On the use of neuronets for simulating the stress-strain behavior of soils

Recurrent computational neural network (R-CNN) approach is used to demonstrate the potential use of this technique in modeling the constitutive behavior of geomaterials. For this purpose, R-CNN was used to simulate the triaxial-based consolidated undrained monotonic stress-strain behavior of Nevada sand and the uniaxial-based cyclic stress-strain response of a fine-grained soil. Overall, R-CNN approach was found efficient in simulating the stress-strain behavior of both soils under various initial confining pressure, density and compaction states.