[cig-commits] r11472 - in long/3D/Gale/trunk: . documentation documentation/images
walter at geodynamics.org
walter at geodynamics.org
Wed Mar 19 01:06:11 PDT 2008
Author: walter
Date: 2008-03-19 01:06:11 -0700 (Wed, 19 Mar 2008)
New Revision: 11472
Modified:
long/3D/Gale/trunk/
long/3D/Gale/trunk/documentation/gale.lyx
long/3D/Gale/trunk/documentation/images/tibet_setup.eps
long/3D/Gale/trunk/documentation/images/tibet_setup.fig
Log:
r2042 at earth: boo | 2008-03-19 01:03:38 -0700
Incorporate Gurnis' edits to the manual
Property changes on: long/3D/Gale/trunk
___________________________________________________________________
Name: svk:merge
- 3a629746-de10-0410-b17b-fd6ecaaa963e:/cig:2036
+ 3a629746-de10-0410-b17b-fd6ecaaa963e:/cig:2042
Modified: long/3D/Gale/trunk/documentation/gale.lyx
===================================================================
--- long/3D/Gale/trunk/documentation/gale.lyx 2008-03-19 00:58:41 UTC (rev 11471)
+++ long/3D/Gale/trunk/documentation/gale.lyx 2008-03-19 08:06:11 UTC (rev 11472)
@@ -1,4 +1,4 @@
-#LyX 1.5.1 created this file. For more info see http://www.lyx.org/
+#LyX 1.5.3 created this file. For more info see http://www.lyx.org/
\lyxformat 276
\begin_document
\begin_header
@@ -166,10 +166,10 @@
\begin_layout Standard
The main audience for Gale is research scientists interested in modeling
- tectonic processes on long time scales.
+ tectonic processes on long geological time scales.
Examples of problems that can be solved are the development of tectonic
- structure associated with extension and compression, especially where localizat
-ion is important.
+ structures associated with extension and compression, especially where
+ localization is important.
You do not have to be an expert in finite element analysis or scientific
computing to use this software.
\end_layout
@@ -256,7 +256,7 @@
EAR-0406751.
However, most of the software components below Gale have been developed
by the Victoria Partnership for Advanced Computing (VPAC) and Monash University.
-
+ Some of the support for VPAC has come from subawards from CIG.
\end_layout
\begin_layout Standard
@@ -291,7 +291,7 @@
\end_layout
\begin_layout Standard
-CIG has developed Gale in response to community demand by building on existing
+CIG developed Gale in response to community demand by building on existing
work by VPAC and Louis Moresi's group at Monash University.
The code is being released under the GNU General Public License.
\end_layout
@@ -323,7 +323,6 @@
You only have to make sure that you are consistent.
For example, if you give velocities in cm/year, make sure that your viscosities
and lengths also use cm and years.
-
\end_layout
\begin_layout Subsection
@@ -408,36 +407,18 @@
.
Gale simulates creeping flows, so acceleration terms are neglected and
material motion evolves through a series of equilibria.
- If your boundary condition has a time component, then you may infer a time.
+ If your boundary condition has a time dependent component, then you may
+ infer a time.
For example, if the boundaries move inwards at 1 mm/sec, then the solution
when the boundaries have moved 5 mm would correspond to 5 seconds.
\end_layout
\begin_layout Standard
-So far, these equations are simple, taking no account of gravity or temperature.
- So this is mostly useful when simulating laboratory experiments.
-\end_layout
-
-\begin_layout Standard
-Eqs.
-
-\begin_inset LatexCommand ref
-reference "eq:simple momentum conservation"
-
-\end_inset
-
- and
-\begin_inset LatexCommand ref
-reference "eq:continuity"
-
-\end_inset
-
- are solved by assuming a simple Newtonian fluid.
- Then
+Assuming a simple Newtonian fluid, we can write
\begin_inset Formula $\tau$
\end_inset
- can be written in terms of the rate of strain tensor
+ in terms of the rate of strain tensor
\begin_inset Formula $D$
\end_inset
@@ -455,12 +436,8 @@
\end_inset
is the viscosity.
+ Using standard finite-element techniques, eqs.
-\end_layout
-
-\begin_layout Standard
-Using standard finite-element techniques, eqs.
-
\begin_inset LatexCommand ref
reference "eq:simple momentum conservation"
@@ -572,7 +549,20 @@
\end_layout
\begin_layout Standard
-Gravity is accounted for by adding a body force term to eq.
+Equations
+\begin_inset LatexCommand ref
+reference "eq:simple momentum conservation"
+
+\end_inset
+
+ and
+\begin_inset LatexCommand ref
+reference "eq:continuity"
+
+\end_inset
+
+ do not include the effect of gravity.
+ Gravity is accounted for by adding a body force term to eq.
\begin_inset LatexCommand ref
reference "eq:simple momentum conservation"
@@ -630,7 +620,18 @@
\end_layout
\begin_layout Standard
-Temperature is included by solving the energy equation
+Equation
+\begin_inset LatexCommand ref
+reference "eq:body force"
+
+\end_inset
+
+ does not explicitly include the effect of temperature and heat transfer.
+ Temperature can be implicitly included by using a temperature dependent
+ viscosity and/or modifying the gravitational force to have a thermal buoyancy
+ term.
+ To make the simulation completely self consistent, we solve the energy
+ equation
\end_layout
\begin_layout Standard
@@ -903,7 +904,7 @@
\end_layout
\begin_layout Itemize
-Mac OS X 10.4.6 (G4 and G5)
+Mac OS X 10.4.6 (G4, G5, and Intel)
\end_layout
\begin_layout Itemize
@@ -2450,8 +2451,7 @@
n: the force of gravity, which by default is set to 1 (if you are using
cgs, for example, the force of gravity must be changed to 980), and the
normal velocities on all boundaries except the top, which are set to zero.
- Beyond that, you only need to add where different materials are placed
- initially.
+ Beyond that, you only need to specify where to place materials.
\end_layout
\begin_layout Subsection
@@ -5946,8 +5946,8 @@
\end_inset
) should be as easy as changing the various parameters to the right number.
- In practice, it can be difficult to get right, because you may have to
- change many different parameters at once to ensure stability.
+ In practice, it can be quite difficult, because you may have to change
+ many different parameters at once to ensure a stable solution.
To make that transition easier, use the sample input files in
\family typewriter
input/examples/tibet.xml
@@ -5975,7 +5975,7 @@
placement H
wide false
sideways false
-status collapsed
+status open
\begin_layout Standard
\align center
@@ -6007,30 +6007,181 @@
\end_layout
\begin_layout Standard
-The 3D input file layers a uniform thickness of crust onto a mantle, but
- then imports a realistic topography from a data file.
- At the moment, this location is hard coded in
-\family typewriter
-src/StGermain/Discretisation/Mesh/src/
-\newline
-SurfaceAdaptor.c
-\family default
+The thickness of the mantle in the center is set to isostatically compensate
+ the thicker crust.
+ There is also a nonlinear temperature gradient, going from 273 K at the
+ surface to 1333 K at the bottom.
+\end_layout
+
+\begin_layout Standard
+The crust has a density of
+\begin_inset Formula $2800kg/m^{3}$
+\end_inset
+
+, a coefficient of thermal expansivity (
+\begin_inset Formula $\alpha$
+\end_inset
+
+) of
+\begin_inset Formula $3\cdot10^{-5}K^{-1}$
+\end_inset
+
+, a thermal diffusivity of
+\begin_inset Formula $10^{-6}m^{2}/s$
+\end_inset
+
+, a radiogenic heating rate (
+\begin_inset Formula $Q$
+\end_inset
+
+) of
+\begin_inset Formula $10^{-12}K/s$
+\end_inset
+
+, and a radiogenic decay timescale (
+\begin_inset Formula $\lambda$
+\end_inset
+
+) of
+\begin_inset Formula $0s$
+\end_inset
+
.
- It needs a data file
+ The crust's viscosity is modeled with a Frank-Kamenetskii temperature dependent
+ viscosity (see section
+\begin_inset LatexCommand ref
+reference "sub:Frank-Kamenetskii"
+
+\end_inset
+
+), with the viscosity going from
+\begin_inset Formula $10^{25}kg\, m^{-1}s^{-1}$
+\end_inset
+
+ at the surface to
+\begin_inset Formula $10^{20}kg\, m^{-1}s^{-1}$
+\end_inset
+
+at the bottom.
+ The crust's yielding behavior is modeled with a Mohr-Coulomb rheology,
+ with a cohesion of
+\begin_inset Formula $4.4\cdot10^{6}kg\, m^{-1}s^{-2}$
+\end_inset
+
+, weakening to
+\begin_inset Formula $4\cdot10^{5}kg\, m^{-1}s^{-2}$
+\end_inset
+
+.
+ The internal angle of friction is
+\begin_inset Formula $30$
+\end_inset
+
+ degrees.
+\end_layout
+
+\begin_layout Standard
+The mantle has a density of
+\begin_inset Formula $3300kg/m^{3}$
+\end_inset
+
+, a coefficient of thermal expansivity (
+\begin_inset Formula $\alpha$
+\end_inset
+
+) of
+\begin_inset Formula $3\cdot10^{-5}K^{-1}$
+\end_inset
+
+, a thermal diffusivity of
+\begin_inset Formula $10^{-6}m^{2}/s$
+\end_inset
+
+, and radiogenic heating rate (
+\begin_inset Formula $Q$
+\end_inset
+
+) of
+\begin_inset Formula $8.48\cdot10^{-13}K/s$
+\end_inset
+
+, and a radiogenic decay timescale (
+\begin_inset Formula $\lambda$
+\end_inset
+
+) of
+\begin_inset Formula $0s$
+\end_inset
+
+.
+ The mantle is assumed to be purely viscous with the same parameters as
+ the crust.
+\end_layout
+
+\begin_layout Standard
+The top cover image shows the strain rate invariant after the model has
+ extended 30 km.
+ The resolution is 2048x512, and we used a direct solver (Mumps).
+ The most prominent faults occur near where the crust thickens, although
+ smaller faults occur throughout the crust.
+ The depth of the faults is limited by the relatively low viscosity deeper
+ in the crust.
+
+\end_layout
+
+\begin_layout Standard
+The 3D input file models a region 1000km x 1000km.
+ Topography is imported from a data file.
+ Underneath, the crust extends further down 32 km, and the mantle is 68
+ km thick beyond that.
+ The location of the input file is hard coded in
+\end_layout
+
+\begin_layout LyX-Code
+
\family typewriter
+src/StGermain/Discretisation/Mesh/src/SurfaceAdaptor.c
+\end_layout
+
+\begin_layout Standard
+It needs a data file
+\family typewriter
ascii_topo
\family default
.
This file is very large, so it is not distributed with Gale.
- You may download it at
+ You may download the file used to make the picture at
\begin_inset LatexCommand htmlurl
target "geodynamics.org/~walter/ascii_topo"
\end_inset
.
+ It covers a region of the Tibetan plateau under extension.
\end_layout
+\begin_layout Standard
+The material properties are the same as in the 2D case with one exception.
+ The crust cohesion weakens to
+\begin_inset Formula $4\cdot10^{6}kg\, m^{-1}s^{-2}$
+\end_inset
+
+ rather than
+\begin_inset Formula $4\cdot10^{5}kg\, m^{-1}s^{-2}$
+\end_inset
+
+.
+ This softens the viscosity jump when material yields, allowing iterative
+ solvers to converge.
+\end_layout
+
+\begin_layout Standard
+The bottom cover image shows the strain rate invariant after the model has
+ extended 24 km.
+ The resolution is 128x128x16, and we used an iterative solver (GMRES).
+ The fault locations are determined by the variations in topography.
+\end_layout
+
\begin_layout Chapter
Modifying Gale
\begin_inset LatexCommand label
@@ -11129,6 +11280,11 @@
\end_layout
\begin_layout Subsubsection
+\begin_inset LatexCommand label
+name "sub:Frank-Kamenetskii"
+
+\end_inset
+
Frank-Kamenetskii
\end_layout
@@ -14235,12 +14391,8 @@
\end_layout
\begin_layout Section
-
+VTK Files:
\family typewriter
-VTK
-\family default
-Files:
-\family typewriter
.vts, .pvts, .vtu, and .pvtu
\family default
(Visualization)
@@ -14704,14 +14856,14 @@
\end_layout
\begin_layout Standard
-Clast derived a simple analytic solution for the pressure and velocity fields
- for a circular inclusion under simple shear
+Schmid and Podladchikov
\begin_inset LatexCommand cite
key "Clast"
\end_inset
-.
+ derived a simple analytic solution for the pressure and velocity fields
+ for a circular inclusion under simple shear.
The file
\family typewriter
input/benchmarks/circular_inclusion/README
@@ -14760,10 +14912,6 @@
\end_layout
-\begin_layout Standard
-
-\end_layout
-
\end_inset
@@ -14775,9 +14923,30 @@
\end_layout
\begin_layout Standard
-Figure xxx plots the error in the pressure along the x axis as you change
- resolution and distance to the boundary.
- As you increase resolution and move the boundary away, the numerical solution
+We set the viscosity of the inclusion to 10 and the viscosity of the background
+ to 1.
+ This creates a discontinuity in the viscosity at the surface of the inclusion,
+ so the error is concentrated around that surface.
+
+\end_layout
+
+\begin_layout Standard
+Figure xxx plots the error in the velocity along the x axis within the inclusion
+ for different resolutions.
+ While convergence is good deep inside the inclusion, it is not so good
+ near the surface on the inclusion.
+\end_layout
+
+\begin_layout Standard
+Figure xxx plots the error in the pressure along the diagonal for different
+ resolutions.
+ The errors are large near the boundary between the inclusion and the rest
+ of the viscous medium.
+
+\end_layout
+
+\begin_layout Standard
+As you increase resolution and move the boundary away, the numerical solution
converges to the analytic solution.
\end_layout
@@ -14944,72 +15113,6 @@
\end_layout
\begin_layout Section
-\begin_inset LatexCommand label
-name "sec:Falling-Cylinder"
-
-\end_inset
-
-Falling Cylinder
-\end_layout
-
-\begin_layout Standard
-This benchmark is very similar to the falling sphere case and uses the same
- input file.
- We can write down an expression for the drag
-\begin_inset LatexCommand cite
-key "Landau & Lifschitz"
-
-\end_inset
-
-
-\end_layout
-
-\begin_layout Standard
-\begin_inset Formula \[
-F=\frac{4\pi\eta u}{\log\left(3.70\eta/uR\right)},\]
-
-\end_inset
-
-
-\end_layout
-
-\begin_layout Standard
-and the buoyancy force
-\end_layout
-
-\begin_layout Standard
-\begin_inset Formula \[
-F=\pi R^{2}g\delta\rho.\]
-
-\end_inset
-
-
-\end_layout
-
-\begin_layout Standard
-For the same parameters as above, a nonlinear solve gives a velocity of
-
-\end_layout
-
-\begin_layout Standard
-\begin_inset Formula \[
-u=0.0140.\]
-
-\end_inset
-
-
-\end_layout
-
-\begin_layout Standard
-This solution is subject to the same caveats regarding the finite viscosity
- of the cylinder, finite Reynolds number, and boundary effects.
- It turns out, though, that the boundary effects are much stronger in this
- case than in 3D.
- As the Figure xxx shows, the boundary has to be moved extremely far away
- to see convergence to the analytic solution.
-\end_layout
-
-\begin_layout Section
Relaxation of Topography
\end_layout
@@ -15087,8 +15190,9 @@
\end_inset
shows the results of a low resolution run.
- Even this run is not that small (128x256), because we need fairly high
- resolution to be able to accurately resolve the small (1%) height difference.
+ Even this run is not particularly small (128x256), because we need fairly
+ high resolution to be able to accurately resolve the small (1%) height
+ difference.
Also note that we use symmetry to only simulate half of the wavelength.
\end_layout
@@ -15204,7 +15308,7 @@
\begin_inset Float figure
wide false
sideways false
-status collapsed
+status open
\begin_layout Standard
\align center
@@ -15223,7 +15327,8 @@
\end_inset
-Error in the height at the trough scaled with resolution
+Error in the height at the trough with the high resolution error scaled
+ by 4 and the medium resolution error scaled by 2.
\end_layout
\end_inset
Modified: long/3D/Gale/trunk/documentation/images/tibet_setup.eps
===================================================================
--- long/3D/Gale/trunk/documentation/images/tibet_setup.eps 2008-03-19 00:58:41 UTC (rev 11471)
+++ long/3D/Gale/trunk/documentation/images/tibet_setup.eps 2008-03-19 08:06:11 UTC (rev 11472)
@@ -1,9 +1,9 @@
%!PS-Adobe-2.0 EPSF-2.0
%%Title: tibet_setup.fig
%%Creator: fig2dev Version 3.2 Patchlevel 5
-%%CreationDate: Mon Mar 10 11:16:12 2008
+%%CreationDate: Tue Mar 18 15:49:40 2008
%%For: boo at earth (Walter Landry,,,)
-%%BoundingBox: 0 0 341 117
+%%BoundingBox: 0 0 356 141
%Magnification: 1.0000
%%EndComments
%%BeginProlog
@@ -433,8 +433,8 @@
/pageheader {
save
-newpath 0 117 moveto 0 0 lineto 341 0 lineto 341 117 lineto closepath clip newpath
--14.1 272.1 translate
+newpath 0 141 moveto 0 0 lineto 356 0 lineto 356 141 lineto closepath clip newpath
+-14.1 295.7 translate
1 -1 scale
$F2psBegin
10 setmiterlimit
@@ -515,12 +515,100 @@
n 450 2700 m 225 2700 l 225 4050 l 450 4050 l
cp gs /PC [[1.00 1.00 1.00] [0.00 0.00 0.00]] def
15.00 15.00 sc P5 [16 0 0 -16 15.00 180.00] PATmp PATsp ef gr PATusp
+% Polyline
+7.500 slw
+gs clippath
+2730 2657 m 2730 2505 l 2670 2505 l 2670 2657 l 2670 2657 l 2700 2537 l 2730 2657 l cp
+eoclip
+n 2700 2790 m
+ 2700 2520 l gs col0 s gr gr
+
+% arrowhead
+n 2730 2657 m 2700 2537 l 2670 2657 l col0 s
+% Polyline
+gs clippath
+2670 3193 m 2670 3345 l 2730 3345 l 2730 3193 l 2730 3193 l 2700 3313 l 2670 3193 l cp
+eoclip
+n 2700 3060 m
+ 2700 3330 l gs col0 s gr gr
+
+% arrowhead
+n 2670 3193 m 2700 3313 l 2730 3193 l col0 s
+% Polyline
+gs clippath
+4350 3287 m 4350 3135 l 4290 3135 l 4290 3287 l 4290 3287 l 4320 3167 l 4350 3287 l cp
+eoclip
+n 4320 3510 m
+ 4320 3150 l gs col0 s gr gr
+
+% arrowhead
+n 4350 3287 m 4320 3167 l 4290 3287 l col0 s
+% Polyline
+gs clippath
+4290 3913 m 4290 4065 l 4350 4065 l 4350 3913 l 4350 3913 l 4320 4033 l 4290 3913 l cp
+eoclip
+n 4320 3735 m
+ 4320 4050 l gs col0 s gr gr
+
+% arrowhead
+n 4290 3913 m 4320 4033 l 4350 3913 l col0 s
+% Polyline
+gs clippath
+4350 2736 m 4350 2685 l 4290 2685 l 4290 2736 l 4290 2736 l 4320 2706 l 4350 2736 l cp
+eoclip
+n 4320 2835 m
+ 4320 2700 l gs col0 s gr gr
+
+% arrowhead
+n 4350 2736 m 4320 2706 l 4290 2736 l col0 s
+% Polyline
+gs clippath
+4290 3114 m 4290 3165 l 4350 3165 l 4350 3114 l 4350 3114 l 4320 3144 l 4290 3114 l cp
+eoclip
+n 4320 3015 m
+ 4320 3150 l gs col0 s gr gr
+
+% arrowhead
+n 4290 3114 m 4320 3144 l 4350 3114 l col0 s
+% Polyline
+gs clippath
+587 4560 m 435 4560 l 435 4620 l 587 4620 l 587 4620 l 467 4590 l 587 4560 l cp
+eoclip
+n 2160 4590 m
+ 450 4590 l gs col0 s gr gr
+
+% arrowhead
+n 587 4560 m 467 4590 l 587 4620 l col0 s
+% Polyline
+gs clippath
+4813 4620 m 4965 4620 l 4965 4560 l 4813 4560 l 4813 4560 l 4933 4590 l 4813 4620 l cp
+eoclip
+n 3060 4590 m
+ 4950 4590 l gs col0 s gr gr
+
+% arrowhead
+n 4813 4620 m 4933 4590 l 4813 4560 l col0 s
+/Times-Roman ff 190.50 scf sf
+5040 3780 m
+gs 1 -1 sc (1 cm/year) col0 sh gr
/Times-Roman ff 317.50 scf sf
-2250 2925 m
+900 3060 m
gs 1 -1 sc (Crust) col0 sh gr
/Times-Roman ff 317.50 scf sf
-2250 3825 m
+900 3780 m
gs 1 -1 sc (Mantle) col0 sh gr
+/Times-Roman ff 190.50 scf sf
+2430 3015 m
+gs 1 -1 sc (52 km) col0 sh gr
+/Times-Roman ff 190.50 scf sf
+4095 3015 m
+gs 1 -1 sc (32 km) col0 sh gr
+/Times-Roman ff 190.50 scf sf
+4095 3690 m
+gs 1 -1 sc (68 km) col0 sh gr
+/Times-Roman ff 190.50 scf sf
+2250 4680 m
+gs 1 -1 sc (1000 km) col0 sh gr
% here ends figure;
pagefooter
showpage
Modified: long/3D/Gale/trunk/documentation/images/tibet_setup.fig
===================================================================
--- long/3D/Gale/trunk/documentation/images/tibet_setup.fig 2008-03-19 00:58:41 UTC (rev 11471)
+++ long/3D/Gale/trunk/documentation/images/tibet_setup.fig 2008-03-19 08:06:11 UTC (rev 11472)
@@ -32,5 +32,34 @@
4950 3375 5625 3375
2 2 0 0 0 7 50 -1 45 0.000 0 0 -1 0 0 5
450 2700 225 2700 225 4050 450 4050 450 2700
-4 0 0 50 -1 0 20 0.0000 4 225 720 2250 2925 Crust\001
-4 0 0 50 -1 0 20 0.0000 4 225 945 2250 3825 Mantle\001
+2 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 1 0 2
+ 0 0 1.00 60.00 120.00
+ 2700 2790 2700 2520
+2 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 1 0 2
+ 0 0 1.00 60.00 120.00
+ 2700 3060 2700 3330
+2 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 1 0 2
+ 0 0 1.00 60.00 120.00
+ 4320 3510 4320 3150
+2 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 1 0 2
+ 0 0 1.00 60.00 120.00
+ 4320 3735 4320 4050
+2 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 1 0 2
+ 0 0 1.00 60.00 30.00
+ 4320 2835 4320 2700
+2 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 1 0 2
+ 0 0 1.00 60.00 30.00
+ 4320 3015 4320 3150
+2 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 1 0 2
+ 0 0 1.00 60.00 120.00
+ 2160 4590 450 4590
+2 1 0 1 0 7 50 -1 -1 0.000 0 0 -1 1 0 2
+ 0 0 1.00 60.00 120.00
+ 3060 4590 4950 4590
+4 0 0 50 -1 0 12 0.0000 4 180 795 5040 3780 1 cm/year\001
+4 0 0 50 -1 0 20 0.0000 4 225 720 900 3060 Crust\001
+4 0 0 50 -1 0 20 0.0000 4 225 945 900 3780 Mantle\001
+4 0 0 50 -1 0 12 0.0000 4 135 510 2430 3015 52 km\001
+4 0 0 50 -1 0 12 0.0000 4 135 510 4095 3015 32 km\001
+4 0 0 50 -1 0 12 0.0000 4 135 510 4095 3690 68 km\001
+4 0 0 50 -1 0 12 0.0000 4 135 720 2250 4680 1000 km\001
More information about the cig-commits
mailing list