Rapid Prototyping in Architecture – Virtual to Physical

Our latest architecture blog entry. By Piet Meijs, 3D Printing Expert at Rietveld Architects.

In earlier posts, I talked about the bigger picture of implementing Rapid Prototyping into the practice of Architecture. But I haven’t really talked about the actual process of going from a design to a physical model. In my last post, I did touch on our workflow a little bit, but I will elaborate more on that here today.

Design Phase
Every model starts with a design. The designs in our office are always project related. This means that the model has to reflect a specific design solution which was created for a specific design assignment with it’s own specific parameters. The design usually starts out as a rough hand sketch which may or may not be refined by hand.

Schematic Phase
Once the initial design sketches are complete, they are then digitalized in CAD software. Depending on the complexity of the design we start with 2D line drawings or we go directly to a simple 3D massing model. These initial setup drawings are intended to lay out the design with the correct dimensions and proportions. These drawings or schematic computer models can be used to check if the program fits properly and to see if the initial massing of the design makes sense. At this point, it is still easy to adapt the design to make the program fit while keeping the initial design intent.

3D Modeling
Once the schematic line drawings or computer models are done, we can start working on the 3D computer model that will be sent to the 3D printer. This part is obviously the most crucial for the print model.

At this point, we have to interpret the design that came out of the Schematic Phase. This means that we have to go back to the design intent from the Design Phase. What was it exactly that we wanted to achieve? Did we want a solid or a transparent building? Is the structure part of the design, or should it be hidden? Is the context important? etc. We always have to look back at the initial design intent and to keep this in mind when making decisions. Otherwise you run the risk of ending up with a design that is so full of compromises that it no longer bears any relation to the initial design assignment.

The next step is to determine what scale the model will be. Unlike with industrial design, architectural models are rarely made on a 1 to 1 scale. More often they are at a scale of 100 or 1000 times smaller then reality. This creates a whole new set of challenges, because when a design gets scaled down that much, certain important elements might disappear, while other non-essential elements might start to clutter up the model. The trick is therefore to understand what the important elements are in order to convey the design intent as it was set out in the Design Phase. The elements that are not important may be left out completely while the important parts might be increased in size to emphasize them. There are no strict rules to this, but we have learned from experience what works and what doesn’t, and when we have a new situation we usually print multiple test models to see which solution works best to communicate what we want it to.

Once we have made all these determinations, we basically created a design for the print model. This model design can be best described as an abstract caricature of the building design. This model is no longer an exact scaled down version of the building design, but does communicate exactly the design intent.

The design of the model is then worked out in either SketchUp or ArchiCAD. Once the basic geometry of the model is done we export the model to Rhino. In Rhino we are able to use grasshopper to apply textures to certain surfaces and we can use mesh-check tools to check if our geometry is clean and watertight.

3D Printing
From Rhino we export the model to an STL file. This file gets loaded into Objet’s software to place the model onto the 3D printer’s build tray. This software also takes care of slicing the model and sending the individual slices to the Objet machine to be printed.

Currently we only use Objet’s Rigid White material for the acutal printing of the model. For us, this was a very obvious choice. Colors in architectural models are very often interpreted as materials. So you run the risk that your client might think he gets a blue building if you show him a model made out of a blue material. White material is most often seen as blank sheet of paper. This works a lot better, because when the model is made this early in the design process, it’s not clear yet what the material should be. White can therefore represent nothing or anything.

When the model is printed, it needs to be cleaned. Objet’s machines uses a liquid model material and a liquid support material. These materials are laid down on the tray in slices and hardened with UV light. The model material hardens to a solid, and the support material hardens to a more flexible state.

The removal of the support material is probably the most low-tech part of the whole process. We use a mix of plastic utensils, paperclips and a water jet to clean our models. The key is to find tools that either bend or break before the model does. That way, you won’t accidentally break the model while cleaning. And if little pieces do break off, they are easily glued back on once the model is fully cleaned.

When everything in the model is white, you run the risk that the model will appear as a solid white mass. To counteract this, we like to paint certain parts of the model a very contrasting color, often red. We usually use the vertical transportation cores for this. They can easily be printed separately, and they provide a nice continuous hole through the model, which helps the cleaning process. This painted element, when put in place, offers a background for other elements. Now the row of thin, white columns no longer looks like one big blur -  you can actually see each column separately.

An added advantage of using a very strong color is that the white model material will appear to be whiter. This has to do with the human eye only being able to see a certain spectrum of colors. The closer the colors are the more sensitive the eye is to color differences. So when off-white is put next to a very saturated red, it will actually look pure-white.

The last and final step is to document all our models with photographs. Most of our models are given to clients and will stay with them. We sometimes trade a new model for the older version to display in our office, but we still want pictures of the most current models. These pictures are not only used for PR on our website and in our iBook, but also in presentations for potential clients.


This post is also available in: Chinese (Simplified), French, Spanish, Korean, Portuguese (Brazil)


  1. Hi Sam,

    We create SketchUp tutorials for designers on our School website (www.go-2-school.com/sketchup-tutorials) and we have a lot of architects that ask us about 3D printing physical architectural models. You mentioned that you will go from SketchUp to Rhino before exporting to STL, but I am curious to find out if you’ve tried using the SketchUp STL extension to take your 3D model directly from SketchUp to the 3D printer? If so, do you like the results, or do you find it better to go through Rhino first?

  2. Since my last question, we decided to pull together a new guide: 3D Printing from SketchUp (http://www.go-2-school.com/sketchup/3d-printing). We included a video tutorial for our SketchUp-to-3D Printer workflow. However, we still have a lot of information we’d love to add to the guide and would appreciate any feedback you have on your workflow.

Leave a Comment