How to solve the staircase boundary problem?

Rieck1991

Hi,
I want to establish a curve wall, and I found the curve wall created by DualSPHysics like staircase and quite rough. How can I solve this problem? Any help will be appreciated.
Best regards.

Alex

Dear user,

We are now working on new GenCase, so that next version will allow to create particles at the locations the user need instead of using the 3D cartesian grid.

Next releases will include this option, but now there is no solution for you. The only way is that you code your own initial state of the particles to create the .bi4 file.

Pierre1

Hi,
I still have this problem, my version of dualSphysics is 4.2. Does it have a new tool to solve this trouble? Thank you.
Best regards

Alex

Dear user

In v4.2 we have included new options, like the one shown in the example RotatedBox.
More info in https://github.com/DualSPHysics/DualSPHysics/wiki/12.-Testcases#1216-extra-examples

But I am afraid that this example is still not what you need. I am so sorry.
We will work on new functionalities that allows to create particles at any exact position defined by the user (like a curve) instead of points in the 3-D lattice

Regards

Pierre1

Thank you for your answer and all the work on dualSphysics.

I import a complex geometry from a stl, so use the rotate tool isn't possible in my case.

Waiting for this new functionalities, I have a new question in order to reduce this staircase problem. When we decline dp, this problem is smaller, but the computation time is too big. It is possible to define differents dp for the boundary and the fluid? (using the mk or the type?)

Regards
(and sorry for my english if it isn't totally exact)

Alex

Dear user,

That functionality is not available now, but we are also working on that.

Regards

rlowe

Dear Alex

For my application I am also finding this issue to be signficant as when waves propagate over a slope there is too much dissipation created by the "staircase" at the bed set by the particle spacing. This is having a noticible effect on results. As I understand functionality to improve the issue will be implemented in a future version. You mention there is a work around by writing initial particle locations in a .bi4 file. Are there any examples, codes, etc. you could direct me to on how this could be done. Depending on how challenging it is, I might have a go.

Thanks
Ryan Lowe
University of Western Australia

Alex

You can try "rotateaxis" after you have created a plane.


Example in RotatedBox
https://github.com/DualSPHysics/DualSPHysics/wiki/12.-Testcases#1216-extra-examples

rlowe

Thanks Alex - I will give that a go. Ryan