3D or not 3D?
We haven’t heard much about 3D printing lately. Maybe it’s quietly become mainstream and therefore not sexy anymore? Or maybe it’s gone quiet because it hasn’t delivered on its early promises? Or maybe the new future it promised has simply been replaced by something else that’s the new future? The truth is probably a bit of all three. Don’t get me wrong. I think 3D printing is great, and have written about it before. “ [c.f. 3D Printing in History”] It produces amazing results in the field of medicine where it can be used to replicate important things like knees and jawbones for people who need them, and research is progressing in incredibly useful directions such as creating artificial human hearts by layering organic materials into what are called organoids.
Making things as large as buildings is still problematic for one very good reason.
Slump is one of those great words we don’t use very often unless we’re concerned with economic downturns, dead bodies, or the amount of shape that concrete will hold once its formwork is removed. A slump test is a test of concrete viscosity and if you want your formwork to be removed as soon as possible, then the less slump the better. On the other hand, if your formwork is complex and you want to be sure the concrete fills every corner, then a more viscous, “slumpy” concrete is preferable. One of the primary advantages of 3D printing is that it needs no formwork. If formwork is used then the nozzle and robot are nothing more than expensive means of delivering concrete or some other type of self-hardening material.
There’s still too much concrete production happening in the world but this principle of slump also applies to any construction material that can be oozed into place via a nozzle attached to a computer-controlled robot arm. As long as the 3D printing of anything involves squeezing a somewhat viscous substance through a nozzle and expecting it to stay in place much like we expect squeezed toothpaste to stay on the brush, then the problem is for the viscosity of the material to be low enough for it to be squeezed through the nozzle yet high enough to maintain its shape without slumping until it has hardened sufficiently. A small amount of true slump can facilitate spreading and creating layers but this needs to be compatible with the speed of hardening and the speed of layering because adding another layer before sufficient hardening has taken place will cause the lower layer or layers to spread. It is necessary for the layers to bond before complete hardening, and so the interval between laying of successive layers is usually around eight minutes plus or minus. However, the sweet spot between viscosity, hardening speed and layering speed is different for every material. In the examples below, the one on the right has a greater consistency of mix and spread but in both, the individual layers forming the vertical surfaces have convex edges caused by material spread and slump so the nozzle doesn’t have to waste time creating too fine a grain. For the time being, the look is a kind of 3D-printed Brutalism.
Working with gravity
In the middle left of the above image you can see some waves formed when a momentary lessening of nozzle volume for a lower layer causes the upper layer to flow into the thinner region below. If this phenomenon could be controlled then it might make for walls more resistant to lateral forces and shear fracture, but what it shows here is that the viscosity is sufficiently high and the hardening sufficiently quick to prevent all the material from flowing too far into the thinner layers below. As long as that’s prevented, it’s possible to make vertical walls fairly easily. It’s a mystery why some houses are called 3D printed houses when only the walls have been.
It’s of course possible for 3D printing technology to make floor slabs but, in a sense, floor slabs are already 3D objects formed relatively instantly and monolithically and not by trying to resist gravity but by making use of it. It’s not a bad process when you think of it, as you get to improve the strength of the concrete by adding aggregate and reinforcement. However, it could be worth investigating 3D printing the slab formwork.
Roofs and upper floor slabs present a different problem as it’s currently impossible to make them without formwork. It’s possible to make them on the ground in the same way as the lowermost floor slab was made, and then use a crane to lift them into place but this wouldn’t be in the spirit of 3D printing. In the same way, putting lightweight roofs onto 3D printed walls isn’t in the spirit of 3D printing either but that’s where we are.
It’s also not really in the spirit of 3D printing to use lintels to span door and window openings either. In the first and second examples above, the window openings extend to the roof. This decreases the wall rigidity. Instead of walls, the third example has multiple fat but rigid columns. The first example beow uses some kind of infill panel above the window frames but this also reduces structural rigidity. The other two images below use dedicated formwork or lintels as one-dimensional permanent formwork to support the 3D printing material until it hardens.
If window frames are going to have to be inserted sooner or later anyway, it might be a good idea to make them double as lintels. Doing so however, would tie a manual process into an automated one and this goes against the rule of sequential construction championed by William Levitt and every house-builder since.
Walls that are curved in plan have an inherent stability and architects such as Antonio Gaudi and Eladio Dieste have used this to good advantage in not only vertical surfaces but also spans created with the use of formwork. [c.f. Not The Sagrada Famiglia, Architecture Misfit #14: Eladio Dieste] 3D printing isn’t up to spans yet but the following three examples make use of the inherent lateral stability of horizontally curved surfaces. The examples on the left and middle uses 3D printing only for the processes 3D printing can benefit, while the example on the right full-height openings, self-stabilizing 3D printed walls and a lightweight roof doesn’t appear an afterthought as its beams link to the ends of the curved walls where the area of bearing surface is larger. It shows not only thought but knowledge that the compressive strength of the fine-aggregate concrete 3D printing uses is about 30% less than conventional concretes.
Dubai seems to be a center of 3D printed building experimentation and this villa with 3D printed walls but a conventionally poured slab and a concrete roof constructed with the use of formwork is typical. The window openings of this house seem too well finished and may be frames as permanent formwork.
This next house uses permanent window and door frame units as lintels but the entrance overhang can’t have been 3D printed without formwork. The line of grey colour along the edge of the soffit appears to be the parapet material. The material and construction of the roof and floor slabs is unclear. Unless gravity relents or such as thing as a 3D printed material that hardens while floating is invented, 3D printing won’t do overhangs or cantilevers.
Working Around Gravity
How then, to construct 3D printed houses without using formwork? One solution is to create the walls horizontally on the ground and to lift them into place in an updated version of the tilt-slab system we last saw in the post on Irving Gill. A variation is to just 3D print the pieces offsite.
Working with Gravity
The 2020 Clay 3D Printed House prototype developed by TECLA Technology and Mario Cucinella Architects gets a lot of things right. The circular walls are inherenetly rigid and have an internal layering for additional rigidity and a vented cavity. Instead of using lintels, the viscosity of the clay mix has been adjusted to allow openings with corbelled sides. Similarly, the walls corbel inwards for as far as gravity permits and the rooftop openings covered by readymade skylights. It’s an intelligent design that acknowledges both the advantages and disadvantages of 3D printing. The modular system for the nozzle arm supports is interesting but the less visible breakthrough is the clay composition that is both viscous enough yet fast-setting and bonding enough to make those corbels happen. I don’t see any structural benefit in making the walls corbel out before they corbel inwards. I suspect it’s been done to provide more useable internal space. Structurally, a cylindrical lower section would have done the job equally well if not better but would not have looked so new. Even so, any plastic settlement of the dome will create lateral outward-acting forces whether the lower part is convex or not. This is a problem as it’s not possible to embed tension chains at a height lower than the door opening. The “woven” structure between the outer and inner shells adds rigidity and lightness to what would otherwise be a massive and, because of that, unstable wall.
The same problem with how to use 3D printing to construct a roof that is even tangentially flat – i.e. a dome – has remained unexplained in the more spectacular proposals of recent years and how to solve this problem remains a good test of innovation as it hits at the fundamental weak spot of 3D printing.
Unless gravity relents or such as thing as a 3D printed material that hardens while floating is invented, 3D printing simply can’t do overhangs, cantilevers or domes.
And so, a 3D printed material that hardens in mid-air was developed. It’s used in a process called cellular prefabrication that uses carbon-reinforced plastic to prefabricate components of this building as curvy variations on a space frame. From what I can work out, these pieces are then somehow stuck or joined together onsite, filled with foam insulation and clad with a conventional cladding or coating material. Visible joints on the inside suggest interior finishes are pre-installed. The process can create large cantilevers without columns as support. These pieces can’t be that heavy so I hope they’ve thought of vertical wind lift. Nevertheless, it’s true that 3D printing is an important part of the process, but for what purpose?
I’m reminded of the words of Povilas Cepaitis, Luis Enrique, Diego Ordoñez and Carlos Piles from 2011 when they described the purpose of their Efficient fabrication system for geometrically complex building elements. [c.f. Prefab Parametric]
I have, oh, five questions relating to the seven words in the statement “everyone deserves to enjoy spatially interesting architecture” and at least the same for “the unquestionable visual attractiveness and spatial quality …” but I’ll leave them for now. I don’t remember how Messrs. Cepaitis, Enrique, Ordoñez and Piles’ 3D printed parametric pieces were connected to create “spatially interesting architecture” but the new carbon-reinforced plastic incarnation iof 2023 has an identical intent. I’m no advocate of aesthetic churn but I can’t see what makes 2023’s spatially interesting architecture any different from that of 2011. Why are people still bothering?
I was going to end it there but as is sometimes the way with blog posts, something I read only an hour before seems relevant. Today it was this Los Angeles Review of Books article.
The author uses the example of crypto currencies and NFT to discuss the relationship between human desire, faith, and the inevitable con that comes along to justify that faith. The author notes that technology usually operates with the technology being identified, then demonstrated, and a market built off the back of that. However, with crypto currencies, the market already exists but the technology has to be continually demonstrated in order to keep people believing it will someday be of value.
• • •