CFD Online Logo CFD Online URL
www.cfd-online.com
[Sponsors]
Home > Forums

Search Results

Register Blogs Community New Posts Updated Threads Search

Showing results 1 to 15 of 15
Search took 0.00 seconds.
Search: Posts Made By: FMDenaro
Forum: Main CFD Forum May 22, 2017, 11:47
Replies: 31
Views: 2,897
Posted By FMDenaro
Ok, this way you complete the pressure matrix...

Ok, this way you complete the pressure matrix entries after convergence in the SOR is reached, right? I think that this boundary values will be used also in the correction step.
Just check the...
Forum: Main CFD Forum May 19, 2017, 14:23
Replies: 31
Views: 2,897
Posted By FMDenaro
Maybe you start you initial guess vector with...

Maybe you start you initial guess vector with zeros? However, if you get convergence to the correct solution that's not a problem but (in order to accelerate convergence) consider to save the...
Forum: Main CFD Forum May 18, 2017, 16:16
Replies: 31
Views: 2,897
Posted By FMDenaro
What do you mean for "singularity "?? You get...

What do you mean for "singularity "?? You get NaN??
Forum: Main CFD Forum May 17, 2017, 15:32
Replies: 31
Views: 2,897
Posted By FMDenaro
Well, I strongly suggest to cancel such...

Well, I strongly suggest to cancel such instruction...While it is correct that you have a solution apart from an arbitrary constant value (actually a function of time), fixing a value is not a...
Forum: Main CFD Forum May 17, 2017, 15:01
Replies: 31
Views: 2,897
Posted By FMDenaro
- Why are you prescribing P(2,2)=1?? - Then,...

- Why are you prescribing P(2,2)=1??
- Then, you compute E from the difference between p at step k and k-1 but this is not correct...you have to apply the norm on the residual of the equations
Forum: Main CFD Forum May 17, 2017, 14:02
Replies: 31
Views: 2,897
Posted By FMDenaro
I give an example of pseudo-code i=2;j=2 ...

I give an example of pseudo-code

i=2;j=2 (2 BC.s)
p(i,j)=..
j=2 (1 BC)
for i=3,nx-2
p(i,j)=...
end
i=nx-1;j=2 (2 BC.s)
p(i,j)=..
Forum: Main CFD Forum May 17, 2017, 11:47
Replies: 31
Views: 2,897
Posted By FMDenaro
I strongly suggest to use a different approach......

I strongly suggest to use a different approach... Assume i=2,nx-1 and j=2,ny-1 are the internal pressure unknowns. You have to split the cycle for nodes where the unchanged Laplace (no BC.s) matrix...
Forum: Main CFD Forum May 16, 2017, 16:44
Replies: 31
Views: 2,897
Posted By FMDenaro
That's the point! In your SOR you have to compute...

That's the point! In your SOR you have to compute only the equations for the internal unknowns and all these equations (once you substitute correctly the BC.s) have no nodes for the pressure at the...
Forum: Main CFD Forum May 16, 2017, 16:20
Replies: 31
Views: 2,897
Posted By FMDenaro
If you want to compute the pressure along the...

If you want to compute the pressure along the walls, you need to compute it explicitly from the Neumann BC after the SOR has reached convergence.
Forum: Main CFD Forum May 16, 2017, 15:53
Replies: 31
Views: 2,897
Posted By FMDenaro
1) For the lid driven cavity is correct that you...

1) For the lid driven cavity is correct that you have at corners a non-symmetric pattern, you have pressure increaing when the flow impact with the wall.
2) Depending on you formulation, you can end...
Forum: Main CFD Forum May 16, 2017, 12:45
Replies: 31
Views: 2,897
Posted By FMDenaro
I give you an example for a 1D simple case: ...

I give you an example for a 1D simple case:

d/dx(dp/dx) =(1/dt) du*/dx

with the Neumann BC (U is the known physical velocity on the boundary)
dp/dx = (1/dt) (u*- U)


Now consider that i=1...
Forum: Main CFD Forum May 12, 2017, 15:03
Replies: 31
Views: 2,897
Posted By FMDenaro
You do not have to discretize the Neumann BC.s......

You do not have to discretize the Neumann BC.s... just substitute the normal derivative into the equation for the nodes close to the frontiers, simplify the entries of the matrix and modify the...
Forum: Main CFD Forum May 12, 2017, 03:06
Replies: 31
Views: 2,897
Posted By FMDenaro
As you can see, the relation used on the BC.s is...

As you can see, the relation used on the BC.s is nothing else that the discrete momentum equation arranged in the form of the Hodge decomposition. This way, the Neumann BC involves the real velocity...
Forum: Main CFD Forum May 9, 2017, 17:31
Replies: 31
Views: 2,897
Posted By FMDenaro
The slides show the interpolation introduced to...

The slides show the interpolation introduced to manage the decoupling in the pressure grid when collocate arrangement is used. But you need to see the BC for pressure.
Forum: Main CFD Forum May 9, 2017, 16:33
Replies: 31
Views: 2,897
Posted By FMDenaro
Actually, fixing the homogeneous Neumann BC, you...

Actually, fixing the homogeneous Neumann BC, you have to correct the source term with the correct normal velocity components. Otherwise, the compatibility condition is not fulfilled.

I can suggest...
Showing results 1 to 15 of 15

 
Forum Jump

All times are GMT -4. The time now is 16:52.