Estimation of Compression Index of Cohesive-Frictional Soil in Southwest Delhi District Using Statistical Correlations

Abstract

Oedometer tests for the determination of compression index are time-consuming and need to be meticulously performed. Hence, empirical equations need to be formulated to predict the value of the compression index. In the current study, regression models have been developed based on correlation studies for lightly overconsolidated cohesive-frictional soil obtained from a site in the Southwest Delhi district, India. Statistical analyses have been performed on 80 observations from consolidation tests conducted on the samples retrieved from 58 boreholes. Based on the correlation analyses, it is found that the existing empirical correlations estimate the compression index erroneously. So, formulating an empirical equation for the study area is important. The soil properties, such as natural water content and liquid limit, are unsuitable variables to predict the compression index. These findings are not consonant with those of previous research as these engage the mentioned parameters in modelling the compression index. The several univariate and multivariate regression equations developed in this study are tested for validation using the coefficient of determina ion. Among the several regression models examined, the logarithmic equation of the compression index contributes to accurate prediction.