system for unmanned aerial vehicle control design and evaluation according to claim 1, wherein discrete boundary forces obtained by an algebraic empirical model of a near-object field are diffused and loaded on a far-field grid by a tight support kernel function to realize momentum conservation and boundary effect feedback, in the process, in order to avoid excessive smoothing under a coarse grid, linear kernels with a radius of 1 after unitization to a grid scale are adopted to ensure clear and stable near-wall actions, and the resolution of the far-field grid is 0.5-2 times the average chord length of a rotor wing of the unmanned aerial vehicle so as to achieve both stability and the resolvable of a flow field structure.

Description

Method for constructing high-efficiency hybrid simulation system for unmanned aerial vehicle control design and evaluation Technical Field The invention relates to a method for constructing a high-efficiency hybrid simulation system for unmanned aerial vehicle control design and evaluation under strong current field interference, which is used for realizing high-efficiency hybrid modeling, simulation and optimization, and belongs to the technical fields of fluid simulation and aerial vehicle dynamics modeling. Background Unmanned aerial vehicles (Unmanned AERIAL VEHICLE, UAV) have found wide use in recent years, playing an important role in a variety of fields, including aerial photography and imaging, agricultural plant protection, logistics delivery, mapping and exploration, environmental monitoring, infrastructure inspection, military use, and the like. With the continuous development of unmanned aerial vehicle technology, the structure and the configuration of the unmanned aerial vehicle also show diversified trends. Currently, fixed wing unmanned aerial vehicles, multi-rotor unmanned aerial vehicles, and hybrid unmanned aerial vehicles that combine fixed wing and multi-rotor characteristics have been widely used. The development trend of the diversification reflects the continuous improvement of the unmanned aerial vehicle in the aspects of functions and applicability. However, achieving stable, high performance flight control for a particular drone design still faces significant challenges. Unmanned aerial vehicles are affected by complex airflow environments in the flight process, and the aerodynamic characteristics of the unmanned aerial vehicle often show nonlinearities and time-varying properties. Particularly when designing a new or atypical configuration of the unmanned aerial vehicle, it is a complex task how to obtain a suitable controller to achieve stable flight. The control algorithm verification is directly carried out on the real unmanned plane, so that the cost is high, the period is long, and potential safety risks exist. In order to solve the problems, the simulation platform becomes an important tool for designing and optimizing the unmanned aerial vehicle controller. The common unmanned aerial vehicle aerodynamic simulation model mainly adopts an algebraic empirical model (algebraic empirical model) and calculates aerodynamic force applied to the unmanned aerial vehicle by modeling the unmanned aerial vehicle state. The parameters of the model are typically calibrated by physical experiments or high-precision computational fluid dynamics (Computational Fluid Dynamics, CFD) simulations. Such models perform well in relatively steady aerodynamic environments such as stable cruising of unmanned aerial vehicles, but are limited in accuracy and applicability in turbulent environments, near-ground or near-wall flights, and complex scenarios such as multi-unmanned aerial vehicle coordination. Theoretically, a more comprehensive Fluid-Structure Interaction (FSI) simulation method can be used for calculating aerodynamic force distribution of the unmanned aerial vehicle in different flight states. The method realizes synchronous simulation of aerodynamic force and body motion by coupling the CFD solver and the rigid body dynamics solver. However, the existing CFD algorithm has low calculation efficiency, and in order to ensure the aerodynamic force calculation precision at the boundary, a high grid refinement is generally required for the area near the surface of the unmanned plane, even though the algorithm such as the Graphic Processor (GPU) acceleration and the multi-resolution lattice boltzmann method (Lattice Boltzmann Method, LBM) is used, the calculation amount is still huge, and the real-time requirement required by the design of the flight controller is difficult to meet. At present, although some researches try to improve the calculation efficiency of the unmanned aerial vehicle flow field by improving a simulation method, realizing high-efficiency and low-delay simulation while maintaining physical consistency is still a key difficult problem in the unmanned aerial vehicle control design field. Particularly, in the aspects of flight control design and verification under the environment of cooperative flight and complex airflow of multiple unmanned planes, the conventional simulation means are difficult to meet the requirements of real-time or interactive simulation. Disclosure of Invention The invention aims to solve the technical problems that the existing fluid-solid coupling (FSI) unmanned aerial vehicle has low simulation calculation efficiency, is difficult to process free open space and multi-machine scenes, has poor near-wall precision and has insufficient controller evaluation mobility. In order to solve the technical problems, the technical scheme of the invention discloses a method for constructing an efficient hybrid simulation system for unmanned aerial vehicle control de