Year

2003

Degree Name

Doctor of Philosophy (PhD)

Department

Faculty of Engineering

Abstract

Ageing infrastructure is a widespread problem, with potentially catastrophic consequences. Reliable structural integrity and remaining life assessment are essential for the resolution of the problem. Ageing railway bridges is a particularly difficult subset of the ageing infrastructure problem. There is a complex array of issues that face integrity and remaining life assessment of railway bridges. The structural conditions of railway bridges may change from site to site, even where bridges are supposedly of the same design. Structural conditions may also change over the life of any particular bridge. Dynamic interaction, between railway vehicles and railway bridge structures, has a dramatic effect on structural response and therefore remaining life. This dynamic interaction often requires complex modelling techniques in integrity assessment. For assessment of railway bridges, different assumptions and different methods of assessment are often used, making repeatable and verifiable assessment difficult. In this research, in order to work toward addressing the problem of ageing steel railway bridges, a clear, repeatable methodology for high-level structural assessment has been formulated. The method integrates state-of-the-art modelling, testing and fatigue code tools, and uses dynamic digital data for testing, modelling and assessment. The method is demonstrated by assessment of the steel girder approach spans of the Mullet Creek Railway Bridge, Dapto, NSW, Australia. The method developed in this research begins with a finite element sensitivity model, that is a finite element model that permits variations in joint fixity and support conditions. This model is tuned and validated, using transient digital data from the dynamic field-test response of a slow moving vehicle of known load, through a dynamic analytical model intermediate step. The dynamic field-test response of normal traffic is then recorded and axle loads are identified from the digitised measured response, using the dynamic analytical model as a transfer function. Identified loads are applied to the tuned finite element sensitivity model and the dynamic stress history generated for virtually any bridge component. Dynamic stress histories are then entered directly into a software system, which estimates remaining life via several international fatigue codes. After verification of the finite element sensitivity model against dynamic field-test results, loading and structural conditions may be adjusted and the integrity and remaining life of the structure evaluated for virtually any combination of structural and loading condition. In the demonstration of the method, three components of the Mullet Creek Railway Bridge approach spans have been chosen for structural integrity and remaining life assessment. Structural conditions and applied loading conditions have been altered and the impact of the changes investigated. From the structural integrity and remaining life assessment, two locations have been identified as fatigue critical and recommendations have been made for structural changes and ongoing inspection. The most significant contribution of this research is expected to be the complete methodology, its clarity and repeatability, its integration throughout and the way in which it deals with the difficult problem of true dynamic response. Other contributions have been made within individual steps of the methodology, where attempts have been made to extend current research. These contributions include: the development of a technique for modelling the dynamic response of structures in finite element software; a dynamic analytical model for beams with rotationally stiffened supports subjected to moving distributed loads; extension of load identification theory to distributed loads on non-simply supported beams; and a comparative study of several key international fatigue codes.

Share

COinS
 

Unless otherwise indicated, the views expressed in this thesis are those of the author and do not necessarily represent the views of the University of Wollongong.