Re: [ff3d-users] Navier Stokes equation

[Stephane Del Pino]

Le Vendredi 7 Janvier 2005 22:08, Bharat Rangan a écrit :
> Hi,
> While solving the NS equations, I used the following equation for pressure
>
> -div(grad(p)) + dx(u1)*dx(u1) + dy(u2)*dy(u2) + 2*dx(u2)*dy(u1) = 0;
>
> I got the following error. unexpected '*'. Expected +,- or =. Is this some
> parser error or am I making a mistake ??
I think you did a mistake ... There are a few things that confuse me here.

If I am right, you want to solve something like
 -div(grad(p))+div(U) = 0
  where U=(u1,u2,u3). Then just write
 -div(grad(p)) = -dx(u1)-dy(u2)-dz(u3);
  If you are solving an equation on 'p', it is the only unknown, so only 'p'
  may be on the left hand side of the equation. If it was a system, you should
  of course write
 -div(grad(p))+dx(u1)+dy(u2)+dz(u3)=0;
  in ff3d, PDEs unknowns have always to be given at the left hand side and
  data at the right hand side.

I really do not understand the terms dx(u1)*dx(u1) + ...
They are non-linear terms. Could you explain them to me?

Finally, as I see, you are not using 'dz' or 'u3'. Are you trying to solve a 
2D problem? In that case, ff3d may not be the right answer, it only solves 3D 
problems up to now ...

Best regards,
Stephane.