This study focuses on the use of artificial neural network (ANN)-based pavement backcalculation tools for analyzing falling weight deflectometer (FWD) data collected from flexible pavement sections. Some of the pavement sites have been part of the Long-Term Pavement Performance (LTPP) program, and a history of pavement materials testing and FWD deflection data already exist in the LTPP database. Pavement backcalculation tools developed in this study have been utilized to predict the pavement layer moduli and critical pavement responses of flexible pavement layers under typical highway pavement loading conditions. Unlike the linear elastic layered theory commonly used in pavement layer backcalculation, nonlinear subgrade soil response models were used in the ILLI-PAVE finite element program to account for the softening nature of fine-grained subgrade soils and hardening behavior of the unbound base materials under increasing stress states. Preliminary investigations have shown that the ANN-based backcalculation models were capable of rapidly predicting the layer moduli and critical pavement responses with low average errors when compared to the models obtained directly from the finite element analyses. In addition to the analyses of large amounts of FWD data using the backcalculation tools developed in this study, predicted pavement layer moduli values were compared with the traditional backcalculation software solutions; the comparison results are presented in this paper. The advantages of using an ANN-based rapid pavement layer backcalculation tool are also discussed.
Samenvatting