Week 3 : TPI does not solve the curvature problem

summary

in the last reaction I was pointed by Rudi to the fact that TPI does not automatically imply a division in smaller parts when curvature gets bigger. triangular tesselation also works for dividing a surface into planes.

the reason for making a smaller division when curvature is larger is mainly for making the elements (with a certain thickness) more plane-like and easier to produce. My assumption was that when a surface gets larger curvature it needs to have a finer division in order to make the elements easier to produce. It would make the components less wedge shaped.
In fact, the divide-option in Grasshopper seems to generate a bigger division at the less curved borders of the surface (figure 2), maybe a variable divsion depending on curvature is not a necessity.

Paul de Ruiter (TOI) advised me to look carefully at the connections/joints between the components. The connections are influencing both the structural and physical performance of the facade. Also he mentioned spray-composites as a possible finish. fibre reinforced resin that can be applied by spraying (and maybe by dipping as well)

the connections are limited by:
- the geometry that can be created by a milling machine (figure 5)
- the surface finishing of the elements (possibility for adding seals for instance)
- the stuctural connection

The aim is to keep the structural connection really flexible, so it works for every variable element configuration (As seen in the protospace pavilion by ONL, connections between the frame and cladding always fit). maybe flexible strips can offer a solution.




figure 1


figure 2


the coarse division could also have elements whit a curved outer surface in order to make it approximate the master surface.

possible disadvantage - elements become wedge shaped, difficult milling? (only in a 2,5-axis machine, or with complex edge detail)




figure 3

a smaller division decreases the difficulties caused by wedge-shape, but more elements are needed, which possibly has negative effect on strength, watertightness and reduces the possibility to add windows to the structure






figure 4
extracted element from a free formed main surface



figure 5

meshing method - Tangent plane intersection

smaller elements when curvature gets bigger. Advantage of avoiding large curvature in milled elements and keeping elements relatively flat.

test of planar intersectionmodel on othogonal grid. not suitable for TPI mesh, because planes do not nescesarily intersect.






main principle of TPI, tangent planes always intersect when placed on a triangular grid unless planes are parallel



triangular grid made with VB-script in grasshopper as in Woojsung-tutorial.



TPI model with tangent planes. intersection of planes maybe needs another script, because intersection of only 2 or 3 planes at once is possible in grasshopper, but planes intersect with 4-6 adjacent planes

maybe the VORONOI option can be usefull

parametric model which devides a free form surface into regular UVdivision. elements can get extracted from the model seperately (green highlighted element is the same as the two in front) and use as a milling-model. The extracted elements schould be provided with possibilities for connecting to each other.

major flaws in this model are:

-boxes are placed on surface in normal direction, which results in warped boxes. one or more flat surfaces would make milling easier because the workpiece doesn't have to rotate or re-clamped.

- when the curvature gets to big, geometry on the surface sticks trough each other. I will look for a solution as in the tagent plane intersection method by Cristian Troche from the "week2" post

iterative process for research

first cycle

the first cycle in the design process is more or less an exploration of how a model can be made for a free-form vertical surface representing a facade fragment. The model is simplified to a square surface without taking the borders into account. this can be a next step in research.

feedback wk2

most of improtance to do iterative process:

define criteria and goal for research. After each cycle of the design process, the results can be tested to the criteria and be improved.



the design criteria will be:
1) products have to be created in a fully automated process.
This means the elements can consist of multiple materials or parts, but have to be assembled without manual labour. for instance in case of applying reinforcing fibres to the resin cover of foam.

2) carry windload and deadload
facade must be able to carry loads, for now I assume a span from floor to floor, like 4m, but later on it can maybe be implemented as a parameter

3) connections wind- and watertight
to get a proper facade, connections should prevent water or outside air from getting in. for instance, rubber seals can be added in one of the final stages of the production process.

week2

research take-off

3-axis milling

5-axis milling

the possibilities for 3d milled shapes with a 5-axis milling machine seem to be endless (until we take a look at de base plane of the workpiece, which stays flat)

free form divided in elements. The 3D or 2,5D milling process can handle any shape, but concave shapes with a large curvature might be more difficult. create smaller elements when curvature increases.

Planar Hexagonal Meshes by Tangent Plane Intersection

Christian Troche

Universität Kassel

division for custom shaped facade-elements.

feedback februari 8th

narrow the perspective to a single technique
- milling rigid foam block to any shape covered in resin or other cover (sort of lost mould principle)

try to imply iterative process
-first try one shape shell, regarding structural and physical qualities. iterate this proces for other shapes

search for limitations in:
- production process
- shapes
- sizes
- structural quality

maybe choose material in accordance to those limitations

research topic / hypthesis

hypothesis

a system of repetitive elements with small differences can be produced with CAM techniques to create a facade or building shell in any shape.

approach

- design an element that can be custom produced in large quantities

- generate a series of models to test the systems applicability and limitations

- extract data from a parametric model that is usefull for CNC modelling

Nice example of file-to-factory work

important remarks:

- structure consists of only two parts nodes and beams, slightly different in shape, so really 'simple' system (or is it called 'double' when talking about two parts?)

- not a load bearing structure. extra stuctural parameters could have been implemented, but the structure had probably lost some of it's elegance by regularity

- triangular shape allows largest freedom of form, but also determines the appearance quite a bit

pieces were cut out and welded together by computerized machines, according to digital output from a parametric model.

parametric modeling usable for CAM-technique

(grasshopper model with boxedcomponents)

possible use for slightly different shaped elements.

The main purpose for digitally manufactured products will be free form shapes I guess. Think of cassette systems for a facade or load bearing structures from repetitive elements in a regular pattern. Really easy to produce data by a parametric model.

first session

two subjects have my interest in particular:

1. can digital design combined with CAM techniques be used for de-standardization of building products?

most building products are manufactured in large quantities to make the production process a repeatable process and more cost efficient. digitally manufactured products could escape the need to be repetitive in form; a programmed robot won't notice whether it makes 1000 different parts or 1000 similar parts.

interesting research questions will be:

-where are building products in different shape, size color or whatever needed?

-which property in particular needs to be different?

-what types of production processes are suitable for making different shaped products?

-how large can the span of variables be in a single production process

2. adaptable structures that change according to changing demands from users or changing structural or environmental demands

adaptable architecture can serve several purposes. Mainly it will react to it's changing environment.

-what are changing environments that ask for a changing building?

for instance a structure loaded by changing wind loads (or other changing loads), a house with changing usage (living, sleeping, working), building in a changing climate (hot, cold, dry or humid)

-what type of change suites that particular changing envirmonment?

welcome

this is my weblog for Stand-up Architecture. The goal of this coarse is to do research on a specific, building technology related topic of interest.