Consequences of fracture mechanics for size effect, crack spacing and crack width in concrete pavements

The ASTM standard C 78 for determining the modulus of rupture f_r (or flexural strength) provides a constant value of a measure of tensile strength which is currently used for designing pavements of all thicknesses, as well as concrete structural members, of all sizes. However, accumulated test data on f_r, as well as analytical studies and numerical simulations, clearly indicate that f_r is not constant but significantly decreases as the structure size increases. This evidence of size effect has so far been ignored. The consequence for thick pavements is increased cracking and thus poorer durability and serviceability, and for large structures even a significant increase of failure probability. The size dependence determined from test data follows a simple equation, which represents a gradual transition between a deterministic (energetic) size effect caused by stress redistribution at small sizes, and the power-law size effect of the classical Weibull statistical theory of strength, which is approached for large sizes. The transition is explained and modeled by the non-local Weibull theory. In this paper, modifications to the current standards that incorporate a simple procedure to obtain the size dependence f_r are proposed. Furthermore, the cohesive fracture mechanics, which is a more fundamental way to predict the effect of pavement thickness, resting on a given subsoil, is described. The energy release calculated according to cohesive fracture mechanics is shown to give the initial crack spacing, which may later increase due to instability of interacting parallel cracks. This is and important design consideration, since crack spacing controls the crack opening, and thus the ingress of water and corrosive agents, as well as transmission of shear stresses. Discussion and examples are limited to the formation of secondary cracks in between stress-relieving notches. This summary paper presents a general apercu of the problem, explains the fracture mechanics approach, and concludes by comparisons with test data.