Development of An Algorithm For Daylight Simulation and Prediction in Buildings Using A Deep Learning Approach
The Journal of Engineering Research · 15.06.2026 · 23 (1) · 1-11
DOI: 10.53540/1726-6742.1322 ↗Kurz gefasst
Fachartikel in der Zeitschrift The Journal of Engineering Research, online erschienen am 15.06.2026. Verfasst von Tolulope David Makanju und 2 weiteren Autor:innen.
Abstract (Originalsprache)
This paper presents an algorithm for estimating daylight in buildings based on the weather parameters. The algorithm was used to simulate daylight in buildings. The simulated result was used to develop a deep-learning predictive model that can help engineers estimate the daylight that will be appropriate during the design stage of a building. The deep learning approach for the prediction of daylight is based on the recurrent neural network Batch Normalisation-long short-term memory (BN-LSTM) model because daylight is a time-series occurrence. The accuracy of the model is less than 10% for the daylight lux level in the building, indicating that at an early stage of design, daylight in the building can be estimated based on sun angle, outdoor visible radiation, and the building parameters. Daylight can be predicted at an early stage in the design, and appropriate control strategies can be designed to control the artificial lighting to avoid glare and save energy in the workspace.
Zitieren
APA
Makanju, T. D., Oyeleye, M. O., & Famoriji, O. J. (2026). Development of An Algorithm For Daylight Simulation and Prediction in Buildings Using A Deep Learning Approach. The Journal of Engineering Research, 23(1), 1-11. https://doi.org/10.53540/1726-6742.1322
Vancouver
Makanju TD, Oyeleye MO, Famoriji OJ. Development of An Algorithm For Daylight Simulation and Prediction in Buildings Using A Deep Learning Approach. The Journal of Engineering Research. 2026;23(1):1-11. doi:10.53540/1726-6742.1322
BibTeX
@article{Makanju2026,
author = {Makanju, Tolulope David and Oyeleye, Matthew Olayinka and Famoriji, Oluwole John},
title = {Development of An Algorithm For Daylight Simulation and Prediction in Buildings Using A Deep Learning Approach},
journal = {The Journal of Engineering Research},
year = {2026},
volume = {23},
number = {1},
pages = {1--11},
doi = {10.53540/1726-6742.1322}
}
EUVIMED