Adaptive and scalable control of air-conditioning and mechanical ventilation (ACMV) systems
Overview
To facilitate energy-efficient buildings, three challenges are required to address regarding the control and optimization of HVAC systems: i) support the uncertain thermal demand of occupants; ii) overcome computational challenge at coordinating zone cooling demand and interactions, especially commercial HVAC system; iii) Lack indoor air quality (IAQ)management. To address such challenges, we have applied reinforcement learning (RL) and decentralized optimization techniques to achieve adaptive and scalable control HVAC systems for saving energy and providing an enhanced indoor environment. Particularly, we have developed a learning framework based on RL technique which can learn optimal stochastic control policy for HVAC (based on historical weather and occupancy data) in response to the uncertain thermal demand. Further, to overcome the computational burden of commercial HVAC system, we developed a scalable and computationally efficient method to optimize the control of commercial HVAC system using decentralized computation. The proposed decentralized method can embed the IAQ index in commercial HVAC control. Our main milestone at simulation level is that our method is scalable to 200-zone buildings and can achieve about 11% electricity cost reduction and enhanced indoor environment (thermal comfort and IAQ simultaneously) within seconds of computation.
Adaptive HVAC Control based on Reinforcement Learning (RL)
Scalable control of commercial HVAC system using decentralized computation
Maintaining Thermal Comfort and IAQ via Two-level Distributed HVAC Control
References
- Yu Yang, Seshadhri Srinivasan, Guoqiang Hu, and Costas J. Spanos, “Distributed Control of Multi-zone HVAC Systems Considering Indoor Air Quality,” IEEE Transactions on Control System Technology, 2020.
- Yu Yang, Guoqiang Hu, and Costas J. Spanos, “Stochastic Optimal Control of HVAC system for Energy-efficient Buildings,” conditionally accepted by IEEE Transactions on Control Systems Technology, 2020
- Yu Yang, Guoqiang Hu, and Costas J. Spanos, “HVAC Energy Cost Optimization for a Multi-zone Building via A Decentralized Approach,” IEEE Transactions on Automation Science and Engineering, 2020.