P1-Process of a Giant Peanut.

Hullo, just thought it might be useful to post a short chronicle of my P1 process findings. Find below many accidents that you can avoid:

1. Laser- Ribs and Surface (Enter Giant Peanut)

• For doubly curved surfaces that meet at a point, the normal-to-surface ribs need to be offset to prevent this terrible joint:
  
• Although acrylic is nice and pretty, it does not bend. You will have to manually create a angled mould and melt each angle individually. Please use thin materials such as thin cardboard, tracing paper, the thinner the better:
• If you like to be smartass and make a surface that goes from the outside to the inside, please remember to offset the thickness of your material when making the model in Rhino:
• A plug-in that helps arrange your ribs automatically in Rhino: RhinoNest. Useful. Saves time and material.

 

2. 3D Printer (Slicing the Centre off Giant Peanut)

• This went well. When creating the .stl file for 3D printing, it might be useful to know the Mesh>Repair Mesh command, as it automatically closes up holes in your mesh object. Terribly convenient.
• To make taller objects than the 3D Printer height of 25, or to save 3D printing time in general (as height of model determines total print time), break the model up into smaller pieces and stack/spread them out so that they cover a shorter height. UHU glue works fine in joining them together. Will not recommend white wood glue.

3. CNC Milling- Mould and Negative Mould (Moduling Peanuts)

• CNC milling may be used with the intention to create moulds in mind.
• If you intend to create a plastic (acrylic) mesh structure to stretch over a mould, vacuum forming will be out of the question. Thus in order to form your (laser-cut) mesh over the mould, you cannot heat it with a hot air blower and press it down with gloves on. Why? Because a) your mould will catch fire, b) it is very, very hot and thus painful even with gloves on, and c) no one has that many hands.
   
  Also the resulting mesh will not conform to the shape properly:
  
• The wise student will then create a negative mould, offset with the thickness of acrylic to be used with an additional 3mm minimum. Preheat industrial oven at 150 deg C, place acrylic mesh on positive mould and place both in oven and set temperature to 180 deg C. When the temperature reaches 180 deg C, immediately remove mould and acrylic and place negative mould on, pressing the two units tightly together while the acrylic cools. (You may sit on it, if wearing thick jeans.)
  
  
  (Please note when lasering the mesh, please etch hairline lines for the module to break off from the rest of the mould, as well as to keep a datum on the Mould and mesh so that they may align well.)

Well, here’s hoping you don’t make the same mistakes. Keep on experimenting! 🙂

P2- Model Image and Algor Simulation

Simple Y-shaped module multiplies to act like a tree-structure. Note density and inter-connectivity increases at higher levels. May be able to create interesting atrium spaces. Horizontal shear is a problem…resolving! 🙂

Algor simulation image on the above, unfortunately mixing vector and surface analysis which generates inaccurate results. Still useful in determining the minimum density of “trees” per floor area, as well as continuity of structure from top to bottom.

P2 Model + Simulation

Vertical louvres to act as sunshades and structure at the same time.

Optimized Ecotect simulation showing cooled wall facade.

P2:Performance based modeling and simulation (Algor)

 

Graphical represenation of stress distribution of the improvised model.

External vertical ribs designed to replace/reinforce the external structural shear wall system.

p2: ecotect

model showing convex-box louvres on n/s facade and horizontal louvres on e/w facade

avg daily direct radiation for south facing facade

P2 Exercise

Ideas for next development: Using grasshopper to twist my structure into vertical louvers. Degree of change (structure to louver) depend on distance from ground. Degree of torsion depend on image sampler with max rotation perependicular to general EAST/WEST. (example of how it will look like hopefully..)

3-Directional ribs

non fabricable to fabricable surface by paneling tool

Dense pattern works like a square pattern. Result in unroll panel error

3 directional rib version #1 with a thicker 3rd rib across to maintain material integrity

3-directional ribs version #2 with equal rib depth achieved by varying slits depth and width

On left- 3 directional rib of equal rib depth

on right- 3 directional rib of unequal rib depth. Deeper 3rd depth

Experiences from Ribbing and Panelling

Hi guys, this is my finished product! And unlike what it illustrates on its facade (notice the word “slacker”), i wish to reiterate that the task was not “slack” in any way.

Firstly, just like most of you guys, i rebuilt a rectangular surface with more control points and made it undulating. The reason why i did not try an extremely funky and curvy form was because i wanted the facade to be composed of rectangular panels rather than triangular ones. Next i created vertical and horizontal ribs.

After that, i opened grasshopper, imported this image that i randomly took off the net and baked it.

I used it to split my panelized surface and to my horror, the command took almost 30 mins! (i thought my computer had died on me!) i realised that the “circles” produced from the image sampler script was too close together and while performing the split command, they caused my panels to be split into hundred and thousands of surfaces.

Subsequently my file size became close to 300mb. And because it was split into so many surfaces, the unroll command obviously did not work. I had to create another copy of the panelized surface, unroll it and align each of the previous panels that was split onto it.

Then again to my terror, i realised my previous panels were unable to align to the unrolled ones because they were not unrolled! (on the plan view they actually look like they fit, but the borders are actually curving in to Z-axis direction) Not only that but the unrolled panels had slightly curved edges. (How naive of me to think that if i did not do an extremely funky and curvy surface then my panels would turn out to have straight edges) Anyway with further investigation, i realised that the edges turned out to be curved because the “faces” options in panelling from grid – panel 2d grid produced surfaces that was slightly curved in the xyz direction. (meaning to say it might look planar but it is not)

Actually it could have still been possible if i was going to use a flexible material for the facade and not engraving a sampled image on it. (Eg: fabric) I then tried to change the option to “add flat faces” but i got a result where the border of the face were not connected to adjacent faces.

So i had to resort to triangulating my surface in the end…

After repeating all the steps again, my file size was close to 700mb! (this was probably something to do with each rectangular surface being “divided” into 2 triangles) But this time i was able to align each panel(split using image sampling) to its corresponding unrolled panel.

Due to the filesize of the file, i did not dare to bring it directly into the laser machine because not only might it hang the computer and the machine but the thousands of circles would take forever to engrave. Therefore I changed the image sampled surfaces (which are vectors) into a raster format by converting it into bitmap using Coreldraw. The ribs were lasered on 2mm black acrylic and the facade on 1.5mm clear acrylic which i had to painstakingly assemble one by one.

This is the Rhino file i would have used if i did not plan to change the image sample’s vector format to raster format.

On hindsight, i would have labelled my components more clearly to prevent myself from having to go back to the file to check for the orientation of each rib and triangulated facade. (Can you imagine workers having to open a Rhino file on worksite to check if they are placing each facade panel correctly for a design that easily consists of hundreds of panels) If i were to do it again, i would have labeled which direction the ribs should face (eg:project north for horizontal ribs and project east for vertical ribs) and also an arrow on each rib pointing towards the sky. And for the triangular facade, i would label each of its 3 points with coorinates such as v1,h1|v2,h2|v1,h2 for ease of assembly.

Welcome to Digital Fabrication in Architecture Studio!

Welcome to Digital Fabrication in Architecture Studio!
A gentle reminder before we begin,
Do post: any sketchy idea, exploration, reference on/around Digital Fabrication in Architecture, which something contribute or hint to Structure, Space, Function, Materiality. It could be just Visual effect, Material sample, Component Assembly, Structural System, Space organization, or Fabrication process.
Don’t post: free form exploration kind of things! You’re the architects who control the design, don’t be servant of technologies.
Best regards, Shinya

Popps, RE-usable PP Shelters for fast assembly and disassembly