Backcasting and forecasting stabilized soil mechanical properties for mechanistic-empirical pavement design

Abstract

The ability to reliably forecast or backcast mechanical properties of chemically stabilized soils to use as design inputs (e.g. Mechanistic-Empirical Pavement Design Guide) is not readily available. In this paper, seven data sets comprised of 1642 laboratory specimens and 206 field cores were used to calibrate numerical relationships from literature to backcast and forecast mechanical properties of chemically stabilized soil commonly used in Mississippi pavements. When relationships from literature were compared to Mississippi materials, unconfined compressive strength (UCS) was underpredicted; however, when regression parameters were optimized to the data sets, prediction accuracy increased and were on average within 89–99% of actual UCS values. Equations were also shown to provide reasonable estimates of elastic modulus (E) and indirect tensile strength (IDT) when parameters were optimized. Accelerated curing to replicate later life pavement properties in laboratory specimens as well as field cores were compared to relationships from literature to determine their applicability to later life mechanical properties. Literature equations calibrated using laboratory cured specimens predicted mechanical properties within the range of properties reported from cores. Based on this analysis, a set of recommended regression parameters are given that are optimized for common Mississippi materials.