Equivalent Loading Frequencies for Dynamic Analysis of Asphalt Pavements

An investigation of the existence of predominant frequency, $f_p$ (or frequencies), and the ability to predict critical pavement responses in asphalt layers using those predominant frequencies has been presented. The study used an extensive database of computed pavement response histories of four different asphalt pavement structures (thin and thick) at two temperatures (21 and 40°C) subjected to a tandem axle load at three speeds (16, 65, and $97\mathrm{km}/\mathrm{h}$) using the pavement response analysis program 3D-Move Analysis. The conversion of load duration to frequency and the validity of using one response component (${\sigma }_{zz}$) alone in the estimation of the pulse time ($t_p$) were explored. Instead of focusing on only ${\sigma }_{zz}$, attention is given to all pavement response components (stresses and strains). In addition, verification of whether the use of a consistent set of $f_p$ value(s) can in fact adequately capture all components of pavement response has also been undertaken. The viscoelastic properties that are representative for the predominant (or equivalent) frequencies from frequency sweep data were used. A major contribution of the study is that the critical responses and performance in a flexible pavement system can be successfully simulated with a static multilayer elastic analysis (within $\pm 10\%$) by assigning elastic modulus to hot-mix asphalt (HMA) based on appropriately selecting the corresponding predominant or equivalent frequency, $f_p$ (or frequencies). Generalized equations for estimating those $f_p$ values for all pavement structures, temperatures, and vehicle speeds under consideration have been developed. Good fitting parameters ($R^2$) were observed for all evaluated cases. The impact of loading frequency on predicted asphalt distresses was assessed for rutting and bottom-up fatigue cracking following the AASHTO mechanistic-empirical pavement design guide approach. The study showed that the proposed procedure for estimating $f_p$ values resulted in rut depths and bottom-up fatigue cracking that are within $\pm 10\%$ of those determined using the viscoelastic dynamic responses.