While earning his PhD at MIT in the early 1960s, Ivan Sutherland PhD ’63 developed Sketchpad, a computer program that allowed users to create images on a screen using a light pen instead of code.
“He really explored this idea of how the computer would change the way we can think and design and create,” explains Caitlin Mueller, associate professor in the Building Technology Program at MIT, where she leads the Digital Structures research group. “Unfortunately, after that piece of work, it became more commonplace for both architects and engineers to think of the computer as replicating analog methods.” In other words, engineers use computers for calculations, and architects use them for drafting.
Mueller’s goal is to employ machine learning to support the design process from both an architectural and engineering perspective. By creating software that generates design alternatives and simulates their performance, she hopes to qualitatively change how buildings are conceived and built. A big part of that is encouraging architects and engineers to work together—every step of the way.
In the traditional building process, a client hires an architect and provides a set of specifications—X square feet, X number of rooms, etc. After finalizing the design, the architect hires an engineer, who typically looks at the design and says the building can be constructed using X amount of steel, for example. There’s often little back and forth. Engineers generally don’t offer large-scale design suggestions in order to, for example, save a substantial amount of steel. As a result, buildings that look great can often prove expensive to build and operate.
That is a wasted opportunity, Mueller says, arguing that engineers should be an integral part of the process from the beginning. The tools she creates make it easier for architects and engineers to work together to find design solutions and assess how changes can influence metrics ranging from the energy needed to heat a building to the cost of labor in construction.
Clients can also evaluate in real time how different designs affect costs, impact the environment, and influence factors such as occupant comfort—giving them better information on which to base decisions. Architects and engineers can further employ Mueller’s tools to ensure that, as designs are changed, a building continues to meet both a client’s requirements, such as number of rooms, and safety regulations, such as required number of egresses.
The tools even work well on less traditional structures. Recently, Mueller’s research team used them to design a community garden trellis system in Somerville, Massachusetts, using wood from culled urban trees. “We generated interesting forms by discerning the intrinsic geometry of the trees’ branches to arrange them in structures that used the material efficiently and effectively,” Mueller says. “We would never have been able to understand how to use this complex geometry or the structural behavior of these forms without the tools we’re developing.”
Bringing architecture and engineering together, and considering engineering problems during the design process, will ultimately lead to buildings that are more cost-effective, more environmentally friendly, and cheaper to build and operate, Mueller says.
“People have long been lamenting the fact that architects and engineers don’t work together,” Mueller says. “Today, both because of the sustainability imperative that’s so serious and the abilities these new tools open up for us, I think in the next 5 or 10 years we’re going to see a big shift in the types of tools companies use.”