Model error correction from truncated modal flexibility sensitivity and generic parameters. II: experimental verification

Abstract

This paper reports on the experimental verification of the truncated modal flexibility sensitivity based model updating method proposed in the companion paper (Wu and Law, Mechanical System and Signal Processing, doi: 10.1016/S0888-3270 (03) 00094-3) with real measurements on a three-dimensional cantilever frame structure in the laboratory. Two schemes of model error correction are presented to improve the initial inaccurate finite-element model of the structure. The first procedure updates the initial model in two stages. The systematic model errors are firstly updated using macrogeneric parameters. The local errors are then improved using elemental eigenparameters. The second procedure improves both the systematic and local errors without a measured baseline reference simultaneously. The first procedure can differentiate the types of errors in the structure while the latter procedure removes the usual requirement in most existing model based updating methods that an accurate finite -element model should be used as reference.