[gschaider]

Originally Posted by heavy_user

just an general advise:

do NOT install cantera in /opt/ !!!!!!!!!!!!!!!!!

guess what...I did...just finished reinstalling the system..
it screws up your kernel !!

With all due respect: I don't see how it would do that on its own. I mean it could have triggered some other problem

Bernhard

[markusrehm]

Hello Nir,

you can install the packages anywhere where you have write permission. So the Thirdparty dictionary is not a bad idea. Usually you indicate this during the configuring process with the preface-option. Have a look at the Sundials installation procedure indicated in the Wiki.

Often it is easiest to install as much as possible through you package manager (yast for linux, aptget for debian).

Markus

[heavy_user]

Hey Guys,

thx for your answers!

I was pretty sure that it would work in /opt as well.
I stuck pretty much to the manual (defining $CANTERA_HOME ...) except for the location /opt/cantera-1.8 .
So after typing make the system was screwed.
So I reinstalled and cp-ed from backup i had taken after it was screwed up.
I wanted to install cantera in another place, but bevore moving it out of opt i had the great idea to "sudo make uninstall" to clean up....
Guess what...system was screwed again!
I am pretty sure that it was cantera...and not a coincident..
If you doubt, give it a try !!

Have a nice weekend!

Best!

[gschaider]

Originally Posted by heavy_user

Hey Guys,

thx for your answers!

I was pretty sure that it would work in /opt as well.
I stuck pretty much to the manual (defining $CANTERA_HOME ...) except for the location /opt/cantera-1.8 .
So after typing make the system was screwed.
So I reinstalled and cp-ed from backup i had taken after it was screwed up.
I wanted to install cantera in another place, but bevore moving it out of opt i had the great idea to "sudo make uninstall" to clean up....
Guess what...system was screwed again!
I am pretty sure that it was cantera...and not a coincident..
If you doubt, give it a try !!

Would love to, but don't have the time

But at least you admit that you compiled/installed/uninstalled as root (or sudo)? Then it would have been an overzealous script/Makefile

That is one of the reasons why when compiling/installing software yourself you should do it as a separate non-privileged user that may only write to the directory where you want to install your software. That way he can't mess up the system. On the other hand normal users can't mess up the installation

Bernhard

[KrisT]

Hi,

while running alternatereactingFoam on the dualInlet case I get

Reading from Cantera-File "/home/tegner/OpenFOAM/tegner-1.5.x/run/AlternateChemistry/Transient/dualInlet/constant/mix.cti" the mixture mix
--> FOAM Warning :
From function canteraMixture::canteraMixture
in file canteraMixture.C at line 70
The thermophysical properties of CANTERA are currently not converted to OpenFOAM. Instead the properties from "OpenFOAM/user-1.5.x/run/AlternateChemistry/Transient/dualInlet/constant/gasThermo" are used

I have seen this asked before, but I couldn't find an answer. Is this due to an error on my part?

Thanks!

[SilPaut]

Originally Posted by KrisT

Hi,

while running alternatereactingFoam on the dualInlet case I get

Reading from Cantera-File "/home/tegner/OpenFOAM/tegner-1.5.x/run/AlternateChemistry/Transient/dualInlet/constant/mix.cti" the mixture mix
--> FOAM Warning :
From function canteraMixture::canteraMixture
in file canteraMixture.C at line 70
The thermophysical properties of CANTERA are currently not converted to OpenFOAM. Instead the properties from "OpenFOAM/user-1.5.x/run/AlternateChemistry/Transient/dualInlet/constant/gasThermo" are used

I have seen this asked before, but I couldn't find an answer. Is this due to an error on my part?

Thanks!

Check the first page of this forum the post by markus on October 21, 2009. :-P
Silvano

[KrisT]

Thanks a lot!

[N. A.]

Hi Bernhard,

Can you recommend what can be done to modify the package to make it work with OpenFOAM 1.6 version.

Thanks,
Nir

[KrisT]

Hi,

I want to modify the transient alternateReactingFoam, and to that end I would like to make a copy of the massfractions.

In alternateReactingFoam the massfractions are created with the line:

PtrList<volScalarField>& Y = composition.Y();

I have vainly tried to make the copy with the following lines (and several other variants of similar type):

PtrList<volScalarField> YOld;
for(label i=0; i<Y.size(); i++) {
    volScalarField * Ytmp = & Y[i];
    YOld[i] = *Ytmp;

Obviously I'm not an expert on c++, and I would really appreciate some help on this.

Thanks!

Edited:

The following seems to be working (correct me if I'm wrong):

PtrList<volScalarField> YOld(Y.size());
for(label i=0; i<Y.size(); i++) {
    YOld.set(i,Y[i]);

[gschaider]

Originally Posted by KrisT

Edited:

The following seems to be working (correct me if I'm wrong):

PtrList<volScalarField> YOld(Y.size());
for(label i=0; i<Y.size(); i++) {
    YOld.set(i,Y[i]);

Now both PtrLists hold the same pointers.
Consequences:
a) (YOld[i]-Y[i])==0 at all times
b) Memory will be freed twice in the end

So you might want to try set(i,new volScalarField(Y[i])) instead

[KrisT]

Originally Posted by gschaider

So you might want to try set(i,new volScalarField(Y[i])) instead

Thanks a lot!

I tried both, and in my very naive test

Info<< "YOld[0].internalField()[45] " << YOld[0].internalField()[45] << "\n" << endl;
Info<< "Y[0].internalField()[45] " << Y[0].internalField()[45] << "\n" << endl;
Y[0].internalField()[45] = 0.15;
Info<< "YOld[0].internalField()[45] " << YOld[0].internalField()[45] << "\n" << endl;
Info<< "Y[0].internalField()[45] " << Y[0].internalField()[45] << "\n" << endl;

the same result were produced:

0.1
0.1
0.1
0.15

I'll stick with your formulation though (since I'm not completely convinced by this test).

Thanks again, much appreciated!

[gschaider]

Originally Posted by KrisT

Thanks a lot!

I tried both, and in my very naive test

Info<< "YOld[0].internalField()[45] " << YOld[0].internalField()[45] << "\n" << endl;
Info<< "Y[0].internalField()[45] " << Y[0].internalField()[45] << "\n" << endl;
Y[0].internalField()[45] = 0.15;
Info<< "YOld[0].internalField()[45] " << YOld[0].internalField()[45] << "\n" << endl;
Info<< "Y[0].internalField()[45] " << Y[0].internalField()[45] << "\n" << endl;

the same result were produced:

0.1
0.1
0.1
0.15

OK. That is not what I would have expected. You learn something new every day

Bernhard

[KrisT]

I'm trying to understand Cantera and how it is linked to OpenFOAM (i.e., I'm at the same time trying to understand OpenFOAM - so please excuse me if my questions are stupid). For instance in createFields.H in Transient/alternateReactingFoam:

volScalarField rho
(
    IOobject
    (
        "rho",
        runTime.timeName(),
        mesh,
        IOobject::READ_IF_PRESENT,
        IOobject::AUTO_WRITE
    ),
    thermo->rho()
);

Here I guess the last line (thermo->rho) signifies that rho is calculated by thermo (in turn defined by Cantera) if it is not supplied as initial data?

A bit further down in the same file:

volScalarField& p = thermo->p();
const volScalarField& psi = thermo->psi();
const volScalarField& T = thermo->T();

1. In the example code supplied with the solver (dualInlet), both p and T are given as initial data - but I can't find any code (similar to rho above) which specifies how they should be read?
2. Why is T declared as const?
3. What is psi?

Finally, what is the best source of information (to read) regarding issues like these?

Thanks

[gschaider]

Originally Posted by KrisT

I'm trying to understand Cantera and how it is linked to OpenFOAM (i.e., I'm at the same time trying to understand OpenFOAM - so please excuse me if my questions are stupid). For instance in createFields.H in Transient/alternateReactingFoam:

volScalarField rho
(
    IOobject
    (
        "rho",
        runTime.timeName(),
        mesh,
        IOobject::READ_IF_PRESENT,
        IOobject::AUTO_WRITE
    ),
    thermo->rho()
);

Here I guess the last line (thermo->rho) signifies that rho is calculated by thermo (in turn defined by Cantera) if it is not supplied as initial data?

A bit further down in the same file:

volScalarField& p = thermo->p();
const volScalarField& psi = thermo->psi();
const volScalarField& T = thermo->T();

1. In the example code supplied with the solver (dualInlet), both p and T are given as initial data - but I can't find any code (similar to rho above) which specifies how they should be read?
2. Why is T declared as const?
3. What is psi?

Finally, what is the best source of information (to read) regarding issues like these?

Thanks

All this is not Cantera-specific but the regular compressible solvers all work that way
@rho: Yep
@1: they are stored and maintained by the thermophysical-model
@2: because the solver is not supposed to modify it
@3: compressibility
@further information: the source code. If you find other documents please let us know

Bernhard

[KrisT]

Great, thanks!

I do understand that the source is a good place to look (as would be any decent book in C++ ;-).

Just one more question, I _guess_ that the expression

thermo->correct();

is responsible (among other things) for updating the temperature after the massfractions have been modified. But the problem here is that I don't find the souce to verify. I have looked (doing "grep correct *") under "canteraThermosChemistry" and "alternateChemistryModels" but that didn't give me anything. But I'm probably looking the wrong way/places?

Anyway, many thanks!

[alfa_8C]

Hello everybody,

I am looking for a steady state version of reactingFoam. Has anybody an idea where to get one? I'm using OF 1.7.1 where no steady state version is available.

Best
Tony

[gschaider]

Originally Posted by KrisT

Great, thanks!

I do understand that the source is a good place to look (as would be any decent book in C++ ;-).

Just one more question, I _guess_ that the expression

thermo->correct();

is responsible (among other things) for updating the temperature after the massfractions have been modified. But the problem here is that I don't find the souce to verify. I have looked (doing "grep correct *") under "canteraThermosChemistry" and "alternateChemistryModels" but that didn't give me anything. But I'm probably looking the wrong way/places?

Anyway, many thanks!

No. That'll be in the $FOAM_SRC/thermoPhysical. Although that in turn calls the Mixture-classes that are in the canter-Lib

[alfa_8C]

Dear Bernhard,

a short question. Would it be a big deal to compile alternateSteadyReactingFoam without those additional capabilites related to cantera, sundials, blas ect.? I need only a steady alternative to reactingFoam. If so, what do I have to pay special attention to?

Best Regards
Tony

[markusrehm]

Hello Tony,

this should be possible but problems arise in the details. Just take the reactingFoam, change the PISO loop to a SIMPLE one as done in the alternateSteady solver and adapt the chemistry.H for steady-state. This was done in the alternate solver with the help of a cell residence time. This is a very very simple approach but you need that value because deltaT in steady-state is arbitrary.

The next issue might occur with the ODE solver. All this is explained in the Berlin Presentation cited in the very first posting.

Markus

[alfa_8C]

Hello Markus,

thanks a lot for your quick reply.

I'll try this, following your hint,

Tony