Green Building Innovation Cluster (GBIC) — IoT-Based Distributed MPC for Commercial HVAC
Research Scientist role at NTU Singapore (2016–2018), Green Building Innovation Cluster. PI: Prof. Rong Su. Project site.
Built and deployed an HVAC control platform in an occupied commercial building at NTU (S1-B1B office area, 85 zones, July 2018). The system implements a Smart-Token Based Scheduling Algorithm (Smart-TBSA) running as distributed MPC across a server plus Raspberry Pi edge fleet, and delivered 12–13% measured energy savings relative to the legacy building automation system without replacing the underlying BAS hardware.
The technical contribution was an IoT-compliant reference architecture (hardware, software, networking, and integration with legacy BAS and a cloud back-end) plus a transformation of the centralized optimization models into a distributed active-control technique that runs on low-cost edge devices. Results published in Applied Energy: An IoT Upgrade for Smart and Scalable HVAC Control in Commercial Buildings (Png, Srinivasan, Bekiroglu, Chaoyang, Su, Poolla — 162 citations).
The “smartness” in Smart-TBSA comes from three additions to the optimization model: recursive online identification of per-zone thermal dynamics, predicted occupant thermal sensation, and cost/comfort trade-offs solved per-zone rather than globally. A centralized legacy BAS becomes a decentralized, adaptive control layer deployed via low-cost IoT devices without hardware replacement.
Structure of the local and central MPC:
Local MPC:
Central MPC:
