ABSTRACT
In this study, multilayer pavement structure is simplified into one layer of equivalent thickness by using Equivalent Layer Theory (ELT). Artificial Neural Network (ANN)-based pavement structural analysis models were developed to find an equivalent thickness and elastic modulus of the modeled pavement system. The synthetic databases used as inputs in ANN forward and backcalculation models were created using MnLayer, a Layered Elastic Analysis (LEA) program. ANN models were trained to obtain the critical responses at the top and bottom of such layers in pavement systems. The multilayered flexible pavements were subjected to a 20-kip of Falling Weight Deflectometer (FWD) load in a circular area with uniform pressure. ANN models were found to represent a useful alternative approach for not only determining equivalent thickness and modulus but also providing close estimate of deflections of a multilayered flexible pavement system.
