Laminated plate damage parameter identification method based on convex model regularization thought

Through the method based on the regularization idea of ​​convex model, the uncertainty correlation of composite laminated plate structure is quantified and regularization parameters are optimized, which solves the problems of uncertainty and disquality in the structural damage identification of composite laminated plate structures, and improves the damage The accuracy and stability of the identification of parameter boundaries.

CN118939906APending Publication Date: 2024-11-12JIANGSU UNIV OF TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410963648.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-18
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

During the damage identification process of composite laminated board structure, direct damage identification of each unit in the structure will lead to huge calculations, and due to noise interference and uncertainty, pathological problems are prone to occur, leading to discomfort in solving the problem in the reverse problem. Qualitatively, the traditional regularization method of damage recognition lacks the effect of correlation factors on regularization parameters.

Method used

Using a method based on the idea of ​​regularization of convex model, by measuring structural response information, quantifying the correlation between uncertainties, initially screening the damage units, establishing the sensitivity matrix and regularization parameters of convex model to solve the equation, and obtaining regularization under uncertain conditions Use convex model theory to eliminate unpleasant qualities and optimize regularized parameter selection.

Benefits of technology

Improve the convergence and robustness of the boundary solution process of damage parameter, and can achieve robust selection of regularized parameters under noise interference, clarify the impact of measurement errors under quantification uncertainty conditions, and enhance the stability of the damage recognition process. and reliability.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118939906A_ABST
    Figure CN118939906A_ABST
Patent Text Reader

Abstract

The invention discloses a laminated plate damage parameter identification method based on a convex model regularization thought, and relates to the technical field of structure damage identification, and the method comprises the following steps: S1, measuring the structure response information of a local measurement point of a composite material laminated plate structure; s2, determining the boundary of the response information of the composite material laminated plate structure by using a convex model theory so as to quantify the correlation between uncertainties; s3, preliminarily screening a damage unit of the composite laminated plate by utilizing a damage positioning index phi NCR; s4, establishing a sensitivity matrix of the structural response information to the unit-level damage parameters; s5, taking the boundary of the convex model of the response information and the sensitivity matrix as input, establishing a convex model regularization parameter solving equation, and obtaining regularization parameters under the uncertain condition of the convex model; and S6, solving an identification boundary of the unit-level damage parameters of the composite material laminated structure by using a convex model regularization method. The method aims to enhance the stability and reliability in the process of solving the damage parameter boundary.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of structural damage identification, and in particular to a laminate damage parameter identification method based on a convex model regularization concept. Background Art

[0002] Fiber reinforced composite laminate structures have been widely used in aviation, aerospace, transportation, electronics, energy and other fields due to their high specific strength and high specific modulus. However, in actual service, composite laminate structures will be damaged by fiber breakage, matrix cracking, delamination and debonding due to impact, tensile and other loads, and the identification process is difficult. In order to simplify the damage identification process, this patent uses unit-level damage parameters to describe the damage. This indicator has the advantages of fewer damage parameters and more damage information.

[0003] In the damage identification process, directly identifying the damage of each unit in the structure will lead to a huge amount of calculation. Therefore, the present invention first locates the damage, preliminarily screens the damaged units, and then performs quantitative damage identification. The damage identification process will inevitably be affected by noise, considering that there will be correlation between uncertainties. In the process of quantitative identification of the degree of damage, it is necessary to invert the matrix, and the matrix inversion will cause ill-posed problems, resulting in the ill-posedness of the inverse problem solution. The regularization method is an effective method to solve the ill-posed problem, but the traditional damage identification regularization method lacks consideration of the influence of correlated uncertainty factors on the selection of regularization parameters. To this end, the present invention proposes a convex model regularization idea to quantify the identification boundary of unit-level damage parameters. Summary of the invention

[0004] The purpose of the present invention is to provide a laminate damage parameter identification method based on the convex model regularization concept to solve the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solution: a laminate damage parameter identification method based on convex model regularization idea, comprising the following steps:

[0006] S1. Measure the structural response information of local measuring points of composite laminate structure;

[0007] S2. Use convex model theory to determine the boundaries of composite laminate structure response information and quantify the correlation between uncertainties;

[0008] S3. Using the damage location indicator Φ CR Preliminary screening of damaged elements in composite laminates;

[0009] S4, establishing a sensitivity matrix of structural response information to element-level damage parameters;

[0010] S5. Taking the convex model boundary and sensitivity matrix of the response information as input, a convex model regularization parameter solution equation is established to obtain the regularization parameter of the convex model under uncertainty conditions;

[0011] S6. Use the convex model regularization method to solve the identification boundary of the unit-level damage parameters of the composite laminate structure.

[0012] Preferably, a damage identification method combining regularization idea with convex model theory is proposed. The dispersion of uncertain damage parameters is quantified through the convex model idea, and the damage boundary of structural damage parameters is obtained under the condition of limited test samples.

[0013] Preferably, a method for solving the regularization parameter taking into account uncertainty factors such as test errors is provided, wherein the method for determining the regularization parameter of the convex model is implemented by the following optimization model:

[0014]

[0015] in, is the partial derivative of the GCV function with respect to the uncertainty parameter, COV(X,X) is the covariance between the uncertain parameters; I is the identity matrix; γ is a non-negative regularization parameter, S is the sensitivity matrix of the response information to the damage parameters of the composite laminate element level, Δe is the response residual vector between the reference model and the actual model, X is the measurement error, and Y r Indicates the reference value of the response, Y a Represents the actual value of the response quantity.

[0016] Preferably, the stiffness matrix of the laminate structure based on the element-level damage parameter can be expressed as follows:

[0017]

[0018] in They represent the in-plane, coupled, out-of-plane and shear stiffness matrices of the composite laminate structure in an undamaged state, is the laminated element stiffness parameter in the in-plane, coupling, out-of-plane and shear directions of the i-th element. A value of 0 indicates complete damage, and a value of 1 indicates no damage. This damage index not only reflects the loss of various bearing capacities of structural components, but also effectively reduces the scale of parameters to be solved. At this time, the stiffness matrix of the structure can be expressed as follows:

[0019] Where m is the number of elements in the laminate, K i Represents the element stiffness matrix of the ith element including the damage parameters.

[0020] Preferably, the uncertainty variable of the laminate structure damage parameter can be represented by a central value and an uncertain boundary, wherein the central value solution formula of the structural damage parameter is:

[0021] Δθ(X c )=((S c ) T S c +γI) -1 S c Δe(X c )

[0022] Among them, Δθ is the residual vector of the relative stiffness coefficient between the reference model and the actual model, that is, the damage parameter, the superscript "c" represents the central value, S is the sensitivity matrix of the structural response information to the unit stiffness coefficient, I is the unit matrix, and Δe is the residual vector of the structural response between the reference model and the actual model;

[0023] The unit-level damage parameter θ obtained by using the convex model regularization idea i The boundary of (X) is expressed as:

[0024]

[0025] in, represents the central value of the damage parameter in the o direction of the i-th unit, Cov(X,X) is the covariance, is the partial derivative of the damage parameter of the i-th unit in the o direction with respect to the uncertainty variable, and “-” and “_” represent the upper and lower bounds of the damage parameter.

[0026] Compared with the prior art, the present invention has the following beneficial effects:

[0027] 1. Aiming at the uncertainty and ill-posedness problems in the process of structural damage identification, the present invention proposes a regularization method based on convex model.

[0028] 2. This method can effectively improve the convergence of the damage parameter boundary solution process by introducing a convex model regularization term. When determining the regularization parameter, the present invention comprehensively considers the influence of the GCV regularization parameter and uncertainty on the dispersion of the selected regularization parameter, thereby achieving the robustness of the regularization parameter value under noise interference.

[0029] 3. In addition, the present invention can also use convex model theory to quantify the impact of measurement errors under uncertainty conditions, which has clearer engineering significance for the damage identification process. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 is a flow chart of the present invention;

[0031] Figure 2A schematic diagram of the numbering of the clamped beams at both ends of the composite laminate of the present invention;

[0032] Figure 3 is the damage location index Φ NCR Schematic diagram for identifying damaged units;

[0033] Figure 4 This is a curve result diagram of selecting the optimal regularization parameter based on the convex model regularization concept according to a specific embodiment of the present invention. DETAILED DESCRIPTION

[0034] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0035] like Figure 1 As shown, the present invention provides a method for identifying damage parameters of composite laminate structures based on convex model regularization, comprising the following steps:

[0036] S1. Measure the structural response information of the local measuring points of the composite laminate structure according to the sensors arranged on the laminate.

[0037] S2. Convex model theory is used to determine the boundaries of the structural response information of composite laminates to quantify the correlation between uncertainties.

[0038] Specifically, step S2 includes: using the convex model regularization idea to describe the uncertainty quantification technology to quantitatively analyze the response information of the local measurement points of the structure, and the uncertainty parameter boundary can be expressed as:

[0039]

[0040] S3. Use damage localization indicators to preliminarily screen damaged units of composite laminates.

[0041] By calculating the modal strain energy before and after structural damage, the damage location index Φ is calculated NCR

[0042]

[0043] in,

[0044]

[0045] Where m represents the number of units, μ(·) and σ(·) represent the mean and standard deviation of the vector, respectively, and the superscripts “d” and “u” represent the damaged structure and the undamaged structure, respectively. According to the definition, Φ NCR The larger the index value, the greater the possibility that the unit is damaged. Units with larger indexes are considered damaged units.

[0046] S4. Establish a sensitivity matrix of structural response information to element-level damage parameters.

[0047] Specifically, for a finite element model of a structure, its stiffness matrix can be expressed as

[0048]

[0049] Where m is the number of elements in the structural finite element model; is the stiffness matrix of the ith element in the o direction, where o represents the direction of damage, including A for in-plane damage and D for out-of-plane damage. Similarly, the changed stiffness matrix K c It can also be expressed as:

[0050]

[0051] in It represents the element-level damage parameter of the i-th element in the o direction affected by the uncertain variables.

[0052] The model correction based on the sensitivity matrix can be expressed as

[0053] SΔθ=Δe

[0054] The sensitivity matrix is ​​written in the corresponding specific form as

[0055]

[0056] The residual vector of the structural response is

[0057]

[0058] The damage parameters of the structure can be expressed as follows:

[0059]

[0060] Where n is the number of structural response parameters and m is the number of damaged elements.

[0061] S5. Taking the convex model boundary and sensitivity matrix of the response information as input, establish the convex model regularization parameter solution equation to obtain the regularization parameter of the convex model under uncertainty conditions.

[0062] Furthermore, the robust regularization parameter solution equation is:

[0063] min V(γ)+ΔV(γ)

[0064] in,

[0065] Where I is the identity matrix, γ is a non-negative regularization parameter, S is the sensitivity matrix of the response information to the damage parameters of the composite laminate element level, and Δe X is the residual vector of the structural response parameters between the reference model and the actual model, is the partial derivative of the GCV function with respect to the uncertainty parameter, COV(X,X) is the covariance matrix between the uncertain parameters; X is the measurement error. The regularization parameter curve is a curve with γ as the horizontal coordinate and V(γ)+ΔV(γ) as the vertical coordinate.

[0066] S6. Using the convex model regularization method, the identification boundary of the unit-level damage parameter of the composite material structure is solved. This includes solving the central value of the damage parameter and solving the boundary of the damage parameter. The formula for solving the median value of the structural damage parameter is:

[0067] Δθ(X c )=((S c ) T S c +γI) -1 S c Δb(X c )

[0068] Among them, Δθ is the residual vector of the relative stiffness coefficient between the reference model and the actual model, that is, the damage parameter, the superscript “c” represents the center value, S is the sensitivity matrix of the modal parameters to the unit stiffness coefficient, γ is a non-negative robust regularization parameter, I is the unit matrix, and Δb is the residual vector of the modal parameters between the reference model and the actual model.

[0069] Furthermore, the damage parameter boundary based on the convex model regularization idea is:

[0070] (θ-θ c ) T C(θ r ,ρ) -1 (θ-θ c )≤1

[0071]

[0072] in,

[0073] Considering the uncertainty and its correlation, these uncertainty parameters can be expressed as denote the tension-compression, coupling, bending and shear stiffness matrices respectively, is the covariance between the ith and jth impairment parameters, is the correlation coefficient between the ith uncertain variable and the jth uncertain variable.

[0074] According to the damage parameter identification method of the composite laminate structure based on the convex model regularization concept proposed in an embodiment of the present invention, the structural response information of the local measuring points of the composite laminate structure is first measured, and then the boundary of the structural response information of the composite laminate is determined by using the convex model theory to quantify the correlation between uncertainties, and then the damage unit of the composite laminate is preliminarily screened by using the damage location index, and then a sensitivity matrix of the structural response information to the unit-level damage parameters is established, and then the convex model boundary and the sensitivity matrix of the response information are used as input to establish a convex model regularization parameter solution equation to obtain the regularization parameter under the uncertainty condition of the convex model, and finally the convex model regularization method is used to solve the identification boundary of the unit-level damage parameters of the composite structure.

[0075] Therefore, the present invention aims at the uncertainty and ill-posed problems in damage identification of composite materials, and proposes a structural damage identification method based on the idea of ​​convex model regularization, aiming to enhance the stability and reliability when solving the damage parameter boundary.

[0076] Example

[0077] In order to more clearly describe the characteristics of the invention and its adaptability to actual engineering structures, the invention identifies the structural damage parameters of the clamped beam at both ends based on the convex model regularization idea. The clamped beam structure is divided into 15 equally divided units, and the unit numbers are as follows: Figure 2 The two-end clamped beam consists of 8 layers, and the ply order is [0° / 90° / 90° / 0° / 0° / 90° / 90° / 0°]. Each layer is 0.2mm thick, 300mm long, and 20mm wide. The elastic modulus of the unidirectional tape in the fiber direction is 185GPa, the elastic modulus in the matrix direction is 10.5GPa, and μ 12 =0.28, μ 23 =0.3, density is 1700kg / m 3 .

[0078] The specific damage condition settings are shown in Table 1:

[0079] Table 1 Damage condition settings

[0080] Damage unit number <![CDATA[θ A ]]> <![CDATA[θ D ]]> Unit 2 0.9 0.85 Unit 8 0.9 0.85 Unit 13 0.9 0.85

[0081] By using the method proposed in the present invention to identify the damage of the laminated plate fixed-support beam structure at both ends of the embodiment, firstly, the damage location index Φ NCRThe damage of the laminate structure unit is preliminarily screened to determine the possible damaged units. NCR When the index is greater than 1, the unit is a damaged unit. Therefore, units 2, 8, and 13 are selected as damaged units. Figure 3 shown.

[0082] Then, based on the damage location, the static displacement response is used as the structural response, and the sensitivity matrix of the structural response information to the unit-level damage parameters is further used. The boundary and sensitivity matrix of the convex model are used as input to establish and solve the regularization parameter equation of the convex model, and select the optimal regularization parameter, such as Figure 4 As shown. Finally, the convex model regularization method is used to quantitatively identify the damage of the composite laminate structure. The structural response noise level is set to 1%, and the correlation between the test information of each measuring point is set to 0. The damage parameter boundaries identified by the convex model regularization idea are shown in Table 2. It can be seen that the proposed method can obtain damage parameters containing true values, so that the damage of the laminate structure can be effectively identified.

[0083] Table 2 Boundary values ​​of damage parameters

[0084]

[0085] In summary, the damage parameter identification method for composite laminated structures based on the convex model regularization idea proposed in the present invention eliminates the ill-posedness caused by solving the inverse problem by introducing a convex model regularization term, optimizes the selection of regularization parameters, and quantifies the influence of measurement errors under uncertainty conditions. It can effectively improve the convergence and robustness in the process of structural damage identification, and has important theoretical significance and application value.

[0086] It is known from common technical knowledge that the present invention can be implemented by other embodiments that do not deviate from its spirit or essential features. Therefore, the above disclosed embodiments are only illustrative in all respects and are not exclusive. All changes within the scope of the present invention or within the scope equivalent to the present invention are included in the present invention.

Claims

1. A laminate damage parameter identification method based on convex model regularization concept, characterized by: The following steps are involved: S1. Measure the structural response information of local measuring points of composite laminate structure; S2. Use convex model theory to determine the boundaries of composite laminate structure response information and quantify the correlation between uncertainties; S3. Using the damage location indicator Φ NCR Preliminary screening of damaged elements in composite laminates; S4, establishing a sensitivity matrix of structural response information to element-level damage parameters; S5. Taking the convex model boundary and sensitivity matrix of the response information as input, a convex model regularization parameter solution equation is established to obtain the regularization parameter of the convex model under uncertainty conditions; S6. Use the convex model regularization method to solve the identification boundary of the unit-level damage parameters of the composite laminate structure.

2. The laminate damage parameter identification method based on convex model regularization concept according to claim 1 is characterized by: A damage identification method combining regularization idea with convex model theory is proposed. The dispersion of uncertain damage parameters is quantified through the convex model idea, and the damage boundary of structural damage parameters is obtained under the condition of limited test samples.

3. The composite laminate structure damage identification method based on convex model regularization concept according to claim 1 is characterized in that: a regularization parameter solution method considering uncertainty factors such as test errors is provided, wherein: The method for determining the regularization parameter of the convex model is implemented by the following optimization model: in, is the partial derivative of the GCV function with respect to the uncertainty parameters, COV(X,X) is the covariance between the uncertain parameters; I is the identity matrix; γ is a non-negative regularization parameter, S is the sensitivity matrix of the response information to the damage parameters of the composite laminate element level, Δe is the response residual vector between the reference model and the actual model, X is the measurement error, and Y r Indicates the reference value of the response, Y a Represents the actual value of the response quantity.