Neural Networks Analysis of Airfield Pavement Heavy Weight Deflectometer Data

摘要

The Heavy Weight Deflectometer (HWD) test is one of the most widely used tests for assessing the structuralintegrity of airport pavements in a non-destructive manner. The elastic moduli of the individual pavement layers predictedfrom the HWD deflection measurements through inverse engineering analysis are effective indicators of pavement layercondition. The primary objective of this study was to develop a tool for backcalculating non-linear pavement layer modulifrom HWD data using Artificial Neural Networks (ANN) for rapid structural evaluation of airfield pavements. A multilayer,feed-forward backpropagation ANN which uses an error-backpropagation algorithm was trained to approximate theHWD backcalculation function. The synthetic database generated using an axisymmetric pavement finite-elementprogram was used to train the ANN. Using the ANN, the Asphalt Concrete (AC) moduli and subgrade moduli weresuccessfully predicted. Apart from the moduli, an attempt was made to predict the critical pavement structural responsesusing ANN models. The final product was used in backcalculating pavement layer moduli and predicting subgradedeviator stresses from actual field data acquired at the Federal Aviation Administration’s National Airport Pavement TestFacility (NAPTF).