SCRATCH PAD
Illustration of the problem on the example
The figure below shows how much more flexibility system can offer if the modules will not be square-identical.
My own questions to the problem:
- How to parametrize the definition of geometry?
- How to define relations between modules?
- Is this problem suitable for parametric study?
So which STEPS do I want to include in my project:
- Definition of the modularized grid (input)
- “Placing” of the existing modules in the grid (creating the building plan)
- “Analysis” of the nodes and placing on the building plan the node elements (columns, X-stabilizing elements, T-stabilizing elements)
- Completing of the plan with beams
- OUTPUT: generation of the output lists (list of elements, dimensions, price)for one store or more if desired (potential development)
 |  |
All these steps were planned to be included in the “prototype application” made most probably as Matlab script to test the functionality. This version of idea for geometry definition includes “recognizing” in defined grid built of rectangular areas, which modules are included in the building and which are not.
UML Diagrams
This is the result of the workshop in which we participated in Luleå last week. Very interesting, but little bit tricky…
The first try:
(2008-11-26)
Other diagrams:
A partial diagram showing dependency of the vertical elements of the structure on the node type (2009-01-13)
The diagram that represents the idea: (not UML)
(2008-10)
Parametrization of the STEP no 3
This part requires revision:
* How can I define the geometry of the building in the parametric way?
* position of “beam-boxes” in some coordinate system in relation to chosen corner or middle of the module?
* Positioning of posts and bracing elements dependent on defined geometry
* values of three corner parameters (a,b,c) decide which element will be placed in this corner
* Relations between neighboring modules and their dimensions \\
If we look at each corner of the each module of the horizontal geometry and assign to it 3 parameters (a,b,c) which can have just 2 values zero (0) or one (1). As seen on the figure above the parameters are placed in clockwise order and the represent presence (1) or absence (0) of neighboring module. Depending on the configuration of these modules, suitable element should be put in considered corner: column, X-shaped or T-shaped stabilizing wall. This will decide about the lengths of beams in particular modules.
All possible variations of these parameters are:
| Case | Param a b c | Element input |
|---|---|---|
| 1 | 0 0 0 | column |
| 2 | 1 1 1 | column |
| 3 | 0 0 1 | T-wall |
| 4 | 1 0 0 | T-wall |
| 5 | 1 0 1 | X-wall |
| 6 | 1 1 0 | X-wall |
| 7 | 0 1 1 | X-wall |
| 8 | 0 1 0 | X-wall |
So, let's illustrate these situations:
Notice 1!
This figure requires some corrections, but the unsymmetrical placing of posts and stabilizing walls is intentional and is a part of the problem. The initial idea was to place columns and stabilizing elements by the edges of modules for outer walls, but for inner walls to put them centrally on the edges between modules (compare cases 1 and 2 on the illustration above). However it appeared that it may not be feasible in reality, so I took a decision to place all the elements centrally.
The idea was processed and the it appears that the problem can be narrowed to 4 types instead of 8:
So this problem I can accept as solved temporarily. (2008/10)
After the first tries of applying the idea in matlab it appeared that the presented above approach is not exactly practical. The improved idea is again presented below (2009/01):
| Type | Param a b c d | Element input |
|---|---|---|
| 1 | 1 1 1 1 | column |
| 2a | 1 0 0 0 | column |
| 2b | 0 1 0 0 | column |
| 2c | 0 0 1 0 | column |
| 2d | 0 0 0 1 | column |
| 3a | 1 1 0 0 | T-wall |
| 3b | 1 0 0 1 | T-wall |
| 3c | 0 1 1 0 | T-wall |
| 3d | 0 0 1 1 | T-wall |
| 4a | 1 0 1 0 | X-wall |
| 4b | 0 1 0 1 | X-wall |
Such a distinction is necessary for using the concept in the Matlab software, to place especially the T-elements with the correct rotation.
(better quailty figure coming soon….)
….in progress….
Feedback
This is a section for people to give feedback.
Petter Andersson writes:
Hypothesis: A parametric approach can provide a flexible way to configure the beam box structure.
Each beam-box can be viewed as an object (Class).
Interfaces for each beam-box are post positions. These are derived by the beam length. The beam length has a >possible interval between 6-8m, where the optimum weight ratio is at 8m.
Possible attributes are;
* Beam length
* Posts
Parameters in the system view are total floor dimensions and wall height.
Interesting thoughts, thank you for support in starting this description and interesting discussion.
Discussion