Local calibration of material characterization models for performance-based flexible pavement design
MetadataShow full item record
The Mechanistic Empirical Pavement Design Guide (MEPDG) method, currently known as Pavement ME, recommends using locally calibrated material characterization models developed from laboratory testing of local materials under specific environmental and traffic loading conditions. The Pavement ME design method offers a more realistic design procedure and reduces the uncertainty that arise from empirical design procedures. This thesis developed a locally calibrated indirect tensile (IDT) strength material model for low temperature cracking predictions of hot mix asphalt (HMA) in Manitoba, Canada. In addition, the research investigated the integration of locally calibrated HMA, and unbound granular material characterization models into the Pavement ME framework to improve the design of flexible pavements. Laboratory IDT testing was conducted on typical HMA mixtures containing extracted binders and varying percentages of reclaimed asphalt pavement (RAP). The laboratory measured IDT strengths were used to calibrate a local IDT strength predictive model for Manitoba. The predictions from the local Manitoba model were compared to the predictions from the global Pavement ME IDT model, and a Michigan calibrated IDT model, using a statistical analysis. It was found that the global Pavement ME IDT strength model, if used without local calibration, produced inaccurate predictions of the IDT strength for Manitoba mixtures. It was also found that binder characterization methods in Level 2 and Level 3 can significantly impact the accuracy of IDT strength predictions. A case study using developed local HMA, base, and subgrade material characterization models in Manitoba were compared to designs using default (Level 3) material input values in Pavement ME design software. The results of integrating the locally calibrated models for HMA, base and subgrade layers demonstrated that the locally calibrated materials model inputs produce lower pavement structural thicknesses with higher reliability in the predicted distresses when compared to the default materials inputs. The effect of using calibrated material inputs was more pronounced for higher traffic loadings. The results of the study demonstrate that the use of calibrated models can potentially produce optimized pavement thicknesses due to improved pavement designs.
- FGS - Electronic Theses and Practica 
- Manitoba Heritage Theses