Structural Damage Detection Using Frequency Response Function

Abstract

In this work, algorithmic approaches to enhance system health monitoring capability are developed in which direct frequency response function data is utilized. The use of direct frequency response functions, as opposed to measured modal parameters, are shown to be one method to address a part of the incomplete measurement problem common to model correlation and system health monitoring; namely the mismatch in the number of measured vibration modes in the measured frequency band in comparison to the number of modes included in the analytical finiteelement model. In addition, there are several advantages to using direct frequency response function measurements, including: (i) reduction in testingtime, inthattypicallydatais requiredat onlyafewnumber of frequency values and (ii) reduction in analysis effort, in thattheskilliariof modal parametercurvefining iseliminated. Of course, it is well recognized that the advantage of reducing measurement noise via modal curve fitting must be addressed by choosing “regions” of the frequency response function which are not adversely corrupted by noise or using other numerical techniques to reduce the effects of noise.