Design Procedure for Pavements on Expansive Soils: Volume 3 (0-4518-1 (V3))

Swelling and shrinkage of sub-grade soils are critical factors contributing to increases in roughness and degradation of serviceability of highway pavements. Existing procedures for predicting swell are largely based on the potential vertical rise (PVR) procedure developed by McDowell in 1956. While the PVR procedure represents a major development in the design of pavements on expansive soils, instances of apparently over-conservative PVR predictions have led some designers to suggest revision or replacement of the existing procedure. This project reviews the basic assumptions of the existing PVR procedure and identifies the likely sources of the questionable predictions that have arisen in the past. An alternative procedure is presented that features rigorous modeling of both the moisture diffusion process that induces changes in suction within a soil mass and the deformations that occur in response to changes in suction. This alternative procedure includes provisions for measuring and/or estimating soil and environmental input parameters necessary for the predictions. A procedure for predicting the impact of soil deformations on pavement performance is also presented. The proposed procedure is applied to three study sections involving Texas roadways on expansive soils, and parametric studies are presented evaluating the effectiveness of various design measures including moisture barriers, lime treatment, and replacement of in situ sub-grade soils with "inert" soils.