A number of uncertainties and random factors play a role in the deterioration process of pavements under the effect of traffic. A probabilistic approach can make allowance for the stochastic nature ofthis process and provide a better rationale for the design methods.An extension in this direction of the empirical design method derived from the aasho road tests for flexible pavements with unbounded granular layers is presented. The probabilistic development addressestwo aspects of the problem. First, the reliability assessment of a pavement section is obtained by considering the shook and finn design equation as a function of two random variables--the expected number of traffic loads and the california bearing ratio (cbr) of the subgrade soil. The formulation is a second-order, second-moment development of this equation. It requires knowledge of only the means and standard deviations of the random variables. This model compares favorably with a statistical analysis of the aasho test results. A sensitivity analysis indicates that the cbr variability has a dramatic influence on the pavement reliability. The uncertainty on the expectedtraffic loads has little effect on the reliability for a large number of axle loads. Hence, the mean value of the expected traffic loadis a sufficient estimate of this parameter. Second, an analytical model is formulated for the coefficients of equivalence of the (unbounded) granular materials of the base and subbase courses, using the theory of stochastic stress propagation in particulate media. It is shown that these coefficients are representative of the ability of the granular course to spread the applied load in a diffusion process. They are expressed as functions of the angle of internal friction of the material, and a modified formulation is derived for the equivalent thickness of the pavement. This paper appears in transportation research record no. 1286, Design and evaluation of rigid and flexible pavements 1990.
Abstract