What is it about?
This paper introduces a custom machine learning (ML) tool that predicts daylighting performance instantly within a design environment. By training a neural network (NN) model on over 500 daylight simulations of office configurations, the model accurately estimates daylight autonomy (DA) values. The tool is seamlessly integrated into Grasshopper, a visual programming plugin for Rhino, making it easy for architects and designers to get daylight predictions without needing to write code or run time-consuming simulations. The training dataset was generated efficiently using high-performance computing (HPC), allowing thousands of parametric configurations to be simulated in parallel.
Featured Image
Why is it important?
Traditional daylighting simulations are accurate but slow, which discourages their use during early design stages when performance feedback is most valuable. Many architects simplify or skip simulations entirely to meet tight deadlines. This research offers a fast, accessible alternative by embedding AI into the same modeling tools architects already use, eliminating the need for coding or separate software. It supports energy-efficient design decisions early on—without sacrificing precision or usability.
Perspectives
From an architectural and sustainability standpoint, this work bridges the gap between performance simulation and real-world design workflows. It empowers non-technical users—like architects and students—with AI-enhanced tools that deliver instant feedback on daylight performance, encouraging broader use of performance-informed design. From a technical perspective, it demonstrates the effective coupling of neural networks with parametric design environments and high-performance computing to streamline otherwise intensive processes. Ultimately, this research contributes to a future where smart, sustainable design is standard—not optional.
Dr. Rania Labib
Read the Original
This page is a summary of: Integrating Machine Learning with Parametric Modeling Environments to Predict Building Daylighting Performance, IOP Conference Series Earth and Environmental Science, September 2022, Institute of Physics Publishing,
DOI: 10.1088/1755-1315/1085/1/012006.
You can read the full text:
Contributors
The following have contributed to this page
