USE OF NEURAL NETWORKS TO DEVELOP ROBUST BACKCALCULATION ALGORITHMS FOR NONDESTRUCTIVE EVALUATION OF FLEXIBLE PAVEMENT SYSTEMS

摘要

Artificial neural networks (ANNs) are valuable computational intelligence tools that are increasingly being used to solve resource-intensive complex problems as an alternative to using more traditional techniques. Recent research at the Iowa State University has focused on the use of artificial neural networks (ANNs) as pavement structural analysis tools for accurate and rapid prediction of pavement layer moduli and critical pavement responses (stresses, strains and deflections) of conventional flexible pavements subjected to highway pavement loadings. In order to develop more robust networks that can tolerate the noisy or inaccurate pavement deflection patterns collected from the Falling Weight Deflectometer (FWD) field tests, several network architectures were trained with varying levels of noise in them. Applied noise levels in deflection basins and pavement layer thicknesses ranged from ± 2% to ± 10% to train the robust ANN models that can account for the variations in deflection measurements and pavement layer thicknesses due to poor construction practices. ANN models then trained with the results from the ILLI-PAVE finite element program solutions have been found to be viable alternatives. The trained ANN-based backcalculation models were capable of rapidly predicting the pavement layer moduli and critical pavement responses and pavement surface deflections with very low average errors compare to those obtained directly from the finite element analyses.