Minimum condition number by orthogonal projection row selection of artificial boundary conditions for finite element model update and damage detection

Abstract

The method of artificial boundary conditions provides a way of significantly expanding the amount of useable data for performing model update or damage detection. It accomplishes this by artificially applying boundary conditions directly to the measured frequency response functions, which can then be curve fit to obtain the modal parameters. A critical task in the application of this method is the selection of modes from a large available set of candidate modes generated by the application of various artificial boundary conditions. An orthogonal projection procedure is used to select modes from which to build a composite sensitivity matrix of minimum condition number for update and damage detection. Both simulation and experiment are used to investigate this method. The experimental study makes use of two beams, identical except for damage installed in one beam. Issues regarding the number and type of artificial boundary conditions, the synthesis of boundary conditions, and test frequency resolution are studied. Both the simulation and the experiment reveal that model updating and damage detection can be accomplished with high accuracy.