CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Main CFD Forum (http://www.cfd-online.com/Forums/main/)
-   -   Weak costraint to solve laminar NS? (http://www.cfd-online.com/Forums/main/103039-weak-costraint-solve-laminar-ns.html)

fruitkiwi June 9, 2012 02:19

Weak costraint to solve laminar NS?
 
Hi, dear all,

Recently go through some article during lunch, have your guys heard about "weak constraint"?

I heard that weak constraint can served as Dirichlet condition to solve the CFD convergence issues. it may be applied as boundary condtion to make the model converge faster.

is there any book i can refer to ?i believe this term can be explained through mathematical, but how about CFD?

Thanks.

cfdnewbie June 9, 2012 05:31

Not sure if I understood you correctly, but was it about weak and strong boundary conditions? Sounds like....

fruitkiwi June 9, 2012 05:46

Quote:

Originally Posted by cfdnewbie (Post 365554)
Not sure if I understood you correctly, but was it about weak and strong boundary conditions? Sounds like....

Thanks for cfdnewbie interest.

Because i am curious what is different if we add in weak constraint. As in CFD, we have inlet and outlet condition. We can Neumann/dirichlet condition, and after few iteration, it may help to solve the problem.

where is the application of this weak constraint term? or does it means that if i combine Neumann condition and weak constraint, if can form a good boundary condition?

cfdnewbie June 9, 2012 05:53

a strong bc formulation means that you prescribe the value of the function directly, a weak bc means that you prescribe the flux that fits to the state you would like to achieve.
Let's say you would like to enforce a no-slip wall. You can then force the noslip through the state directly (strong bc), or indirectly through the fluxes. The latter does not guarantee (for all resolutions) no slip on the wall, but it will make convergence easier.

fruitkiwi June 9, 2012 09:48

Quote:

Originally Posted by cfdnewbie (Post 365560)
a strong bc formulation means that you prescribe the value of the function directly, a weak bc means that you prescribe the flux that fits to the state you would like to achieve.
Let's say you would like to enforce a no-slip wall. You can then force the noslip through the state directly (strong bc), or indirectly through the fluxes. The latter does not guarantee (for all resolutions) no slip on the wall, but it will make convergence easier.

i get it. thanks

FMDenaro June 9, 2012 13:19

I think that the issue has a mathematical implication, prescribing the flux as BC on the boundaries requires to satisfy a compatibility constraint involving the surface integral of the flux, this remember the compatibility constraint in the pressure equation for incompressible flows.

fruitkiwi June 10, 2012 09:49

Quote:

Originally Posted by FMDenaro (Post 365610)
I think that the issue has a mathematical implication, prescribing the flux as BC on the boundaries requires to satisfy a compatibility constraint involving the surface integral of the flux, this remember the compatibility constraint in the pressure equation for incompressible flows.

Hi, FMDenaro,

thanks for your thought, however, I have some confusion.
well, I thought when you prescribe the flow rate or surface integral of the flux(flow rate) on the inlet, CFD solver automatically assign a pressure term on the inlet BC.
If you add in a compatible constraint, will it cause NS equation to be overdefined or overconstraint, and hence results in convergence issues?

i am newbie, sorry for the confusion.


All times are GMT -4. The time now is 12:41.