The use of index parameters to predict soil geotechnical properties

Abstract

This paper presents new attempts to explore the dependency of some key geotechnical properties of soils such as compaction characteristics, hydraulic conductivity, and soil shear strength to their index properties and performance of developed models to predict these properties. To do this, a database of 580 data sets was compiled including the results of grain size distribution, Atterberg limits, compaction, and permeability, measured at different levels of compaction degree (90 to 100 %) as well as consolidated-drained triaxial compression tests. Dependency of each geotechnical property to their index parameters was investigated using an Evolutionary polynomial regression method, to develop prediction models based on the collected database. Investigation of the performance of the developed models indicates that these models are capable of predicting these soil properties with a confidence interval of 95 %. Parametric analyses were also performed on the developed models.