A semi-analytical method for analyzing the dynamic response of composite material compartments

By separating the frame and interface parameters using a semi-analytical method, the problem of parameter compensation in the dynamic analysis of composite material compartments was solved, achieving efficient and accurate dynamic response analysis.

CN123120153APending Publication Date: 2026-11-17BEIHANG UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611499323.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-24
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

In the current technology for dynamic analysis of composite material compartments, the equivalent parameters of the frame and the interface connection parameters compensate for each other, making it difficult to distinguish between the equivalent error of the frame and the contribution of the interface flexibility, which leads to the accumulation of dynamic analysis errors.

Method used

A semi-analytical method is adopted, which separates the independent frame from the interface parameters. By using the boundary energy matrix and coupling potential energy matching, the equivalent parameters of the frame and the interface connection parameters are determined. The dynamic matrix is ​​constructed by combining the first-order shear deformation theory, and the response analysis is carried out.

Benefits of technology

It enables accurate separation of the frame and interface parameters while keeping the structure and connection state unchanged, limits the transfer of errors to the interface, and improves the accuracy and efficiency of dynamic analysis.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123120153A_ABST
    Figure CN123120153A_ABST
Patent Text Reader

Abstract

This invention discloses a semi-analytical method for analyzing the dynamic response of composite material compartment sections, belonging to the field of structural dynamics analysis. Under small deformation linear elastic conditions with unchanged connection states, the composite cylindrical section and the upper and lower connecting frames are represented as a cylindrical shell and an equivalent ring plate, respectively. The ring plate parameters are determined and fixed by matching the boundary energy and inertia of the independent frame. Under the same boundary displacement state, the residual energy at the interface is separated, and the working condition is selected by the ratio of projection error and the equivalent energy deviation of the frame to the residual energy at the interface. The fourth-order symmetric semi-definite interface stiffness matrix is ​​identified using three-dimensional relative translation and relative rotation around the tangent of the common circumference. A semi-analytical model is constructed based on first-order shear deformation theory and energy variation to solve for the absolute acceleration power spectral density and root mean square of the axial random foundation excitation, which are then output after independent verification. This method restricts the transfer of the equivalent error of the frame to the interface parameters and is used for compartment vibration response analysis.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of composite material structural dynamics analysis technology, specifically to a method for analyzing the dynamic response of composite material compartments based on a semi-analytical approach. Background Technology

[0002] The composite material compartment transfers loads through the upper and lower docking frames. The deformation of its cylindrical section, docking frames, and connection interfaces jointly affects the vibration response. Predicting the acceleration power spectral density and root mean square under random foundation excitation is an important basis for compartment structural design and vibration environment assessment.

[0003] However, refined finite element models typically have many degrees of freedom while preserving local connection details, and the cost of multiple response calculations increases with the model size; simplified models, on the other hand, require simultaneous handling of component equivalence and interface compliance. Existing methods for analyzing composite shells include using elastic couplers, displacement expansion, and energy methods to establish semi-analytical models, as well as reducing the system's degrees of freedom through substructure model condensation, providing different approaches to reducing the scale of dynamic analysis.

[0004] When applying the above modeling approach to composite material modules with docking frames, it is still necessary to determine the equivalent stiffness and inertia of the docking frames and the flexibility of the connection interface. If the frame and interface parameters are adjusted simultaneously based solely on the overall dynamic response, the influence of the two types of parameters may compensate for each other, causing the equivalent error of the frame to be included in the interface stiffness. Matching only the frame mass or the stiffness of a single deformation is also insufficient to constrain the energy deviation under combined deformation. Therefore, it is necessary to separate and determine the two types of parameters based on the mechanical equivalence of the components, and to provide a dynamic solution and applicability verification method consistent with this separation process. Summary of the Invention

[0005] This invention provides a method for analyzing the dynamic response of composite material compartments based on a semi-analytical approach, in order to solve the problems of mutual compensation between the equivalent parameters of the frame and the interface connection parameters, and the difficulty in distinguishing the equivalent error of the frame and the contribution of the interface flexibility.

[0006] The analyzed section includes a composite cylindrical section, an upper docking frame, and a lower docking frame, which are represented as cylindrical shell elements and parameterized equivalent annular plate elements at both ends, respectively. Under the condition of small deformation linear elasticity with the connection interface state remaining unchanged, the following steps are adopted.

[0007] S1, obtain the geometry, layup and material parameters of the cylindrical segment and the finite element reference models of the upper and lower local connections; each reference model includes the corresponding frame, adjacent cylindrical segments and finite flexibility connection interfaces, divides the energy components that do not overlap, and establishes independent frame models and independent cylindrical segment models with consistent geometry and material parameters.

[0008] S2, apply the same boundary displacement mode and pairwise combination mode to the independent frame and the corresponding equivalent ring plate, and construct the boundary energy matrix with cross terms from the strain energy; under the mass, centroid, inertia tensor error and constitutive and inertial positive definite constraints, minimize the difference of the energy matrix, and determine and fix the parameters of the upper and lower ring plates respectively.

[0009] In S3, the boundary displacements of the frame and cylindrical section obtained from the local static calibration are reapplied to each independent model. The residual energy at the interface is obtained by subtracting the strain energy of the two independent models from the total local strain energy. Only the working conditions with a positive energy resolution lower limit are retained, and the frame boundary displacement is projected to the equivalent ring plate. Only the working conditions with acceptable projection error and an absolute value of the equivalent energy difference of the frame not greater than a preset proportion of the residual energy at the interface are used in S4. This proportion is greater than zero and less than one. If the projection is acceptable but the energy difference exceeds the limit, the corresponding displacement mode is added to S2 to redetermine, fix the parameters, and recalculate all candidate working conditions.

[0010] S4. Within the same circular connecting line and coordinate system, a four-dimensional interface motion is formed by three-way relative translation and relative rotation around the common circumferential tangent. The rotation angle is scaled using positive characteristic length, and fourth-order symmetric positive semi-definite stiffness matrices with constant values ​​along the circumference are set for the upper and lower interfaces respectively. Under the condition that the identification matrix is ​​of full column rank, the coupling potential energy obtained by integrating the four-dimensional motion quadratic form along the connecting line is used to match the residual energy of the interface, keeping the ring plate parameters unchanged and determining the interface stiffness.

[0011] S5 establishes the strain energy and kinetic energy of three elements based on the first-order shear deformation theory. Using the finite term expansion coefficients of the displacement allowable function as generalized coordinates, and combining interface coupling potential energy and boundary constraints, the mass and stiffness matrices are formed through energy variation, and the damping matrix is ​​set.

[0012] S6, input the steady random foundation acceleration power spectral density transmitted from the lower docking frame along the axial direction of the compartment, construct an inertial load consistent with the mass matrix and the foundation motion influence vector, solve the frequency domain response relative to the foundation motion, reconstruct the absolute acceleration power spectral density at the specified measurement point and direction, and obtain the root mean square by integrating over the preset target frequency band.

[0013] Furthermore, the generalized stiffness and thickness mass moment of the ring plate are taken as constant values ​​within a predetermined radial partition; the interface energy extraction uses the same material, linearization, and boundary displacement state. The interface retains the cross stiffness of the four motion components, and the working conditions are identified by filtering through the normalized rank and condition number. After the parameters are fixed, the frame energy, interface energy, generalized boundary reaction force, and complex transfer function are checked using working conditions not involved in the fitting. The response is output when the deviations of each item are qualified; otherwise, a check failure flag is generated.

[0014] The beneficial effects of this invention are as follows: the parameters of the ring plate are determined by matching the boundary energy and inertia of the independent frame, and the interface energy is separated and the frame parameters are fixed under the same displacement state to identify the connection stiffness, so that the two types of parameters correspond to the mechanical properties of the component and the interface flexibility, respectively; the equivalent deviation of the frame is screened by using the residual energy of the interface as a scale, which can limit the transfer of frame error to interface parameters; the cross energy term and semi-positive definite stiffness constraint take into account the combined deformation and non-negative energy storage; the semi-analytical matrix can be reused for response calculation when the structure and connection state remain unchanged; and the independent verification makes the output correspond to the clear target frequency band and error conditions. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating the overall process of the method of the present invention.

[0016] Figure 2 The diagram shows the compartment structure and equivalent coupling model, where (a) is the compartment analysis object, (b) is the coupling model of the cylindrical shell and the upper and lower equivalent ring plates, and (c) is the common coordinates and four-component relative motion at the connecting line.

[0017] Figure 3 A flowchart for determining the separation of frame and interface parameters is provided, in which (a) is the frame boundary energy and inertia matching, and (b) is the displacement reapplication, interface energy separation, working condition screening and connection stiffness identification.

[0018] In the figure: 101, composite material cylindrical segment; 102, upper docking frame; 103, lower docking frame; 201, cylindrical shell element; 202, upper equivalent ring plate element; 203, lower equivalent ring plate element; 301, upper interface elastic coupler; 302, lower interface elastic coupler; Response measurement points; Basic acceleration. Detailed Implementation

[0019] The present invention will be further described below with reference to the accompanying drawings and embodiments. The composite material compartment analyzed consists of a composite material cylindrical section 101, an upper docking frame 102, and a lower docking frame 103, which are represented by cylindrical shell unit 201, upper equivalent ring plate unit 202, and lower equivalent ring plate unit 203, respectively. An upper interface elastic coupler 301 and a lower interface elastic coupler 302 are respectively provided at the two connection positions. Figure 1 Indicates the calculation process, Figure 2 This indicates the relationship between the structure and the interface motion. Figure 3 This describes the process of separating and determining the parameters of the frame and the interface.

[0020] This implementation uses a small-deformation linear elastic model, maintaining the material state, interface state, and basic constraint form unchanged in a single calculation. Interface flexibility corresponds to recoverable elastic deformation near the selected equilibrium state; conditions involving contact opening / closing, macroscopic slippage, material damage, or significant amplitude correlation are not treated using the same set of constant interface stiffness values. The equivalent ring plate is a mechanical model with generalized stiffness and inertial parameters, whose parameters are not directly equivalent to those calculated by treating the overall height of the docking frame as the thickness of a uniform solid plate.

[0021] Take the axial, radial, and circumferential basis vectors as , , ,satisfy The radius of the reference surface of the cylindrical shell is , length is The axial coordinate x is measured from the bottom end, and the two ring plates adopt... The coordinates, whose reference plane intersects the shell reference plane at the connecting line. , Distinguish between upper and lower interfaces. Figure 2 (b) The separate drawing only illustrates the equivalent elements and coupling relationship, and does not indicate the existence of axial gaps between reference surfaces.

[0022] Input the actual geometry, the thickness and angle of each layer, the elastic constants and density of the material; predetermine the target frequency band. Measurement points and directions, basic input spectrum, damping, radial partitioning, calibration and independent verification set, error limits and normalized positive lower limit. Each element retains three displacements and two cross-sectional rotation fields, and the four components only limit the common motion space of the interfaces. The material and connection state remain unchanged in a single calculation; the same set of linear parameters is not used for working conditions involving contact opening and closing, slippage or damage.

[0023] In step S1, upper and lower local reference models are established respectively. The frame, adjacent cylindrical segments, and the physically flexible connection region are divided into non-overlapping energy sets. Independent frame and independent cylindrical segment models are replicated with the same geometry, material, reference surface, and linearization state, preserving the node correspondence and all exposed connection and truncation boundaries. The truncation position is determined by comparing the results of moving the truncation boundary; numerical penalty constraints, artificial stabilization, and dissipated energy are not included in the elastic performance of the physical interface. The interface adopts a massless coupling approximation, and it must be confirmed that the inertia of the actual connection region is negligible within the target frequency band, and that the mass attribution of each model is consistent.

[0024] Step S2 Figure 3 (a) Implementation. Axial, radial, and circumferential translation and common tangential rotation modes are set at the corresponding inner and outer boundaries according to the same coordinate direction and amplitude scale. The circumferential distribution is taken as... , , , A positive integer is selected. The rotation mode of the solid frame is formed by the cross product of the reference line translation, the rotation angle, and the node offset vector to obtain the nodal displacement; both models use the same remaining boundary conditions and eliminate rigid body motion using the same rules. Let , The dimensionless mode magnitude vector, The mode matrix, including displacement amplitude scale, is obtained by using unit modes and pairwise combinations of modes:

[0025] ,

[0026] ,

[0027] ,

[0028] ;

[0029] For the boundary energy matrix, , For the corresponding unit coefficient vector; Let g be the boundary reaction force at the node, and g be the reaction force at the node. The generalized reaction force of virtual work conjugate. Retaining off-diagonal cross terms, d patterns are adopted. A symmetric matrix is ​​constructed from the energy values ​​of each single mode and its combination. The independent frame matrix is ​​denoted as... The ring plate matrix is ​​denoted as , This is the parameter vector for the ring plate.

[0030] The ring plate has generalized tensile, tension-bending coupling, bending, and transverse shear stiffness matrices A, B, D, S, and thickness mass moment. , , The description states that each parameter takes constant values ​​within a pre-fixed radial partition. Let... For density, For the thickness coordinates relative to the annular reference plane, integrate along the thickness and apply the following objectives and constraints:

[0031] , ;

[0032] , ;

[0033] , , ;

[0034] In the formula It is the Frobenius norm. It is a positive energy scale; composed of A, B, BT The block matrices of D arranged in the top left, top right, bottom left, and bottom right directions, and S are all positive definite. The total mass, centroid, and inertia tensor errors about the same reference point and coordinate system each meet preset limits; mass and inertia are determined by... , and Seeking, Let I be the position of the material point relative to the reference point, and let I be the identity matrix. The centroid is obtained by dividing the first-order mass integral by the mass. Minimize using constrained optimization. Save and fix the parameters of the upper and lower ring plates respectively.

[0035] During implementation, after unifying the strain and curvature to a uniform scale based on characteristic lengths, the positive definite stiffness block matrix can be parameterized using the Cholesky factor. Constrained optimization is employed to solve the problem and check for inertial conditions and mass errors. The radial partition boundaries remain unchanged during the first fitting, and the frame parameters are not used as joint optimization variables for subsequent interface identification.

[0036] Step S3 Figure 3 (b) Implementation. Axial, radial, circumferential, and tangential rotations, as well as mixed static loads, are applied to the local model. Only boundary loads are applied; no volume loads are applied. Extracting the total elastic strain energy under each working condition All connection and truncation boundary displacements of the frame and cylindrical segments are applied one by one to the independent model, with internal nodes in free equilibrium, to obtain the following results. , ,calculate:

[0037] ;

[0038] in, These represent the upper and lower interfaces, respectively. The working condition is numbered; the static calibration working condition applies only boundary loads and not volume loads, the three energy terms correspond to the same material state and the same linearization state, the boundary displacement field includes the displacements on the exposed connecting boundary after separation and the local model truncation boundary; when there is an initial preload, each energy term adopts the incremental displacement quadratic energy relative to the same equilibrium state; it will satisfy... The operating conditions are retained as candidate calibration operating conditions, among which This is the lower limit of energy resolution.

[0039] For candidate working conditions, the motion of the frame is represented as a displacement vector at the same sampling points on the corresponding boundaries. ;by The mapping from the ring plate boundary pattern to the sampled displacement is represented by least squares projection. Cross-sectional rotation is represented by displacement across thickness nodes, maintaining consistency in the dimensions of the displacement norm. The mode coefficients are obtained from the projection. Displacement constraints are generated at the corresponding boundaries of the equivalent ring plate with fixed parameters, while the internal degrees of freedom remain in free equilibrium. The elastic strain energy is then calculated. ; reconstructed displacement at the sampling point Compared with the original displacement The difference is used to calculate the projection error, and the selection criteria are:

[0040] ,

[0041] ;

[0042] in For displacement scale, For projection error limits, The preset ratio is used. Conditions with excessive projection error are not included in the identification; if the projection is qualified but the energy difference exceeds the limit, the displacement mode is added to the S2 matching set, the parameters are re-determined, the projection and energy deviation of all candidate conditions are recalculated, and then the parameters are fixed; only conditions that pass the two screenings enter S4.

[0043] Step S4 First press Figure 2 (c) Unify the movement of both sides to the same connecting line and common coordinates, and define three relative translations and relative cross-sectional rotations around the common circumferential tangent, defined in the same positive direction, based on the shell side and the reduced plate side:

[0044] , , ,

[0045] For example, the displacement of the shell mid-surface along Recorded as Displacement along the mid-surface of the plate Recorded as The axial orientation of the shell's radial normal changes as follows: The radial direction of the plate's axial normal changes as follows ,but Both the upper and lower ring plates are made of For the normal direction, the node offset is first converted to the connecting line according to the small rotation relationship; the translation or rotation of the same rigid body on both sides corresponds to... The interface coupling potential energy is:

[0046] ;

[0047] in, For the corresponding circumferential connecting lines, For arc length elements, characteristic length Before calibration, the radial width of the corresponding ring plate can be used. A fourth-order symmetric positive semi-definite matrix with constant values ​​along the circumference of the same interface is defined for both the upper and lower interfaces; stiffness is set only for the above four components, and the off-diagonal cross terms obtained from calibration are retained. When using meters and Newtons, its elemental dimensions are N / m. 2 The common rigid body motion check is used to eliminate mapping errors; the applicability of the four-component model is still checked independently as described later.

[0048] according to Arrange ten independent parameters, and then... Operating Condition 1 Each scaling component is denoted as The energy identification coefficients are as follows:

[0049] ,

[0050] ;

[0051] Each working condition is in accordance with The coefficient rows are arranged in the corresponding order, and the identification row vectors of each calibration condition are combined to form an energy identification matrix. ; composition of remaining energy at the interface For non-zero columns, normalize by column 2 norm, and calculate rank and condition number using singular value decomposition. If a zero column exists, the rank is less than ten, or the condition number exceeds the preset upper limit, add unidirectional and mixed-directional working cases, and re-filter the new working cases through S3. After achieving ten columns of full rank and a qualified condition number, apply the preset positive weights. Solution:

[0052] ;

[0053] The weights can be preset to equal positive numbers, keeping the ring plate parameters constant during identification. At least ten operating conditions are only a necessary condition for full rank; the directional combinations or circumferential distribution need to be changed to generate independent information. Repeated double-amplitude loading in the same direction is not acceptable. The generalized linear force at the interface is... Before comparing it with the finite element reaction force, it must be projected as virtual work in the same mode, and it is not directly subtracted from the nodal forces one by one.

[0054] Step S5 employs a first-order shear deformation displacement field: the tangential displacement varies linearly along the thickness, while the normal displacement remains constant along the thickness. Represents the displacement of the mid-surface in three dimensions. This indicates its two tangential components. The membrane strain is obtained by representing the corresponding change in normal direction and using the linear strain relationship between cylindrical and polar coordinates. curvature and transverse shear strain The energy of each unit is:

[0055] ,

[0056] ;

[0057] As the unit reference plane, the cylindrical shell and the annular plate are respectively taken as , The dot sign indicates the time derivative, and kinetic energy is retained. Translational-rotational inertial coupling. The A, B, and D coordinates of the cylindrical segment are derived from the in-plane stiffness along the thickness after transformation of the coordinates of each ply. , , The weighted integral is used to obtain S, which is obtained from the transverse shear stiffness and the shear correction factor used; the ring plate uses a fixed parameter S2.

[0058] Cylindrical shell Ring plate , , Inner and outer radii. Axial or radial direction is used. , and sine auxiliary term , , ; Circumferential adoption , , , The closed-period condition is satisfied. The displacements of each mid-plane and the rotation angles of the cross sections are expanded by the tensor product of the basis functions in two directions, and the coefficients are summed into q. The integral is performed by radial partitioning, and the truncation number and the order of integration are added to check convergence.

[0059] , ;

[0060] It consists of the boundary values ​​of the basis functions, local coordinate transformation, and cross-sectional rotation direction transformation. (The two...) The three element strain energies correspond to the stiffness and boundary constraint stiffness assembly, which in turn affect the variational formation of total potential energy and kinetic energy. , The lower connection frame installation boundary is subject to consistent axial foundation movement, while other constraints are set according to the actual installation. When using relative coordinate constraints for element elimination, the mass, stiffness, load, and measuring point reconstruction undergo the same transformations. Damping matrix. Pre-given, available The damping coefficients are determined by the damping ratio at two given frequencies, and the verification model maintains the same damping convention.

[0061] Step S6: Input the one-sided basic acceleration spectrum with Hz as the variable. ,structure ,make .by The relative coordinate constraint elimination matrix, The mass matrix before elimination. Let the coefficient vector of the rigid body translation along the common axis of the three elements be defined as follows: and Determine the basic influence vector and solve:

[0062]

[0063] in, , , These are the mass matrix, stiffness matrix, and damping matrix after the basic constraint treatment is completed. This refers to the fundamental motion influence vector under the same generalized coordinate system, where i is the imaginary unit. For virtual relative generalized displacement amplitude; specify measurement point The specified direction is determined by the reconstructed row vector. The virtual amplitude of its absolute acceleration, power spectral density, and root mean square are respectively:

[0064] ;

[0065] ;

[0066]

[0067] An asterisk indicates complex conjugation. The projection coefficient of the basic rigid body translation in this direction. and The input analysis frequency band used in this embodiment is the lower and upper limits of the preset target frequency band, and the spectrum and integration are uniformly measured in Hz. The axial response is taken as... Radial and circumferential responses are taken The fundamental direct term is used to recover absolute acceleration. Frequency points in the resonance neighborhood are encrypted, and spectral integral convergence is checked. With the model and connection state unchanged, the unit fundamental transfer function can be reused to calculate other stationary input spectra in the same frequency band.

[0068] After the parameters are fixed, press Figure 1 Perform independent verification. With fixed ring plate parameters and connection stiffness matrix, select boundary displacement combinations not involved in parameter determination, and verify the equivalent ring plate energy and independent frame energy, interface coupling potential energy and residual interface energy obtained in step S3, as well as the generalized boundary reaction force of the corresponding model under the same displacement mode. Within the preset target frequency band, using dynamic conditions not involved in parameter adjustment and verification frequencies, under the same excitation, constraints, damping, and measurement point directions, verify the complex transfer function of absolute acceleration to foundation acceleration of the semi-analytical model and the finite element reference model of the compartment. The normalized deviation of each verification quantity is:

[0069] , ;

[0070] in, and These are the value to be checked and the corresponding reference value. For the same dimension, a preset positive lower limit is used. Energy, generalized reaction force, and complex transfer function are expressed as absolute value, L2 norm, and complex modulus, respectively. When the deviations of all static load cases and all specified measuring points at each verification frequency point are not greater than their respective preset limits, the absolute acceleration power spectral density and root mean square acceleration within the target frequency band are output, and the target frequency band is recorded. Otherwise, a failure flag containing over-limit load cases, verification quantities, and deviation values ​​is generated. When adjusting the model based on the verification results, the adjusted model is saved as a new parameter version, and an independent verification is performed again using load cases that did not participate in this parameter adjustment.

[0071] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for analyzing the dynamic response of a composite cabin section based on a semi-analytical method, characterized in that, The composite material section includes a composite material cylindrical section, an upper docking frame, and a lower docking frame. The method, designed for small deformation linear elastic conditions where the connection interface remains unchanged, includes the following steps: S1, obtain the geometric parameters, ply parameters and material parameters of the cylindrical segment, obtain the upper and lower local connection finite element reference models that respectively include the corresponding docking frame, adjacent cylindrical segments and finite flexibility connection interface, divide the energy components of the frame, cylindrical segment and connection interface that do not overlap, and establish independent frame model and independent cylindrical segment model that are consistent with the geometric and material parameters of the corresponding parts in the local connection finite element reference model. S2, the cylindrical segment is represented as a cylindrical shell unit, and the upper and lower connecting frames are represented as parameterized equivalent ring plate units located at both ends of the cylindrical shell unit; corresponding boundary displacement modes and pairwise combination modes are applied to each independent frame model and the corresponding equivalent ring plate unit, and a boundary energy matrix containing cross terms is constructed from the elastic strain energy. Under the constraints that the total mass, center of mass position and inertia tensor error meet the preset limits and the equivalent constitutive matrix and inertia matrix are positive definite, the ring plate parameters are determined with the goal of minimizing the difference of the boundary energy matrix, and the upper and lower ring plate parameters are fixed respectively. S3: Apply multiple sets of static calibration conditions to each locally connected finite element reference model, extract the boundary displacement fields of the frame and cylindrical segment under each condition, and reapply them to the corresponding independent frame model and independent cylindrical segment model respectively; subtract the elastic strain energy of the two independent models from the total elastic strain energy of the locally connected finite element reference model to obtain the interface residual energy, and only use the condition with energy greater than the preset positive energy resolution lower limit as the candidate calibration condition; project the frame boundary displacement field of the candidate calibration condition and apply it to the equivalent ring plate element with fixed parameters, and only use the condition with projection error meeting the preset limit and the absolute value of the energy difference between the ring plate and the independent frame model not greater than the preset proportion of the interface residual energy for step S4, where the preset proportion is greater than zero and less than one; for the condition with qualified projection error but excessive energy difference, add its boundary displacement mode to the matching condition of step S2, redetermine and fix the ring plate parameters, and then execute this step again; S4. The motions of the two sides in each working condition selected in step S3 are unified to the same circumferential connecting line and the same interface coordinate system, forming a relative motion vector composed of three-way relative translation and relative rotation angle around the common circumferential tangent. The relative rotation angle is multiplied by a positive characteristic length to obtain a dimensionally unified four-dimensional vector. A fourth-order symmetric positive semi-definite connection stiffness matrix with constant values ​​along the circumferential direction of the corresponding connecting line is set for the upper and lower interfaces respectively. The coupling potential energy obtained by integrating the quadratic form formed by the four-dimensional vector and the connection stiffness matrix along the connecting line is used to match the remaining energy of the interface. Under the condition that the energy identification matrix is ​​full column rank, the connection stiffness matrix of the two interfaces is determined. The ring plate parameters remain unchanged during the identification process. S5. Based on the first-order shear deformation theory, the strain energy and kinetic energy of the cylindrical shell element and two equivalent ring plate elements are established. The finite term expansion coefficient of the displacement allowable function is used as the generalized coordinate. The strain energy, kinetic energy, interface coupling potential energy and boundary constraints are combined. The mass matrix and stiffness matrix of the compartment are formed through energy variation, and the damping matrix is ​​set. S6, input the stationary random foundation acceleration power spectral density passed from the lower docking frame along the axial direction of the compartment, construct the generalized inertial load consistent with the mass matrix and the influence vector of foundation motion, solve the frequency domain response relative to the foundation motion, reconstruct the absolute acceleration power spectral density in the specified direction at the specified measurement point, and integrate it in the preset target frequency band to obtain the root mean square of acceleration.

2. The method according to claim 1, characterized in that, The boundary displacement modes in step S2 are defined on the corresponding inner and outer boundaries with the same coordinate direction and amplitude scale, including axial, radial, and circumferential translations and rotations about the tangent of the circle. Their circumferential distribution uses constant terms, sine terms, and cosine terms. The diagonal elements of the boundary energy matrix are determined by twice the strain energy of each unit mode, and the off-diagonal elements are determined by subtracting the strain energy of each individual mode from the superimposed strain energy of the two modes. The corresponding models use consistent remaining boundary conditions and constraints to eliminate rigid body motion. Represents the nodal boundary displacement, where The dimensionless mode magnitude vector, The mode matrix includes displacement amplitude scales, and is calculated according to... reaction force at the node boundary Converted to generalized boundary reaction force .

3. The method according to claim 2, characterized in that, The parameterized equivalent ring plate uses a generalized tensile stiffness matrix A, a tension-bending coupling stiffness matrix B, a bending stiffness matrix D, a transverse shear stiffness matrix S, and a mass moment in the thickness direction. , , The description states that each parameter takes a constant value within a predetermined radial partition; Using the larger value between the Frobenius norm of the boundary energy matrix of the independent frame model and a preset positive energy scale as the normalization scale, the normalized Frobenius norm of the difference between the boundary energy matrices of the equivalent ring plate and the independent frame model is minimized; (A, B, B) T The generalized tensile-bending stiffness block matrix formed by D and the constraint S are both positive definite matrices, and the mass moment in the thickness direction satisfies: , , ;in, , , For density, The thickness coordinates are relative to the reference plane of the ring plate; the total mass, the position of the center of mass, and the inertial tensor about the same reference point and the same coordinate system respectively satisfy the preset error limits in step S2.

4. The method according to claim 3, characterized in that, In step S3, for the interface and calibration conditions Calculate the total elastic strain energy of the locally connected finite element reference model. Elastic strain energy of the independent frame model with the extracted boundary displacement field reapplied. And the elastic strain energy of the independent cylindrical segment model Determine the remaining energy at the interface using the following formula: ; in, These represent the upper and lower interfaces, respectively. The working condition is numbered; the static calibration working condition applies only boundary loads and not volume loads, the three energy terms correspond to the same material state and the same linearization state, the boundary displacement field includes the displacements on the exposed connecting boundary after separation and the local model truncation boundary; when there is an initial preload, each energy term adopts the incremental displacement quadratic energy relative to the same equilibrium state; it will satisfy... The operating conditions are retained as candidate calibration operating conditions, among which This is the lower limit of energy resolution.

5. The method according to claim 4, characterized in that, In step S3, the boundary displacement field of the candidate calibration condition is projected onto the boundary motion space of the equivalent ring plate, and the projected displacement is applied to the equivalent ring plate to calculate the elastic strain energy. The projection error is obtained by dividing the L2 norm of the displacement difference at the same boundary sampling point before and after projection by the larger value between the L2 norm of the displacement before projection and the preset positive displacement scale. The corresponding working condition is applied to step S4 only when the projection error is not greater than a preset limit and satisfies the following formula: , ; in, The preset ratio is used; for working conditions where the projection error is acceptable but the above formula does not hold, the boundary displacement mode is added to the matching working condition in step S2. After re-determining the ring plate parameters, the projection error and equivalent energy deviation of all candidate calibration working conditions are recalculated, and then the ring plate parameters are fixed; working conditions where the projection error exceeds the limit are not included in the connection stiffness identification.

6. The method according to claim 1, characterized in that, In step S4, the cylindrical shell element intersects the reference plane of each equivalent annular plate element at the corresponding circumferential connection line. The interface coordinate directions are, in order, the axial, circumferential, and radial directions of the compartment. The relative motion vector and scale transformation are as follows: , , The first three components represent the relative translations of the two sides along the aforementioned directions, while the fourth component represents the relative cross-sectional rotation around a common tangent and defined in the same positive direction. The characteristic length is given; the relative motion vector corresponding to the translation or rotation of the same rigid body on both sides is zero; the interface coupling potential energy is: ; in, For the corresponding circumferential connecting lines, Let the integral element be the arc length. The fourth-order symmetric positive semi-definite connection stiffness matrix is ​​constant along the circumferential direction of the same interface. The upper and lower interfaces... The coupler is determined separately; the coupler sets stiffness only for the four relative motion components and retains the off-diagonal cross terms determined by calibration.

7. The method according to claim 6, characterized in that, Will The four diagonal elements and six independent off-diagonal elements are arranged into a ten-dimensional parameter vector. , will the The a-th scaling relative motion component of a calibration condition is denoted as The energy identification coefficients are as follows: , ; According to The corresponding order of the parameters is used to form an energy identification matrix by assembling the identification row vectors of each calibration condition. And the remaining energy at the corresponding interface is used to form a vector. ;right The non-zero columns are normalized according to the column 2 norm, and the rank and condition number of the normalized matrix are calculated. When there is a zero column, the rank is less than ten, or the condition number exceeds the preset upper limit, static calibration conditions in one direction and mixed direction are added, and step S3 is performed to screen the new conditions. After reaching a full rank of ten columns and the condition number is not greater than the upper limit, in Under symmetric positive semi-definite constraints, minimize with preset positive weights. and The weighted squared residuals between them are determined. .

8. The method according to claim 7, characterized in that, In step S5, the mid-plane displacement and cross-sectional rotation of the three structural units are expanded using the tensor product of the basis functions in two directions in the plane. The axial direction of the cylindrical shell and the radial direction of the ring plate are given by the basis functions composed of cosine series and sine auxiliary terms. The circumferential direction is given by the constant term and the sine and cosine terms of the integer circumferential wave number to satisfy the closed period condition. Motion mapping by interface Determine the interface stiffness contribution: ; in, Let the generalized coordinate column vector be composed of the expansion coefficients of the finite terms. The interface motion mapping matrix is ​​formed by the boundary values ​​of the basis functions, local coordinate transformation, and cross-sectional rotation direction transformation; the two The strain energy corresponds to the stiffness and boundary constraint stiffness of the three structural units, and a mass matrix is ​​formed by the kinetic energy, which retains the first-order mass moment in the thickness direction. The corresponding translational and rotational inertial coupling terms.

9. The method according to claim 8, characterized in that, In step S6, the power spectral density of the single-sided foundation acceleration with frequency f as the independent variable is used. Construct virtual foundation acceleration amplitude and order Solve in the relative fundamental motion coordinate system: ; in, , , These are the mass matrix, stiffness matrix, and damping matrix after the basic constraint treatment is completed. This refers to the fundamental motion influence vector under the same generalized coordinate system, where i is the imaginary unit. For virtual relative generalized displacement amplitude; specify measurement point The specified direction is determined by the reconstructed row vector. The virtual amplitude of its absolute acceleration, power spectral density, and root mean square are respectively: , , ; in, These are the projection coefficients of the basic rigid body translation in the specified direction; the asterisk indicates the complex conjugate. and These are the lower and upper limits of the preset target frequency band, respectively.

10. The method according to claim 9, characterized in that, With fixed ring plate parameters and connection stiffness matrix, select boundary displacement combination cases not involved in parameter determination, and respectively verify the equivalent ring plate energy and independent frame energy, interface coupling potential energy and interface residual energy obtained in step S3, as well as the generalized boundary reaction force of the corresponding model under the same displacement mode; within the preset target frequency band, using dynamic cases not involved in parameter adjustment and verification frequency points, under the same excitation, constraint, damping, and measurement point direction, verify the complex transfer function of absolute acceleration to foundation acceleration of the semi-analytical model and the finite element reference model of the compartment; the normalized deviation of each verification quantity is: , ; in, and These are the value to be checked and the corresponding reference value. For the same dimension, a preset positive lower limit is used. Energy, generalized reaction force, and complex transfer function are represented by absolute value, L2 norm, and complex modulus, respectively. When the deviation of all static working conditions and all specified measuring points at each check frequency point is not greater than their respective preset limits, the absolute acceleration power spectral density and root mean square acceleration in the target frequency band are output and the target frequency band is recorded. Otherwise, a failure flag containing the over-limit working condition, check quantity, and deviation value is generated.