Page 91 - Mechanics of Asphalt Microstructure and Micromechanics
P. 91

84   Ch a p t e r  Th r e e


                     Contact Normal          Branch Vector            Orientation
          Region  f(n) ob  f(n) pre  Error, %  f(n) ob  f(n) pre  Error, %  f(n) ob  f(n) pre  Error, %
          1        4     3.163 20.9       7     7.628   9.0    4      2.754    31.2
          2        8     6.819 14.8       8     8.375   4.7    3      4.536    51.2
          3        4     6.134 53.4      14    12.038  14.0    7      6.039    13.7
          4        8     5.685 28.9       6     8.588  43.1    2      3.309    65.5
          5       10    10.554  5.5      13    12.512   3.8    4      4.389     9.7
          6       12    12.516  4.3      10     9.595   4.0    4      3.727     6.8
          7        6     6.983 16.4       9     5.884  34.6    3      4.290    43.0
          8        9    10.164 12.9       8     8.176   2.2    9      6.209    31.0
          9       12     9.849 17.9       7     9.014  28.8    5      7.205    44.1
         10        4     6.169 54.2       5     4.899   2.0    4      2.647    33.8
         11       13    12.463  4.1       8     8.436   5.5    6      4.366    27.2
         12       12    11.501  4.2       7     7.853  12.2    1      2.528   152.8
         Total
         average 102   102      19.8    102   101      13.7
        TABLE 3.4  Predicted versus observed distributions of the fabric quantities.

              2D dataset. It is therefore not meaningful to make any comparison between the 3D tech-
              niques and the 2D techniques. First, a particle in contact may not be shown as in contact
              in 2D. Similarly, for particle orientation, a cutting plane may not be consistent with the
              orientation of the longest Feret diameter of the particle. The only fabric quantity of the
              three that might be estimated with reasonable accuracy from 2D datasets (three orthogo-
              nal planes) is the branch vector distribution (Wang, et al., 2001a). Nevertheless, with
              particles of irregular shapes, the mass center coordinates may be significantly different
              from the centroids of a revealed cross-section (Wang, et al., 2001a).
                 It should be noted that many researchers have developed 2D techniques to quantify
              these fabric quantities not because the 2D techniques are reliable, but because of the
              lack of 3D techniques.

              3.4.2.9  Fabric Quantities and Their Use in Constitutive Models
              The characterized fabric quantities can be directly used in constitutive models. This sec-
              tion presents a review on how fabric tensors are quantitatively related to stress, strain,
              and yielding function.

              Stress-Fabric Tensor Relation
              A direct application of the fabric quantities is to relate average stress (effective stress)
              with contact force and fabric tensors including contact normal tensor (J ij ) and branch
              vector tensor (F ij ). Equations 3-17 and 3-18 (Mehrabadi and Nemat-Nasser, 1983) pres-
              ent the relationship of how the average stress of s ij  in a granular assemble is related to
              the contact normal tensor and the branch vector tensor, respectively.
                                           −
                                           σ =  α J  +  α J J                    (3-17)
                                            ij
                                           −    0  ij  1  ik kj
                                           σ =  β F  +  β J J                    (3-18)
                                            ij
                                                0  ij  1  ik kj
                 a 0 , a 1  and b 0 , b 1  are functions of the invariants of the stress tensor.
                 Other relations include those defined by Tobita (1989), Tobita and  Yanagisawa
              (1992), and Bagi (1996).
   86   87   88   89   90   91   92   93   94   95   96