DEVELOPMENT OF A TIRE/PAVEMENT CONTACT-STRESS MODEL BASED ON ARTIFICIAL NEURAL NETWORKS

This report presents the first world-wide tire/pavement contact-stress model based on artificial neural networks (ANNs). Developed by the authors at the Pennsylvania Transportation Institute, Pennsylvania State University, this model represents the first mathematical representation of real, measured, contact stress at wide ranges of vertical loads and inflation pressures for two types of tires. The developed ANN model has the capability of generating complex stress distribution patterns under a tire at any given load and inflation pressure for a specific tire type used for the ANN training. The information given in this report is considered to be an important contribution to the ongoing efforts to improve tire/pavement contact-stress modeling and analysis. The neural network representation of a tire contact-stress distribution is named "Neuro-Patch Model." The neural network model has been trained using precise measured three-dimensional contact-stress distribution patterns obtained from low-speed rolling tire tests conducted by the University of California at Berkeley, and data have been supplied by the Federal Highway Administration. In this study, two types of tires, namely Goodyear 11R22.5 radial-ply and Goodyear 10.00x20 bias-ply truck tires, were modeled at different inflation pressures ranging from 520 to 920 kPa and vertical loads ranging from 26 to 56 kN.