Neural networks analysis of silty sand reinforced by carpet wastes

This paper discusses artificial neural networks application for predicitng the strength parameters of randomly distributed fiber-reinforced soil. Five strength parameters and three affecting factors have been identified from analyzing the results of triaxial drained tests. Two neural networks, namely, Multilayer Perceptron and Radial-Basis Function as well as a mathematical model were examined using the data obtained fom 125 triaxial tests. The reults show that all models can achieve a high level of accuracy although the Radial-Basis Function method outperformed tthe other two methods