Difference between grad and div on boundaries?

tschenkel · August 8, 2022, 09:28

Hi,

I am trying to implement a source term for a transport equation (shear stress driven particle migration).

The original term is implemented using fvc::laplacian, but I need to move to an anisotropic model (multiply with a nonisotropic diagonal matrix).

I have implemented the following:

sourceCorig = fvc::laplacian(linKc*Kc*sqr(a)*H, H*gammaDot);
sourceCHgamma = fvc::div(linKc*Kc*sqr(a)*H*fvc::grad(H*gammaDot));
sourceCHgammaZ = fvc::div(linKc*Kc*sqr(a)*H*fvc::div(H*gammaDot*Z)) ;

where Z is for the moment the identity matrix.

If my maths is right, these should all be identical.

However, while there are some discrepencies between laplacian() and div(grad()), the div(div( *I) does not behave at the wall (right side in the graph).

Am I missing something here? Do I need to treat the wall separately?

Tibo99 · August 22, 2022, 14:07

Maybe this could help?

discrepancy in fvc::grad(U)

tschenkel · August 24, 2022, 04:55

Quote:

Originally Posted by Tibo99

Maybe this could help?

discrepancy in fvc::grad(U)

Thanks. I'll need to look into that. But it seems to be something different (to do with BCs?).

Tibo99 · August 24, 2022, 08:26

Because you use the Laplacian and grad in the calculation of the BC, I just pointed this thing out in case that could may affect the results....

I didn't want to lead you toward something that could waste your energy in the wrong path....

It's really subtle and people using openfoam may be surprised when they find out about this information.

Good luck with your work!

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
upwind scheme for grad and div	Yujuan	OpenFOAM Running, Solving & CFD	0	June 26, 2020 04:55
A simple question about adding Div and Grad of a source term to the momentum equation	adkar	Main CFD Forum	5	May 23, 2016 17:31
Numerical schemes for grad and div	Tobi	OpenFOAM	6	March 11, 2013 21:24
calclulate div and grad just on specified cells -Local Time stepping	mm.abdollahzadeh	OpenFOAM Programming & Development	0	January 25, 2013 06:32
Possible Incompressible Turbulence Model Bugs	gocarts	OpenFOAM Bugs	12	November 26, 2009 14:58

August 22, 2022, 14:07		#2
Tibo99 Senior Member René Thibault Join Date: Dec 2019 Location: Canada Posts: 114 Rep Power: 6	Maybe this could help? discrepancy in fvc::grad(U)

August 24, 2022, 08:26		#4
Tibo99 Senior Member René Thibault Join Date: Dec 2019 Location: Canada Posts: 114 Rep Power: 6	Because you use the Laplacian and grad in the calculation of the BC, I just pointed this thing out in case that could may affect the results.... I didn't want to lead you toward something that could waste your energy in the wrong path.... It's really subtle and people using openfoam may be surprised when they find out about this information. Good luck with your work!