Learning Forum

[javat33489]

Hi all.

I assembled a test model and solve the problem of drilling with deformable nozzles. Drill rigid body. Plank for drilling - SPH/

I need to solve the problem in 1 second, but with this time, the calculation time on the PC is very objectionable. And so I took advantage of a life hack and increased the density. Units in the SI system (mm, Pa, s).

I increased the density of all elements by 1e9.

But now my nozzles are very distorted during rotation. I also get negative volum or dt errors.

Why?

I did the same task on planing (without rotation) everything was successful.

I attached screenshots of the results from the workbench (LS-Dina ACT), but I solved the problem both in prepost and in the workbench, just to make it clear, the problem is the same

 

[Jim Day]

Increasing density to such a large extent in a dynamic problem will produce nonphysical nonsense. The inertia of the elements is so large that it becomes very difficult to move the material being drilled. The corresponding forces on the drill are unrealistically large.

[Jim Day]

As a starting point, use the actual densities, model the drill parts as rigid (so the drill parts do not control time step) and coarsening the SPH mesh as necessary to give a time step that will allow you to run the simulation in a feasible length of time. Consider whether it's really necessary to run for a full second.

[javat33489]

Thank you for your responses.

Yes, I started to reduce the density closer to the real one. I'd really like not to run the calculation for a whole second. The task in milliseconds is solved well without increasing the density.

But how can I evaluate the condition of the drill if I make very small movements?

[javat33489]

>>As a starting point, use the actual densities, model the drill parts as rigid (so the drill parts do not control time step)

If I make the drill and bit deformable (rigid), how can I evaluate its state (stress, heat)?

[Jim Day]

It appears the spatial extent of the SPH workpiece is too small. Broaden it so the outermost portions are relatively undisturbed as drilling proceeds.

[Jim Day]

Drill/bit as rigid is only intended as an experiment to see if the drilling operation can be successfully and reasonably simulated. If the answer to that question is yes, then the next step would be to reintroduce deformability of the drill/bit and look at ways to reduce the run time other than scaling up the mass unreasonably.

[javat33489]

Thank you for your response.

>>But how can I evaluate the condition of the drill if I make very small movements?

[javat33489]

I have tested many options and read many scientific papers. I need your help.

My goal is to analyze the state of the deformable nozzle. I need to carry out the calculation so that the calculation time is not 1000 hours and at the same time I can tell how the nozzle heated up and how it deformed and how long it will last.

Previously, I changed the density and calculation time to reduce the time, as well as reduce the size of the FFS particles. The only option that led to a normal solution time (12 hours of PC solution) is the calculation time of 0.01 sec, the density is multiplied by 1000, and all units are in the SI system. But the stresses at the same time on concrete are 24k MPa and on the nozzle 10k MPa, this is too much.

Tell me how to make it so that the calculation time does not take more than two days, while I could assess the condition of the nozzle? How can I make the part to be drilled infinite for calculation so that I can mentally drill 1000 meters? I can connect the Weller curve if needed for the nozzle material.

I need a sequence of actions. Calculation time, how long is it enough to turn the drill with a nozzle to assess its condition?

In some works, they make a turn of several degrees and at the same time they can give an assessment of the condition of the nozzles. How can I do that? Explain to me please in detail and concisely? I think a lot of people are looking for this answer. It will be very helpful.

[Loic Ivaldi]

Hello,

Could you please tell us the following :

- How many CPU do you use for the simulation ? Can you increase this number ?

- Use the real densities and set TSSFAC=0.9 in *CONTROL_TIMESTEP then give us the timestep.

- Then, make your drilling tool rigid and give us the new time step.

Best,

Loïc