Development of Prediction Model Using Artificial Neural Network to Predict Transverse Crack Length in Wet and Dry No Freeze Climatic Regions

This study addresses the critical issue of transverse cracking in the United States interstate highway system, which spans approximately 222,000 miles according to the Federal Highway Administration. Transverse cracking, a major distress type in flexible pavements, is influenced by factors such as temperature, flexibility, moisture infiltration, cycling mechanisms, aging, and construction quality. The research employs an artificial neural network (ANN) to develop a predictive model for transverse crack length, achieving a remarkable R2 score of 0.91 for goodness of fit. The resulting equation from the model offers a valuable tool for predicting transverse crack lengths, providing insights that can significantly contribute to proactive maintenance strategies and the overall longevity and performance of interstate highways.