Dynamic Parameter Characterization for Railway Bridges Using System Identification

Abstract

The problem of structural health monitoring of railway bridges is an ongoing research area, and numerous researchers have worked in this domain. The fundamental problem of condition assessment can be addressed by identifying the dynamic parameters like fundamental frequencies and mode shapes of tested structures and validating numerical models using these identified parameters. The entire problem can hence be broken down into two major areas, one of identifying the dynamic parameters in a robust manner from site data and the other of creating updated numerical models which exhibit convergence with these. In the present paper, the first of these areas is addressed. Various system identification techniques are presented along with their results to illustrate the robustness of the various techniques. A novel method is also developed to identify the mode shapes using sparse sensor applications. In this paper the model updation techniques developed elsewhere by the authors is used in conjunction with system identification to as a synthesized approach to identification of mode shapes. All the techniques are illustrated by application on an existing in-service railway bridge to verify these techniques for real structures. In this paper, the concept of global and local modes is introduced where the structural modes involving the movement of the entire structure are termed as global modes, whereas the modes involving only certain elements are termed as local modes. It is shown that a system identification technique is possible which identifies and differentiates these.