how to keep the same material behaviour after dynamic relaxation for new boundary?

Dear all,

I try to explain my question through these 2 shots below. It help to figures out what I am trying to say:

We have 4 coordinates: A, B, C and D. At the beginning I will create a surface according these 4 points and then find out the optimized curvature by dynamic relaxation (kangaroo).

During the simulation of dynamic relaxation I might setting up some of parameter, for example rest-length.

The ‘result/ definition/ parameter` which I get after the dynamic relaxation should/ must be kept and applied for the next coordinates: in my case when I shift the point C!

In principle we could treat this surface like a physical material which has material behavior its own and it deforms when the boundary-conditions changed.

I am not sure if this way to think makes sense. I was just thinking it is important for a simulation in which we use only one material-property for differently conditions.

If I use kangaroo to simulate this, I get the new optimized surface after dynamic relaxation after I moved one of these 4 Point, instead of ‘translated the old properties to new geometry‘.

I hope I didn’t confused u guys…

Replies to This Discussion

Permalink Reply by Alejandro REBOREDO on September 19, 2013 at 12:44pm

I think you could relax with desired material parameters first, and then, with the resulting mesh from kangaroo output, relax again with springs_from_mesh and rest lenght factor =1

This could give you an idea of how the relaxed geometry could behave while varying edges and support point of main structure.

Just my Idea

hope this could help.

Permalink Reply by Daniel Piker on September 19, 2013 at 3:20pm

Hi Frank,

I'm not sure I entirely follow your question, but I think the confusion here is over the difference between form-finding and analysis/simulation.

For form-finding we often use fictitious material properties, such as zero-rest-length springs, which do not necessarily correspond to the real material the structure will be built from.

This is the case for both digital form finding tools like Kangaroo and for form-finding with actual physical models.

For instance, using chain nets to find shapes for masonry structures, or soap film for fabric - it is not that those materials accurately mimic the true building materials at all, but just that certain aspects of their behaviour can cleverly be made use of to find geometric forms well suited to the structural properties of the actual building material and the loads that will be applied to it (hanging chains and pure compressive forms, zero mean curvature and uniform fabric tension).

It really is a sort of ingenious subterfuge - in a way we are tricking nature into giving us an answer to its own challenge by asking it a different, seemingly separate, question.

Now back to your question - you can indeed use the output from the form-finding run as the input to a further analysis run, as Alejandro suggests, taking the resulting lengths from the the first as the rest lengths for the second (perhaps multiplied by some factor <1 to give some pre-stress).

Permalink Reply by Alejandro REBOREDO on September 19, 2013 at 4:33pm

Daniel: Thanks for your explanation and, of course, for clarifying my point of view.

This will be very useful for what I'm working on. ;-)

Regards

Ale

Permalink Reply by Jon on September 23, 2013 at 4:34am