Evaluation of Subgrade Strength and Pavement Designs for Reliability

Reliability is an important factor in flexible pavement design to consider the variability associated with the design inputs. In this study, subgrade strength variability and flexible pavement designs are evaluated for reliability. Six existing pavement section's design data are studied using probabilistic, AASHTO, and mechanistic-empirical pavement design guide (MEPDG). Parameters such as mean, maximum likelihood, median, coefficient of variation, and density distribution function of subgrade strength (R value) are determined. Design outputs are compared in terms of reliability and thickness using these design procedures. It is shown that the AASHTO provides higher reliability values compared to the probabilistic procedure. All the existing pavements fail in the MEPDG distress reliability such as rutting and top-down cracking reliabilities. Currently, New Mexico Department of Transportation uses a single design R value to deal with variability associated with subgrade strength in flexible pavement design. It is shown in this study that single design R value for a roadway section does not yield an effective design regarding target reliability, while the subsectioning procedure based on coefficient of variance of R value is a better way to deal with the subgrade variability. An assessment of minimum R value for making the decision of subexcavation is also presented. It is shown that increasing the minimum R value for subexcavation is not always the proper solution to meet design reliability; rather it yields an inefficient design for requiring higher frequencies of subexcavation. Finally, the reliability of the flexible pavement design is evaluated by varying hot mix asphalt properties. Alternative designs are recommended for the existing pavement thicknesses by modifying material and subgrade properties to mitigate different distresses.