The nascent additive construction industry is slowly starting to take shape as an increasing number of start-ups appear on the scene with techniques for 3D printing large-scale structures. The latest is a London-based company called Ai Build, which aims to make additive construction smarter and more accessible through the use of artificial intelligence and affordable materials.
As with many additive construction endeavors, Ai Build's entry into the field begins with a 3D-printed pavilion. As an ornamental building, a pavilion is the perfect large, yet nonfunctional structure for demonstrating the possibilities of 3D-printed architecture, as there is no need to meet critical requirements for a building that might be used by people, as with an office or a home.
Unveiled at the GPU Technology Conference in Amsterdam at the end of September, the Daedalus Pavilion is a structure made from 48 different pieces 3Dprinted from Formfutura PLA filament over the course of three weeks. The structure, measuring approximately 16.4 ft(5m) wide by 16.4 ft (5m) deep by 14.8 ft(4.5 m) tall, was then shipped in pieces and assembled on site in just a matter of hours.
As beautiful as the initial structure may be, the Daedalus Pavilion is only the beginning for the start-up as Ai Build explores the use of large-scale 3D printing and computer vision for more advanced construction methods.
Cam received a bachelor's degree in architecture from Istanbul Technical University in 2009 before going on to obtain a master's from the Architectural Association Design Research Lab in 2012. Cam explained that, in pursuing architecture, it was often necessary to translate one’s designs into something that could be reasonably executed in the physical world using existing technology. Large-scale 3D printing, however, presented the opportunity to materialize even the most complex designs.
About one year ago, Cam founded Ai Build with his partners using their own funds. “Traditionally, 3D printing is used for smaller objects, mostly for prototyping and some medical applications, but it's not very common in the construction industry, so we really want to make it a bit more accessible by developing hardware and software and making things easier for large-scale 3D printing,” Cam said. “When we started the company, the initial idea was that why don’t we find an easier way? Why don't we transform construction in the 3D printing space so that we can design things more freely?”
The firm has since been working on the ability to 3D print large-scale structures using methods that would both enable the freedom of design and lower the cost of construction. To do so, Ai Build relies on a KUKA industrial robotic arm outfitted with a custom extruder designed and built inhouse.
Cam explained that Ai Build is primarily working with plastics, like the PLA and ABS filament traditionally used with desktop 3D printing, because it is much easier to print with than materials like concrete and metal. An issue that may immediately come to mind is the durability and integrity of a large-scale plastic structure.
That is something that the firm is working on, according to Cam, but, at the moment, he does not feel comfortable explaining how Ai Build plans to address it. So far, however, Cam and his team have conducted some early experiments.
Aside from the use of plastic, Ai Build is contributing some other unique elements to the field of additive construction. Ai Build has listed a legacy project called aiPort on its website, a concept that would see machine vision used for smart home devices, but the start-up has since put that project on hold to focus entirely on large-scale 3D printing. With that in mind, the team has begun work on incorporating ideas of artificial intelligence and computer vision to the large-scale 3D printing process.
Some companies have begun referring to the generative design algorithms used to model 3D-printed structure as “artificial intelligence.” While Ai Build does use generative design to create its 3D models, this is not the “AI” that the firm intends to employ.
“Today, it's very common that industrial robots are being used in manufacturing. It is especially common in the aerospace and automotive industries and, in 3D printing, robots have started being used for a while now,” Cam explained.
“The problem with these industrial robots by default is that they cannot see,” he continued. “That becomes problematic, especially when you scale up the production. The way they work is that they take a digital file from the computer and then they blindly execute it. If anything goes wrong while the production is taking place, the robot doesn't see that and then keeps executing the same file that came from the computer initially.”
To solve this issue, Ai Build is developing computer vision for its construction process. This entails outfitting a 3D printing robot with sensors that allow it to monitor the fabrication process and detect errors that occur. The robot will then be able to compensate for the error in the following layers.
“The materials we are using now are quite easy to work with, but there is still a degree of unpredictability in the material,” Cam said. “You can anticipate what the material will do to a degree, but you cannot know 100 percent. The idea is to create a feedback loop between the physical environment and the digital environment.”
So far, Ai Build is still in the very early stages of advancing this technology, but, over the course of the next few months, the start-up intends to focus primarily on developing its Ai platform. In the meantime, the company should have a new structure to reveal in the very near future and, next week, plans to bring the Daedalus Pavilion to Digital Construction Week in London.