 |
At Schmidlin I was part of a small group responsible for developing future strategies in Visualization/Simulation, Mathematical Modeling and Process. All the the 6 members of the group collaborated on all fields but we were each responsible for a couple of tasks. My main investigations were Visual Communication, Simulation, Knowledge Management and tools coding for sculptural geometries. I was also involved in a lot of discussions surrounding Lean Process and Supply-chain Optimization, more specifically how to map and monitor process.
The building projects I were involved in included Campus des Wissens for Novartis by Frank Gehry and Burj Dubai by SOM. |
|
|
| This was a tender I was involved in. My part was to investigate and process the Catia geometry given to us by the architect and look at it with the eyes of Facade engineering. The early work included visual inspection of the geometry and flagging of all kinds of exceptions, geometrical discrepancies and omissions, and to get a general feel for the complexity of the project. |
|
 |
The second step was to read out all the geometry into sorted lists so that it could be processed with our custom tools. In most cases the geometries we worked on were relatively rectilinear or otherwise rational and could easily be recreated in code. In this case it was not possible due to the sculptural nature of the geometry and so the mathematical grid had to be extracted from the geometry.
For this task I wrote some tools in Rhino and Catia. The first tool was used to add meta-data like glass type, element ID, etc. to surfaces. The second tool extracted information like points, edges, meta-data, etc then sorted this and wrote the data to CSV files for further processing. |
|
 |
The final part of my involvement was to study similarities in corners and try and come up with a reusable parametric solution. The really expensive parts of the building were the many complex corners, some with as many as 8 different gutters and closures coming together.
For this I worked in Catia developing a skeleton rig. The rig guided sheet metal folding which in term guided all other components making up the corner. I am not certain that there was any benefit in developing a parametric solution as it is quite time consuming and has low reusability. It may have been a better to approach them from a sculptural point of view using Rhino. |
|
 |
| I worked for quite some time on investigating various tools and methods for visualization and simulation in relation to facades and the products developed by Schmidlin. Included here is the concept tree but I removed all the example files to conserve space and for copyright issues. |
|
 |
| A part of my work was to model and render details for various projects. The purpose of these visualizations was not photo realism, but to as clearly as possible express form and function to clients and suppliers. |
|
 |
| The group was responsible for making 3D prints using a Stratasys Dimension FDM machine. We used it to make prototypes of functional parts such as brackets and clamps, sometimes combined with existing parts. The advantage of the Stratasys machine was its use of ABS plastic which meant parts that were quite durable and that could be mechanically machined to add threads, etc. I wanted to write a custom postprocessor for catia NC machining so that wa could print directly and have control over build-patterns but never really got too far.
|
|
 |
| An early project when joining Schmidlin was to build a new public website. This included a completely new graphic concept, a reworking of the content structure and the development of a new back end based on the CMS Typo3. The really interesting aspect of this project was how much it exposed various internal discrepancies in information coherence and how hard it can be for a large organization that has existed for a long time to fully understand what sells. The graphic, functional and structural aspects were completed but the collecting of contents proved very difficult. As a result the website was never finished. |
|
 |
| An important part of the development of new modeling methods was to teach these to others. For this the group developed a 4 part Catia training course in methods of facade modeling at Schmidlin. My part of this was to develop the documentation. In relation to this we looked at creating an internal forum and chat-based support-system. The great advantage of a forum is that it will gather useful questions and answers and make them available to more users by going beyond the traditional person to person support. |
|
 |
| FEMO needed a way to collaboratively develop our ideas. We looked around for software and found Freemind, an Open Source mind mapping tool. Using this we created a linked structure with a central Group-Tree with links to individual Member-Trees. This way we could work on our own concepts in our personal tree, while other members of the group could see what someone was working on. At each or our knowledge transfer meetings we would discuss concepts we had been working on and if the group deemed the concept mature enough it would be moved to the Group-Tree. It was a good way of collaboratively going from the unstructured to the more structured and formalized. |
|
 |
| The original initiative was to create a handbook of how facades were developed at Schmidlin. The handbook would contain brief descriptions of all steps in the process as well as knowledge surrounding it. It quickly became apparent it was not the optimum way for Knowledge Capture and that the "Handbook" was in fact part of a wider Information Space. The gateway to authoring would have been too narrow and update-cycles too time-consuming. I looked at using MediaWiki as a platform for capturing the content and facilitate update-cycles and topic evolving. I also looked at automatic generation of PDF documents for a topic using XML and style sheets. |
|
 |
| For the work in Supply-chain Optimization I was looking for tools or platforms that would allow us to visualize multidirectional graphs with many different types and weight-values attached to nodes and links. The idea was to find a way of visualizing complete process maps in order to locate and understand possible problems or regions to improve. Pretty ambitious idea that would require a lot of honesty from all parts of the process. I looked at Thinkmap, a graph api that could read XML from a database, as a promising alternative. |
|