Structural evaluation with a falling weight deflectometer (FWD) plays an essential role in the maintenance and rehabilitation of highways. Research on the influence of bedrock on the results of FWD tests and dynamic back-calculations for asphalt pavements has been a hot topic. In this study, the viscoelastic characteristics of the asphalt surface layer were described by the modified Havriliak-Negami (MHN) model rather than the layered elastic theory computing framework. The effects of bedrock on the deflection time history curves of viscoelastic asphalt pavement under the FWD test were investigated and analyzed by the spectral element method with fixed-end boundary conditions (B-SEM). The effects of bedrock on parametric back-calculation results for the viscoelastic asphalt surface layer and other elastic layers were explored by dynamic back-calculations. To exclude the impact of bedrock and improve the back-calculation accuracy, support vector regression optimized by the gravitational search algorithm (GSA-SVR) was adopted to predict bedrock depth. The viscoelastic responses and dynamic parametric back-calculation results were significantly influenced by bedrock beneath asphalt pavements, and the influence decreased (but was not eliminated) with increasing depth of bedrock. The GSA-SVR model provided excellent predictions of bedrock depth, and it can be applied for pavement performance assessment in road projects.
Read full abstract