[NightWing]

Hello All,

I am trying to find out some non-dimensional parameters such as Reynolds no, Grashof no and Rayleigh number for my test case.

I am trying to model a flow through a glass tube through which water is flowing. I have used Boussinesq approximation for the analysis (natural convection taking place inside the tube). I have given a fixedGradient condition on the tube walls in my 0/T directory.

TubeWalls
(
type fixedGradient;
graidentValue uniform 700;
);

I approach the problem in this way as described below:
I apply a heat flux on the tube walls (gradient i mentioned above; Q/A = -k dt/dx, using this relation). due to the flux given, the water inside takes in the heat from walls and gets heated up and will begin to move upwards to the top side of the tube like the natural convection phenomenon.

Now i know after running the case., i can print out these non-dimensional numbers by making necessary changes in my solver.

1) for Grashof no , Gr = g*beta*(T-Tref)*pow(D,3)/pow(nu,2) <-- basic formulae , D is diameter of tube and rest are thermal properties of water

what i did is

volScalarField Gr
(
IOobject
(
"Gr",
runTime.timeName(),
mesh,
IOobject::NO_READ,
IOobject::AUTO_WRITE
),
mesh,
dimensionedScalar("Gr", dimensionSet(0, 0, 0, 0, 0), scalar(0))
);

Then i added the formulae to calculate Gr as

Gr = g*beta*(T-Tref)*pow(D,3)/pow(nu,2)

Note: here i had a problem, since i have g here as a vectorfield and Gr is defined as volumeScalaField. to overcome this i just assigned the value of g (9.81) in there.

Is there any alternative way to put 'g' in the equation??

Suppose if i want to find Re like the Gr way i followed.
Re should require U value. In this case U is vectorField and my Re is volScalar. Hence it will conflict. Is there any suggestions?


Note: Is the procedure i followed to compute my Gr is ok or is it the right way?

Now I want to calculate Reynolds number. I have some old thread in which people suggested that, it is possible to get Reynolds number from Paraview. May i know, how it is done. I tried reading the user guide but i couldnt get a hand on it.

By using Paraview, Calculator option, i tired to calculate Re and I used cell point and entered the formulae like

(U*D)/(nu); ( i used D and nu as the constant value which i inputed directl)

It gave a Re value which is like 1400, where Re based on my velocity is 1000.

To calculate Re, Should i integrate something before using the formulae directly?

[NightWing]

Based on it, when i compute the Gr value, I get combination of positive and negative values and its highly random values.

What may be the reason?

[nimasam]

Hi friend

Is there any alternative way to put 'g' in the equation??

in the grashof number, g represents the value of gravity acceleration, so you can use mag(g) in your code, for Reynold number use the same method for U

[NightWing]

Thank you Nima

The procedure what i followed, is it correct?

And I suppose to us magU <-- I should first use foamCalc to find magU? is it that way?

[nimasam]

it depends

if you want to calculate Re as a post processing procedure. then you should
1-foamCalc mag U
2-open paraview
3- use filter calculator

but also you can add some extra pieces of code in your solver then it write Re in time folders for you. for code you should use mag(U)

[NightWing]

Yes, Indeed. Right now I am trying to compute the Grashof no and Re based on a piece of code written in the solver.

I have defined a separate header file as parameterCalc.H. In that I have defined Gr and Re as

Gr = (mag(g)*beta*(T-Tref)*pow(L,3)/pow(nu,2);

Re = (mag(U)*Din)/nu; where Din is the diameter of the inlet.


Moreover, I have defined Re and Gr as a volScalarField (is this the correct way? like is it right to define these Non-dimensional parameters as a volScalarField?)

Info << \n Reading field Gr\n" << endl;
volScalarField Gr
(
IOobject
(
"Gr",
runTime.timeName(),
mesh,
IOobject::NO_READ,
IOobject::AUTO_WRITE
),
mesh,
dimensionedScalar("Gr", dimensionSet(0, 0, 0, 0, 0), scalar(0.0))
);

Similarly for Re in the createFields.H

But when i see my run-time files, i see wide variation of values in the results written at the write intervals like:

Gr <--- File obtained due to writing out values

dimensions [0 0 0 0 0 0]

internalField nonuniform List<scalar>
61274
(
18667.6
10940.3
2135.76
-0.518128
0.240865
...
....
......
......
-0.000886643
...
.....

like these values
What could be the possible reason for negative values ?

[nimasam]

use this expression grashof difinition

max(T-Tref,1E-30*Tref);