Data-Based Real-Time Moisture Modeling in Unsaturated Expansive Subgrade in Texas

Abstract

Moisture variations significantly influence the strength and stiffness of expansive subgrade soils, shortening the service lives of pavements and increasing the associated maintenance costs. Accurate measurements of soil moisture can be obtained through soil sampling and testing, but the process can be extensive and costly. Empirical models can accurately predict the moisture variations in an expansive subgrade in a shorter period of time, with lower accompanying costs. The objective of the current study was to develop moisture models, using real-time field monitoring data from two hot mix asphalt roads in North Texas. The collected data were analyzed in a statistical environment to solve two first degree Fourier series. The solution produced a moisture variation model that captured variations associated with seasonal effects and temporary variations associated with rainfall. The outputs of this model were within 90% of the values measured on site. Application of the developed models will facilitate noninvasive estimations of the response of soil strength and stiffness properties to variations in moisture.