Abstract:
Superload (SL) vehicles have unique axle configurations and high axle weights, indicating the potential for a low number of SL applications to cause significant fatigue damage to a jointed plain concrete pavement (JPCP). The current JPCP fatigue model in the AASHTO Pavement Mechanistic-Empirical (ME) Design Guide is unable to account for damage caused by SLs because the stress prediction models within it do not consider these unique axle configurations. As a result, it is not possible to account for the potential fatigue damage accumulation caused by SL applications or identify critical conditions that contribute to damage accumulation using Pavement ME. To address this, a critical, SL stress-prediction model was developed that can be used along with the current fatigue damage model. Typical SL configurations in Pennsylvania were identified based on available permit data, and a database of critical tensile stresses generated by these SLs for various JPCP structures was developed using finite element analysis. This database was used to train a series of artificial neural networks (ANNs), which predict critical tensile stress as a function of SL axle configuration, temperature gradient, and pavement structure. The method for determining fatigue damage accumulation using the ANNs to predict stresses is then presented.