Nondestructive Flexible Pavement Evaluation Using ILLI-PAVE Based Artificial Neural Network Models

摘要

Artificial neural networks (ANNs) were used in this paper to develop an improved and more accurate approach for backcalculating pavement layer moduli from Falling Weight Deflectometer (FWD) test data collected in the field. For this purpose, critical pavement responses were computed by the ILLI-PAVE finite element program widely used and proven to be effective for the analysis of flexible pavement systems with the considerations of the nonlinear aggregate base and subgrade soil behavior. The ANN models were then trained to map the nonlinear functional relationships between the FWD deflections, layer properties, and the critical pavement responses.