1
\chapter{Supplementary topics}
2
\label{supplementary-considerations}
4
\section{Frame invariance and flux contribution from acceleration}
5
\label{acceleration-objectivity}
7
In our earlier treatment \citep{growthpaper}, the constitutive
8
relation for the fluid flux had a driving force contribution arising
9
from the acceleration of the solid phase,
10
$-\rho_0^\mathrm{f}\bF^{\mathrm{T}}\frac{\partial \bV}{\partial t}$.
11
This term, being motivated by the reduced dissipation inequality, does
12
not violate the Second Law and supports an intuitive understanding
13
that the acceleration of the solid skeleton in one direction must result in
14
an inertial driving force on the fluid in the opposite
15
direction. However, as defined, this acceleration is obtained by the
16
time differentiation of kinematic quantities,\footnote{And not in terms
17
of acceleration {\em relative to fixed stars} for e.g., as discussed
18
in \cite[][Page 43]{TruesdellNoll:65}.} and does not transform in a
19
frame-indifferent manner. Unlike the superficially similar term
20
arising from the gravity vector,\footnote{Where every observer has an
21
implicit knowledge of the directionality of the field relative to a
22
fixed frame, allowing it to transform objectively. Specifically, under
23
a time-dependent rigid body motion imposed on the current
24
configuration carrying $\bx$ to $\bx^+ = \bc(t) + \bQ(t)\bx$, where
25
$\bc(t) \in \mathbb{R}^3$ and $\bQ(t) \in \mbox{SO}(3)$, it is
26
understood that the acceleration due to gravity in the transformed
27
frame is $\bg^+ = \bQ^\mathrm{T}\bg$ and is therefore
28
frame-invariant. However, $\ba^+ = \ddot{\bc} + 2\dot{\bQ}\bv +
29
\ddot{\bQ}\bx + \bQ\ba$ , and is therefore not frame-invariant.} the
31
term presents an improper dependence on the frame of the
32
observer. Thus, its use in constitutive relations is inappropriate,
33
and the term has been dropped in \mbox{Equation (\ref{fluidflux})}.
35
%% \todo{Cite Einstein's general relativity here.}
37
\section{Stabilisation of the simplified solute transport equation}
38
\label{stabilisation-solute-transport}
40
In weak form, the SUPG-stabilised method \citep{Paper6} for
41
Equation~(\ref{morestdform}) is,
45
&\int_{\Omega} w^{\mathrm{h}} \left(
46
\frac{\mathrm{d}\rho^{\mathrm{s}^{h}}}{\mathrm{d}t} +
47
\bm^f\cdot\mathrm{grad}\left[\frac{
48
\rho^{\mathrm{s}^{h}}}{\rho^f}\right] \right)
49
d\Omega\\ &+\int_{\Omega} \left( \mathrm{grad}
50
\left[w^{\mathrm{h}}\right] \cdot \bar{\bD^\mathrm{s}} \mathrm{grad}
51
\left[ \rho^{\mathrm{s}^{h}}\right] \right)\ d\Omega\\ +&
52
\sum_{\mathrm{e}=1}^{\mathrm{n_{el}}} \int_{\Omega_{\mathrm{e}}}
53
\tau \frac{\bm^{f}}{\rho^f} \cdot \mathrm{grad}
54
\left[w^{\mathrm{h}}\right] \left(
55
\frac{\mathrm{d}\rho^{\mathrm{s}^{h}}}{\mathrm{d}t} +
56
\bm^f\cdot\mathrm{grad}\left[\frac{
57
\rho^{\mathrm{s}^{h}}}{\rho^f}\right] \right) \ d\Omega\\ -&
58
\sum_{\mathrm{e}=1}^{\mathrm{n_{el}}} \int_{\Omega_{\mathrm{e}}}
59
\tau \frac{\bm^{f}}{\rho^f} \cdot \mathrm{grad}
60
\left[w^{\mathrm{h}}\right]
61
\left(\mathrm{div}\left[\bar{\bD^\mathrm{s}}\ \mathrm{grad} \left[
62
\rho^{\mathrm{s}^{h}}\right]\right]\right) \ d\Omega\\ = &
63
\int_{\Omega} w^{\mathrm{h}} \pi^\mathrm{s} \ d\Omega +
64
\int_{\Gamma_{\mathrm{h}}} w^{\mathrm{h}} h \ d\Gamma\\ +&
65
\sum_{\mathrm{e}=1}^{\mathrm{n_{el}}} \int_{\Omega_{\mathrm{e}}}
66
\tau \frac{\bm^{f}}{\rho^f} \cdot \mathrm{grad}
67
\left[w^{\mathrm{h}}\right] \pi^\mathrm{s} \ d\Omega,
68
\label{stabilizedmassbal}
72
\noindent where quantities with the superscript $\mathrm{h}$ represent
73
finite-di\-men\-sion\-al approximations of infinite-dimensional field
74
variables, $\Gamma_{\mathrm{h}}$ is the Neumann boundary, and this
75
equation introduces a numerical stabilisation parameter, $\tau$, which
76
we calculate from the $\mathrm{L}_{2}$~norms of element level
77
matrices, as described in \cite{tezduyarsupg}.
83
% TeX-master: "thesis"