Characterization of Unbound and Stabilized Materials and Improved Consideration of Their Effects on Pavement Performance

Abstract

The United States has more than 2.7 million miles of paved roads and highways that require approximately $165 billion spending each year. Pavement design, construction, maintenance, and management techniques are critical factors for optimizing this massive budget that are constantly being evolved. To control the design and maintenance of new and existing pavements, the American Association of Transportation Officials’ (AASHTO) currently follow a design guide named AASHTOWare Pavement Mechanistic-Empirical (ME) Design. The design guide provides a methodology for the analysis and performance prediction of pavements and overlays. Although the performance of pavements is known to be closely related to properties of the subgrade and underlying layers (i.e., base and/or subbase), some recent research studies indicate that the performance predicted by this methodology shows a low sensitivity to the properties of underlying layers and does not always reflect the extent of the anticipated effect. To overcome these limitations, this study proposes several enhancements, as needed, to the Pavement ME Design procedures to better reflect the influence of subgrade and unbound layers (properties and thicknesses) on the performance of pavements. These enhancements include several modifications of the models contained in Pavement ME Design such as (a) development of an artificial neural network (ANN) based soil water characteristics curve (SWCC) prediction model of base and subgrade; (b) development of a mechanistic-empirical equilibrium suction (ue) model for subgrade; (c) development of a ANN based resilient modulus (MR) model of base ; (d) development of a new shear strength (τ) and permanent deformation (εp) prediction model; (d) development of a modified modulus of subgrade reaction (k) model. The sensitivity of base and subgrade layers are evaluated on the basis of both flexible and rigid pavements performance using the developed models and compared with the predicted performance from Pavement ME design models. The results clearly show that the developed models have better sensitivity to moisture and interface bonding on both rutting and fatigue cracking performance compared to the Pavement ME design models. Attaining uniform construction of the required specification quality is another key factor to ensure the performance of pavement. To develop more efficient quality control (QC) evaluation methods, this study develops a quick and accurate, non-destructive method for determining reliable values of the in-place as compacted base course modulus. Simple laboratory test methods are incorporated with the ground penetrating radar (GPR) scans to determine the resilient modulus of the base layer through a mechanistic-based approach. Research efforts have also been undertaken to develop and calibrate the mechanistic-based models for predicting the construction quality of stabilized base materials. A mechanistic-empirical model is developed to predict the percentage of stabilizer in the base layer from electrical conductivity readings in the laboratory and further incorporated with GPR scans to estimate the stabilizer content of the base layer in the field.