[zhulianhua]

Hi! FOAMers

Does anybody know if there's a zero second derivative boundary condition already implemented in OpenFOAM ?

I wan't to let the fvc::grad() correct the gradient at the boundary face cells to the gradient of the next layer cells.


Regards,

Lianhua

[titanchao]

Hey, Lianhua,

Could you define a scalar field that is the first-order derivative of your desired function and then impose zeroGradient BC of the defined scalar field.

Best of luck!

[zhulianhua]

Hi, titanchao!

Thanks for you kindly reply.

Yes, I use the lsGrad scheme, from its source code, I can see the returned gradient field is initialized with "zeroGradient" by default.

Suppose I want get the gradient g of the scalar volume field s, I do not care about the boundary value of the g, but the it's value in the cells adjacent to the boundary faces.

The problem is: When lsGrad scheme process these cells adjacent to the boundary, it uses the values of s on the boundary faces.
However I want the boundary values of s to be linearly extrapolated form the inner cells (which is zero 2nd order B.C. I think).

Until now, I can't find how to extrapolate linearly the inner cell values of s to the boundary faces.
I found some people asked this question on this forum, but the replies say it's non-sense to do that physically!


Regards,

Lianhua

Originally Posted by titanchao

Hey, Lianhua,

Could you define a scalar field that is the first-order derivative of your desired function and then impose zeroGradient BC of the defined scalar field.

Best of luck!

[egp]

Originally Posted by zhulianhua

Hi! FOAMers

Does anybody know if there's a zero second derivative boundary condition already implemented in OpenFOAM ?

I wan't to let the fvc::grad() correct the gradient at the boundary face cells to the gradient of the next layer cells.


Regards,

Lianhua

Search before asking! While the SNR has gotten very low on the forum, there are some nuggets out there for you to find!

http://www.cfd-online.com/Forums/ope...condition.html

[zhulianhua]

Thanks, Eric!

The post in the link is what I'm looking for and really a nugget!

Sorry, I should have done a good search work.


Regards,

Lianhua