A method for frequency optimization of stiffener structures based on level set and variable density method

By using a method based on level sets and variable density, the width, spacing, and volume of the mesh stiffener structure are optimized, solving the problems of discontinuity and poor manufacturability in the optimization of stiffener structures in the prior art, and achieving optimization effects with high manufacturability and high fundamental characteristic frequencies.

CN117390913BActive Publication Date: 2026-11-17HUAZHONG UNIV OF SCI & TECH
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202311232139.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-09-22
Publication Date
2026-11-17
Estimated Expiration
2043-09-22

AI Technical Summary

Technical Problem

Existing mesh-reinforced structures suffer from problems in shape and topology optimization, such as incomplete reinforcement path representation, difficulty in controlling wall thickness and spacing, and a single optimization objective. In particular, they are difficult to fully utilize material properties under complex vibration conditions.

Method used

A method based on level set and variable density is adopted to optimize the width, spacing and volume of stiffener structure through finite element mesh generation, level set function solution, gradient constraint and characteristic frequency optimization, and construct an optimization model to maximize the basic characteristic frequency.

Benefits of technology

This achieves good continuity and high manufacturability of the stiffener structure, enabling it to adapt to complex vibration environments and improving the applicability and design flexibility of engineering applications.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117390913B_ABST
    Figure CN117390913B_ABST
Patent Text Reader

Abstract

This invention belongs to the technical field of structural optimization and discloses a frequency optimization method for stiffener structures based on level set and variable density methods. The method includes the following steps: S1. Divide the design domain of the stiffener structure to be optimized into a finite element mesh and solve for the level set function at the center point of each element; S2. Obtain the element density of each element and construct constraints on the stiffener spacing to ensure uniform width and spacing between adjacent stiffeners; S3. Solve for the fundamental characteristic frequencies of the stiffener structure; S4. Using the maximization of the fundamental characteristic frequencies of the stiffener structure as the optimization objective, the width, spacing, and volume of the stiffeners as constraints, and weighting coefficients and projection threshold parameters controlling the stiffener width as design variables, construct an optimization model for the stiffener structure, and solve this optimization model until a preset termination condition is met, thereby achieving optimized design. This invention solves the problem of shape and topology optimization for stiffener structures.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention belongs to the technical field of structural optimization, and more specifically, relates to a method for optimizing the frequency of stiffener structures based on level sets and variable density methods. Background Technology

[0002] Grid stiffened structures are lightweight reinforced structures composed of a matrix and stiffeners attached to it, characterized by a repeating grid pattern formed by intersecting stiffeners. In engineering applications, grid stiffeners are widely used to improve structural performance. They possess excellent mechanical properties such as high specific strength, high specific stiffness, and high stability, as well as capabilities for vibration reduction and buckling resistance. Currently, grid stiffened structures are widely used in aerospace, automotive, shipbuilding, and construction industries.

[0003] Traditional mesh-reinforced structures are of uniform density, typically composed of rows or columns of straight columns or ribs. However, in practical applications, different parts of the structure bear different loads. Especially when considering conditions involving complex vibrations, to fully utilize material properties and avoid resonance, the natural frequencies of the mesh-reinforced structure should be optimized to determine the optimal density distribution of the material, maximizing the natural frequencies of the mesh-reinforced structure for the same weight. The fundamental (first-order) characteristic frequency, as a key element of the structure's natural frequencies, is also of great significance as an objective function for optimization.

[0004] In recent years, curved mesh stiffener structures have attracted increasing attention. Recent studies have shown that the design optimization of stiffeners composed of curved and non-uniform meshes has greater potential in improving structural performance because the stiffener arrangement is more flexible, and curved mesh stiffeners can adjust the force transmission path according to the different requirements of different local areas of the structure, thus having greater potential in improving stiffness, natural frequency, and buckling load.

[0005] Currently, the shape and topology optimization methods used for mesh stiffener structures mainly have the following problems: (1) The stiffener path representation often does not include cross-sectional information such as stiffener height and width, and is only a geometric abstraction; (2) The stiffener wall thickness and the spacing between stiffeners cannot be controlled, which will lead to practical manufacturing difficulties. Even if additive manufacturing can manufacture them, there is a minimum wall thickness limit. In addition, uneven changes in wall thickness may lead to additional thermal deformation and thermal stress; (3) For the sake of simplicity, many stiffener structure optimizations are currently optimized with compliance as the target, and optimization of natural frequencies is less common. Summary of the Invention

[0006] To address the aforementioned deficiencies or improvement needs of existing technologies, this invention provides a frequency optimization method for reinforcing rib structures based on level sets and variable density methods, solving the problems of shape and topology optimization of mesh reinforcing rib structures.

[0007] To achieve the above objectives, according to the present invention, a method for frequency optimization of stiffener structures based on level sets and variable density methods is provided, the method comprising the following steps:

[0008] S1 divides the design domain of the mesh stiffener structure to be optimized into a finite element mesh, thereby dividing the design domain into multiple elements. Multiple nodes in the finite element mesh are selected as nodes of the radial basis function. Two different weight coefficients are assigned to each node. The level set function at the center point of each element is obtained using the weight coefficients and the radial basis function.

[0009] S2 uses the level set function to solve for the element density of each element, and uses the gradient constraint of the level set function to construct the constraint conditions for the width and spacing of the stiffeners, so that the width and spacing between adjacent stiffeners are uniform.

[0010] S3 sets the height of the mesh stiffener structure to be optimized and solves the height of each element in combination with the element density. Using the height of each element, the characteristic equations of the stiffness matrix and mass matrix of each Mindlin plate element are constructed, and the basic characteristic frequencies of the mesh stiffener structure to be optimized are solved.

[0011] S4 takes maximizing the basic characteristic frequency of the mesh reinforcing rib structure to be optimized as the optimization objective, the width, spacing and volume of the reinforcing ribs as constraints, and the weighting coefficient and the projection threshold parameter controlling the width of the reinforcing ribs as variables. In this way, an optimization model of the mesh reinforcing rib structure to be optimized is constructed, and the optimization model is solved until the variable values ​​that meet the preset termination conditions are obtained, thereby realizing the optimized design of the mesh reinforcing rib structure to be optimized.

[0012] More preferably, in step S1, the level set function is performed according to the following relationship:

[0013]

[0014]

[0015] Wherein, φ(x) e ) and ψ(x e ) represents the two obtained level set functions; m is the number of radial basis functions; α i and β i These are the weighting coefficients of the i-th radial basis function; x e These are the coordinates of the center point of the unit; qi These are the node coordinates of the i-th radial basis function; These are radial basis functions.

[0016] More preferably, in step S2, the unit density is determined according to the following steps:

[0017] S21 uses the cosine function to transform the level set functions φ(x) and ψ(x) to obtain the functions Φ(x) and Ψ(x);

[0018] S22 uses a differentiable max function to combine Φ(x) and Ψ(x) to obtain the density function ρ(x);

[0019] S23 calculates the density function using an approximate Heaviside function, thereby obtaining the corresponding element density.

[0020] More preferably, in step S22, the density function is performed according to the following relationship:

[0021]

[0022]

[0023] Where ρ(x) is the density function, Φ(x) and Ψ(x) are functions obtained by cosine transformation of the level set function, and χ is a constant.

[0024] More preferably, in step S23, the unit density is determined according to the following relationship:

[0025]

[0026] Where, x e This represents the center point of the element; here, the parameter value of the center point represents the parameter value of the entire element. The element density is obtained after calculation using the approximate Heaviside function; ρ(x) e H(x) is the density function; H(x) is the approximate Heaviside function.

[0027] More preferably, in step S2, the constraints on the width and spacing of the reinforcing ribs are as follows:

[0028]

[0029]

[0030] Among them, g φ g ψ The resulting gradient constraint function; p wξ is a parameter greater than 0 in the p-norm condensation method; n is the total number of finite element elements; ξ is the upper limit of the two gradient constraints after condensation. and These are the corresponding functions of the magnitudes of the gradients of the two level set functions at the center point of each unit.

[0031] More preferably, in step S3, the characteristic equation is performed according to the following relationship:

[0032]

[0033] Where λ1 is the eigenvalue corresponding to the fundamental characteristic frequency ω1, i.e. λ1 is the eigenvector corresponding to the eigenvalue; K is the global stiffness matrix; M is the global mass matrix.

[0034] More preferably, in step S3, the stiffness matrix and mass matrix of each Mindlin plate element are determined according to the following relationship:

[0035]

[0036] M e =M0(h em )

[0037] Among them, K e and M e These are the stiffness matrix and mass matrix of each element; K b0 K is the bending stiffness matrix of a unit thickness element. s0 h is the shear stiffness matrix per unit thickness element. e It is the unit height h(x) e (abbreviation of ); M0 is the mass matrix of a unit thickness element under Mindlin plate element theory, h em h is another expression for the unit height. m (x e (abbreviation of ).

[0038] More preferably, in step S4, the optimization model is performed according to the following relationship:

[0039] Find: α i ,β i η (i = 1, 2...m)

[0040] max J=λ1

[0041]

[0042]

[0043] gφ g ψ ≤ξ

[0044] δ min ≤α i ,β i ≤δ max

[0045] η min ≤η≤η max

[0046]

[0047] Where, α i β i η and λ1 are design variables, representing the weighting coefficients of the radial basis functions and the threshold projection parameter of the approximate Heaviside function, respectively; m is the number of radial basis functions; ω1 is the fundamental (first-order) characteristic frequency, and λ1 is the eigenvalue corresponding to the fundamental characteristic frequency, i.e. λ1 is the eigenvector corresponding to the eigenvalue; K is the global stiffness matrix; M is the global mass matrix; g φ g ψ These are the functions obtained by transforming the magnitudes of the gradients of the level set functions φ(x) and ψ(x) and then performing p-norm condensation; ξ is the upper bound of the two gradient constraints after condensation; δ min δ max These represent the design variables α and β, respectively. i and β i The upper and lower limits; η min η max These represent the upper and lower limits of the design variable η, respectively. V is the allowable total volume of the reinforcing rib; V is the total volume of the reinforcing rib.

[0048] More preferably, in step S4, the preset termination condition is determined according to the following relationship:

[0049] ω err ≤δ ω or k≥k max

[0050] Where, ω err The basic characteristic frequency error; k is the current iteration number; δ ω k is the lower limit of the fundamental characteristic frequency error; max This represents the upper limit of the number of iterations.

[0051] In summary, the technical solutions conceived by this invention have the following beneficial effects compared with the prior art:

[0052] 1. This invention uses contour lines of the level set function to represent the stiffener path, and uses variable density methods and cosine functions to assign cross-sectional information such as stiffener thickness and width, rather than a geometrically abstract curve of the stiffener; this invention combines the magnitude of the gradient of the level set function to establish two gradient constraints, thereby ensuring uniform stiffener width and spacing; finally, this invention uses the fundamental characteristic frequency as the optimization objective to study the structural vibration problem, rather than the commonly used compliance as the optimization objective, and obtains the optimized mesh stiffener structure by maximizing the fundamental characteristic frequency;

[0053] 2. This invention is based on a variable density method using level set functions. The element density is variable, and by utilizing level set functions, cosine functions, differentiable max functions, smooth approximate HeaViside functions, and p-norm condensation, a mesh-reinforced structure with good structural continuity, excellent stiffener layout, and good manufacturability can be obtained. Unlike the base structure method, it does not require pre-defining matrix elements, improving design flexibility. The relevant parameters (λ1, λ2, γ, η) of the level set function can be easily adjusted to obtain mesh-reinforced structures with different widths and spacings. Gradient constraints are added to the level set function to meet manufacturing constraints, greatly improving the manufacturability of the mesh-reinforced structure. The method is simple, easy to implement, highly applicable, and easily integrated with existing finite element software.

[0054] 3. This invention optimizes the basic characteristic frequency of the grid stiffener structure to obtain a grid stiffener structure with good continuity, non-uniformity, high manufacturability, and high basic characteristic frequency that can adapt to complex vibration environments. This solves the technical problems of missing and discontinuous cross-sectional information and poor manufacturability in existing grid stiffener structures, and improves the applicability in engineering fields. Attached Figure Description

[0055] Figure 1 This is a flowchart of a method for optimizing the basic characteristic frequencies of a mesh stiffener structure based on level set and variable density method, according to a preferred embodiment of the present invention.

[0056] Figure 2 This is a schematic diagram of a specific structural design domain involved in a method for optimizing the basic characteristic frequency of a mesh stiffener structure based on level set and variable density method, constructed according to a preferred embodiment of the present invention.

[0057] Figure 3 This is an initial design schematic diagram of a mesh reinforcing rib structure constructed according to a preferred embodiment of the present invention;

[0058] Figure 4 This is a schematic diagram of the optimized result of the mesh reinforcing rib structure constructed according to a preferred embodiment of the present invention;

[0059] Figure 5It is the iterative curve of the objective function constructed according to a preferred embodiment of the present invention. Detailed Implementation

[0060] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0061] This invention divides the design domain of a mesh-reinforced structure into several elements, uniformly sets design points within the domain, and defines two weighting coefficients as design variables at each design point. Combining these weighting coefficients and radial basis functions, two level set functions are obtained. The level set functions are then transformed and combined using cosine functions, differentiable max functions, and approximate Heaviside functions to obtain the element density. Through finite element analysis, the element height is calculated using the element density, and the Young's modulus and mass of the mesh-reinforced structure on that element are obtained using the element height. Stiffness and mass matrices are then established, and the characteristic equation is solved. Constraint functions for the stiffener volume and the uniformity of its width and spacing are constructed. Based on Mindlin plate theory, the optimization objective is to maximize the fundamental characteristic frequency. The radial basis function expansion coefficients describing the stiffener layout and the projection threshold controlling the stiffener width are used as design variables. Constraints include stiffener volume, manufacturing constraints, and upper and lower limits of the design variables. The fundamental characteristic frequency optimization design problem of the mesh-reinforced structure is defined. The sensitivity of the objective function and constraint functions to the design variables is calculated, and the design variables are updated based on the sensitivity until the optimization termination condition is met, resulting in the optimal mesh-reinforced structure.

[0062] like Figure 1 and Figure 2 As shown, this invention provides a method for optimizing the fundamental characteristic frequencies of a mesh-reinforced structure based on level sets and variable density methods. The method is illustrated using the optimization problem of maximizing the fundamental characteristic frequencies of a rectangular beam mesh-reinforced structure with concentrated mass and fixed on both sides as an example. Within a given 4m × 2m rectangular structural design domain D, there are two symmetrically distributed holes with a radius of 0.4m, located at a distance of 2 / 3 from the axis of symmetry. Assume that the Young's modulus of the matrix and reinforcing material is E = 200 GPa, Poisson's ratio is υ = 0.3, and mass density is ρ0 = 7800 kg / m³. 3 The square concentrated mass region is located at the center of the axis of symmetry, with dimensions of 0.16m × 0.16m and a mass density of 2 × 10⁻⁶ times the set value ρ₀. 5 times. Figure 3The initial design of the grid reinforcement structure for the embodiment is a traditional uniform grid reinforcement structure. The method provided by this invention is used to optimize the above-mentioned rectangular beam grid reinforcement structure to maximize its basic characteristic frequency.

[0063] The variable density method based on level sets for optimizing the basic characteristic frequency design of mesh stiffener structures provided by this invention mainly includes the following steps:

[0064] Step 1: Divide the structural design domain of the mesh stiffener structure to be optimized into n elements, and select m nodes q in the structural design domain. i As nodes of radial basis functions, and as nodes q of each radial basis function. i Assign weight coefficient α i β i Combined with the weighting coefficient α i β i The radial basis functions yield two level set functions φ(x) at the center point of each element. e ), ψ(x) e ); n and m are both positive integers and n > m.

[0065] Specifically, Figure 2 The given structural design domain D is divided into 400×200 square elements; 41×21 nodes q of radial basis functions are uniformly selected within the design domain D. i And in each q i Define two different weighting coefficients α at the point. i β i As a design variable in optimization, combined with the weighting coefficient α i β i The radial basis functions yield the two horizontal set function values ​​φ(x) at the center point of each element. e ), ψ(x) e The specific calculation formula is as follows:

[0066]

[0067]

[0068] Where m is the number of radial basis function nodes, in this embodiment For radial basis functions, we choose compactly supported radial basis functions, whose expression is:

[0069]

[0070] The symbol (·) + Let max{0} represent the value of the radial basis function node. The parameter r can be determined based on the coordinates q of the radial basis function node. i =(x i yi ) and the coordinates of the unit center point x e =(x e y e The calculation is as follows:

[0071]

[0072] Where ε is a small positive number, here taken as 10. -4 ;d min Here, we take the support radius of CS-RBFs as 1.6.

[0073] Based on the initial level set function φ 0 (x e ), ψ 0 (x e The initial weight coefficients are obtained by solving the problem. Specifically, in this embodiment, let the level set function φ 0 (x e ), ψ 0 (x e The values ​​of ) are respectively equal to the coordinates of the center point (x) e y e ) of x e and y e .

[0074] Step two: Transform the level set functions φ(x) and ψ(x) using the cosine function to obtain functions Φ(x) and Ψ(x), and combine Φ(x) and Ψ(x) using a differentiable max function to obtain ρ(x). Based on ρ(x), obtain the cell density through an approximate Heaviside function.

[0075] Specifically, the functions Φ(x) and Ψ(x) are obtained by transforming the two level set functions φ(x) and ψ(x) using the cosine function. The specific calculation formula is as follows:

[0076]

[0077]

[0078] Where parameters λ1 and λ2 are the wavelengths of the cosine function, which control the dimensions and spacing of the reinforcing ribs in the two directions of the mesh reinforcing rib structure, respectively. Here, λ1 = λ2 = 0.28 is selected.

[0079] Combining Φ(x) and Ψ(x) using differentiable max functions:

[0080] ρ(x) = max{Φ(x), Ψ(x)}

[0081] The expression for a differentiable max function is:

[0082]

[0083] Among them, the function abs χ (xy) is calculated using the following formula:

[0084]

[0085] χ should be a small constant, such as χ = 0.01;

[0086] Based on the obtained ρ(x), the element density can be further obtained. The calculation formula is:

[0087]

[0088] Where H(x) is the approximate Heaviside function, and its specific calculation formula is:

[0089]

[0090] In the formula, the parameters γ and η control the smoothness and step point of the Heaviside function, respectively, and are taken as 7 and 0.8.

[0091] Step 3: Calculate the gradients of the two level set functions at the center point of each unit. model Then according to the calculation formula They were obtained respectively Then respectively p-norm condensation yields g φ g ψ Then, the constraint relation g can be obtained. φ ≤ξ、g ψ ≤ξ, where ξ is a positive number less than 1;

[0092] Specifically, the gradient of the level set function at the center point of each unit is calculated. and The modulus, specifically calculated as follows:

[0093]

[0094]

[0095] in:

[0096]

[0097]

[0098] In the formula The specific calculation formula is as follows:

[0099]

[0100]

[0101] In the formula, m is the number of radial basis functions; α i β i The weighting coefficients of the radial basis functions; x e q represents the coordinates of the unit center point; i d represents the node coordinates of the i-th radial basis function; min x is the support radius of the radial basis function; e y e x represents the coordinates of the unit's center point; i y i These are the coordinates of the radial basis function nodes.

[0102] According to the calculation formula They were obtained respectively Then respectively p-norm condensation yields g φ g ψ The specific calculation formula is as follows:

[0103]

[0104]

[0105] In the formula, n is the total number of finite element elements; parameter p w It is generally a positive integer not less than 2, and the value should not be too large. Here, we take 12.

[0106] Thus, the constraint relation g can be obtained. φ ≤ξ、g ψ ≤ξ, where ξ is a positive number less than 1, and here we take 0.3.

[0107] Step four, perform finite element analysis: through element density The height h(x) of each unit is calculated. e Establish the stiffness matrix K for each Mindlin plate element. e and mass matrix M e By assembling K e The overall stiffness matrix K is obtained, and then assembled by M. e The overall mass matrix M is obtained, and then the eigenvalues ​​λ1 are solved according to the characteristic equation formula.

[0108] The formula for calculating the height of each unit is:

[0109]

[0110] In the formula, h b Let h be the height of the matrix. b =0.2; h s Let h be the height of the mesh reinforcing rib. s =0.4; parameter p is the penalty factor, here p=3, according to Mindlin plate theory, the stiffness matrix K of each element e The formula for calculation is:

[0111]

[0112] Among them, K b0 K is the bending stiffness matrix of a unit thickness element. s0 h is the shear stiffness matrix for a unit thickness element. e It is h(x) e K is an abbreviation for finite element theory. According to finite element theory, K... b0 and K s0 The formula for calculation is:

[0113]

[0114]

[0115] In the formula, κ is the cross-sectional shear correction factor, which is taken as κ = 5 / 6 for matrix cross-sections; A e B represents a unit area region. f D is the bending strain matrix. f B is the bending elasticity matrix; c Let D be the shear strain matrix. c Let be the shear elasticity matrix. The above matrix can be calculated using methods such as isoparametric transformation and numerical integration in finite element analysis. Mainstream finite element analysis books and materials provide detailed explanations, so the relevant derivations and calculation processes will not be elaborated here.

[0116] Mass matrix M of each unit e The formula for calculation is:

[0117] M e =M0(h em )

[0118] In the formula, M0 is the mass matrix of a unit thickness element under the Mindlin plate element theory, and its expression is:

[0119]

[0120] In the formula, ρ is the element density; A is the area of ​​the element region; h is the element thickness, which is 1 here; and N is the shape function matrix.

[0121] h emHere is another expression for the height of each element, used to solve the mass matrix:

[0122]

[0123] In the formula, h em It is h m (x e () is an abbreviation for ); the parameter q is the penalty factor, which is taken as q=1 here.

[0124] By assembling the element stiffness matrix K e The overall stiffness matrix K is obtained, and the mass matrix M of the assembled unit is then used. e The overall mass matrix M is obtained, and then according to the formula... Solve for the eigenvalue λ1. ω1 is the fundamental characteristic frequency, which is used to complete the structural finite element analysis.

[0125] Step 5, define the mesh-stiffened structure optimization design problem: the design variable is the design point q. i The weighting coefficient α at the location i β i And the projection threshold parameter η, the design objective is to maximize the basic characteristic frequencies of the mesh stiffener structure, and the design constraint is the characteristic equation. Orthogonal normalization equation for mass and gradient constraint g φ ≤ξ、g ψ ≤ξ and α i β i The upper and lower bound constraints of η and the upper and lower bound constraints of volume constraints The mathematical expression of this optimization design problem is as follows:

[0126] Find: α i ,β i η (i = 1, 2...m)

[0127] max J=λ1

[0128]

[0129]

[0130] g φ g ψ ≤ξ

[0131] δ min ≤α i ,β i ≤δ max

[0132] η min ≤η≤ηmax

[0133]

[0134] Where, α i β i η and m are design variables, m is the number of radial basis functions; ω1 is the fundamental (first-order) characteristic frequency, and λ1 is the eigenvalue corresponding to the fundamental characteristic frequency, i.e. λ1 is the eigenvector corresponding to the eigenvalue; K is the global stiffness matrix; M is the global mass matrix; g φ g ψ φ(x) and ψ(x) are the gradient constraint values ​​for the level set functions, respectively, and ξ is the upper bound of the gradient constraint; δ min δ max These represent the design variables α and β, respectively. i and β i The upper and lower limits, η min η max These represent the upper and lower limits of the design variable η, respectively. V is the allowable total volume of the stiffener; V is the total volume of the stiffener, expressed as follows:

[0135]

[0136] Where n is the total number of finite element elements; h is the unit density; p is the penalty factor; h s For the thickness of the reinforcing rib; A e Let e ​​be the area of ​​the region in the e-th unit;

[0137] Step 6: Calculate the objective function λ1 with respect to the design variable α. i β i Sensitivity of η Calculate the gradient constraint function g φ For design variable α i Sensitivity Gradient constraint function g ψ For design variable β i Sensitivity Volume constraint function V on design variable α i Sensitivity Volume constraint function V on design variable β i Sensitivity Sensitivity of volume constraint function V to design variable η Specifically, using λ1 and h e h em The partial derivative relationship, h e h em and α i βi The partial derivative relationship of η and the eigenvector corresponding to each unit Calculate the objective function λ1 with respect to the design variable α i β i The sensitivity of η is specifically calculated using the following formula:

[0138]

[0139]

[0140]

[0141] in:

[0142]

[0143]

[0144]

[0145]

[0146]

[0147]

[0148]

[0149]

[0150]

[0151]

[0152]

[0153]

[0154] in:

[0155] A=tanh(γη), B=tanh[γ(ρ e -η)],C=tanh[γ(1-η)]

[0156] A′=1-A 2 B′=1-B 2 C′=1-C 2

[0157]

[0158]

[0159]

[0160]

[0161]

[0162]

[0163]

[0164] In the formula, n is the total number of finite element elements; The eigenvectors corresponding to the eigenvalues ​​of the unit; T denotes matrix transpose; e represents the e-th unit; h e It is h(x) e ) is an abbreviation for the thickness of each element; ρ e It is ρ(x) e () is an abbreviation for the element density function; K b0 K is the bending stiffness matrix of a unit thickness element. s0 M0 is the shear stiffness matrix per unit thickness element; M0 is the mass matrix per unit thickness element; h em This is another expression for the element thickness; p and q are penalty factors; h s The stiffener thickness is γ; η are threshold projection parameters, controlling the smoothness and step point of the Heaviside function, respectively; N is the shape function matrix. yes Φ is an abbreviation for element density; e It is Φ(x) e The abbreviation of Ψ e It is Ψ(x) e φ is an abbreviation for cosine transformation, representing the level set function after cosine transformation; e It is φ(x) e abbreviation of ) ψ e It is ψ(x) e ) is an abbreviation for two generated level set functions; λ1 and λ2 are the wavelengths of the cosine function; For compactly supported radial basis functions.

[0165] The constraint function V on the design variable α i Sensitivity The formula for calculation is:

[0166]

[0167] The constraint function V on the design variable β i Sensitivity The formula for calculation is:

[0168]

[0169] Sensitivity of constraint function V to design variable η The formula for calculation is:

[0170]

[0171] In the formula, A e The area of ​​a single cell is the ratio of the total area of ​​the design domain to the number of grid cells.

[0172] constraint function g φ For design variable α i Sensitivity The formula for calculation is:

[0173]

[0174]

[0175] constraint function g ψ For design variable β i Sensitivity The formula for calculation is:

[0176]

[0177]

[0178] In the formula, pw is a positive integer not less than 2; e represents the e-th unit; The gradient of the two level set functions at the center point of each unit The modulus; m is the number of radial basis functions; yes abbreviation, yes Abbreviation; x, y are direction coordinates;

[0179] Step 7, based on sensitivity Solve the basic characteristic frequency optimization design problem of the mesh stiffener structure defined in step five, and update the design variable α. i β i , η.

[0180] Specifically, based on the objective function λ1 and the constraint functions V and g φ g ψ and their sensitivity to design variables. The design variable α is updated using the Moving Asymptotes (MMA) method. i β i , η.

[0181] Using the updated α i β i The new level set function value φ(x) at the center point of the element is obtained by finding η. e ), ψ(x) e and unit density By plotting a grayscale image of the design domain height h(x), the new mesh stiffener structure is obtained.

[0182] Step 8: Repeat steps 2 through 7, each repetition being called an iteration, until the preset optimization termination condition is met, obtaining the optimal design of the mesh stiffener structure. The optimization termination condition is as follows:

[0183]

[0184] Where, ω err The basic characteristic frequency error is given by k, where k is the current iteration number and δ is the value of δ. ω Here, k is taken as the lower limit of the basic characteristic frequency error, which is 0.2%. max This is the upper limit for the number of iterations; here we take 500.

[0185] The optimization results of the preferred embodiment of the present invention are as follows: The optimized mesh reinforcing rib structure is as follows: Figure 4 As shown, the basic characteristic frequency value is approximately 11.09, which is 12% higher than the initial design. The optimized mesh stiffener structure has smooth and clear boundaries, which can meet manufacturing constraints. Figure 5 This is the iterative curve of the objective function of the optimization model for the maximum fundamental eigenfrequency of the mesh-reinforced structure. As can be seen from the figure, the objective function eventually converges to a relatively stable value.

[0186] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A stiffener structure frequency optimization method based on level set and variable density method, characterized in that, The method includes the following steps: S1 The design domain of the mesh stiffener structure to be optimized is divided into finite element meshes, thereby dividing the design domain into multiple elements. Multiple nodes in the finite element mesh are selected as nodes of the radial basis function. Two different weight coefficients are assigned to each node. The level set function at the center point of each element is obtained using the weight coefficients and the radial basis function. S2 uses the level set function to solve for the element density of each element, and uses the gradient constraint of the level set function to construct the constraint conditions for the width and spacing of the stiffeners, so that the width and spacing between adjacent stiffeners are uniform. S3. Set the height of the mesh stiffener structure to be optimized and solve the height of each element in combination with the element density. Use the height of each element to construct the characteristic equations of the stiffness matrix and mass matrix of each Mindlin plate element, and solve the basic characteristic frequency of the mesh stiffener structure to be optimized. S4. Taking the maximization of the basic characteristic frequency of the mesh stiffener structure to be optimized as the optimization objective, the width, spacing and volume of the stiffeners as constraints, and the weight coefficient and the projection threshold parameter controlling the width of the stiffeners as variables, an optimization model of the mesh stiffener structure to be optimized is constructed. The optimization model is solved until the variable values ​​that meet the preset termination conditions are obtained, thereby realizing the optimization design of the mesh stiffener structure to be optimized. In step S2, the constraints on the width and spacing of the reinforcing ribs are as follows: in, , The resulting gradient constraint function; is a parameter greater than 0 in the p-norm condensed method; n is the total number of finite element elements; It is the upper limit of the two gradient constraints after condensation; and These are the corresponding functions of the magnitudes of the gradients of the two level set functions at the center point of each unit; In step S3, the characteristic equation is performed according to the following relationship: in, The fundamental characteristic frequency The corresponding eigenvalues, i.e. ; Eigenvalues The corresponding feature vector; The overall stiffness matrix; For the overall quality matrix; In step S3, the stiffness matrix and mass matrix of each Mindlin plate element are determined according to the following relationship: in, and These are the stiffness matrix and mass matrix of each element, respectively; The bending stiffness matrix for a unit thickness element. The shear stiffness matrix for a unit thickness element. It is the unit height. abbreviation; This represents the mass matrix of a unit thickness element under Mindlin plate element theory. It is another expression for cell height. The abbreviation of .

2. The stiffener structure frequency optimization method based on level set and variable density method of claim 1, wherein, In step S1, the level set function is performed according to the following relation: in, and Let m be the two level set functions obtained; m is the number of radial basis functions. and These are the weight coefficients of the i-th radial basis function; These are the coordinates of the center point of the unit; These are the node coordinates of the i-th radial basis function; These are radial basis functions.

3. The method for optimizing the frequency of stiffener structures based on level sets and variable density method as described in claim 1 or 2, characterized in that, In step S2, the unit density is determined according to the following steps: S21 uses the cosine function to apply the level set function , The function is obtained by transformation. , ; S22 uses a differentiable max function to describe the... , Combining the results yields the density function. ; S23 Calculate the density function using an approximate Heaviside function to obtain the corresponding cell density. .

4. The method for optimizing the frequency of stiffener structures based on level sets and variable density method as described in claim 3, characterized in that, In step S22, the density function is performed according to the following relationship: in, Let the density function be... and The function obtained by performing a cosine transform on a level set function. Take a constant value.

5. A method for optimizing the frequency of a stiffener structure based on level sets and variable density method as described in claim 3 or 4, characterized in that, In step S23, the element density is determined according to the following relationship: in, This represents the center point of the element; here, the parameter value of the center point represents the parameter value of the entire element. The element density is obtained after calculation using the approximate Heaviside function; It is the density function; It is an approximate Heaviside function.

6. A method for optimizing the frequency of a stiffener structure based on level sets and variable density method as described in claim 1 or 2, characterized in that, In step S4, the optimization model is performed according to the following relationship: in, , , , where are the design variables, and are the weight coefficients of the radial basis functions and the threshold projection parameters of the approximate Heaviside function, respectively; m is the number of radial basis functions; The fundamental characteristic frequency, The eigenvalues ​​corresponding to the basic characteristic frequencies, i.e. ; Eigenvalues The corresponding feature vector; The overall stiffness matrix; For the overall quality matrix; , These are the level set functions , The function obtained by transforming the magnitude of the gradient and condensing it with p-norm; This represents the upper limit of the two gradient constraints after condensation; , These represent design variables. and The upper and lower limits; , These represent design variables. The upper and lower limits; It is the allowable total volume of the reinforcing ribs; It is the total volume of the reinforcing ribs.

7. The method for frequency optimization of stiffener structures based on level set and variable density method as described in claim 6, characterized in that, In step S4, the preset termination condition is determined according to the following relationship: or in, The basic characteristic frequency error; This represents the current iteration number; This is the lower limit of the error in the basic characteristic frequency. This represents the upper limit of the number of iterations.

Citation Information

Patent Citations

  • Noise-reducing device and method for reducing noise

    CN102132339A

  • Energy finite element topological optimization method based on parameterized level set

    CN112446163A