Experimental behaviour of concrete box culverts — comparison with current codes of practice

Abstract

It is now accepted that current expeditious models for determining earth pressures on flexible underground structures under compacted layers do not include several technical nuances of the soil–structure interaction. Thus, these models are not capable of delivering an optimized design. The present paper compares the results from the well-known American Association of State Highway and Transportation Officials (AASHTO) model with two different numerical models — a user-friendly elastic model and a more robust finite element model — and with results retrieved from a full-scale monitoring of a concrete box culvert, 5.5 m high and 3.77 m width, over which a 15 m high embankment was built. This structure was instrumented selectively, over a period of almost 1 year, during which several parameters were recorded, including earth pressures and structural deformation. Results have shown that the two most significant drawbacks associated with the use of the AASHTO model are the inadequate evaluation of vertical pressure on the top slab and the coefficient of earth pressure, which results in a significant overestimation of the lateral pressures and, consequently, in an overall inefficient design of the structure.