Steady Flow Input

alexmees

Hello,

I've just began using DualSPHysics and I would really appreciate any help you could give me.
I've run the Test Cases provided and did a few simulations of my own but still could not find how to create a steady flow entering and exiting the simulation. I know there is a Periodicity sample included which re-inputs the exiting particules but I have not found where in the XML does it actually does that. Is it Incz? How does it changes X then (in a 2d case)? When I open it in DPI in comes out empty. (?)

Ideally I would like to re-input the particules but change their velocities (as a parabolic profile). Is that even possible?

I suppose it's all really simple but not for me!

Thanks!

alexmees

I have just read an old post saying it can't be done yet...
Any news regarding that?

Mojtaba

Dear Alexmees,

You should change the source code by your own to define a specific section to assign a steady velocity to the fluid particles in a simulation. It isn't a default capability in the xml or compiled files.
By the way you could do this in JSphcpu.Cpp file by assigning a velocity to set of fluid particles in a specific and constant Z position (Or/And X Pos.) in each time step for your 2D case.

Regards,
Mojtaba

alexmees

So that takes care of assigning a velocity profile in the input but is it possible to continually generate new particules in a given section? I imagine it would be needed to check for existing particules next to the inputing region as not to "break" the program with a too sparse or too dense region. Any other technique that you could think of would be more adequate?
If I can make anything close to a functioning change to the source code how could I share it?

Alex

alexmees

Just thought of a way to approximate a parabolic profile in simple geometry. Create layers/slices of a certain heigth each one with a different initial velocity, for example, create several slices of a box volume each one with a fraction of the total heigth and increasing velocity.
This would give a discrete approximation to an parabolic profile with needing to change anything in the source code. With a little more work you could make modifications for a 3d case.

Alex

Mojtaba

You can't generate new particles during simulation time.

Mojtaba