An Enhanced Pavement Temperature Prediction Model Incorporating Improved Initial and Boundary Conditions

Abstract

Pavement temperature affects pavement performance, distress development, and the urban heat island effect. Accurate pavement temperature prediction models support pavement design, maintenance, and mitigation of the urban heat island effect. However, existing numerical models often oversimplify surface boundary and initial temperature conditions, producing notable errors. This study developed an enhanced prediction model through: (1) refined initial temperature conditions, (2) improved surface heat flux calculations, and (3) a new method for estimating effective rainfall mass in rainfall-induced heat flux. The model was validated using typical pavement structures and temperature data from the Long-Term Pavement Performance (LTPP) program. Results indicated that initial conditions had a lasting impact on predictions, especially at deeper layers. Significant relationships were observed between differences in surface heat flux and both the prediction start time and error dissipation. The proposed preconditioned-initial condition improved accuracy for both hot and cold days. The new rainfall-induced heat flux calculation method effectively supplemented surface heat flux estimation during rainfall events, reducing temperature prediction discrepancies. Overall, the enhanced temperature prediction approach developed in this study demonstrated broad applicability across various regions and seasons and achieved higher accuracy than existing methods, even when the impact of initial errors was not significant.