[TemC]

Good morning foamers,

I'am working with simpleFoam, and I just implemented a new viscosity model, based on the so called "regularization of Papanastasiou". In order to do that, I have adapted the Herschel-Bulkley model already available in OpenFoam.

After that, I succesfully managed to compile the necessary library, so that my new viscosity model is now valid and available.

Now when I try to run a case with my new viscosity model, I always got an error message ("attachment 1") that I don't understand...

That's the reason why I come here to ask if anybody has already encountered this type of error while running a case.

May it be linked to the way I formulated the viscosity model? ...

- "attachment 2" is the original formulation of the Herschel-Bulkley modeil in OpenFoam;

- "attachment 3" is the formula I needed to implement ;

- "attachment 4" shows how I put it in my "newViscosityModel.C" file.

Of course any comment/suggestion will be well appreciated.

Thanks in advance for your time, and have a nice weekend.

Regards.

[agustinvo]

Hi,

for the error you get, there is a divided by zero over there. Are you sure that for the first iterations, you get a non zero diffusivity?

[TemC]

Morning Sir,

First of all thank you for your quick reply.
The initial formulation of the Herschel-Bulkley model is the following :

// * * * * * * * * * * * * Private Member Functions * * * * * * * * * * * *

Foam::tmp<Foam::volScalarField>
Foam::viscosityModels::HerschelBulkleySimple::calc Nu() const
{
dimensionedScalar tone("tone", dimTime, 1.0);
dimensionedScalar rtone("rtone", dimless/dimTime, 1.0);
tmp<volScalarField> sr(strainRate());

return
(
min
(
nu0_ ,
( tau0_ + k_* rtone * ( pow(tone * sr(), n_)) ) / (max(sr(), dimensionedScalar ("VSMALL", dimless/dimTime, VSMALL)))
)
);
}

In this formulation, the denominator "max(...)" is supposed to avoid the division by zero, if the strain rate "sr()" is too small or equal to zero.
That's why I picked up this tip in order to build the formulation I need.

Now the difference is that, in the formulation I have to implement, there is no parameter "nu0_" (since the formula is supposed to work whatever the value of the strain rate). This implies that in my "return" block, I don't have a sub-block "min(nu0_, apparentViscosity).

Regarding your remark, I'am now wondering if I shouldn't write the expression of my apparentViscosity (attachment 4) differently...

Regards.

[TemC]

Hi foamers,

Hope you had a good weekend. I'am still stucked with the difficulty to use a different version of the original Herschel-Bulkley model in OpenFoam.

By trying different equations for the apparent viscosity, I found out that whatever the expression I use, I got the same error message when I launch a simulation. Here are some examples of some expressions that I use ( in the file "newHerschelBulkley.C" ) :

a).

return
(
k_*rtone*(pow(tone*sr(),n_))
/ max(sr(), dimensionedScalar ("VSMALL", dimless/dimTime, VSMALL))
);
}

b).

return
(
(
(k_*rtone* ( pow(tone*sr(),n_)))/ (max(sr(), dimensionedScalar ("VSMALL", dimless/dimTime, VSMALL)))
+ (1.0 - exp(-m_*sr()*tone))*((tau0_) / (max(sr(), dimensionedScalar("VSMALL", dimless/dimTime, VSMALL))))
)
);
}

c).

return
(
min
(
(m_*tone*tau0_),
((1.0-exp(-1.0*m_*tone*sr()))*tau0_ + k_*rtone*pow(tone*sr(), n_))
/(max(sr(), dimensionedScalar ("VSMALL", dimless/dimTime, VSMALL)))
)
);
}

Only the initial formulation of the model seems to work, which is :

return
(
min
(
nu0_,
(tau0_ + k_*rtone*pow(tone*sr(), n_))
/(max(sr(), dimensionedScalar ("VSMALL", dimless/dimTime, VSMALL)))
)
);
}

And even a slightly modification of this formulation leads to the same error message. Does anybody have a comment/suggestion about that?

PS: I compile all my modified models within my "run" directory, I don't touch anything in the source code ( "src/transportModels/..." )

Thanks in advance, and have a nice week.

Regards.

[agustinvo]

Hello,and what do you have in your transportProperties? If the equations are ok, the input is the problem.

[TemC]

Morning Sir,

Thanks for your reply.
I send in attachments my "constant/transportProperties" file. As you can see, I always take the same parameters for all the viscosity models. With the original Hershel-Bulkley model, the simulation runs very well with those parameters. But once I pick another model, the simulation fails.

PS: Each time I attempt to use a new viscosity model, I make sure to include the right library into the file system/controdict.

Another thing that I noticed is is that when I use the regularization parameter "m_", even when I change its value (300, 600, 1000 for example), I still got the same error...

Regards.

[TemC]

Hi foamers,

Just to say that I found out what I was doing wrong:

Each time I copied the original Herschel-Bulkley model into my run directory, after making the required modifications, I used to execute the command "wclean" before creating the library with "wmake libso".

I realized that without using the "wclean" command, by creating the library right after modifying the files, I don't get the error while launching the simulation with the new viscosity model.

Thanks again for your comments/suggestions.

Regards.