Prediction of unconfined compressive strength of cement stabilized pavement materials

Abstract

This paper evaluated and justified suitable prediction equations for unconfined compressive strength (UCS) of cement stabilized pavement materials. Equations by multiple linear regression analysis were also proposed. Data including pavement recycling base and subbase layers were collected from the Bureau of Material, Analysis, and Inspection, Department of Highways, Thailand. Multiple linear regression was conducted using the SPSS software. Three equations obtained from multiple linear regression and past researches were compared. Two criteria were considered in the selection of equations in order to check the accuracy and precision of each equation. First criterion was based on the predicted UCS vs. actual UCS graph. The slope and interception values of each graph were evaluated. Second criterion used two indices: ranking distance (RD) and ranking index (RI). For a good correlation, both these indices shall approach zero. By considering the statistical analysis outputs from the two criteria, a relationship between UCS and California Bearing Ratio – water cement ratio [CBR/(w/c)] proposed by Sunitsakul in 2012 was justified as a suitable equation to predict the UCS of cement stabilized pavement materials. The study also confirmed that the relationship proposed by Sawangsuriya in 2016 for fine-grained soil with plasticity index was not appropriate for the prediction of UCS because the majority of the samples were non-plastic.