Basilisk in 3D

[Stephane Popinet]

Dear all,

I am pleased to announce the latest release of Basilisk. Much progress
has been made during the summer holidays. The quadtree implementation
has been generalised to octrees (the same implementation now does
bitrees, quadtrees and octrees). This also includes MPI parallelism.

At the same time, most solvers have been generalised to 3D, including
Navier-Stokes and Volume-Of-Fluid.

The octree implementation has been tested on up to 16384 cores on
Occigen for large Poisson problems (up to 2048^3). The results are here:

http://basilisk.fr/src/test/mpi-laplacian.c

A new example illustrates the combination of adaptive octree, solid
boundaries (and gfsview-batch).

http://basilisk.fr/src/examples/sphere.c

The next step is to implement height-functions in 3D (there is already a
testable but incomplete implementation) and we should be very close to
the capabilities of Gerris. For a summary see:

http://basilisk.fr/Features
http://basilisk.fr/src/COMPATIBILITY

Other significant changes are:

* scalars are properly typed. For example the common mistake

scalar s[];
foreach()
s = 0.;

will now give a compilation error.

* Built-in memory tracing:

http://basilisk.fr/src/README.mtrace

enjoy!

Stephane

PS: the random basilisk.fr server deconnection issues should now be fixed.

[felixh...@hispeed.ch]

tnx a lot! I'm very pleased and look forward to do my 3D-jobs in basilisk. Greeeeaaaat job! :-)

[Bruno SARKIS]

Hello Stephane,

I am currently trying to compute 3D simulations on Basilisk.
It is in a 3D box and I am trying to set boundaries.

The purpose is to create a simple shear flow (SFF) of the following form :

u=z.ex

How to set the boundaries ? There are 6 borders : left-right, back-front and tom-bottom.
But what are the corresponding spatial directions ?
Is it correct that :
- left-right is x-direction
- back-front is y-direction
- top-bottom is z-direction
?

Inspired by http://basilisk.fr/src/examples/sphere.c, I tried :

u.n[left]=dirichlet(z);
u.n[right]=neumann(0.);

p[left]=neumann(0.);
pf[left]=neumann(0.);
p[right]=dirichlet(0.);
pf[right]=dirichlet(0.);
 
But it does not work.

Thank you in advance.

Yours faithfully

Bruno Sarkis

[Antoon van Hooft]

Hallo Bruno,

The 3D can be seen as an extension of the 2D.
in 2D:
left/right is associated with the x direction.
top/bottom --> y direction

in 3D, front/back are 'added' along with the z-direction.

I do not spot an error in your boundary conditions. How did you find out that is does not work? 



Op maandag 25 juli 2016 11:04:22 UTC+2 schreef Bruno SARKIS:

[Bruno SARKIS]

bHello, and thank you for your answer.

I finally created such flow successfully.
But I still have a problem for setting front/back boundaries.

My aim is to immerse an hyperelastic boundary in this shear flow.
How to make sur they will not be any unstability due to the lack of boundary condition in z direction ?

Thank you in advance.

Yours faithfully,

Bruno