triangular mesh frame optimization

Hello. I'm an architecture student with little experience who just graduated from school. But I decided to give a shot at a small scale real life competition. I intend to use triangular mesh frame to support the windows and the whole building. The best example to illustrate this is the Blob in Eindhoven (see attached image), although I don't intend to use a blobby form.

But the main question here is the optimization of the structural mesh frame. It is easy to make a mesh in rhino for any given surface by simply typing "mesh", but first the individual shape and size of each mesh face will vary, and second the connection vertex between different faces are all different too. As a result, it will be incredibly hard and costly to build in real life.

So, as many others have already did in this forum, I'm wondering ways of optimizing mesh by limiting the shape and size of each face. And also can someone shed me some light on how does real life "meshing" work (examples of connections and construction methods)?

From searching this forum, I found lots of similar threads. Many of them ask how to cover a curvy surface using identical equilateral triangles. The answer seems that a smooth covering is impossible as long as the surface isn't developpable, and only approximations can be made using kangaroo.

Daniel had some good scripts regarding this http://www.grasshopper3d.com/forum/topics/triangulation-scrip-with and

http://www.grasshopper3d.com/forum/topics/folded-plane-subdivided-i...

And I also encountered a script that can approximate a curved surface with limited set of triangles, but cannot find it anymore.

But all of the scripts above have their problems in real-life fabrication, as the first script has triangles approximately equilateral but not really so, and the second script has to change the border of the surface. I'm just wondering how did people do this kind of thing in real life, like the Blob in Eindhoven or Ghery's abu dhabi guggenheim museum's long transparent tubes. Did they just simply prefabricated everything or did they have everything optimized to the max? How did they do that?

Sorry for this long and demanding question lol. But any direction of research is appreciated. :) Thanks!

Attachments:

blob eindhoven.jpg, 124 KB
abu dhabi guggemheim ghery.JPG, 127 KB

Replies to This Discussion

Permalink Reply by peter fotiadis on March 11, 2015 at 6:31am

For "do" read, in fact ... find something simple and cost effective (not very easy: the find bit, he he, we've lost the plot long time ago with all that ultra complex things).

This is very easily doable and by aligning panels "on top" avoids pockets of rainwater that greatly affect the look of polycarbonate/lexan over time. Use only polyurethanes AND never silicons (unless they are structural silicons) .Bright Flats are replaced with L cold extrusions that look a bit cooler AND allows to to align them properly very easily. fixing bolts are of self penetrating type (wind negative pressures are nothing in this occasion).

Attachments:

170_DEMO_56B_V4.3dm, 1.5 MB

Permalink Reply by shuo zhang on March 11, 2015 at 4:17pm

Hi Peter! I didn't expect this. You basically did my part of the job. I was wondering how could i make the thing more "pro" and here it is lol, after waking up at 11 and having 0 progress all day. I checked your work on behance (i assume it's you by the way you talk), and they are so fresh. You seem to be very interested in the technical/mechanical/structural part of architecture, which I think is crucial but often overlooked in modern schools (at least in NA). From my limited experience today's arch schools are driven by either graphical aesthetic (people crave for artsy/beautiful renderings) or historical nostalgia (so called anthropology of architecture or phenomenology). I have deep interest in the technical part of architecture too (the reason I applied to arch school) but found myself very limited in this side of knowledge. I'm really glad that you helped me and I look forward to learn more from you.

All the bests,

Shuo

Permalink Reply by peter fotiadis on March 12, 2015 at 1:33am

Well...it's me (Behance), but I've stopped adding new/latest stuff because in fact ... I hate Internet/publicity/myself, he he.

Some last notes before walking the walk (i.e. the rabbit hole):

1. In the past Architecture was the art of sketching: some "idea" with pencils/crayons + vellum paper (or with some computer) > then "others" trying to make this happen. This in general is known as top-to-bottom approach. Naive and dangerous (for the reputation/reception/acceptance of Architects/Architecture) to the max.

2. These days we work both ways: whilst some work on some "idea" (called it: "assembly") others (in sync mode) resolve the bits and nuts of that "idea" - up to 1:1 level of detail (called it "components"). This is the bottom-to-top approach. Make this your way: NEVER proceed in something whist's not EVERY bit of that something is well addressed (with at least 3-5 ways).

3. The emergence of parametric (GH, Generative Components, Dynamo) in AEC (an approach well known in MCAD word many years ago, mind) made things ... worst: the tremendous topology exploitation capabilities blinded people's mind and they are completely sucked up by the forest forgetting/by passing the critical fact that there's no forest without trees.

4. That's expected: is in the human nature to follow/admire the blink/glam and omit/skip the humble. It's the easy way you know, he he.

5. The tremendous growth of countries the likes of UAE/China/Russia made AEC things ... even worst: lot's of cash available > make us some encomium to Vanity, forget Modesty. You can replace "Vanity" with "New Frontiers" ... if you like fooling yourself.

Some Academics are not capable to understand all that: if they could they would potentially operate in the field (where the pink color is rarely used) and not in fishbowl(s). Some Academics believe that an "idea" is the 99% of the whole whilst actually is less than 1%. But on the other hand anyone can do Architecture (even Architects, he he).

That said (Vanity crisis) you want some other "component" options for this case of yours? (starting with "some" dollars more and ending with the mortgage the house/sell wife+kids option).

take care (and kill them all)

Permalink Reply by peter fotiadis on March 12, 2015 at 3:19am

On the other hand ... we live once > hail Vanity (and the art of pointless) > should I modify this hybrid (rigid LBS + cables + planar) for your thingy? (it costs less that Nigeria's GNP > bargain > order it now and save up to 0.0001%).

Attachments:

234_UAE_T34_Custom_012A.pdf, 1.4 MB

Permalink Reply by shuo zhang on March 12, 2015 at 3:51pm

Hi Peter, for some reason I can't reply to your latest posts so I'll just reply here.

Man what we've discussed has been so far out of topic for this noble forum lmfao. But I'll just continue for a bit.

That vanity crisis that you've talked about isn't that much of a bad thing in my opinion. We are humans after all. Being a Chinese immigrant to Canada I kinda experienced the two different atmospheres of a burgeoning economy and a stable developed country. China, UAE etc. are like the America of the 20th century. Rapid development is embedded in a whole lot of enthusiasm and fever, often blind, but surely hot. People outside of this rage may look at it with a rational eye, seeing all the foolishness in it's deployment, just like how the West is looking at China/UAE/India now. But in the end it is during this kind of fever that we do great things, Egyptian pyramids, Babylonian hanging gardens, the Parthenon and the acropolis, and the most recent Las Vegas, "Delirious New York" has all been through the hot phase, blinded by it, yet driven by it. What Rem Koolhas missed in his book was exactly this kind of atmosphere, or vanity crisis. You acknowledged it in your last post btw :)

Of course I'm pro environment and sustainability. But to what extend can we elongate our life as a human species anyway? We have one Earth, diminishing resources, yet we are reproducing with exponential rates. At one point there has to be an end (or a new beginning), but let's just live fully the life of the human species. (I may be very naive here as maybe technology will save us, but that's another topic).

Also I completely agree with you about the "idea" part. I think people are weighing too much weight on the "idea" itself, shows like ted talk make it look like an idea can automatically generate every "component", while it is only 1% of the work, as you said.

And back to topic (I feel really bad spamming irrelevant personal rants on a tech forum). What I'm tying to do now is to optimize the position, size, length of the steel tubes (the triangulation) so that it is physically sound. I'm trying to use kangaroo for this, but I don't know how to optimize the behaviour of solid steel tubes when most tutorials about relaxation are for fabrics, which can be mimicked by springs. Do you have any idea? And thank you for the hybrid proposal, as much as I wanna see it i think it's fine for now :) .

May the force be with you.

All the bests,

Shuo

Permalink Reply by shuo zhang on March 12, 2015 at 5:12pm

for clarity, I attached a picture of what I'm thinking of. first, I'm not sure whether it is the right way of doing it . But I'm thinking of determining first the main frame of the structure (how?), and then to use it as a constraint for the form searching of the other rods. Is it right?

I'm thinking that the main frame will be optimal when after the loading (triangular polycarbonates + self weight) it will not create concentrated loads on any particular point/nod of the structure.

I'm thinking that given the length of the rods constraint and the weight constraint kangaroo should be moving the position of the nods until an equilibrium is obtained but I don't know how to do it.

I may be totally wrong.

Attachments:

mesh frame.jpg, 3 MB

Permalink Reply by peter fotiadis on March 13, 2015 at 1:05am

Hmm ... K can "equalize" things but here this is not the primary concern since "randomness" is better than "equality" (as a design orientation - in this particular case - so to speak):

Note: this thingy is very small and we can "safely" substitute "finest structural solution ever" with "gimme the next size tube" : crude but it's working, he he. But if it collapses ... we can use the next next size tube (trial and error , he he).

PS: Of course a "thick" MERO could yield a LBS capable to withstand Katrina^2.

Provide your def/data and I'll try some freaky ideas.

Permalink Reply by shuo zhang on March 13, 2015 at 6:57pm

Hi Peter. Yeah maybe I'm overthinking too much. I'm still trying to learn kangaroo so after this whole day my definition is pretty laughable and not worth posting. Here is the structure of my rhino frame, I'll keep working on it with kangaroo and see if I can make something out of it (what's equal strain triangle in kangaroo btw? I'm trying to use that but don't fully understand it, went to things like FEM and technical stuff like that but they are too deep for me... My idea is to try to optimize the frame in some sort of "scientific" way.

Attachments: