Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > comp.soft-sys.math.mathematica > #6873 > unrolled thread

Re: Laplace equation with gradient boundary conditions

Started byOliver Ruebenkoenig <ruebenko@wolfram.com>
First post2011-12-01 12:23 +0000
Last post2011-12-02 12:29 +0000
Articles 2 — 2 participants

Back to article view | Back to comp.soft-sys.math.mathematica

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.


Contents

  Re: Laplace equation with gradient boundary conditions Oliver Ruebenkoenig <ruebenko@wolfram.com> - 2011-12-01 12:23 +0000
    Re: Laplace equation with gradient boundary conditions Tom Wolander <ultimni@hotmail.com> - 2011-12-02 12:29 +0000

#6873 — Re: Laplace equation with gradient boundary conditions

FromOliver Ruebenkoenig <ruebenko@wolfram.com>
Date2011-12-01 12:23 +0000
SubjectRe: Laplace equation with gradient boundary conditions
Message-ID<jb7rjp$j16$1@smc.vnet.net>

On Tue, 29 Nov 2011, Tom Wolander wrote:

> I have bought Mathematica 8 a week ago and this is my first post on
> this board.
> My main purpose for the purchase was to work on PDEs, specifically on
> the heat equation.
> As one of the first tests I wanted to solve a steady state temperaure
> distribution on a rectangular domain with a radiative boundary
> condition on one face (flux=0 on the other 3). I made sure to have
> continuity in corners.
> This is a rather easy exercice of a radiating wall - I have solved
> many of similar and more complex problems "by hand" many years ago.
> Unfortunately I failed with NDSolve in Mathematica and the tutorials
> are of no help despite some 4 hours I spent in there.
> I found only one rather esotherical hint somewhere deep in one "Issue"
> section on a command which seemed to say that NDSolve could work only
> with Cauchy boundary conditions.
> If this were true, then use of Mathematica 8 would be excluded for
> virtually any work in thermics where the boundary conditions are
> always of the (non Cauchy) convection/radiation type.

Tom, could you be a little more specific in writing down the boundary
condition you want? I assume it is not a generalized Neumann condition. Is
it more like a Sommerfeld condition?

Thanks,
Oliver

> In other words the elementary steady state problem (Laplace equation)
> with flux conditions on boundaries can't be solved?
>
> It might be that this issue has been already discussed but I couldn't
> find a relevant thread by using search.
>
> Could somebody help me by answering whether Laplace equation with
> Robin like BC can't really be solved?
> And if it can be done, what have I missed to make NDSolve work?
>
>

[toc] | [next] | [standalone]


#6900

FromTom Wolander <ultimni@hotmail.com>
Date2011-12-02 12:29 +0000
Message-ID<jbagc7$229$1@smc.vnet.net>
In reply to#6873
Oliver

It is really an easy exercice of a rectangular wall (axb) isolated on
3 sides and radiating on the 4th.
In steady state f is solution of Laplace = 0 with BC  df/dx(a,y)=df/
dx(0,y)=df/dy(x,0)=0 and df/dy(x,b) = S(x) - k.[f(x,b)]^4 . Same
problem is obtained with a convective case just by replacing k.
[f(x,b)]^4 by k.f(x,b).
S(x) is the incoming radiation and can serve to adjust continuity if
that's what one wants to do.
So it has nothing to do with Sommerfeld conditions. The convective
case is just a Robin BC and the radiating case is a generalised Robin.

Andrzej

It is neither clearly nor obviously stated that NDSolve can't solve
the Laplace equation. DSolve even uses it as an example of symbolical
solution. I agree that after having spent hours and/or for somebody
familiar with resolution algorithms it becomes clear but I can assure
everybody that for a first time user it is a long and unpleasant
journey. I would have expected that in the tutorial (part dealing with
classification of PED) there would be a warning with glowing red
characters following the definitions of PED : "Beware ! NDSolve is
unable to solve most of these PED with the exception of : {list of
conditions}".

Mark

Thanks. I have come empirically to the same conclusion and you
explained me why. I have also got some results by looking at the
asymptotic behaviour of the time dependent equation. This is not
something one would do spontaneously when working by hand - the
Laplace equation in 2D is easy so there is no incentive going to the
time dependent case. I agree, will have to wait for Version 9 to deal
with the problems I wanted to deal.
Thanks for the link. Even if I knew the mathematics, I understood a
bit more about Mathematica :)

[toc] | [prev] | [standalone]


Back to top | Article view | comp.soft-sys.math.mathematica


csiph-web