Model Update Using Modal Contribution to Static Flexibility Error

A new technique for the parametric correction of full-order analytical stiffness matrices from reduced-order modal measurements is presented. The proposed algorithm corrects model parameters by minimizing a matrix residual formed at measurement degrees of freedom only. The residual is the difference between the measured modal contribution to the static flexibility matrix and the corresponding analytical modal contribution to the static flexibility matrix. Analytical expressions are developed for the flexibility matrix error residual gradient in terms of modal sensitivities found via Nelson's method. By utilizing a flexibility-based matrix error residual, the effect of poorly modeled inertia properties on stiffness parameter estimates is greatly reduced. Unlike technlques that utilize only static data, this technique is suitable for use on unconstrained structures. In addition, the method avoids the problems of mode selection, determining modal correspondence, and eigenvector expansion or model reduction. Numerical simulation results are presented for Kabe's model as well as a finite element model of a welded frame structure.

[1]  Scott Cogan,et al.  Updating finite element models using static deformations , 1992 .

[2]  W. C. Mills-Curran,et al.  Comment on "Eigenvector Derivatives with Repeated Eigenvalues" , 1990 .

[3]  Lee D. Peterson,et al.  Method for Structural Model Update Using Dynamically Measured Static Flexibility Matrices , 1997 .

[4]  K. Park,et al.  Second-order structural identification procedure via state-space-based system identification , 1994 .

[5]  Lee D. Peterson,et al.  Method for determining minimum-order mass and stiffness matrices from modal test data , 1995 .

[6]  A. Kabe.  Stiffness matrix adjustment using mode data , 1985 .

[7]  L. Peterson,et al.  Measurement of static flexibility matrices for experiments with incomplete reciprocity , 1995 .

[8]  Richard B. Nelson,et al.  Simplified calculation of eigenvector derivatives , 1976 .

[9]  John E. Mottershead,et al.  Finite Element Model Updating in Structural Dynamics , 1995 .

[10]  R. Lane Dailey,et al.  Eigenvector derivatives with repeated eigenvalues , 1989 .

[11]  Lee D. Peterson.  Efficient computation of the eigensystem realization algorithm , 1995 .

[12]  K. F. Alvin,et al.  TIME AND FREQUENCY DOMAIN PROCEDURE FOR IDENTIFICATION OF STRUCTURAL DYNAMIC MODELS , 1997 .

[13]  L. Peterson,et al.  Comparing Model Update Error Residuals and Effects on Model Predictive Accuracy , 1998 .

[14]  F. J. Soeiro,et al.  Structural damage detection based on static and modal analysis , 1989 .

[15]  Michael I. Friswell,et al.  Direct Methods using Modal Data , 1995 .

[16]  Masoud Sanayei,et al.  Damage assessment of structures using static test data , 1991 .

[17]  I. U. Ojalvo,et al.  Efficient Computation of Mode-Shape Derivatives for Large Dynamic Systems , 1986 .

[18]  W. C. Mills-Curran,et al.  CALCULATION OF EIGENVECTOR DERIVATIVES FOR STRUCTURES WITH REPEATED EIGENVALUES , 1988 .

[19]  L. Peterson,et al.  Estimation of reciprocal residual flexibility from experimental modal data , 1996 .