[aspect-devel] Fwd: problem with ASPECT - the iterative Stokes solver did not converge

[Wolfgang Bangerth]

-------- Original Message --------
Subject: 	problem with ASPECT - the iterative Stokes solver did not converge
Date: 	Thu, 24 Jul 2014 07:10:18 -0400
From: 	Petar Glisovic <pglisovic at gmail.com>
To: 	Wolfgang Bangerth <bangerth at math.tamu.edu>
CC: 	Alessandro Forte <alessandro.forte at tomo.ig.erdw.ethz.ch>



Dear Wolfgang,

My name is Petar Glisovic and I am currently a postdoctoral researcher
working with Alessandro Forte on numerical modelling of convection in
the Earth's mantle.

I have experienced a problem running ASPECT. It seems that the iterative
Stokes solver did not converge after three days of the job running on 48
CPUs - please see attached log files (/aspect.e2477653, aspect.o2477653
& log.txt/).

It also seems, after a numerous of intensive tests, that the critical
subsection is Material model. Namely, if I use the option with simple
model as follows:
subsection Material model
    set Model name = simple
    subsection Simple model
        set Thermal expansion coefficient = 4e-5
        set Viscosity = 1e22Â
    end
end
everything is OK - the code runs without problems.

However, my goal is to use the depth-dependent (i.e., radial) viscosity
profile (Mitrovica & Forte, 2004; please see attached file:
/mitro-forte-2004-radial-visc-new.txt/) jointly with lateral variations
in viscosity (see below /shell_simple_3d_new.prm/). For example, to use
something similar to the Steinberger model:
subsection Material model
    set Model name = Steinberger
    subsection Steinberger model
        set Bilinear interpolation = true
        set Compressible = true
        set Data directory =
/RQusagers/glisovic/aspect/data/material-model/steinberger/
        set Latent heat = false
        set Lateral viscosity file name =
temp-viscosity-prefactor.txt
        set Material file names = pyr-ringwood88.txt
        set Radial viscosity file name =
mitro-forte-2004-radial-visc-new.txt # ***iso-visc.txt
        set Reference viscosity = 1e22  #1e21
    end
end
** *Please note that I have tried a constant viscosity profile
(iso-visc.txt) but without success - there is no convergence in the
iterative Stokes solver.

Please could you help me to solve this convergence problem with ASPECT?

Thank you very much for your help and time.

Best regards,
Petar

-- 
Petar Glisovic, Dr.
postdoctoral fellow*

*
*address*: GEOTOP, Université du Québec à Montréal, CP 8888,
succursale Centre-Ville, Montréal,Québec, Canada H3C 3P8 *|
ema**il**:Â * pglisovic at gmail.com <mailto:pglisovic at gmail.com>Â &Â
glisovic.petar at courrier.uqam.ca
<mailto:glisovic.petar at courrier.uqam.ca>*Â Â |phone:Â Â *1-514-987-3000
#3592 & #4600*Â Â | fax:Â Â *1-514-987-3635


-------------- next part --------------
-----------------------------------------------------------------------------
-- This is ASPECT, the Advanced Solver for Problems in Earth's ConvecTion.
--     . version 1.2.pre
--     . running in DEBUG mode
--     . running with 48 MPI processes
--     . using Trilinos
-----------------------------------------------------------------------------

Number of active cells: 6144 (on 3 levels)
Number of degrees of freedom: 215962 (156774+6930+52258)

*** Timestep 0:  t=0 years
   Solving temperature system... 0 iterations.
   Rebuilding Stokes preconditioner...
   Solving Stokes system... 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: shell_simple_3d_new.prm
Type: application/octet-stream
Size: 3257 bytes
Desc: not available
URL: <http://lists.geodynamics.org/pipermail/aspect-devel/attachments/20140725/f04d3dce/attachment-0004.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: parameters.tex
Type: application/x-tex
Size: 220492 bytes
Desc: not available
URL: <http://lists.geodynamics.org/pipermail/aspect-devel/attachments/20140725/f04d3dce/attachment-0001.tex>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: parameters.prm
Type: application/octet-stream
Size: 134553 bytes
Desc: not available
URL: <http://lists.geodynamics.org/pipermail/aspect-devel/attachments/20140725/f04d3dce/attachment-0005.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: aspect.o2477653
Type: application/octet-stream
Size: 1443 bytes
Desc: not available
URL: <http://lists.geodynamics.org/pipermail/aspect-devel/attachments/20140725/f04d3dce/attachment-0006.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: aspect.e2477653
Type: application/octet-stream
Size: 1896 bytes
Desc: not available
URL: <http://lists.geodynamics.org/pipermail/aspect-devel/attachments/20140725/f04d3dce/attachment-0007.obj>
-------------- next part --------------
	5.36043453e+21	0.0000000e+00
	5.36043453e+21	7.7000000e+01
	5.36043453e+21	2.1700000e+02
	5.36043453e+21	3.9700000e+02
	5.36043453e+21	5.9700000e+02
	5.36043453e+21	6.3200000e+02
	5.36043453e+21	6.6700000e+02
	5.36043453e+21	8.4700000e+02
	5.36043453e+21	9.9700000e+02
	5.36043453e+21	1.1970000e+03
	5.36043453e+21	1.3970000e+03
	5.36043453e+21	1.5970000e+03
	5.36043453e+21	1.7970000e+03
	5.36043453e+21	1.9970000e+03
	5.36043453e+21	2.1970000e+03
	5.36043453e+21	2.3970000e+03
	5.36043453e+21	2.5970000e+03
	5.36043453e+21	2.7970000e+03
	5.36043453e+21	2.8880000e+03
-------------- next part --------------
	1.0000000e+22	0.0000000e+00
	1.0000000e+21	7.7000000e+01
	1.0000000e+20	2.1700000e+02
	5.0000000e+20	3.9700000e+02
	6.9767293e+20	5.9700000e+02
	3.0000000e+20	6.3200000e+02
	1.0401614e+21	6.6700000e+02
	3.3977561e+21	8.4700000e+02
	7.9951017e+21	9.9700000e+02
	1.5184593e+22	1.1970000e+03
	2.9019065e+22	1.3970000e+03
	5.6430236e+22	1.5970000e+03
	1.0000000e+23	1.7970000e+03
	5.0000000e+23	1.9970000e+03
	5.0000000e+22	2.1970000e+03
	2.0000000e+22	2.3970000e+03
	8.0000000e+21	2.5970000e+03
	2.0000000e+21	2.7970000e+03
	2.0000000e+20	2.8880000e+03