Design of cable-stayed footbridges under serviceability loads

Abstract

During the last decades structural engineers have proposed and developed lighter, slenderer and longer footbridges, frequently innovating from the structural and aesthetic viewpoints, as well as implementing in their design the latest technical developments in relation to material types, strengths, and light-weights. As a result, some of these footbridges have prompted serviceability responses with unexpectedly large and excessive magnitudes (such as the London Millennium Bridge in the UK or the passerelle Leopold Sedar Senghor in Paris, France), which have in turn generated a response in the structural engineering community, focusing their attention in gaining understanding about this phenomena. The extensive research that has been published during the last fifteen years has emphasised the lack of understanding of the nature and magnitude of the actions transmitted by pedestrians while walking on structures that may moderately move, such as footbridges. Hence, the inadequate performance of some footbridges is related to an unrealistic and insufficient representation of the pedestrian loading and design scenarios and to the simple application of some questionable design rules available for the design of these footbridges in service. The existing design criteria for footbridges is not underpinned by sophisticated numerical models which account for the different phenomena and issues that research in different fields has already identified. With the aim to address this deficiency, the current research work focuses on the development of a more accurate and realistic representation of the loads transmitted by pedestrians while walking, a model capable of accounting for intra- and, inter-subject variability, as well as pedestrian-structure and pedestrian-pedestrian interaction, and on its application in order to gain understanding about the structural performance under this loading. In order to include these characteristics, the model adopts a non-deterministic approach and combines results and proposals of a wide range of research fields. The research work of this thesis first investigates the performance of girder bridges and proposes a simple method that captures the response that would have been obtained with a more sophisticated model in a very accurate manner. Secondly, the sophisticated loading model is applied to a set of cable-stayed footbridges which define the structural typology by means of very comprehensive parametric studies, gaining clear understanding about the structural behaviour and performance of these bridges under pedestrian loading. This load model, its implementation in finite element models that represent the cable-stayed footbridges and the reproduction in the dynamic analyses of the nonlinear nature of the loads of each pedestrian has been performed combining Abaqus, Matlab and Fortran software packages. Based on this method, the research work generates detailed analyses of the performance of cable-stayed footbridges that represent this bridge typology to evaluate the sensitivity of that serviceability performance to multiple modifications of geometrical and structural characteristics of these bridges (involving every structural element of these footbridges). According these extensive and accurate analyses, there are several main conclusions that can be extracted (from the 35 conclusions listed in the conclusion chapter) and should be considered for the design of footbridges in general and cable-stayed footbridges in particular: 1. Exclusively load models of lateral loads that include the interaction of pedestrians with the movement of the footbridge can realistically reproduce the effects of pedestrians on structures in this lateral direction. 2. Codes and guidelines proposing a simplified evaluation method must consider that, despite the fact that pedestrians do not walk at frequencies above ∼ 2.5 Hz, the vertical loads of a pedestrian flow have important components well above this frequency. 3. The mass of the deck corresponds to the main parameter that controls the response of footbridges. 4. The damping ratio is a factor of utmost importance in the performance of footbridges and designers should seek to increase it. 5. The bearing arrangement of the footbridge has an essential role in its performance in service, in particular in lateral direction: lateral displacements and rotations of the deck at support sections must be restrained. 6. The coupling of vertical, lateral and torsional modes leads to a drastic change of the magnitude of the response. Hence it should be avoided. 7. The reduction of the dynamic deflections is not necessarily related to a reduction of the accelerations, therefore the assessment of the serviceability performance of footbridges through dynamic deflections is not reliable. 8. The dynamic events must be used to evaluate the response of different structural elements of the footbridges (deck, pylons, etc.) as a static analysis with variable loads of 5kN/m 2 does not always describe the largest stresses at these elements. This point is relevant for safety, as it involves ULS verifications.