Page 45 - IJAMD-2-1

International Journal of AI for
Materials and Design

ORIGINAL RESEARCH ARTICLE
Machine learning-based model predictive

control for multizone building automation: A
case study

Pradeep Shakya 1 , Shiva Sreenivasan 2 , Baskaran Krishnamoorthy 1 ,
2
Shiyu Yang 3 , and Man Pun Wan *
1 Sustainable Built Environments, Energy Research Institute @ Nanyang Technological University
(ERI@N), 1 Cleantech Loop, Singapore
2 Thermal and Fluids Division, School of Mechanical and Aerospace Engineering (MAE), Nanyang
Technological University (NTU), 50 Nanyang Avenue, Singapore
3 Department of Fluid Dynamics, Institute of High Performance Computing (IHPC), Agency for
Science, Technology, and Research (A*STAR), 1 Fusionopolis Way, Singapore

Abstract
In Singapore’s hot and humid climate, air-conditioning and mechanical ventilation
(ACMV) systems account for over 60% of commercial building energy consumption,
driving efforts to enhance energy efficiency through predictive control strategies
*Corresponding author: such as model predictive control (MPC) to overcome the limitations of conventional
Man Pun Wan reactive building automation systems. This paper presents a multizone MPC system
(mpwan@ntu.edu.sg) designed to optimize energy consumption and thermal comfort in a commercial
Citation: Shakya P, Sreenivasan S, building’s ACMV system in Singapore. The system was implemented in a multi-use
Krishnamoorthy B, Yang S, test building with real occupancy and a deployment area of approximately 850 m ,
2
Wan MP. Machine learning-based
model predictive control for partitioned into six learning zones, two office spaces, and three open spaces. The
multizone building automation: A ACMV system serving the deployment area consisted of two primary air-handling
case study. Int J AI Mater Design. units and 16 fan coil units, where chilled water was supplied to the cooling coils,
2025;2(1):39-53.
doi: 10.36922/ijamd.8161 and conditioned air was distributed through motorized diffusers. To facilitate
predictive control, data-driven thermal prediction models were developed for
Received: December 24, 2024 each zone using a non-linear autoregressive exogenous network with exogenous
Revised: February 7, 2025 inputs trained on historical data and disturbances. Thermal comfort optimization
Accepted: February 20, 2025 was guided by the predictive mean vote, which was targeted at 0, representing
thermal neutrality (as per ASHRAE 55 standards), and constrained within a
Published online: March 5, 2025 range of −0.5 – 0.5. Performance comparisons demonstrated that the MPC
Copyright: © 2025 Author(s). system achieved over 42% energy savings compared to the original thermostat-
This is an Open-Access article based control while enhancing thermal comfort. Despite its advantageous
distributed under the terms of the
Creative Commons Attribution control performances, challenges for large-scale deployment remain, including
License, permitting distribution, implementation costs, scalability, and model accuracy. Future work can address
and reproduction in any medium, these challenges by developing comfort models that leverage existing building
provided the original work is
properly cited. sensors.
Publisher’s Note: AccScience
Publishing remains neutral with Keywords: Model predictive control; Coordinated multisystem control; Energy saving;
regard to jurisdictional claims in
published maps and institutional Thermal comfort; Visual comfort; High-performance building
affiliations.

Volume 2 Issue 1 (2025) 39 doi: 10.36922/ijamd.8161