[Lydia]

Hello REEF3D modeller,

I'm trying to simulate a lock filling process with floating ship with REEF3D.

For that need a large upstream channel with higher water level, a lock with floating box (=ship) with lower water level. In between there should be a solid box (=gate) that has to move with prescribed motion.

I added a little sketch for easier understanding =>plan view

How can I insert the gate.stl as rigid body and define a motion in z-direction?


Any hints are highly appreciated!


Best regards
Lydia


P.S. My current state: a simple box with floating ship and two waterlevels is running successfully. Next step: rigid gate with prescribed motion. Then: open upstream channel with fixed upstream water level. Final aim: Rigid gate with two smaller valves opened with prescribed motion to fill the lock.

[Felix_Sp]

Hey there,

as asked in many threads (e.g. here) REEF3D so far cannot simulate multiple independent floating bodies. As the 6DOF solver is also the only way for prescribed motion of solid bodies(X 210 and X 211), what you want to accomplish is so not yet possible with the 6DOF solver.

From here I am just guessing what could work in the framework of REEF3D, but I have not yet tried such approaches. Maybe someone else has other creative ideas as well .

• Simulate the gate motion with the 6DOF solver and measure the discharge via P 67. You might need to add a line in gage_discharge_x.cpp as so far the gage is only printing out the iteration, but not the simulation time.
• In a second simulation, use this information for the simulation of the ship motion as a boundary condition together with the patch boundary conditions P 421 and P 422. P 421 is the obtained discharge and P 422 is the gap in between the gate and the bottom (PatchBC)
• Depending on your setup, you might need multiple PatchBCs.

I am unsure how good such an approach is, as the velocity profiles and turbulent quantities at the gate of both simulations are not identical.

• Otherwise, I think only having no ship motion and assuming the ship as a solid, non-moving body will work.

Hope this helps somehow, even though the answer is I guess not very satisfactory. Maybe someone else has a better idea.

[Lydia]

Hi Felix,
hi there,


thanks for your reply!
I'm a newbie with REEF3D and your comments and ideas are very helpful for me.

If there are further ideas how to solve my lock filling problem, they are very welcome!

Usually, I use OpenFOAM for modelling but the meshing concept with structured grid and level-set geometry representation in REEF3D and the open source concept seem worth testing cases, that are difficult with OpenFOAM's body-adapted unstructured grids.

Also very interesting are the sediment/scour simulation options. They are also on my testing list. So I might be back here soon, sharing questions and experiences.


Regards
Lydia

[Felix_Sp]

Hey Lydia,

REEF3D is a cool toolbox from my perspective. However, it is still a little rough around some edges, as it is younger than OpenFOAM and has yet a smaller community which helps debugging.

In regard to OpenFOAM: A colleague of mine did some heave decay tests with overset meshes, but I have no experience with that. As I only have experience with more "static" cases in OpenFOAM, I cannot really say how good this approach works.

If you are testing open source toolboxes and have access to a good GPU, you might also take a look at DualSPHysics. An SPH method is obviously different from the mesh based approaches. However, they have implemented as many floating bodies as you want. Your problem should be reproducible in DualSPHysics. The problem with SPH is obviously whether you can actually compute it with high enough accuracy, as the method is yet quite expensive in regard to computing power.

Felix

[valgrinda]

Hi Lydia,

as Felix mentioned, multiple floating bodies with prescribed motion are not supported at the moment. What is your timeframe for your simulations?

I have motion file support for 6DOF on my ToDo list, I will have a look at how this can be extended to multi-body.

[KevinMcTaggart]

Hello Hans. Do you know when rigid motion for a single body might be implemented? Thank you.