Regression Models to Estimate Critical Porosity of Soils from Basic Soil Properties Based on Percolation Theory
© 2017, Springer International Publishing AG. The rigidity of geo-materials (e.g., soils and sediments) reduce significantly, as their porosity increases to a critical value, the critical porosity. Knowledge of critical porosity can be useful in assessing the deformational characteristics of these materials. Estimation of critical porosity from experiments can be challenging especially at low effective stress conditions. In this work an effort is made to estimate this characteristic porosity from the basic soil properties (textural and plastic properties) inherited from the parent rock using percolation theory. Uniaxial compression experiments were performed in a low-stress oedometer (up to 2.5 MPa) on field soil samples to estimate their percolation parameters, the critical porosity and critical exponent from their textures based on percolation theory. The textural properties including fines content, mean grain size, the fractal dimensions and the sand to clay ratio. Regression models were developed to predict the critical porosities from soil textural properties that significantly explained in the range 85–90% of the variability in the model. In addition, the relations between the consistency limits of the soils and their critical porosities were investigated and the results indicate that the liquid limit, plastic limit, plasticity index and clay content explained 82–85% of the variability with critical porosity. Textural properties explained 64–79% of variability with the critical exponent, while the consistency limits explained 65–75% of the variability and with clay content explaining 80% of the variability. The provided prediction models may be useful in providing first-hand information about the critical porosity of soils in the field environment.
Volume / Issue
Start / End Page
Electronic International Standard Serial Number (EISSN)
International Standard Serial Number (ISSN)
Digital Object Identifier (DOI)