changing the order of constructing IC, BC and ST

Norihiro_Watanabe2 · May 29, 2012, 4:14pm

Hi developers,

I'd like to hear your opinions about whether my change causes any problems in OGS or not.

In current codes, ICs are constructed after BCs and STs are constructed (in CRFProcess::Create()). Now I'd like to reverse this order (i.e. first ICs, then BCs and STs) to construct BC data based on IC.

I hope this change doesn't harm existing functionality but I'm not really sure. so before running all benchmarks, I'd like to ask you if there is any case that this change cause troubles.

Thank you in advance,
Norihiro

···

--
Norihiro Watanabe
Department of Environmental Informatics

Helmholtz-Zentrum fï¿½r Umweltforschung GmbH - UFZ
Helmholtz Centre for Environmental Research GmbH - UFZ
Permoserstraï¿½e 15 / 04318 Leipzig / Germany
Telefon +49 341 235 1890
norihiro.watanabe@ufz.de / www.ufz.de

Sitz der Gesellschaft: Leipzig
Registergericht: Amtsgericht Leipzig, Handelsregister Nr. B 4703
Vorsitzender des Aufsichtsrats: MinDirig Wilfried Kraus
Wissenschaftlicher Geschï¿½ftsfï¿½hrer: Prof. Dr. Georg Teutsch
Administrativer Geschï¿½ftsfï¿½hrer: N.N.

Cmcdermo · May 29, 2012, 4:23pm

Hi Norihiro

Can you justify more why you want to do this please so we can understand.

The BC's are scaled into the global matrices and account is taken of the fact whether they are Neuman or Dirchlet, so there might be a problem here, i.e. they are not just added on top..at least that is an experience I have made...Not sure if it is still so.

I would advise making sure you write out the equation systems and checking that they are doing what you want them to do before submitting a change like that.

Best regards
Chris McDermott

···

--------------------
Dr. habil. C. I. McDermott M.Sc.
Senior Lecturer Hydrogeology and Coupled Process Modelling
School of Geosciences
Edinburgh University
Tel. ++44 (0) 131 6505931
Mob: ++44 (0) 7583431065

Norihiro Watanabe wrote:

Hi developers,

I'd like to hear your opinions about whether my change causes any problems in OGS or not.

In current codes, ICs are constructed after BCs and STs are constructed (in CRFProcess::Create()). Now I'd like to reverse this order (i.e. first ICs, then BCs and STs) to construct BC data based on IC.

I hope this change doesn't harm existing functionality but I'm not really sure. so before running all benchmarks, I'd like to ask you if there is any case that this change cause troubles.

Thank you in advance,
Norihiro

--
The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.

Norihiro_Watanabe2 · May 29, 2012, 5:14pm

Hi Chris,

Thank you for your replying.

Yes I can explain you more why I want to do this. Our situation is that we have to run simulations using other simulator's result as initial conditions. In addition, we need to use parts of the result for Dirichlet boundary conditions. To do this, I'd like to introduce the new distribution type "INITIAL" in BC file. This keyword can be used to keep values of nodes located on given polylines or surfaces be same as their initial values.

To implement the above feature, I changed CBoundaryConditionsGroup::Set() so that it constructs Dirichlet BC data (i.e. CBoundaryConditionNode) based on initial values at given nodes via calling CRFProcess::GetNodeValue(). In order to let CRFProcess::GetNodeValue() returns initial values, I need to setup IC beforehand. That's why I would like to reverse the order of constructing IC and BC. I think this modification doesn't affect how to apply IC, BC and ST into equation systems.

Best,
Norihiro

···

On 29.05.2012 18:23, Christopher McDermott wrote:

Hi Norihiro

Can you justify more why you want to do this please so we can understand.

The BC's are scaled into the global matrices and account is taken of the
fact whether they are Neuman or Dirchlet, so there might be a problem
here, i.e. they are not just added on top..at least that is an
experience I have made...Not sure if it is still so.

I would advise making sure you write out the equation systems and
checking that they are doing what you want them to do before submitting
a change like that.

Best regards
Chris McDermott

--------------------
Dr. habil. C. I. McDermott M.Sc.
Senior Lecturer Hydrogeology and Coupled Process Modelling
School of Geosciences
Edinburgh University
Tel. ++44 (0) 131 6505931
Mob: ++44 (0) 7583431065
http://www.geos.ed.ac.uk/homes/cmcdermo

Norihiro Watanabe wrote:

Hi developers,

I'd like to hear your opinions about whether my change causes any
problems in OGS or not.

In current codes, ICs are constructed after BCs and STs are
constructed (in CRFProcess::Create()). Now I'd like to reverse this
order (i.e. first ICs, then BCs and STs) to construct BC data based on
IC.

I hope this change doesn't harm existing functionality but I'm not
really sure. so before running all benchmarks, I'd like to ask you if
there is any case that this change cause troubles.

Thank you in advance,
Norihiro

--
Norihiro Watanabe
Department of Environmental Informatics

Helmholtz-Zentrum fï¿½r Umweltforschung GmbH - UFZ
Helmholtz Centre for Environmental Research GmbH - UFZ
Permoserstraï¿½e 15 / 04318 Leipzig / Germany
Telefon +49 341 235 1890
norihiro.watanabe@ufz.de / www.ufz.de

Sitz der Gesellschaft: Leipzig
Registergericht: Amtsgericht Leipzig, Handelsregister Nr. B 4703
Vorsitzender des Aufsichtsrats: MinDirig Wilfried Kraus
Wissenschaftlicher Geschï¿½ftsfï¿½hrer: Prof. Dr. Georg Teutsch
Administrativer Geschï¿½ftsfï¿½hrer: N.N.