ARTIFICIAL NEURAL NETWORKS FOR ASSESSING PERMEABILITY CHARACTERISTICS OF SOILS

Artificial Neural Network (ANN) has been widely used for solving many problems in many areas. Predictions using the neural network are becoming more popular. ANN has been applied in geotechnical engineering also to predict pile capacity, settlement, liquefaction, soil properties and many. Permeability is one of the most important soil properties which is essential in solving large number of engineering problems. Grain size distribution and density are known to influence the permeability of sandy soils. Although the relationships between grain size distribution and permeability has been quantified by some researchers, the influence is not quantified. The correlations between coefficient of permeability and other soil properties individually are common among Geotechnical engineers. But establishing a correlation by assessing the coefficient of permeability of any soil type using all other soil properties is as such impossible generally. The paper presents the method of determining the permeability of soils, factors affecting the permeability of soils, existing correlations practiced and a model for assessing the coefficient of permeability modelled with the optimal input physical parameters.