Resumen
Considerable research efforts are made to develop methods to predict mechanical behavior of bituminous mixtures from properties of basic components. In particular, such methods would be most useful for the design of mixtures containing Reclaimed Asphalt Pavement (RAP) material. The objective of the study is the simulation of Linear ViscoElastic (LVE) behavior of bituminous mixtures containing RAP from LVE properties of base constituents. Five mixtures were produced and tested, a reference mixture produced only with 35/50 binder, a mixture containing 100% RAP and three mixtures produced with 35/50 binder and different RAP contents (20%, 40% and 60%). All mixtures had the same granular size distribution curve. Complete LVE characterization was carried out on all materials. DSR and Tension/Compression (T/C) tests were performed on base and RAP-extracted binders (from -30 °C to +70 °C and from 0.01Hz to 30Hz). T/C tests were carried out on mixtures (from -25 °C to 40 °C and from 0.001Hz to 10Hz). Test results on both binders and mixtures were successfully fitted with 2S2P1D (2 Springs, 2 Parabolic elements, 1 Dashpot) model. LVE behavior of mixtures containing RAP was simulated by using as input data only LVE properties of base binders (35/50 and RAP-extracted binders) and of the reference mixture (produced with 35/50 binder, without RAP). In order to do this, two existing procedures, previously developed at the ENTPE, were used conjunctly. The first procedure was applied to predict LVE behavior of binder blends of pure base and RAP-extracted binders over the whole range of frequency and temperature. This procedure allows estimating 2S2P1D parameters of binder blends from those of base bitumens, according to their proportions in the blend. Therefore, LVE properties of blends can be predicted over the whole range of frequency and temperature. The second procedure used is SHStS (Shift, Homothety, Shift in time, Shift) transformation. This analytical tool is useful to determine the relationship between LVE behavior of mixtures and their corresponding binders. Experimental data of pure 35/50 binder and its corresponding mixture (without RAP) were used to calibrate SHStS transformation, depending mainly on aggregate skeleton. LVE behavior of mixtures produced with RAP was then estimated from simulations of LVE properties of blends, having the same proportions of base binders, obtained with the first procedure. Simulations of LVE behavior of mixtures containing RAP were finally compared to T/C test data. Successful correspondence was found between predicted and experimental results. Small discrepancies observed can be reasonably attributed to incomplete blending of base and RAP binders within the mixture. Therefore, as a first approximation, the proposed procedure can be used to predict LVE behavior of mixtures produced with RAP, over the whole range of frequency and temperature, from LVE properties of base binders and of a reference mixture (with the same granular size distribution) are known.