To all of our dear followers that were eagerly awaiting updates: apologies for the long time since the previous post.
After the analysis we made a masterplan, mostly based on the first part of the analysis (that you can read a few posts back). Keypoints in the plan are the many connections with the area west of the Stadiumpark, that is now completely blocked out by the infrastructure. The public parts of the stadium will mainly be oriented towards the west.
After a lot (!) of discussion and sketches on how to proceed, we decided to go with a very literal interpretation of the masterplan. The housing structure of Rotterdam-Zuid will be continued towards the stadium (see masterplan), and then wrapped around it. The stadium becomes a social housing project. A very literal continuing of the urban fabric. We believe that thit is a clear statement, after all design is politics.
The square opens up towards the stadium, and the space underneath the overarching twists is a very interesting public space. When there is no match going on, the space under the stands is part of a connection between east and west. There can even be a bicycle-path.
Because the curved shape is quite complex, we decided to split it up in 3 parts: the ends ("slabs"), the twist, and the vertical housing. In the curvature they form an unity, but thus far during the progress we treated them more or less seperately.
The slabs consist out of a high-density low-rise social housing complex. It is three layers of housing, and it is one of the subjects for the computational design (later more on this subject). Bottom-line is that it should be as dense as possible, with as much daylight in the bottomlevel as possible, with as much coherence with the rest of the "ribbon" as possible. These are two of a few partly parametric, partly random determined (density, size, number of subdivisions) configurations of the dwellings. The reference from Madrid (couple of posts down) may be more clear now.
The twist is a transition area that is a real challenge. The architectural language of the twist, combined with the slab is quite strong, and we don't want to do too much concessions. Horizontal floors are mandatory, but by creating curved facades, it is in our opinion possible to keep the twisting lines in. Here is an visualization of the twist and the slab+twist seen from the waterside.
The third part is important. It is where the conflict between the different scale (housing and stadium) is most evident. For the largest part we want to seperate the trafficflows. Only some big plaza's are cut out of the housing skin. During matchtime (approx. 3% of the time) are these plaza's closed down for the 'villagers'. Or, what is more appropriate: The village is closed down for the hooligans.
Potential problems:
- The roofstructure, on which we did not yet decide :x The support can ruin our cavity, or can penetrate the unity of the facade. The longer we postpone decisions about the roof, the more it will be problematic later on.
- The twist. We have a clear visual in mind. Chances are that we need to do concessions. It is such an unconvenient form for a building, that our layout of this area should be really well done.
- The daylight in the housing zone. High density does usually not compute with this.
- The acoustics and the vibrating and trembling of the stadium in relation to the housing.
- ...enough more. There is a lot of things to be solved. But we feel like we are getting somewhere.
Bart, concerning the discussion we had on your GH model connected to Galapagos: I had a quick look at the random component in GH - I confirm I really doubt you can run Galapagos on it - in fact for the reasons we already discussed, which I confrim. Simplified: introducing such a random input and/or operation 'confuses' the search algorithm. For a more detailed explanation: you can have a look at the blog by David - http://www.grasshopper3d.com/profiles/blogs/evolutionary-principles , bottom part, look at the continuity of solutions - if you have a random input and or a random reduction of your dwelling units as a matter of fact you have a fitness landscape (the trend of your daylight performance or other performance you are looking at) that has a high degree of noise or chaos. (Let me know if you want to discuss deeper this topic). As a consequence, I confirm you should input variable integers and also avoid the random reduce. Michela
ReplyDelete