dolfin team mailing list archive

[Chong Luo <luo.chong@xxxxxxxxx>]

To answer my own question, this can be done by adding a function v which is f(x,y)-g(x,y) on \Gamma, and zero elsewhere, where g(x,y) was the original boundary function. 

Chong

--- On Thu, 3/24/11, Chong Luo <luo.chong@xxxxxxxxx> wrote:

> From: Chong Luo <luo.chong@xxxxxxxxx>
> Subject: Modify function on a given boundary
> To: dolfin@xxxxxxxxxxxxxxxxxxx
> Cc: luo.chong@xxxxxxxxx
> Date: Thursday, March 24, 2011, 12:12 PM
> Suppose we want to change a function
> u on the \Gamma part of the boundary to a given function
> u=f(x,y), and keep the function u unchanged at the rest of
> the mesh nodes. 
> 
> Is there a simple way to do this? 
> 
> Thanks!
> 
> 
>       
>