Real-Time Non-Destructive Evaluation of Airport Pavements Using Neural Network Based Models

Real-Time Non-Destructive Evaluation of Airport Pavements Using Neural Network Based Models

Kasthurirangan Gopalakrishnan
Copyright: © 2010 |Pages: 23
DOI: 10.4018/978-1-60566-800-0.ch007
OnDemand:
(Individual Chapters)
Available
$37.50
Current Special Offers
No Current Special Offers
TOTAL SAVINGS: $37.50

Abstract

Nondestructive test (NDT) and evaluation methods are well-suited for characterizing materials and determining structural integrity of airfield pavement systems. The Heavy Weight Deflectometer (HWD) test is one of the most widely used NDT impulse device for assessing the structural condition of airport pavements in a non-destructive manner. Through inverse analysis of HWD deflection data (more commonly referred to as backcalculation), the structural stiffness parameters of the individual airport pavement layers are, in general, determined using iterative optimization routines. In recent years, Artificial Neural Networks (ANN) aided inverse analysis has emerged as a successful alternative for predicting pavement layer moduli from HWD deflection data in real-time. Especially, the use of Finite Element (FE) based pavement modeling results for training the ANN aided inverse analysis is considered to be accurate in realistically characterizing the non-linear stress-sensitive response of underlying pavement layers. The development of an effective tool for real-time backcalculation of flexible airfield pavement layer moduli based on HWD test data is discussed in this Chapter. The ANN-based backcalculation tool is validated using actual field data acquired from a full-scale, state-of-the-art airport pavement test facility.
Chapter Preview
Top

Introduction

Objective

The current study described in this Chapter focused on developing an effective tool for real-time backcalculation of flexible airfield pavement layer moduli based on Heavy Weight Deflectometer (HWD) test data. A multi-layer, feed-forward network which uses an error-backpropagation algorithm was trained to approximate the HWD backcalculation function. The developed ANN-models were used in backcalculating pavement layer moduli from actual HWD test data acquired at the National Airport Pavement Test Facility (NAPTF). The NAPTF was constructed to generate full-scale testing data to support the investigation of the performance of airport pavements subjected to new generation aircraft. The NAPTF test details are discussed under a separate section.

The results from this study were also compared with those obtained using a conventional ELP-based backcalculation program. It is noted that this was an initial study with limited scope specifically targeted towards the backcalculation of pavement layer moduli from HWD data acquired at the NAPTF. However, the results highlight the potential for extending this concept for developing generic ANN-based models which would be useful in the analysis of routine HWD test data collected at flexible airfield pavements. This could be accomplished by training the ANN models developed in this study over a broad range of input values.

Complete Chapter List

Search this Book:
Reset