Hybrid neural network and finite element modeling of sub-base layer material properties in flexible pavements

Abstract

This paper introduces a new concept of integrating artificial neural networks (ANN) and finite element method (FEM) in modeling the unbound material properties of sub-base layer in flexible pavements. Backcalculating pavement layer moduli are well-accepted procedures for the evaluation of the structural capacity of pavements. The ultimate aim of the backcalculation process from non-destructive testing (NDT) results is to estimate the pavement material properties. Using backcalculation analysis, in situ material properties can be backcalculated from the measured field data through appropriate analysis techniques. In order to backcalculate reliable moduli, unbound material behavior of sub-base layer must be realistically modeled. In this work, ANN was used to model the unbound material behavior of sub-base layer from experimental data and FEM as a backcalculation tool. Experimental deflection data groups from NDT are also used to show the capability of the ANN and FEM approach in modeling the unbound material behavior of sub-base layer. This approach can be easily and realistically performed to solve the backcalculation problems.