Bridge Condition Evaluation by In Situ Dynamic Data

The performance of vibration-based damage identification techniques for current condition evaluation of existing bridges is studied through their application to dynamic response of an existing concrete bridge. First, in situ dynamic tests were performed and modal parameters (mode shapes and frequencies) were identified. As previous sets of in situ dynamic parameters were not available, the reference dynamic parameters were obtained from a finite element model of the bridge. Comparing modal parameters of the finite element bridge model and the in situ experimental results through damage identification techniques, the bridge current condition was examined. That evaluation results showed that techniques like strain energy and flexibility curvature present promising results for current condition evaluation of existing bridges. INTRODUCTION Civil infrastructures begin to deteriorate as soon as they are in service. Maintaining safe and reliable civil infra-structure for daily use is a topic that has received considerable attention in literature in recent years. Most of the current damage detection methods are either visual or localized experimental methods such as acoustic, ultrasonic, magnetic field and radiographic methods. For now, no current experimental method is general enough to be applicable to all the different portions of a structure as they have their own applicability limitations. Moreover, they often require that the location of the damage be known a priori and that the portion of the structure being inspected to be readily accessible. Subjected to such limitations, these methods generally detect damage on or near the surface of the structure. The need for more global damage detection methods that overcome the above limitations and which can be applied to complex structures has led to the development of methods that examine variations in the vibration characteristics of the structure. Over the past 30 years, detecting damage in a structure from modifications in dynamic parameters has received considerable attention. The basic premise of global damage detection is that modal parameters, notably natural frequencies and mode shapes are the functions of the physical properties of the structure (mass, damping and stiffness). Therefore, variations in physical properties (stiffness or flexibility) will cause modifications in modal properties. In damage detection, the modal properties most commonly used are eigenfrequencies and mode shapes. Early attempts to use frequency shifts to detect and localize damage include those by Vandiver 1975 [1], 1977 [2] and Adams et al. 1978 [3]. At about the same time, researchers started using mode shape changes for damage detection purposes (West 1984 [4], Yuen 1985 [5]). Since then, many techniques have been *Address correspondence to this author at the Research Center on Concrete Infrastructures (CRIB), Department of Civil Engineering, Université Laval, Québec (Québec), G1V 0A6, Canada; E-mail: [email protected] developed using frequency and mode shape variations to locate and quantify damage. These techniques have been well documented in the extensive literature reviews, published by Siddique et al. 2007 [6], Zhou et al. 2007 [7], Doebling et al. 1996 [8] and Salawu 1997 [9]. A major limitation of such damage detection techniques is that, frequency shift and mode shape variation are generally not very sensitive for local and moderate level of damages. The somewhat low sensitivity of these parameters requires either very precise measurement or large levels of damage. Tests conducted on the I-40 Bridge (Farrar et al. 1994 [10]) have also demonstrated this point. Pandey et al. 1991 [11] proposed using curvature mode shapes as a means of locating structural damage. In their research, curvature mode shape was shown to locate damages adequately, in cases where traditional damage localization techniques, such as the Coordinate Modal Assurance Criterion COMAC (West 1984), had failed. Although vibration-based damage identification techniques have been used successfully in a number of cases (Doebling and Farrar 1998 [12]), there are still some limitations associated with their application for current condition evaluation of existing bridges. In fact, in North America it is common knowledge that a significant number of bridges is old and requires repairing or strengthening. In most cases, the in situ dynamic properties of these bridges can be registered, but seldom is a previous set of records available for comparison purposes (reference state). Furthermore, the original (undamaged) state of these existing bridges is even more rarely available since it is not common practice to assess dynamic properties of new structures. At first, this article presents a quick review of existing vibration-based damage identification techniques. In order to evaluate the efficiency of these techniques for bridge condition evaluation, a numerical model of an existing concrete bridge is presented and some damage scenarios are introduced into it. Following the discussion regarding the performance of these techniques, the most efficient ones will be Bridge Condition Evaluation by In Situ Dynamic Data The Open Acoustics Journal, 2008, Volume 1 55 applied to the in situ dynamic data of an existing bridge for current condition evaluation. VIBRATION-BASED DAMAGE IDENTIFICATION TECHNIQUES: A REVIEW In this section, some current vibration-based damage identification techniques will be reviewed. In practice, due to the difficulty involved in exciting the higher modal frequencies (need of high quantity of energy) of a structure, the techniques selected herein require only a small number of mode shapes and/or frequencies. Mode Shape Curvature In formulating the eigenvalue problem, Pandey et al. 1991, assumed that structural damage affects only the stiffness matrix and not the mass matrix. For the undamaged condition the eigenvalue problem is expressed as: K i M ( ) i { } = 0 { } (1) where K [ ] and M [ ] are respectively the stiffness and mass matrices, i and i { } are the i th eigenvalue and the i eigenvector. Similarly, the eigenvalue problem for the damaged condition is: