< Back to Blog

January 2020

,

Aremi MacDonald

Is the future generative design? A Q&A with industry experts

“With computational design, buildings become measurable and you are able to verify performance and compare different designs in metrics.”

- André Agi, Computational Architect, Link Arkitektur

 

Generative design is changing the way architects and urban planners approach the design process. Using sophisticated machine learning and algorithms, an iterative process generates optimised design solutions based on given building parameters and constraints. This enables architects to find the best match to fit their needs. 

Franz Forsberg and André Agi are lead computational architects focused on sustainability and urban design. They aim to discover new data-driven tools and systems to improve and incorporate into existing architectural workflows.

Ingrid Cloud took a moment to speak with these industry experts to find out just what the future of this new design process means for architects and urban designers alike.

André Agi and Franz Forsberg from Link Arkitektur

André Agi and Franz Forsberg from Link Arkitektur: finding new ways of solving old problems.

 

Ingrid Cloud: So, what exactly is generative design?

Franz: As designers, we are always looking for novelty to solve a problem in a new way, or to create a solution that has not been thought of before. The competitive quest for novelty, while still guaranteeing the design will perform as intended, creates a challenge which is not easily solvable through traditional design processes.

IC: Interesting. And a new way of thinking of the generative design process is comparing it to human evolution. Drawing parallels between producing multiple architectural solutions the same way the evolutionary process generates new iterations of a living species. Each new version learning and adapting from the previous.

Franz: Yes, the organisms that exist in nature represent a large selection of new formal solutions that go far beyond the imagination of the most creative human designer. At the same time, different organisms are uniquely fitted to their surroundings’ requirements. So, it seems that nature is capable of producing an infinite amount of shapes and solutions, which are both innovative and highly performative. 

IC: What types of software and plugins are currently on the market?

Franz: Universities and research programs, private companies, developers and all kinds of people are developing tools and sharing them on the internet. A large amount of them are open-source or free, and you can use them to customise your scripts. It offers a lot of flexibility, and it is a very low investment compared to other CAD software and IT infrastructure. However, it’s not easy for lower-skilled users to dive into this process and make something that works on a large scale. For wind analysis, there are a lot of small plugins to use for very simple wind tunnels and geometries, but they shouldn’t be used for documentation. If it’s not a solution like Ingrid Cloud, you would need knowledge in CFD to run an accurate simulation.

André: There’s also a greater need for trust in plugins to address more complex matters like environmental analysis and CFD. You need to be able to trust that the algorithms being used are widely recognised within academia and verified against benchmarks like Ingrid Cloud. Ingrid Cloud offers a system with higher accuracy and higher trust. Because CFD simulations are heavy, using Ingrid Cloud speeds up the process and provides support and feedback that allows us to integrate efficiently throughout the design process.

IC: Because the design process is constantly evolving, is data-driven design and computational analysis creating better solutions for the industry?

André: Data and coding is getting more adapted by the industry, which places demand on architects to acclimate to new tools and methods. I wouldn’t say every architect should learn to code because it’s quite demanding, but with computational design, buildings become measurable and you’re able to verify performance and compare different designs in metrics.  

IC: So what does all of this mean? What does the future of architectural design look like?

André: As designs become more complex, we’re going to need new tools to manage that complexity, and we’re going to need to deliver innovative solutions to them [architects]. Tools that give us answers throughout the design process. Technological innovations have always existed in the design community and have changed the way we design buildings. The use of data and algorithms has brought up a new way for technology to change our design process and offers a system to create more solutions in less time.

Iteration is the new black

Iteration is the new black: using computational tools to iterate around a defined constraint and, based on that, present a series of design alternatives is a key concept for the generative design approach.

Computational tools for smarter design 

Whether you’re a computational architect or a novice in CFD, Ingrid Cloud offers a tool for architects and urban designers at every skill level, providing fully automated and intuitive CFD wind simulations throughout every stage of the design process. Its user-friendly interface and sophisticated algorithms mean it has the power to do the heavy lifting, providing fast and reliable solutions as architects look to iterative and generative design tools as the future of design optimisation.

It’s not about design. It’s about creating a computational process which is more integrated and efficient, allowing interoperability and automation. Designing and co-working with engineers becomes easier, allowing better quality architecture to emerge.

- Franz Forsberg, Computational Architect, Link Arkitektur

We currently offer a fully-integrated add-in for Autodesk Revit software which you can download here. Check out our benchmark cases for an in-depth view on how we used an iterative process to compute turbulent flow solutions. Read more about our benchmarks here: https://www.ingridcloud.com/product-tour/benchmarks/