UTILIZING FALLING WEIGHT DEFLECTOMETER DATA FOR ASPHALT PAVEMENT RUTTING PERFORMANCE PREDICTION

Abstract: Rutting is a prevalent pavement distress that manifests depressions in wheel paths, adversely affecting ride quality and safety, particularly in high temperatures and under repeated heavy traffic. In wet, non-freeze climatic regions of the United States, the phenomenon of rutting significantly exacerbates the hazards associated with hydroplaning by facilitating the accumulation of water, thereby reducing tire traction, and increasing the risk of skidding. This study presents five Artificial Neural Network (ANN) models to predict rut depth in asphalt pavements in the wet, no-freeze climatic regions of the United States. These models employ Falling Weight Deflectometer (FWD) data, alongside pavement, traffic, and environmental factors, using datasets from the Long-Term Pavement Performance (LTPP) database. Artificial Neural Network (ANN) models were constructed and rigorously assessed, and Model 5 revealed the highest correlation between the predicted and the observed rut depths. A sensitivity analysis was also performed to identify the most influential input parameters affecting rutting prediction. This analysis helps refine the model and highlights critical factors for maintenance. The models provide a reliable tool for highway agencies to predict future rut depths, aiding in effective pavement management and maintenance strategies. Integrating the model into pavement management systems can help extend pavement lifespan and improve roadway safety.