The new construction that is projected to take place between 2020 and 2040 plays a critical role in embodied carbon emissions. The change in material selection is inversely proportional to the budget as the project progresses. Given the fact that early-stage design processes often do not include environmental performance metrics, there is an opportunity to investigate a toolset that enables early-stage design processes to integrate this type of analysis into the preferred workflow of concept designers. The value here is that early-stage environmental feedback can inform the crucial decisions that are made in the beginning, giving a greater chance for a building with better environmental performance in terms of its life cycle. This paper presents the development of a tool called LearnCarbon, as a plugin of Rhino3d, used to educate architects and engineers in the early stages about the environmental impact of their design. It facilitates two neural networks trained with the Embodied Carbon Benchmark Study by Carbon Leadership Forum, which learns the relationship between building geometry, typology, and construction type with the Global Warming potential (GWP) in tons of C02 equivalent (tCO2e). The first one, a regression model, can predict the GWP based on the massing model of a building, along with information about typology and location. The second one, a classification model, predicts the construction type given a massing model and target GWP. LearnCarbon can help improve the building life cycle impact significantly through early predictions of the structure's material and can be used as a tool for facilitating sustainable discussions between the architect and the client.

The process of architectural design aims at solving complex problems that have loosely defined formulations, no explicit basis for terminating the problem-solving activity, and where no ideal solution can be achieved. This means that design problems, as wicked problems, sit in a space between incompleteness and precision. Applying digital tools in general and artificial intelligence in particular to design problems will then mediate solution spaces between incompleteness and precision. In this paper, we present a study where we employed machine learning algorithms to generate conceptual architectural forms for site-specific regulations. We created an annotated dataset of single-family homes and used it to train a 3D Generative Adversarial Network that generated annotated point clouds complying with site constraints. Then, we presented the framework to 23 practitioners of architecture in an attempt to understand whether this framework could be a useful tool for early-stage design. We make a three-fold contribution: First, we share an annotated dataset of architecturally relevant 3D point clouds of single-family homes. Next, we present and share the code for a framework and the results from training the 3D generative neural network. Finally, we discuss machine learning and creative work, including how practitioners feel about the emergence of these tools as mediators between incompleteness and precision in architectural design.

No abstract