Page 320 - Mechanics of Asphalt Microstructure and Micromechanics
P. 320

312   Ch a p t e r  N i n e


              lower part of the specimen, but reaches the peak value near the shear plane, which
              agrees with the experimental observations (Figure 9.21c).
                 The comparison between simulations and experimental measurements of both dis-
              placement and rotation indicate that particle shape is a significant factor for DEM simu-
              lations. This is consistent with the DEM rotational formulations.


        9.5 DEM Applications for AC

              9.5.1  Historic Use of DEM
              Since 1971, when Cundall (1971) originally introduced DEM and used it in the analysis
              of rock-mechanics problems, the method has received considerable attention in study-
              ing granular materials. In addition to the granular assemblies, DEM has been extended
              to simulating solid materials using bonded contact models. The early-day application
              of this method included several hundred particles; while in the last 20 years DEM has
              evolved to the capability of simulating a system of a few million particles. The applica-
              tion has been introduced in different fields and also in the asphalt mixture field (Mee-
              goda and Chang, 1994; Meegoda and Chang, 1995; Papagiannakis et al., 2002; Kim et al.,
              2008; Kim and Buttlar, 2005; Gantt and Gatzke, 2005).
                 As with the growth of DEM, many codes have been developed for simulating gran-
              ular and solid materials, including the early codes Ball (Cundall and Strack, 1979) and
              Trubal (Chen and Hung, 1991; Ng and Dobry, 1994; Washington and Meegoda, 2003),
              universal distinct element code (UDEC) (Dickens and Walker, 1996; Zhao et al., 2008),
              and discrete element code in three dimensions (3DEC) (Hart et al., 1988). In the AC area,
              the particle flow code (PFC) in 2D and 3D (PFC2D/3D, Itasca Consulting Group, 2005)
              has been widely used by researchers in AC (Buttlar and You, 2001; Dai and You, 2007;
              Liu and You, 2008; You and Buttlar, 2006; Abbas et al., 2005, 2006, 2007). The major rea-
              sons for PFC3D’s wide use in AC is its commercial availability, flexibility for clumping
              mechanism, and viscoelastic contact models. This literature review focuses on collect-
              ing work done in simulating asphalt-based materials using the DEM method. Table 9.11
              presents the major applications and their features. It should be noted that only essen-
              tially new applications were cited in Table 9.11.
                 Early in these research studies, asphalt mastics were modeled as an elastic material
              (Abbas et al., 2005; Buttlar and You, 2001; Collop et al., 2004, 2006, 2007; Dai and You,
              2007; You and Buttlar, 2004, 2005, 2006). With these non-viscous models, however, the
              time-dependent properties such as relaxation, dynamic modulus, and phase angles
              cannot be predicted. Nevertheless, one of the earliest applications of DEM in AC (Mee-
              god and Chang, 1994, 1995; Chang and Meegod, 1997, 1999) utilized a Maxwell ele-
              ment. Throughout the literature review, the Burger model has been used exclusively
              for modeling the viscoelasticity of asphalt binders or mastics in mixtures (Collop et al.,
              2006, 2007; Abbas et al., 2007; Liu et al., 2009). With these viscoelastic models, time-de-
              pendent behavior of asphalt mixtures was simulated, such as the dilation, dynamic
              modulus, and phase angles. In addition to the contact models above, there was the bi-
              linear cohesive model developed by Kim and Buttlar for simulating fractures in as-
              phalt mixtures, using the displacement-softening model provided in PFC 2D (Itasca
              Consulting Group, 2005; Kim and Buttlar, 2005; Kim et al., 2008). A large portion of the
              literature DEM simulations have been in stiffness prediction (Abbas et al., 2007; Dai
              and  You, 2007; Liu and  You, 2008;  You and Buttlar, 2004, 2006), fatigue modeling
   315   316   317   318   319   320   321   322   323   324   325