A method, system, device and medium for automatic optimization of aircraft shape based on discrete adjoint method

Abstract

The present invention belongs to the technical field of optimization design in aircraft design, and discloses an automatic optimization method for aircraft shape based on a discrete adjoint method, comprising a model pre-processing part, a model parameterization part, a dynamic mesh part and a adjoint optimization part. The model pre-processing part mainly includes model simplification and generation of space and surface meshes. The model parameterization part establishes a connection between the model surface mesh and the space coordinate points by using the FFD parameterization method, laying a good foundation for the next part of the dynamic mesh. The dynamic mesh part realizes the change of the model surface mesh by perturbing the design variables and updates the space mesh. The adjoint optimization part obtains the sensitivity matrix of a large number of design variables at a relatively low cost, provides a forward direction for the gradient-based optimization algorithm, and further obtains the local optimal solution of the aircraft; thus, the calculation time and cost are greatly reduced, and the design efficiency is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of optimization design in aircraft design, and relates to an aircraft shape automatic optimization method, system, equipment, and medium based on a discrete adjoint method. Background Art

[0002] With the rapid development of my country's aerospace and computer technology, high performance, high efficiency, and high reliability have gained widespread attention in aircraft optimization design. Aircraft aerodynamic shape optimization design aims to solve engineering design problems using mathematical optimization concepts. With the support of flow field numerical simulation, optimization design methods based on aerodynamic shape parameterization methods, surrogate model technology, and optimization search strategies provide a broad design space for aircraft aerodynamic shape design.

[0003] However, during the development of aerodynamic optimization design systems, key considerations must be given to the selection of aerodynamic optimization methods, the integration of optimization modules, data transmission interface processing, large-scale parallel computing methods, and the visualization of result data. For example, the selection of aerodynamic optimization methods, such as genetic algorithms, often requires thousands of analytical calculations to determine the optimal design solution. This results in unacceptable computational costs for high-precision, multi-design variable optimization processes, significantly limiting the application of optimization technology in aircraft design systems.

[0004] Numerous researchers have proposed various design solutions for the efficient optimization of aircraft design. Jameson first proposed the continuous adjoint method, and Baysal et al. developed the discrete adjoint method by deriving the adjoint equation from the discrete form of the flow governing equation. This method is advantageous in that it is independent of the number of design variables, reduces computational time and cost, and improves design efficiency. However, few studies have applied this method to the aerodynamic layout of airfoils or even complex three-dimensional configurations. Summary of the Invention

[0005] The purpose of the present invention is to provide a method, system, equipment and medium for automatic optimization of aircraft shape based on discrete adjoint method, which solves the problems of low computational efficiency and huge computational cost when aircraft optimization design faces multiple design variables.

[0006] The present invention is achieved through the following technical solutions:

[0007] A method for optimizing an aircraft shape based on a discrete adjoint method comprises the following steps:

[0008] Step 1: Perform pre-processing based on the current aircraft characteristics to obtain a model mesh, which includes the original surface mesh and the original space mesh;

[0009] Step 2: Based on the aircraft layout, construct the FFD control frame and set some control points as design variables. Preliminarily give the design variable variation range and complete the geometric parameterization of the original surface mesh using the FFD method to obtain the parameterized surface mesh.

[0010] Step 3: Randomly assign values to the design variables, deform the parameterized surface mesh, and update the original spatial mesh using the dynamic mesh method to obtain an updated spatial mesh; the updated spatial mesh and the parameterized deformed surface mesh constitute an updated model mesh;

[0011] Step 4: Solve the updated model grid in the flow field solver to obtain the flow field solution vector, residual vector, first objective function value and first constraint function value. Pass the flow field solution vector and residual vector to the adjoint solver for solution to obtain the adjoint solution vector, second objective function value and second constraint function value.

[0012] The first objective function value, the first constraint function value, the second objective function value, and the second constraint function value constitute a sensitivity matrix;

[0013] Step 5: Use the sequential quadratic programming method to search for the optimal solution with the sensitivity matrix as the direction, perturb the updated model grid, and form a new iterative estimation configuration;

[0014] Step 6: Evaluate the new iterative estimated configuration. If the iterative stopping condition is met, the optimization stops and the corresponding optimization results are output. If the condition is not met, return to step 2, generate new design variable values, and re-deform the spatial grid.

[0015] Furthermore, in step one, the pre-processing is geometric repair.

[0016] Furthermore, in step 2, an FFD control body is established on the original surface mesh so that the surface mesh is completely contained in the control body frame. Several FFD control points are defined along each side of the control body. The FFD control points are design variables. Subsequently, a preliminary upper and lower range of the design variables is given as 0.05m. The control frame and the surface mesh are frozen to establish a mapping between the surface mesh and the control frame, complete the parameterization of the surface mesh, and obtain the parameterized surface mesh.

[0017] Furthermore, in step 4, the specific forms of the first objective function and the second objective function are: The specific forms of the first constraint function and the second constraint function are:

[0018] Among them, A and B are weight coefficients, C L is the lift coefficient, is the target lift value of the current aircraft, C Dis the drag coefficient, The target drag value of the current aircraft.

[0019] Furthermore, in step 5: based on the sequential quadratic programming method, the objective function and the constraint conditions are combined to construct an augmented objective function, where the objective function is a quadratic function and the constraint function is a linear function, thereby transforming the constrained optimization problem into an unconstrained optimization problem;

[0020] The form of the sequential quadratic programming algorithm is:

[0021] Minimize

[0022] Subject to

[0023] Where d is the search direction, x k is the current iteration point, H k is a positive definite matrix, f(x) is the objective function, g j (x) is a constraint function, and the search direction is derived from the sensitivity matrix provided by the adjoint solver.

[0024] Furthermore, in step 6, the criterion for determining the iterative stop condition is whether the values of the objective function and the constraint function meet the expected interval; the expected interval is designed as follows: lift coefficient C L The value of does not decrease or increase by 10%;

[0025] Or resistance coefficient C D The value of does not increase or decreases by 5%;

[0026] If the objective function calculated by the updated model mesh satisfies the expected interval, the optimization is stopped and the corresponding optimization results are output; if not, the process returns to step 2, adjusts the design variable range, reassigns the design variables, and deforms the parameterized surface mesh again.

[0027] Furthermore, the optimization results include flow field information, aerodynamic information and updated model grid.

[0028] The present invention also discloses an aircraft shape optimization system based on a discrete adjoint method, comprising:

[0029] The model pre-processing module is used to perform pre-processing according to the current aircraft characteristics to obtain the model mesh, which includes the original surface mesh and the original space mesh;

[0030] The model parameterization module is used to construct the FFD control frame based on the aircraft layout, set some control points as design variables, preliminarily give the design variable variation range, complete the geometric parameterization of the original surface mesh using the FFD method, and obtain the parameterized surface mesh;

[0031] The dynamic mesh processing module is used to randomly assign values to the design variables, deform the parameterized surface mesh, and update the original spatial mesh using the dynamic mesh method to obtain an updated spatial mesh; the updated spatial mesh and the parameterized deformed surface mesh constitute the updated model mesh;

[0032] A solving module is used to solve the updated model grid in a flow field solver, obtain a flow field solution vector, a residual vector, a first objective function value, and a first constraint function value from the flow field solver, and pass the flow field solution vector and the residual vector to an adjoint solver for solving to obtain an adjoint solution vector, a second objective function value, and a second constraint function value;

[0033] The first objective function value, the first constraint function value, the second objective function value, and the second constraint function value constitute a sensitivity matrix;

[0034] The adjoint optimization module is used to optimize the sensitivity matrix using the sequential quadratic programming method, perturb the updated model grid, and form a new iterative estimated configuration;

[0035] The evaluation module is used to evaluate the new iterative estimated configuration. If the iterative stopping conditions are met, the optimization stops and the corresponding optimization results are output. If the conditions are not met, new design variable values are generated in the model parameterization module and the spatial grid is deformed again.

[0036] The present invention also discloses a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the aircraft shape optimization method based on the discrete adjoint method are implemented.

[0037] The present invention also discloses a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the steps of the aircraft shape optimization method based on the discrete adjoint method are implemented.

[0038] Compared with the prior art, the present invention has the following beneficial technical effects:

[0039] The present invention discloses an automated optimization method for aircraft shape based on a discrete adjoint method, comprising a model pre-processing part, a model parameterization part, a dynamic mesh part, and an adjoint optimization part. The model pre-processing part mainly includes model simplification and generation of spatial and surface meshes. The model parameterization part establishes a connection between the model surface mesh and spatial coordinate points by using an FFD parameterization method, laying a good foundation for the next part, the dynamic mesh. The dynamic mesh part realizes the change of the model surface mesh by perturbing the design variables and updates the spatial mesh. The adjoint optimization part obtains the sensitivity matrix of a large number of design variables at a relatively low cost, provides a forward direction for the gradient-based optimization algorithm, and thus obtains the local optimal solution of the aircraft.

[0040] A discrete adjoint method was introduced. By solving the adjoint equation of the fluid control equation, the sensitivity matrix of a large number of design variables can be obtained at a low computational cost. This provides search directions for the gradient optimization algorithm, thereby obtaining the local optimal solution for aircraft optimization. This greatly reduces computational time and cost, improves design efficiency, and provides an effective solution for solving complex engineering optimization design problems with multiple states, multiple objectives, and multiple constraints.

[0041] The Free-Form Deformation (FFD) method was introduced, which makes the optimization process independent of the topological structure of the shape. It is applicable to both structured and unstructured grids, can maintain the geometric properties of the initial shape, such as continuity, smoothness, and topological relationship, and has good local deformation capabilities.

[0042] The invention introduces a dynamic mesh deformation technology based on inverse distance weighted interpolation (IDW algorithm) with strong deformation capability and high mesh quality, which makes the dynamic mesh part of the invention have strong stability, accuracy, ease of use, efficiency and parallelism.

[0043] The FFD parameterization method, IDW dynamic mesh technology, discrete adjoint method, and sequential quadratic programming algorithm are connected to form an automated process, simplifying manual operations and making it more concise and efficient.

[0044] The present invention takes the advantages of high efficiency of discrete adjoint method, makes up for the shortcomings of traditional optimization method, ignores the limitation of the number of design variables, and combines free surface deformation method, dynamic mesh technology, sequential quadratic programming method and various interface scripts to form a complete and efficient automated high-precision optimization method, which greatly improves the calculation efficiency while ensuring accuracy, and solves the current problem of time-consuming and labor-intensive calculation due to the large number of design variables. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 A flow chart showing the main steps of the aircraft aerodynamic optimization design platform of the present invention;

[0046] Figure 2(a) shows the surface mesh and FFD control volume of the wing-body assembly before deformation; Figure 2(b) shows the surface mesh and FFD control volume of the wing-body assembly after deformation;

[0047] Figure 3 is the initial FFD control volume;

[0048] Figure 4 is the FFD control body after deformation;

[0049] Figure 5 is the initial wing-body assembly surface mesh;

[0050] Figure 6 is the surface mesh of the wing-body assembly after deformation;

[0051] Figure 7 The cross-sectional changes of the wing at different positions along the span before and after deformation;

[0052] Figure 8 The mesh of the deformed front and rear surfaces of the DPW-5 CRM wing-body assembly;

[0053] Figure 9 The cross-sectional shapes of the DPW-5 CRM wing-body assembly at different positions along the span before and after deformation.

[0054] Figure 10 The spatial grid topology of the DPW-5 CRM wing-body assembly before and after deformation;

[0055] Figure 11 The spatial grid topology near the wingtip before and after deformation of the DPW-5 CRM wing-body assembly. DETAILED DESCRIPTION

[0056] In order to make the purpose, technical solutions and advantages of the present invention more clear, the following is a further detailed description with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. That is, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments.

[0057] The components described and illustrated in the drawings and embodiments of the present invention may be arranged and designed in a variety of different configurations. Therefore, the detailed description of the embodiments of the present invention provided in the following drawings is not intended to limit the scope of the claimed invention, but merely represents a selected embodiment of the present invention. All other embodiments derived by those skilled in the art based on the drawings and embodiments of the present invention without inventive effort shall fall within the scope of protection of the present invention.

[0058] The features and performance of the present invention are further described in detail below with reference to the embodiments.

[0059] like Figure 1 As shown, the present invention discloses an automated optimization method based on a discrete adjoint method, comprising the following steps:

[0060] Step 1: Perform pre-processing based on the current aircraft characteristics to generate the corresponding structured or unstructured grid and obtain the model grid;

[0061] Step 2: After obtaining the model mesh based on pre-processing, the free-form deformation method (FFD) is used to establish the corresponding control frame and control points (i.e., set the optimization variables) for the model mesh, and determine the appropriate variable range.

[0062] Combined with the FFD method, the specific measures to establish the corresponding control frame and control points for the model grid are as follows:

[0063] Establish an FFD control volume around the target geometry to be deformed, and define several FFD control points along each side of the control volume; -1 (x, y, z) = (s, t, u) to solve the logical parameters (s, t, u) of any point on the target geometry in the FFD control volume; change the position of the control point in the FFD control volume to achieve the deformation of the FFD control volume; and solve the displacement of each point on the target geometry through (x, y, z) = F(s, t, u).

[0064] The FFD method establishes an FFD control volume (FFD Control Lattice) around the geometric shape that needs to be parameterized. By moving the vertices (also called "control points") of the FFD control volume, the hypercurved space enclosed by the FFD control volume is deformed and mapped to the deformation of the target geometry that has been embedded in the FFD control volume. Mathematically, the mapping relationship for deformation is usually constructed in the form of a tensor product. The geometric shape designer reconstructs the hypercurved space enclosed by the FFD control volume by moving the FFD control points. Each geometric point within the FFD control volume (that is, any point on the target geometry) changes its position by the reconstruction of the hypercurved space, that is, the deformation of the target geometry is achieved.

[0065] In their 1986 paper, Sederberg and Parry first proposed the FFD deformation method, in which Bernstein polynomials were used to define the geometric mapping relationship of the hypercurved space enclosed by the FFD control volume. In their method, the FFD control volume is a parallelepiped, and the FFD control points are evenly distributed along each side. The FFD deformation achieves a continuous mapping F:R 3 →R 3 →R 3, from the actual space R(x,y,z) to the parameter space R(s,t,u) and then to the deformed actual space R'(x,y,z).

[0066] The following uses Bernstein-based FFD technology to establish a geometric shape parameterization program system for the wing-body assembly, and conducts parameter-driven geometric shape deformation tests on the Drag Prediction Workshop-5 Common Research Model configuration.

[0067] As shown in Figure 2(a), an FFD control volume is established around the wing grid (blue grid) of the initial DPW-5 CRM wing-body assembly. As the positions of the control points in the FFD control volume change, the wing geometry deforms, as shown in Figure 2(b). Figure 3 As shown, the wing geometry before deformation is as follows Figure 5 As shown, the deformed control volume is Figure 4 The geometric shape of the deformed wing is shown in Figure 6 As shown in the figure. The cross-section of the wing at different positions along the span before and after deformation is as follows: Figure 7 As shown in the figure, it can be seen that the application of FFD technology in the parameterization of the wing-body assembly can also produce a good parameter-driven deformation effect on the wing, and the torsion angle of each section along the span can also be changed.

[0068] After establishing the control frame and control points, basic parameters such as the Reynolds number, Mach number, flight altitude, temperature, reference area, moment center, and aerodynamic center need to be set in conjunction with the aircraft's flight conditions. At the same time, since the aircraft's aerodynamic performance indicators such as lift, drag, and lift-to-drag ratio are largely dependent on the selected first and second objective functions and the first and second constraint functions, the objectives and constraints generally have conflicting effects on performance indicators. Therefore, it is necessary to select the primary indicator as the target and treat the remaining design indicators as constraints. The ideal design state is then achieved by setting the function weights.

[0069] Step 3: Randomly assign values to the design variables, deform the parameterized surface mesh, and use the dynamic mesh method to update the original spatial mesh to obtain an updated spatial mesh. The updated spatial mesh and the parameterized deformed surface mesh constitute the updated model mesh. The dynamic mesh method mentioned here refers to a mesh deformation algorithm based on inverse distance weighting interpolation (IDW). The deformation principle is as follows:

[0070] The basic principle of the grid deformation algorithm is: if the CFD space calculation grid is Its surface mesh is have The surface grid corresponds to the aerodynamic shape in the optimized design. When the aerodynamic shape changes, becomes The spatial grid Need to be updated according to the mesh deformation algorithm To correspond to the new surface mesh. Space mesh and surface mesh The relationship is set as:

[0071]

[0072] The deformed surface mesh and spatial grid The mapping relationship is:

[0073]

[0074] If only the displacement change of each grid point is considered, the spatial grid change can be solved based on the surface grid change. If this functional relationship is set to W, then:

[0075] G′ V -G V =W(G′ S -G S ) (3)

[0076] The basic steps of the Inverse Distance Weighting Interpolation (IDW) algorithm to solve the displacement of spatial grid points following the surface grid are:

[0077]

[0078] Where N is the number of interpolation base points, u j represents the surface mesh displacement at grid point j, w j is the functional relationship between the surface mesh change and the spatial mesh change at the interpolation point j, and is calculated as follows:

[0079]

[0080] Among them, p is a constant with a value greater than 1. When the value of p is large, it is mainly affected by the surrounding base points. When the value of p is small, it is affected by a wider range of interpolation base points. is the spatial grid coordinate, Represents the spatial grid coordinates of the interpolation point j.

[0081] A dynamic mesh calculation example test was conducted on the DPW-5CRM wing-body assembly configuration. The initial and deformed surface meshes and the cross-sectional shapes at different positions along the span before and after deformation are shown in the figure. Figure 8 、 Figure 9 As shown, the spatial grid topology before and after deformation is as follows Figure 10 、 Figure 11 As shown in Figure 2, it can be seen that after the wing shape of the wing-body assembly is deformed and geometrically twisted, the spatial mesh automatically generated by the IDW dynamic mesh technology still has good tracking performance for the deformed shape.

[0082] Step 4: After generating the disturbed spatial grid, perform the solution calculations in the flow field solver and adjoint solver respectively. The flow field solution vector, residual vector, first objective function value, and first constraint function value are obtained in the flow field solver.

[0083] The flow field solution vector and the residual vector are passed to the adjoint solver for solving, and the adjoint solution vector, the second objective function value and the second constraint function value are obtained.

[0084] Step 5: After the solution is completed, Sequential Quadratic Programming (SQP) will optimize based on the sensitivity matrix provided by the adjoint algorithm and generate a new iterative estimated configuration.

[0085] The specific steps are: Based on the sequential quadratic programming method (SQP), the objective function and the constraints are combined to construct an augmented objective function, where the objective function is a quadratic function and the constraint function is a linear function, thereby transforming the constrained optimization problem into an unconstrained optimization problem. The general form of the sequential quadratic programming algorithm is:

[0086] Minimize

[0087] Subject to

[0088] Where d is the search direction, x k is the current iteration point, H k is a positive definite matrix, f(x) is the objective function, g j (x) is the constraint function. The search direction here comes from the sensitivity matrix provided by the adjoint solver.

[0089] In summary, at the iteration point x k First, the matrix H k The corresponding quadratic programming subproblem is reconstructed and solved, and the optimal solution of the subproblem is used as the search direction d for the next iteration. Then, a constrained one-dimensional search is performed on the objective function of the original nonlinear constrained optimization problem in this direction to obtain the next iteration point xk , and judge whether the convergence accuracy is satisfied. Repeat the above process until the iteration point x k Finally, the termination criterion is met and the optimal solution of the original nonlinear constrained optimization problem is obtained.

[0090] Step 6: Evaluate the current iterative estimated configuration. If the iterative stopping condition is met, the optimization stops and the corresponding optimization results are output. If the condition is not met, return to step 3, generate new design variable values, and re-deform the spatial grid.

[0091] Specifically, the optimization results include flow field information, aerodynamic information, and updated model grid.

[0092] Specifically, the objective function and constraint function calculated by the updated model mesh are compared with the iteration stop condition. The iteration stop condition is whether the values of the objective function and constraint function meet the expected range of the optimization design. For example, the iteration stop condition is to increase the lift coefficient of the objective function by 10%. If the objective function calculated by the updated model mesh reaches this goal, the optimization is stopped and the corresponding optimization results are output at the same time; if the condition is not met, return to step 2, reassign the design variables, and deform the parameterized surface mesh again.

[0093] The expected range is designed as: lift coefficient C L The value of does not decrease or increase by 10%;

[0094] Or resistance coefficient C D The value does not increase or decreases by 5%.

[0095] The present invention also discloses an aircraft shape optimization system based on a discrete adjoint method, comprising:

[0096] The model pre-processing module is used to perform pre-processing according to the current aircraft characteristics to obtain the model mesh, which includes the original surface mesh and the original space mesh;

[0097] The model parameterization module is used to construct the FFD control frame based on the aircraft layout, set some control points as design variables, preliminarily give the design variable variation range, complete the geometric parameterization of the original surface mesh using the FFD method, and obtain the parameterized surface mesh;

[0098] The dynamic mesh processing module is used to randomly assign values to the design variables, deform the parameterized surface mesh, and update the original spatial mesh using the dynamic mesh method to obtain an updated spatial mesh; the updated spatial mesh and the parameterized deformed surface mesh constitute the updated model mesh;

[0099] A solving module is used to solve the updated model grid in a flow field solver, obtain a flow field solution vector, a residual vector, a first objective function value, and a first constraint function value from the flow field solver, and pass the flow field solution vector and the residual vector to an adjoint solver for solving to obtain an adjoint solution vector, a second objective function value, and a second constraint function value;

[0100] The first objective function value, the first constraint function value, the second objective function value, and the second constraint function value constitute a sensitivity matrix;

[0101] The adjoint optimization module is used to optimize the sensitivity matrix using the sequential quadratic programming method, perturb the updated model grid, and form a new iterative estimated configuration;

[0102] The evaluation module is used to evaluate the new iterative estimated configuration. If the iterative stopping conditions are met, the optimization stops and the corresponding optimization results are output. If the conditions are not met, new design variable values are generated in the model parameterization module and the spatial grid is deformed again.

[0103] The aircraft shape optimization method based on the discrete adjoint method of the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROMs, optical storage, etc.) containing computer-usable program code.

[0104] If the aircraft shape optimization method based on the discrete adjoint method of the present invention is implemented as a software functional unit and sold or used as a standalone product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention can also implement all or part of the process steps in the above-mentioned method embodiments by using a computer program to instruct the relevant hardware. The computer program can be stored in a computer-readable storage medium. When executed by a processor, the computer program can implement the steps of each of the above-mentioned method embodiments. The computer program includes computer program code, which can be in source code form, object code form, executable file, or some intermediate form. Computer-readable storage media include permanent and non-permanent, removable and non-removable media, and can use any method or technology to store information. The information can be computer-readable instructions, data structures, program modules, or other data. It should be noted that the content of the computer-readable medium can be appropriately expanded or reduced based on the requirements of legislation and patent practice in certain jurisdictions. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunications signals. Among them, the computer storage medium can be any available medium or data storage device that can be accessed by the computer, including but not limited to magnetic storage (such as floppy disks, hard disks, magnetic tapes, magneto-optical disks (MO)), optical storage (such as CDs, DVDs, BDs, HVDs, etc.), and semiconductor storage (such as ROM, EPROM, EEPROM, non-volatile memory (NANDFLASH), solid-state drives (SSDs)), etc.

[0105] In an exemplary embodiment, a computer device is also provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the steps of the aircraft shape optimization method based on the discrete adjoint method are implemented. The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0106] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered by the scope of protection of the claims of the present invention.

Claims

1. A method for optimizing aircraft shape based on discrete adjoint method, characterized in that: The steps include: Step 1: Perform pre-processing based on the current aircraft characteristics to obtain a model mesh, which includes the original surface mesh and the original space mesh; Step 2: Based on the aircraft layout, construct the FFD control frame and set some control points as design variables. Preliminarily give the design variable variation range and complete the geometric parameterization of the original surface mesh using the FFD method to obtain the parameterized surface mesh. Step 3: Randomly assign values to the design variables, deform the parameterized surface mesh, and update the original spatial mesh using the dynamic mesh method to obtain an updated spatial mesh. The updated spatial mesh and the parametrically deformed surface mesh constitute an updated model mesh; Step 4: Solve the updated model grid in the flow field solver to obtain the flow field solution vector, residual vector, first objective function value and first constraint function value. Pass the flow field solution vector and residual vector to the adjoint solver for solution to obtain the adjoint solution vector, second objective function value and second constraint function value. The first objective function value, the first constraint function value, the second objective function value, and the second constraint function value constitute a sensitivity matrix; Step 5: Use the sequential quadratic programming method to search for the optimal solution with the sensitivity matrix as the direction, perturb the updated model grid, and form a new iterative estimation configuration; Step 6: Evaluate the new iterative estimated configuration. If the iterative stop condition is met, the optimization stops and the corresponding optimization results are output. If the conditions are not met, return to step 2, generate new design variable values, and deform the spatial grid again; In step 4, the specific forms of the first objective function and the second objective function are: , the specific forms of the first constraint function and the second constraint function are ; Among them, A and B are weight coefficients, is the lift coefficient, is the target lift value of the current aircraft, is the drag coefficient, is the target resistance value of the current aircraft; In step 6, the criterion for determining the iteration stop condition is whether the values of the objective function and the constraint function meet the expected interval; The expected range is designed as: lift coefficient The value of does not decrease or increase by 10%; or drag coefficient The value of does not increase or decreases by 5%; If the objective function calculated by the updated model mesh satisfies the expected interval, the optimization is stopped and the corresponding optimization results are output; if not, the process returns to step 2, adjusts the design variable range, reassigns the design variables, and deforms the parameterized surface mesh again.

2. The aircraft shape optimization method based on the discrete adjoint method according to claim 1, characterized in that: In step 1, the pre-processing is geometry repair.

3. The aircraft shape optimization method based on the discrete adjoint method according to claim 1, characterized in that: In step 2, an FFD control body is established on the original surface mesh so that the surface mesh is completely contained in the control body frame. Several FFD control points are defined along each side of the control body. The FFD control points are design variables. Subsequently, a preliminary upper and lower range of the design variables is given as 0.05 m. The control frame and the surface mesh are frozen to establish a mapping between the surface mesh and the control body, complete the parameterization of the surface mesh, and obtain the parameterized surface mesh.

4. The aircraft shape optimization method based on the discrete adjoint method according to claim 1, characterized in that: In step 5: Based on the sequential quadratic programming method, the objective function and the constraints are combined to construct an augmented objective function, where the objective function is a quadratic function and the constraint function is a linear function, thereby transforming the constrained optimization problem into an unconstrained optimization problem; The form of the sequential quadratic programming algorithm is: ； ； Where d is the search direction, is the current iteration point, is a positive definite matrix, is the objective function, is a constraint function, and the search direction is derived from the sensitivity matrix provided by the adjoint solver.

5. The method for optimizing an aircraft shape based on a discrete adjoint method according to claim 1, characterized in that: The optimization results include flow field information, aerodynamic information and updated model grid.

6. An aircraft shape optimization system for implementing the aircraft shape optimization method based on the discrete adjoint method according to any one of claims 1 to 5, characterized in that: include: The model pre-processing module is used to perform pre-processing according to the current aircraft characteristics to obtain the model mesh, which includes the original surface mesh and the original space mesh; The model parameterization module is used to construct the FFD control frame based on the aircraft layout, set some control points as design variables, preliminarily give the design variable variation range, complete the geometric parameterization of the original surface mesh using the FFD method, and obtain the parameterized surface mesh; The dynamic mesh processing module is used to randomly assign values to the design variables, deform the parameterized surface mesh, and update the original spatial mesh by combining the dynamic mesh method to obtain the updated spatial mesh; The updated spatial mesh and the parametrically deformed surface mesh constitute an updated model mesh; A solving module is used to solve the updated model grid in a flow field solver, obtain a flow field solution vector, a residual vector, a first objective function value, and a first constraint function value from the flow field solver, and pass the flow field solution vector and the residual vector to an adjoint solver for solving to obtain an adjoint solution vector, a second objective function value, and a second constraint function value; The first objective function value, the first constraint function value, the second objective function value, and the second constraint function value constitute a sensitivity matrix; The adjoint optimization module is used to optimize the sensitivity matrix using the sequential quadratic programming method, perturb the updated model grid, and form a new iterative estimated configuration; The evaluation module is used to evaluate the new iterative estimated configuration. If the iterative stop condition is met, the optimization stops and the corresponding optimization results are output; If the conditions are not met, new design variable values are generated in the model parameterization module and the spatial grid is deformed again.

7. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the aircraft shape optimization method based on the discrete adjoint method as claimed in any one of claims 1 to 5 are implemented.

8. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the aircraft shape optimization method based on the discrete adjoint method as claimed in any one of claims 1 to 5 are implemented.

Images