Estimation of atterberg limits of soils by artificial neural networks

Abstract

Soils intended to support structures, pavements or other loads should be evaluated by geotechnical engineers to predict their behavior under conditions of applied forces and varying humidity. Soil mechanics tests in geotechnical laboratories measure particle size distribution, shear strength, moisture content, and the potential for expansion or contraction of cohesive soils. Atterberg limit tests determine the moisture content at which fine-grained clay and silt soils transition between solid, semi-solid, plastic and liquid states. Artificial neural networks (ANN) is an information processing technology inspired by the information processing technique of the human brain. In this study, the determination of the Atterberg limits, which are frequently used for the determination of soil properties in geotechnical engineering and their comparison with experimental data, were determined and the results showed a very similar pattern.