Neumann Condition on Curved Boundary using Finite Difference

Sashankh · February 28, 2011, 01:08

Hi,

I am trying the solve the Poisson equation in a domain with curved boundaries using the Finite Difference Method (second order accurate). I need to apply the Neumann condition on the curved boundary. I have used bilinear interpolation to do this but this causes the resultant scheme to be only first order accurate. Could someone please tell me how one may apply the Neumann condition on a curved boundary with second order accuracy? Any links to material would be greatly appreciated!

Thanks

DoHander · February 28, 2011, 08:39

Hello,

you can find a second order FD scheme (left or right) in any good CFD book. Check the book of Ferziger and Peric, especially chapter 7. The idea is to fit a 3rd order polynomial through 3 points of your grid (first point on the boundary and the next 2 inside your mesh).

Do

Sashankh · February 28, 2011, 16:48

Thanks. I did try using two different interpolations:

w = ax^2 + bxy + cy^2 + dx + ey + f

and w = ax + bxy + cy + d (bilinear interpolation)

They both work but cause my scheme to converge at 'first order' though my finite difference scheme is of second order.

I will try what you just said and see if it works.

February 28, 2011, 01:08	Neumann Condition on Curved Boundary using Finite Difference	#1
Sashankh New Member Join Date: Mar 2009 Posts: 16 Rep Power: 17	Hi, I am trying the solve the Poisson equation in a domain with curved boundaries using the Finite Difference Method (second order accurate). I need to apply the Neumann condition on the curved boundary. I have used bilinear interpolation to do this but this causes the resultant scheme to be only first order accurate. Could someone please tell me how one may apply the Neumann condition on a curved boundary with second order accuracy? Any links to material would be greatly appreciated! Thanks

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Neumann Pressure Boundary Condition with UDF	Balder	Fluent UDF and Scheme Programming	8	July 6, 2012 11:21
Airfoil boundary condition	Frank	Main CFD Forum	1	April 21, 2008 18:36
[Commercial meshers] Trimmed cell and embedded refinement mesh conversion issues	michele	OpenFOAM Meshing & Mesh Conversion	2	July 15, 2005 04:15
Finite difference and Periodic Boundary Condition	Linda	Main CFD Forum	2	May 27, 2003 10:52
The Boundary Condition about the Flat Plate	boing	Main CFD Forum	1	January 6, 2002 16:53

February 28, 2011, 08:39		#2
DoHander Senior Member Join Date: Nov 2009 Posts: 411 Rep Power: 19	Hello, you can find a second order FD scheme (left or right) in any good CFD book. Check the book of Ferziger and Peric, especially chapter 7. The idea is to fit a 3rd order polynomial through 3 points of your grid (first point on the boundary and the next 2 inside your mesh). Do

February 28, 2011, 16:48		#3
Sashankh New Member Join Date: Mar 2009 Posts: 16 Rep Power: 17	Thanks. I did try using two different interpolations: w = ax^2 + bxy + cy^2 + dx + ey + f and w = ax + bxy + cy + d (bilinear interpolation) They both work but cause my scheme to converge at 'first order' though my finite difference scheme is of second order. I will try what you just said and see if it works.