AbstractThe development of steel reinforced concrete in the early 20,h century led to a decline in masonry arch bridge building. Traditional masonry arch construction is time consuming and involves considerable formwork usually in the form of timber centring, as well as traditional skills. It is also limited by the spans at which the design is efficient. As a result, reinforced concrete arches and slab and beam bridges became common. However, durability issues associated with corrosion of steel have become a major and costly problem and design can now favour steel-free structures that are either reinforced with FRP or unreinforced. Experience has shown that unreinforced compression structures, such as arches are durable structures that have stood the test of time. The FlexiArch bridge system is a sustainable alternative for short span crossings and contains no steel reinforcement or mortar joints.
This thesis presents detailed research into the behaviour and analysis of straight span FlexiArch systems through laboratory models and non-linear finite element analysis. Eight third-scale arches were constructed and tested. Five third-scale arches were based on a 5 m x 2 m (span x rise) full-scale bridge and three third-scale arches on a 10 m x 2 m geometry. The arches were monitored through lifting, loading of the arch ring only, backfilling, and under full test loading. The variables in the eight arches were the backfill type, arch ring thickness, solid and hollowcore voussoir design, and span to rise ratio. The experimental investigation demonstrated that higher peak loads were achieved in the arches of higher arch ring thickness; the rings with solid voussoirs sustained higher loads than rings with hollowcore voussoirs, and doubling the arch span (shallower profile) typically halved the peak load. The importance of accurate material properties for use in numerical prediction was highlighted in the literature and control tests were carried out to determine the material properties.
The non-linear finite element analysis investigated three areas. Firstly, models were constructed to replicate those tested in the laboratory and to further compare key variables. Secondly, a detailed parametric study in material properties demonstrated that many of the parameters had a linear relationship with the arch predicted peak load. Finally, the NLFEA model was validated through modelling four arch bridges presented in the literature and predicted accurate behaviour when compared to tests.
|Date of Award
|Desmond Robinson (Supervisor) & Su Taylor (Supervisor)