Testing and Fretting Fatigue Analysis of Bridge Stay Cable Wires at Saddle Supports

High-strength steel cables are one of the principal components of cable-stayed bridges. In these structures, the cables transfer the gravitational loads at the deck to the bridge tower structures or “pylons”. Reliable anchoring of the cables is a primary design consideration for cable-stayed bridges. Traditionally, the cables were directly anchored to the bridge deck and the pylons. A newer approach, employing so-called saddle systems, has become more popular in recent decades. With this approach, the cables are anchored to the bridge deck on one side of the pylon, go over a radial surface at the pylon, and are finally anchored to the bridge deck on the other side of the pylon. Material and anchoring costs of saddle systems are lower than the traditional approach, as smaller pylons are required for saddle systems and the cables do not require anchoring at the pylons. A primary design consideration of saddle systems is fretting fatigue failure of the cables at the saddle supports. Despite this fact, very limited previous research can be found in the literature on this topic. Also, the existing standards for saddle systems are rather simplistic. These standards require large-scale fatigue tests to evaluate the saddle systems and do not offer a calculation-based design procedure. With this in mind, the main objectives of the current thesis are: to undertake initial efforts to develop a calculation-based framework to evaluate the fretting fatigue behaviour of cables at saddle supports, to explore a possible framework for probabilistic analysis of this problem, and to design a more economical small-scale fretting fatigue test setup and use it to evaluate the fretting fatigue behaviour of typical bridge cable wires. Several parameters affect the fretting fatigue behaviour of cables (e.g., the relative displacement between the cable and saddle, and the contact force between the cable and saddle). In this thesis, closed-form equations for evaluating these critical parameters are first discussed. Then, an FE model is developed to evaluate the accuracy of these equations. The developed FE model is then used to evaluate the effect of wear on these critical parameters. Overall, the results of the FE model are shown to be close to the results obtained by calculation. However, a higher difference is seen between the results at the points where the cable first meets the saddle. Following the determination of the critical parameters, a multiaxial stress approach based on the Smith-Watson-Topper (SWT) parameter is used to evaluate the fretting fatigue life of the cable wires. A set of large-scale tests previously performed at TU Berlin is used as an example. The predictions based on the SWT parameter are shown to be in good agreement (i.e., fatigue lives and overall trends are estimated with reasonable accuracy) with the tests performed at TU Berlin. In order to extend this approach to a probabilistic framework, several practical approaches aimed at limiting the need to perform time-consuming FE analyses are then explored. These approaches include the use of Monte Carlo simulation (MCS) with fretting maps or employing the multiplicative dimensional reduction method (M-DRM). The results of these approaches are then compared, and the challenges and benefits of each approach are presented. The results obtained using both methods are reasonably close to each other. Finally, an analysis is performed to evaluate the sensitivity of the prediction results to the main model parameters. It is shown that the uncertainties in the contact force and fatigue strength coefficients have the highest sensitivity factors. Following the completion of these analytical studies, a small-scale fretting fatigue test setup was designed to evaluate the fretting fatigue behaviour of bridge stay cable wires. Two different bridge cable types, namely: galvanized and bare, were used for these tests. In these tests, the critical parameters affecting fretting fatigue life were varied. It was found that the bare wires have a better fretting fatigue performance in comparison with the galvanized wires. Following the completion of the fretting fatigue tests, plain fatigue tests were performed to evaluate the fatigue performance of the wire material. After the experimental work, a microstructure analysis was performed to evaluate the microhardness of the wires and observe defects at the surface and core of the wires using SEM photography. Irregular microstructures were found at the surface of the galvanized wire. However, the bare wire had a uniform microstructure at the surface. Following the experimental work, the SWT parameter-based approaches were applied to the tests performed at the University of Waterloo. However, these approaches have limitations in that they do not account for wire defects and their influence on the fatigue life predictions. Given the presence of significant defects in the wires, a linear elastic fracture mechanic (LEFM) approach is lastly employed to study possible effects of these defects on the fretting fatigue life of the wires. The LEFM results are shown to be in good agreement with the test results