Short Communication
Efficient methods for determining modal parameters of dynamic structures with large modifications

https://doi.org/10.1016/j.jsv.2006.06.002Get rights and content

Abstract

The paper presents efficient methods for determining the modified modal parameters (natural frequencies and mode shapes) in a structural dynamic modification analysis when structural modifications are relatively large. Based on the developed theory that can provide an exact relationship between the modifications of structural parameters (stiffness and mass) and the associated modal parameters, an efficient iterative computational procedure is proposed for determining the modified eigenvalues and the corresponding eigenvectors for complex structural systems. A high order approximation approach is further presented from the exact relationship and compared with the existing first-order approximation approach and the proposed iterative procedure. From the results for the given numerical example, it is shown that even in the cases with a large modification of structural parameters the proposed iterative procedure can provide exact predictions of the modified modal parameters after only a few iterations, and the high order approximation approach can give excellent estimates. The computation of the modified modal parameters does not require the knowledge of the original or modified structural parameters, and only a limited knowledge of the original modal data may be sufficient in a dynamic reanalysis for complex structures.

Introduction

Structural dynamic modification (SDM) techniques are very useful for rapidly analysing and applying the effect of structural changes on the dynamic response of structures. That is, given changes in structural parameters, such as stiffness or mass, SDM techniques can efficiently determine the corresponding changes in modal parameters, such as natural frequencies and mode shapes, without solving the generalised eigenvalue problem for the modified dynamic system. Therefore, SDM techniques, which often incorporate finite-element analysis techniques, can greatly reduce the computational effort and increase the efficiency of reanalysis during a structural optimisation process, especially in the cases where complex mechanical and structural systems are considered.

There are various SDM techniques available for the dynamic reanalyses of a structural system, such as sensitivity analyses based on eigenvalue and eigenvector derivatives [1], [2] and Rayleigh quotient iteration [3]. However, the efficiency of sensitivity analysis methods is limited because these methods are complicated and may be only suitable for small modifications of structural parameters [4]. It may be difficult to determine some high-frequency modes using Rayleigh quotient iteration because the predictions of those modes exceeded the limit bound of the Rayleigh quotient [2].

It should be pointed out that the first-order sensitivity analysis techniques and the truncated Taylor's expansion approximation approaches often used for estimating the modified modal parameters may perform properly when the changes in structural parameters from the initial model to the modified model are small. However, for the cases with relatively large modifications of structural parameters, the first order or the truncated Taylor's expansion approximations may be inaccurate. To avoid these shortcomings, a perturbation theory was developed [5], which can provide an exact relationship between the modifications of structural parameters and the associated modal parameters and can be applied to model updating and inverse structural damage identification [6], [7]. Here, based on the developed perturbation theory, an efficient iterative computational procedure is proposed to provide exact predictions of the natural frequencies and the corresponding mode shapes for the modified dynamic system modelled with a large number of degrees of freedom (DOF). A high order approximation approach is also presented without iterative procedures required, which can give excellent estimates of the modified modal parameters. The results of a numerical example show that the iterative procedure presented here converges quickly for evaluating the modified modal parameters and finally gives the exact solution even when structural modifications are large. The proposed computational techniques successfully avoid adopting Taylor series expansion procedure and then the derivatives of modal parameters are not needed. Only limited information on the analytical or experimental modal data of the original structure is required in calculations. Therefore, the proposed methods are well suited for complex structures with a large number of DOFs, especially for the cases where the knowledge of the modified structural parameters is not available and the modified modal data cannot be obtained by solving the generalised eigenvalue problem.

Section snippets

Basic equations for SDM analyses

Assume that modal parameters λi and φi are the ith eigenvalue and the corresponding mass normalised eigenvector of the original dynamic system with structural parameters of the global stiffness matrix K and the global mass matrix M, where i ranges from 1 to N and N represents the total number of DOFs for the system. Then, the ith eigenvalue λi* and the corresponding eigenvector φi* for the modified system can be given by λi*=λi+Δλi,φi*=φi+Δφi,where Δλi and Δφi represent the modifications of ith

Numerical example

A rectangular thin plate model shown in Fig. 1 is utilised to demonstrate the effectiveness of the proposed techniques for calculating the modal data for the modified dynamic system by introducing various levels of structural modifications and comparing different computational approaches. The rectangular aluminium plate is 1000 mm long, 600 mm wide and 10 mm thick, with material properties of Young's modulus E=6.89×1010 N/m2, Poisson's ratio υ=0.30 and density ρ=2796 kg/m3. The thin plate is

Conclusions

An improved iterative procedure is proposed for efficiently determining the eigenvalues and the corresponding eigenvectors for a dynamic system with large modifications of structural parameters and a large number of DOFs present. A high order approximation approach is also presented without iterative procedures involved. From the results for the given thin plate bending model problem, it has been shown that the convergence of the proposed iterative procedure can be achieved rapidly, leading to

References (8)

  • W.M. To et al.

    Structural modification analysis using Rayleigh quotient iteration

    International Journal of Mechanical Sciences

    (1990)
  • R.L. Fox et al.

    Rates of changes of eigenvalues and eigenvectors

    AIAA Journal

    (1968)
  • A.D. Belegundu et al.

    Iterative methods for design sensitivity analysis

    AIAA Journal

    (1988)
  • J. Wang, W. Heylen, P. Sas, Accuracy of structural modification techniques, Proceedings of the Fifth International...
There are more references available in the full text version of this article.

Cited by (29)

  • A system response prediction approach based on global transmissibilities and its relation with transfer path analysis methods

    2017, Applied Acoustics
    Citation Excerpt :

    Hence, the aim of the paper is to derive the mathematical relationships between the proposed method and the above TPA methods and investigate their potential application in improving TPA methods. For the sake of completeness, it must be mentioned that there are some other approaches already reported in literature to address structural modification issues, such as structural modification or reanalysis methods [17–21]. In essence, the proposed method is equivalent to the one in Ref. [17] since the global transmissibility can be expressed in terms of FRFs.

  • Response prediction for modified mechanical systems based on in-situ frequency response functions: Theoretical and numerical studies

    2017, Journal of Sound and Vibration
    Citation Excerpt :

    This paper focuses on the structural modification stage of the above analysis procedure. So far, there are some approaches already reported in literature to address structural modification issues, such as structural modification or reanalysis methods [8–11]. Most research on these methods has been focused on obtaining dynamic characteristics of the modified system based on the determination of the behavior of the original structure and a delta dynamic stiffness matrix [12].

  • Improved inverse filter for the correction of distorted measured cutting forces

    2017, International Journal of Mechanical Sciences
    Citation Excerpt :

    It should be mentioned that obtaining the transfer function of the cutting system is one premise of this study. For rigid cutting system, the required transfer function can be measured before actual cutting starts; while for flexible system, whose modal parameters may dynamically change as cutting continues, the in-process transfer function should be modified by structural dynamic modification scheme [20,21] based on the static signal of modal parameters measured before cutting. In this study, three milling tests with the same depth of cuts and feed but different spindle speeds (see Table 1) are carried out in Hardinge GX710plus CNC machining center.

  • An exact block-based reanalysis method for local modifications

    2015, Computers and Structures
    Citation Excerpt :

    It has been proved to be feasible for static reanalysis, structural optimization, eigenvalue-problem, nonlinear analysis, dynamic reanalysis and sensitively analysis [24–28]. For large changes, in recent years, Chen presented an efficient method for determining the modified modal parameters in structural dynamic modifications [29]. Yang et al. suggested an adaptive static reanalysis method based on the Neumann series and epsilon algorithm when structural modifications are relatively large [30].

View all citing articles on Scopus
View full text