Design of continuous prestressed concrete bridges poses considerable difficulties to engineers because of the large number of parameters involved and their complex interaction. The vast majority of literature on prestressed concrete covers analysis techniques in detail, but lacks design knowledge. The growth of high speed computers has revolutionised the analysis techniques, but has had little impact on design - except to speed up the task of repeated-analysis. Design is often perceived as an iterative process of generation, evaluation and modification of trial solutions. In order to minimise the complexities, it is imperative to develop rational design guidelines.
The main objective of the work described in this dissertation has been to develop good tools for design of prestressed concrete bridges. Design principles have been formulated by a careful study of the design sequence and the associated intricacies introduced by segmental construction. Of the numerous constraints that need to be fulfilled, at each transfer stage and the final service stage, the governing design criteria are first identified, and are incorporated from the start of the design process. This helps to simplify the design task by reducing the number of constraints that need to be tackled. A logical design approach is then formulated, and the design is carried out sequentially in modules. First, suitable cross-section dimensions are selected that ensure the existence of a feasible Magnel diagram. Second, the prestress force variation is determined, such that a valid cable profile can be found. Third, the cable layout, that satisfies the governing stress conditions is found. Finally, the prestress distributions for the intermediate transfer stages are determined, starting from the final transfer stage. The principles developed are applied to both prismatic and non-prismatic bridge beams. The design approach prescribed here ensures that the design evolves in a logical sequence.
The design principles are incorporated in an expert system, BRIDEX. It is argued that expert systems for design should be based on fundamental principles, wherever possible, rather than relying on rules of thumb. BRIDEX is so written that the computer does all the numerical processing, but all major decisions are taken by the designer, guided by the values suggested by the expert system.
[Cambridge University | CUED | Structures Group | Geotechnical Group]
This page is maintained by email@example.com (last update 10 November 2003)