Sustainable Architecture Leans into Artificial Intelligence


SOURCE: NOW.NORTHROPGRUMANN.COM
JUN 10, 2022

Exploring the Possibilities

Increasingly, architects are using AI-leaning software tools in a similar way, calling on algorithms to cull the world of architectural possibilities quickly and efficiently for design approaches that help to meet the growing demand for sustainable architecture and green technology.

“In architecture, AI is generally synonymous with generative design — or, as I like to call it, ‘optioneering,'” explains Dan Stine, director of design technology at Lake|Flato Architects, San Antonio, Texas. “Our software tools use algorithms that generate a large number of design options based on parameters we define, then rank those options according to how well they meet our criteria. Ultimately, we select the option that works best for a given project.”

Pursuing Sustainability

HMC Architects describe sustainable architecture as the use of design strategies that reduce the environmental impact of a building. Sustainable architecture principles consider every aspect of the planning and construction process (for new and remodeled properties alike) including:

  • Choice of building materials
  • Siting
  • HVAC and plumbing design
  • How the building incorporates renewable energy and green technology
  • Choice of landscaping
  • Stormwater management

According to Ryan Goold, an associate and architect at Boulder, Colorado-based Sopher Sparn Architects, AI embedded in modern architectural design tools gives his firm an early jump on meeting clients’ sustainability goals.

“We use AI-based generative design tools to help us decide on the size and appearance of a building (an activity called massing) and how best to orient it on a site to take advantage of passive design strategies such as solar heat gain, natural ventilation and daylighting,” he notes.

Sustainability goals come into sharper focus in the next step, Goold adds, when he and his team use AI-leaning tools to analyze the environmental impact of the project over its design life.

“The environmental analysis not only helps us validate or adjust our massing strategy but also informs the building assemblies and insulation values required to meet or exceed the relevant energy code,” he advises.

Generating Options

As Stine explains, in conjunction with building massing and orientation decisions, architects also rely increasingly on generative design tools to produce multiple floor plan options for projects such as multifamily housing.

“Our urban planning studio is using software that takes inputs such as zoning requirements, building setbacks, height limits, and the number and type of housing units, parking requirements, etc. and then generates many examples of how a project of say, 500 units could look — in just seconds,” he says.

He adds that this same software can also provide cost data for the prospective land developer — information that could help that developer decide if and when to purchase the land and if the proposed project could generate enough income to be profitable.

Optimizing Energy Performance

Stine notes that one of the more powerful AI-leaning tools Lake|Flato is using involves energy modeling, a process that heretofore has been a rather tedious mix of estimates and guesswork.

“In the past, we’ve had to input an R-value (insulating effectiveness) for every aspect of the project — walls, roof, windows, etc. — then do a calculation for energy consumption and hope we hit our target,” he explains. By contrast, their new software tool uses the project design model and location (anywhere in the world) to automatically generate a range of R-values for every aspect of the project for 45-degree increments of building orientation.

“We can now dial in the desired energy performance of a building as if we were using a mixing board in a sound studio, pushing sliders to select from hundreds of options while also helping our clients meet their sustainable architecture goals,” says Stine.

Keeping It Real

In some cases, according to Stine, architects may be using AI tools without even realizing it. He points to the computer graphics cards that his firm uses to create realistic renderings of projects to share with clients.

“Real-time rendering is a very resource-intensive activity that has to calculate reflections and shadows and animate the subtle motions of water and leaves bustling in the wind,” Stine explains. “Our graphics cards use a special chipset that supports AI to render a scene at low resolution then render it at much higher resolution using neural network training.”

In many cases, he claims, the AI-inspired results are actually better than renderings his team has produced using traditional graphics tools.

Putting Trees First

Stine highlights another tool that helps to produce sustainable architecture using software to analyze aerial photography of trees and vegetation to predict precisely where tree trunks are located within a canopy of trees.

“Knowing the location of trees on a site allows us to design structures that integrate well with that environment, a key tenet of our focus on sustainable architecture,” he notes. “We can include this information in our design strategy to produce buildings that our client will love but won’t require the removal of trees they’re fond of.”

Searching for Starting Points

For all of the advancements in AI-flavored architectural tools, Goold notes that there is still no algorithm that can search a database of previous projects to identify and import examples of good starting points for projects requiring sustainable architecture and green technology.

“We’re using generative design tools to lay out projects on a site,” he clarifies, “but the actual types of units are still mostly in our brains or on paper from past projects. We can open up those old 3D models and bring them into a new design, but in terms of having an AI tool that can search an archive and find exactly what we need, we’re not there yet.”

Measuring the Impact

So, what impact is the use of AI and generative design tools having on the architecture profession? Could it begin to disrupt the career trajectories of new or even more senior architects?

According to Goold, AI-based tools definitely inform design strategies that architects use to meet sustainability and green technology goals, but “as architects, we’re reluctant to give up control when it comes to design.”

And unlike some professions where AI has disrupted or displaced more junior level jobs, he adds, “Younger members of our staff just out of school often have broader familiarity with these newer emerging tools, which can really help the firm improve its processes.”

Stine agrees that AI-leaning software is an increasingly important part of architects’ tool set, but he believes that there will always be room in any project for both younger and more seasoned design professionals.

“The entry-level designer might have better graphic design skills or know more about virtual reality tools, but those tools alone can lead to unworkable or dead-end design solutions,” he advises. “That’s where a senior designer who understands how design and architecture work in the real world can guide a project back to a successful outcome.”

Creating a Sustainable Future

Looking forward, Stine expects to see AI tools for architecture continuing to mature and move from their largely generative design status today to more critical decision-making roles in the future.

He argues that, to date, architects have used AI-like tools to essentially validate their assumptions about design. And even though these assumptions often track with what appeals to the human eye esthetically, AI tools can provide a smarter, more cost-effective approach to sustainable architecture.

“Many of the rule-of-thumb assumptions we’ve used over the years — for example, how deep a beam should be compared to its length — are ideas we can’t use anymore,” he suggests. “We need to be designing things optimally, and not oversizing beams and HVAC systems because it creates waste and embodied carbon. With better AI design and simulation tools, we can get there — instead of just relying on our assumptions.”

Delivering the Spice of Life

Still, Stine admits there will always be some design features that appeal to the esthetic and emotional sensibilities of humans, even if they fail to meet the sustainability requirements of AI algorithms.

“Sometimes it’s okay to give an apartment or office a little extra square footage or bump out its façade or add some other interesting visual relief to its design,” he suggests. “To a computer, the added feature makes no sense, but to the human who’s going to live or work in that space, it makes all the difference and delivers that proverbial spice of life.”