Detecting damage in structures using Wavelet Analysis Salgado, R. and Cruz, P.J.S. Civil Engineering Department, University of Minho, Guimarães, Portugal. ABSTRACT Wavelet Analysis for detecting damage in structures is evaluated in the present paper. The methodology for detecting damage consists in obtaining the acceleration response at different locations along the beam before and after damage. Then, mode shapes of the structure are obtained and Wavelet Analysis is applied to detect changes in mode shapes. Dynamic simulations of Euler Bernoulli cracked beams were also carried out. These simulations tried to take into account the conditions that data would be obtained in real structures. Effects of severity of damage are investigated. Results show excellent detection of damage in ideal situations; however, this method is too sensitive to inaccuracies in experimental data. INTRODUCTION Everyday, civil engineering structures are closer to their service life and some structural damage can be expected. For this reason, there is a growing interest for developing techniques for detection of incipient damage. Health Monitoring System (HMS) has emerged as the group of strategies to monitor, detect, assess and make subsequent maintenance decisions. The main idea of HMS consists in monitoring the dynamic response of the structure and to process this information using some analytical methods for detecting damage. Many methods for detecting damage using vibration response have been proposed. Those methods compare dynamic characteristics of the structures such as natural frequencies, mode shapes and damping ratios. A review of those methods was carried out by Sohn et al [1]. _____________ Rolando Salgado, Ph. D. Student, University of Minho, Azurém 4800-058, Guimarães, Pt. Paulo Cruz, Associate Professor, University of Minho, Azurém 4800-058, Guimarães, Pt. E-mail: rolando@civil.uminho.pt