Hi to all the experts,
I’m looking to build this project as a case study and am interested to know what is your approach here to solve the curved glass panels facade (not the rotating fins) so the panels will be planar.
Attached is one approach, but it is leaving a small overlap between the panels. I can space them, but there might be a better approach.
Raffles City_01.gh (25.5 KB)
The surface shape and panel sizes are only a rough guess.
Appreciate your help!
Hey Yoav, I think if you want to build it planar, there is no way around some spacing issues…
Here is my try.
I used the tangent at the midpoint of each segment’s top and bottom boundary. Then I used only the lower tangent to rebuild the top and lower boundary to get a planar result. The gaps are still there, just spaced evenly and are less obvious.
Hope that helps you.
Raffles City_01_Mit_Tangent.gh (29.8 KB)
Hi Felix,
Thank you for your reply! That’s a very nice solution. In this case, allowing for a non-rectangular panel helps keep the lines continuous.
But I wonder if that is the best solution in terms of cost - judging from the pictures, it seems that the panels are not continuous but a bit staggered especially in the curved areas.
Any thoughts on how they have approached this problem?
The irony of this is that Rhino was probably used in some part to develop the original paneling.
Anyway, there are a lot of issues when looking into paneling complex shapes. Here is an older thread that addresses how complex the problem can become and how easy it is to “overconstrain” the problem:
Hi Scott,
Thank you for your answer and the reference links, these are very interesting topics. Indeed I realize this may be too complex to resolve in a forum discussion.
Just for clarity, by saying Rhino was used in some part, do you mean there was a manual approach or some sort of optimization to the surface itself?
Many thanks
Probably both! This is never a pure science. Always a balance constraints of construction system, budget and team knowledge. So there is some manual changing of the shape to allow the range of panels to fit.
I expect you have been on these pages to find out more specifics how they think about paneling with Rhino.
The interview about the project you are looking at: STH BNK by Beulah - UNStudio
UNStudio has provided a two-part article to introduce various digital geometric block strategies used for UNStudio projects. The articles describe how block strategies have been used for past UNStudio projects and are still extensively used for UNStudio projects today.
Geometric blocks are not a new topic in 3D modeling, but through increasingly available parametric plugins it continues to provide an efficient means of controlling a wide range of elements. Effective block strategies can decrease costs by enabling the fabrication of many of the same components for construction.
This study reiterates some key qualities of blocks in the context of digital design and how they can be intelligently applied. It then analyses the specific block strategies applied to the facade design of two of UNStudio’s projects.
Not sure this is helpful, but here is a lot of detail on how a project like this is calculated and fabricated and all the levels of engineering that go into it:
https://wiki.mcneel.com/webinars/morpheus#the_morpheus_hotelfrom_design_to_production
Thank you very much for your help, Scott! all are very good resources, I’ll dig around 
As Scott says, you will not be able to keep the over all form of the building and form the panels planar without over-constraining the geometry. My tip would be to look at the perimeters of the of consecutive floor plans to see how the envelope develops on its way upwards, because from the detail it seems apparent, that this is not a glass curtain, but ribbons of structural glazing that, at the base, sit ontop the of the floor to run past and attach to the outer facing side of the floor above. In other words, the perimeter of the floor the overhead story determines how the glass panels run below. Where there are no floors to support the individual panels, the building shape will inevitably be developable to allow for planar surfaces.
Hi @Yoav_Ronat & @Volker_Rakow
@scottd If you can answer this question and have time, I would greatly appreciate it.
I am reaching out to you two, because I am having difficulty with optimizing the panel on my facade. I am attempting to follow along on the method, but I do not know how to rebuild my surface curves where my panels are located.
I attempted to get a line from the point on the curve and to its intersection with the previous point’s tangent line but that does seem to decrease the deviation (the distance from the potential floor line)
Can someone please assist me with getting the panels to be with a 1/16" to 1/4" distance from the what is the floor line as of now. I am not asking for a script (I have part of it, but I don’t know how to do the next part of my script and the why so I can learn how to do it in an office setting [post grad looking for work])
Please see my script attached to the post.
240930 - Facade Script Post.gh (2.0 MB)
I will add a screenshot of my attempt at rebuilding a curve to decrease the deviation amount. Thank you, I hope to hear from someone soon because I am not feeling good about this right now.
I do not have the plug-ins that you used to create your script.
Is it possible to simplify the problem so that it can be expressed in vanilla Grasshopper? Or to post some images to demonstrate what it is that you are trying to achieve?
I managed to get view of what what you are trying to do from the parts of your script that do work:
This is a vastly more difficult and complex problem than the rebuilding of the Raffles City building poses. In the Raffles City building, we are essentially dealing with the terracing of a purely convex form. The contouring of the floor levels determine the subdivisioning of the vertical facade to each floor independent of what happens on the floor above and below. This fact is then hidden by the purposeful staggering of the facade divisions in the horizontal direction.
In your problem you are going to have to decide how this facade is constructed because this goes hand in hand with how you decide to break down the facade. I assume this is supposed to be a glass curtain that lies infront of the structural elements behind? Which means alot of steel to imitate an organic form.
@Volker_Rakow Thank you for spending some time to look at my script. Raffles City. Building in a horizontal way is what I am attempting to do in my project.
The one plugin that is used is paneling tools.
Yes this surface is going to be mainly glass face that is attached floor to floor. I was hoping that the smaller horizontal panel mullions can be one mullion or it will be covered by cladding using the museums somaya method.
I also want to use this one facade as a test for the remaining parts of the building.
I think my project will be a mix of the pair of top and bottom curve, and the phase tower method.
Using the raffle city method would splitting the curves to where their curvature is the same and bring the curves that are further away from my main curve direction (to the right and north)? Would that work.
My main question would be how to script the part of the raffle city facade that decreases the deviation? I can make a test script or you can for you to look at. I can upload one within a few hours.
I am unclear as to what your intentions are. The constructive solution of using bands of structural glazing to envelope the Raffles City building is different from the steel frame supported triangular grid of the Museo Soumaya, construction images of which can be found here: Museo Soumaya has a Secret — Geometrica .
What I can say about decreasing the deviation between the floor profile in plan and the run of the fassade based on uniform panel widths is that in practice, the problem is attacked from both ends: an idealised outer form is conceptualized, the glass panels attempted to be fit to the floor perimeters, and then the floor perimeters adjusted to better fit what the glass panel widths will allow. Or, the glass panels have individual members that are not standard size make up for the difference in expected width. Either way, some trick is used to provide the appearance of regularity while the reality is that not everything is uniform.
@Volker Rakow I should not have mentioned museo soumaya. That was unnecessary information. That is later down the line.
I forgot about the tangent post from @felix.brunold . Yes, I think I am going to subdivide further into horizontal strips but have the same panel width applied to the different strips. I am going to use the paneling tools divide distance or length component to get straight even lines and then work backwards to get the curved panels. Afterwards, I will make the straight tangent panels (using something similar to @felix.brunold method), and I hope that this will work.
I will update you hopefully tomorrow on how it went. Thank you for responding to my post. I am grateful to you.
@Yoav_Ronat I am working on this type of facade for an old academic project to put in my portfolio, and I am working to get the panels to have whole inch (or maybe in your case centimeter or meter) dimensions. Were you able to get the panels to be whole dimensions in your case study project? If so, how were you able to achieve that in your script?
