One of our interests, as discussed in the introduction, is in porous
materials like cement-based materials and in particular gypsum plaster.
Therefore, we chose to use the measured elastic moduli of gypsum crystals for
the solid backbone moduli. If the backbone moduli tensor is isotropic, then
the Young's modulus E of the porous composite will scale with the magnitude of
Es. The shape of the E/Es vs. porosity graph, for an isotropic backbone, will not
depend at all on the value of
s in 2-D [8],
and only mildly so in 3-D [2].
However, since gypsum plaster is the material that we are looking to model, we have to deal with the gypsum crystal elastic moduli tensor, which is not isotropic. The gypsum crystal has monoclinic symmetry, so that its 3-D elastic moduli tensor contains 13 independent constants. The measurement of these constants using an acoustic method has been published [13]. As the crystal anisotropy would be very difficult to handle computationally when the crystals are overlapping, and also because we are ultimately interested in real porous materials, which are usually isotropic, we will make the assumption of an isotropic tensor for the solid backbone. An angular average, using eq. ( 11) with the full gypsum elastic moduli tensor, leads to [14]:

The corresponding values of the Young's modulus E and the Poisson's
ratio
are:

Equations (19) and (20) are the isotropic average of the 3-D tensor. If we
want to study 2-D models, a 2-D equivalent must be found. As was discussed in
section 2.1, there are two choices for making 2-D moduli from the 3-D moduli:
plane strain or plane stress. The plane strain values of the Youngs modulus
and Poisson ratio are: Es = 50 GPa,
s = 0.45. For all
the 2-D studies, the plane strain values of the 2-D moduli were used. The
plane stress values were used in the data to be presented in Fig.
12 in Section 5.