ABSTRACT
The urgency of climate change and the significant contribution of buildings to it have been leading the recent shift to performance-driven design methodologies in the Architecture, Engineering and Construction (AEC) industry. Despite the constant efforts for the incorporation of building performance simulation tools into current design workflows, they still suffer from demanding time frames due to computational demands and lack of domain knowledge by designers.
Real-time performance feedback is crucial at early design stages as well as to enable the exploration of unprecedentedly large design spaces that computational design methods have enabled. We propose a deep-learning approach for the prediction of environmental simulations which constitutes the basis for our intelligent framework for resilient design. Our results in predicting solar radiation simulations show a great level of accuracy and speed and serve as a proxy for further integration of environmental performance feedback in our intelligent design framework that allows informed sustainable design decisions.
