What is it about?
This paper introduces the novel application of Physics-Informed Neural Networks (PINNs) within the built environment. By leveraging the principles of the Radiative Transfer Equation (RTE) and the diffusion equation, the research presents an approach to integrating physical laws with machine learning (ML) for the precise and instant prediction of daylighting performance in buildings. The architecture of both PINNs model was significantly enhanced through Bayesian Optimization executed on a High-Performance Computing (HPC) environment, which reduced the computational time from an estimated 128 h to approximately 16 h per model. This process facilitated the efficient execution of 5000 optimization iterations for each model. For the RTE-based PINN, the best Mean Square Error (MSE) and Mean Absolute Error (MAE) were achieved in iterations 2416 and 2930, respectively. As for the Diffusion-based PINN, the best MSE and MAE were achieved in iterations 4384 and 1300, respectively. The final PINNs models showed high accuracy; the RTE-based model's MAE of approximately 0.70 and an MSE of 1.59, alongside the diffusion-based model's lower MAE of 0.55, although with a higher MSE of 17.36. Moreover, the paper presents an approach of integration of PINNs into 3D modeling environments which facilitates access to PINNs by the non-technical user.
Featured Image
Why is it important?
This research is important because it addresses a major bottleneck in sustainable building design: the time and computational intensity required for accurate daylighting simulations. Traditional simulation methods can be too slow for early design stages, limiting their practical use. By integrating Physics-Informed Neural Networks (PINNs) with physical laws and accelerating them using Bayesian optimization and high-performance computing, this work makes it possible to generate fast, accurate daylighting predictions. This empowers architects and engineers to make informed design decisions earlier in the process, ultimately leading to more energy-efficient, comfortable, and human-centric buildings—without needing advanced technical expertise or costly software tools.
Perspectives
This project is important to me because it bridges my passion for sustainable architecture with cutting-edge technology. As someone who has spent years teaching and researching building performance, I’ve seen how the slow, complex nature of traditional simulations can hold back creative and timely decision-making. By applying physics-informed neural networks, I wanted to make high-accuracy daylighting analysis faster and more accessible—especially for architects and designers who may not have a technical background. My goal is to empower early design decisions that improve energy efficiency and occupant well-being without sacrificing speed or usability.
Dr. Rania Labib
Read the Original
This page is a summary of: Utilizing physics-informed neural networks to advance daylighting simulations in buildings, Journal of Building Engineering, April 2025, Elsevier,
DOI: 10.1016/j.jobe.2024.111726.
You can read the full text:
Contributors
The following have contributed to this page
