Comparison between Measured Slab Profiles of Curled Pavements and Profiles Generated Using the Finite Element Method

Abstract

The advancement in the processing speed of personal computers, as well as the more user-friendly graphical interfaces starting to be employed in pavement finite element analysis programs, has resulted in a more frequent use of the finite element method (FEM) for pavements analysis. This increase in reliance on the FEM has amplified the need to ensure these finite element models accurately predict the response of the slab to applied and environmental loads. This paper will focus on the ability of currently available finite element programs to accurately predict the response of rigid pavements to temperature gradients. The study entailed collecting surface profile measurements for a range of pavement designs under different temperature conditions. Each pavement design was then modeled using finite element analysis. Temperature gradients representing the conditions under which the slab profiles were measured were applied to the finite element models. Comparisons were then made between the measured and calculated slab shape. A comparison of the length of the unsupported portion of the slab determined using surface profile measurements and the FEM was performed. The FEM under-estimated the portion of the slab fully-supported only 12 percent of the time. The ability of the FEM to accurately estimate curvature appears to be a function of the pavement design and the magnitude of the gradient. Curvature was accurately determined using the FEM only 14 percent of the time with curvature being under-estimated 70 percent of the time. The FEM could not accurately estimate the larger curvatures. The curvatures calculated using slab profiles generated using the FEM tend to overestimate the curvatures calculated using measured profiles more frequently for thinner slabs and granular bases. The FEM predicted higher curvatures than were measured more frequently for unrestrained slabs than restrained.