[markc]

Hi All,

I am trying to modify interFoam solver to my needs.
It is necessary for the new solver to perform an initialisation step (include"initialise.H"). Here, from time = 0 the p field is needed to perform some tasks.
Now my question is: how to make it always reading this time =0?
I tried many things:
In initialise.H add:
//volScalarField p
(
IOobject
(
"p",
runTime.timeName(0),
mesh,
IOobject::MUST_READ,
IOobject::NO_WRITE
)
);

This compiles but when running it reads from the startTime stated in controldict anyway.
Uncommenting the line volScalarField p gives compiler error: p already declared. That is correct, this is done with createrFields.

I moved around with the location in the main solver of the initialisation step, without succes.

I think it is really something trivial, however, not for me apparently.
A more general question (as I have similar questions more often): anyone has advice as where to find some good basic C-programming reading, except for the user manal?

Thanks in advance,

Mark

[wenterodt]

Hi Mark,

isn't the pressure in interFoam called pd? This field is created in createFields.H, you can initialize it right there.

Also when you specify IOobject::MUST_READ, the solver will attempt to read the field upon start. Maybe the combination IOobject::NO_READ, IOobject::AUTO_WRITE is what you are looking for?

Regards,

Tammo

[markc]

Hello Tammo,

Thanks for your reply. Yes, you are right, pd is usualy used in interFoam. However I created an extra field p which contains hydrostatic pressure. However, this does not matter here. I initialised this in createFields.H as with other fields.
The problem is simply that I can only read it once and that it always uses the startTime from the controlDict, as usual. However I want to read one field (p) one more time but at a different time (always "0" instead of "startTime"). I feel a bit dump about it because I suspect it is really something simple for experienced OF-ers.

Brgds,

Mark

[wenterodt]

How about copying your p-file to "constant" and then use runTime.constant()?

[markc]

Hello Tammo,

I think I found the problem.
The code snippet I posted in the 1st message does not contain the word mesh just before the final bracket.
Now I added this word and it seems to work.
The structure now is:
solver.C
int main(int etc etc)
-
-
#include "createFields.H"
etc etc
#include "initialise.H" (extra header file)
This header file contains amongst others:
(
IOobject
(
"p",
runTime.timeName(0),
mesh,
IOobject::MUST_READ,
IOobject::NO_WRITE
),
mesh
);

back to solver.C
and further.

It seems to work.
Thanks for your advice,

Mark

[markc]

Recall: no it did not work. In this setup "initialise.H" still seems to work with the timestep which is stated as "startTime" instead of "0".
We continue trying...
Mark

[wenterodt]

1. Did you give the runTime.constant() idea a shot?
2. how about (this is not pretty but might work) creating a field pConst with IOobject::AUTO_WRITE, from which you make a copy "p" (which you can alter as you like) whenever you need it ? The solver should then always find a pConst in the actual time-directory.

Any good?

[markc]

Hello Tammo,

1. Not yet, will try asap.
2. Will keep that as emergency exit.

Thanks for your advice,

Mark

[markc]

Hello All,

This did the trick:
>>>
# include "addTimeOptions.H"
# include "setRootCase.H"
# include "createTime.H"

// Get times list
instantList Times = runTime.times();

// set startTime and endTime depending on -time and -latestTime options
# include "checkTimeOptions.H"

runTime.setTime(Times[startTime], 0);
# include "createMesh.H"
# include "createFields.H"
createfields e.g.:
>>>
Info<< "Reading field pd\n" << endl;
volScalarField pd
(
IOobject
(
"pd",
runTime.timeName(),
mesh,
IOobject::MUST_READ,
IOobject::AUTO_WRITE
),
mesh
);
<<<
<<<
Brgds,

Mark