ARTIFICIAL NEURAL NETWORK-BASED METHODOLOGIES FOR RATIONAL ASSESSMENT OF REMAINING LIFE OF EXISTING PAVEMENTS. DEVELOPMENT OF A COMPREHENSIVE, RATIONAL METHOD FOR DETERMINATION OF REMAINING LIFE OF AN EXISTING PAVEMENT

Most mechanistic-empirical methods for determining the remaining life of an existing pavement rely on the use of deflection-based nondestructive evaluation (NDE) devices. This report describes a methodology based on Artificial Neural Network (ANN) techniques to estimate the remaining life of flexible pavements given the occurrence of two possible failure modes: rutting and fatigue cracking. The ANN techniques are also used to develop models that predict the critical strains at the interfaces of the pavement. The inputs to all the models are the best estimates of the thickness of each layer and the surface deflections obtained from a Falling Weight Deflectometer test. Uncertainty in these variables is accounted for by the proposed methodology. The report also describes an approach to the production of pavement performance curves using the results of the ANN models.