questions on latest release

Daniel,

First off, many thanks for all the amazing tools you've provided. I can't thank you enough for sharing such phenomenal work. I want to ask you in this latests release if the base units are still in meters. In past discussions http://www.grasshopper3d.com/group/kangaroo/forum/topics/units-mass...

you mention how to calculate forces such as pressure and how they require a divide by 6 step. Is this still the case in the latest release. Also, you metion a membrane force component and I was wondering what the status of that is. We are trying to calculate near zero stretch inflatables with interior baffles (think air mattress) and apply particular loads such as concrete onto these pnuemoforms. We can get very close but are never really able to get the units to work properly and achieve zero stretch and therefore we don't have complete faith in the digital simulations we are able to generate. We also can't seem to get the pressure value to numerically represent the real forces we plan to use on an architectural scale. Typically we are just adjusting a slider until the simulation looks right rather than being able to supply the actual pressure we plan on using. You mention limitations in using 1d spring systems and it seems like this is one of them. This file shows some of the unit conversions we are attempting inflate.ghx

The unit conversions we have seems to work on a simple mesh shape but when we have more complex meshes with internal baffles, the mesh gets wildly distorted. Can you look over the file and verify the conversions are correct. If you can provide any additional insight it would be really helpful.

Replies to This Discussion

Permalink Reply by Daniel Piker on December 30, 2013 at 5:38am

Hi,

Thanks for some interesting questions.

From Kangaroo 0.095 on, the pressure force no longer needs to be divided by 6. The value in the PressureLevel input should be given simply in Pascals. I do plan to add some more documentation about the units.

There are actually 2 different types of triangular 2d elements I've been working on, and I decided to hold them back from this release to avoid confusion until I have them both working correctly.

One is for simulating soap-film like behaviour, and tries to reduce its area to zero, to give true minimal surfaces. However, I realized that the cotan weighting option I've already added to the Laplacian smoothing component is exactly equivalent, so perhaps one doesn't really need both.

The other is a constant strain triangle, which (unlike the soap-film element) also allows simulation of shear resistance and non-zero rest area. Still working on this one - will post any new developments here.

However, since you talk about "near zero stretch", I suspect neither of these elements really matter in this case, as I get the impression (and correct me if I misunderstand) that you are mainly interested in the interaction between pressure and the concrete load, and the exact amount of stretching is not important as long as it is very small.

Because of the nature of the relaxation process, things will always stretch some amount, as they do in reality. To make the stretch of some elements small, the stiffness of those elements should be high relative to the other forces in the simulation, and this may necessitate using some high (fictitious) mass values to avoid numerical instability.

Finally, this release also contains the GasVolume component, which I think might be an easier to use alternative here. It is similar to the Pressure force, except also taking into account the volume of the gas (following Boyle's law).

RSS