Should Torroja’s prestressed concrete Alloz aqueduct be thought of as a beam or a shell?

Abstract

This paper examines the structural action of Eduardo Torroja’s Alloz aqueduct, completed in 1939, to see whether we should think of it as acting as a beam or a shell. This is of interest regarding the Alloz aqueduct itself, but also in the design of similar structures in the future, where we must have a simple conceptual understanding of how we want it to work.We apply two alternative approaches available at that time, before computers. Firstly, the membrane theory of shells, effectively assuming the aqueduct walls are infinitely flexible in bending, and secondly, the Euler–Bernoulli ‘plane sections remain plane’ elementary beam theory. We also review Torroja’s calculations which were based on an elaboration of the Euler–Bernoulli beam theory know as the Griffith–Taylor theory for the bending of cantilevers, although we are uncertain as to why he decided to use the Griffith–Taylor theory for a thin walled structure.Both the membrane shell and Euler–Bernoulli beam theory require a prestress to be applied along the longitudinal edges of the channel. However, the level of prestress in the Alloz aqueduct is consistent with the beam theory, which seams the most appropriate approach.Whether or not a structure of this type acts as a shell depends upon the thickness of the wall. The thinner the wall, the more it act as a shell. The wall thickness of the Alloz aqueduct is sufficient for it to act mainly as a beam.