One for Many: Transfer Learning for Building HVAC Control

被引:56
作者
Xu, Shichao [1 ]
Wang, Yixuan [1 ]
Wang, Yanzhi [2 ]
O'Neill, Zheng [3 ]
Zhu, Qi [1 ]
机构
[1] Northwestern Univ, Evanston, IL 60201 USA
[2] Northeastern Univ, Boston, MA 02115 USA
[3] Texas A&M Univ, College Stn, TX USA
来源
PROCEEDINGS OF THE 7TH ACM INTERNATIONAL CONFERENCE ON SYSTEMS FOR ENERGY-EFFICIENT BUILDINGS, CITIES, AND TRANSPORTATION, BUILDSYS 2020 | 2020年
基金
美国国家科学基金会;
关键词
Smart Buildings; HVAC control; Data-driven; Deep reinforcement learning; Transfer learning; PREDICTIVE CONTROL; COMFORT;
D O I
10.1145/3408308.3427617
中图分类号
TP39 [计算机的应用];
学科分类号
081203 ; 0835 ;
摘要
The design of building heating, ventilation, and air conditioning (HVAC) system is critically important, as it accounts for around half of building energy consumption and directly affects occupant comfort, productivity, and health. Traditional HVAC control methods are typically based on creating explicit physical models for building thermal dynamics, which often require significant effort to develop and are difficult to achieve sufficient accuracy and efficiency for runtime building control and scalability for field implementations. Recently, deep reinforcement learning (DRL) has emerged as a promising data-driven method that provides good control performance without analyzing physical models at runtime. However, a major challenge to DRL (and many other data-driven learning methods) is the long training time it takes to reach the desired performance. In this work, we present a novel transfer learning based approach to overcome this challenge. Our approach can effectively transfer a DRL-based HVAC controller trained for the source building to a controller for the target building with minimal effort and improved performance, by decomposing the design of neural network controller into a transferable front-end network that captures building-agnostic behavior and a back-end network that can be efficiently trained for each specific building. We conducted experiments on a variety of transfer scenarios between buildings with different sizes, numbers of thermal zones, materials and layouts, air conditioner types, and ambient weather conditions. The experimental results demonstrated the effectiveness of our approach in significantly reducing the training time, energy cost, and temperature violations.
引用
收藏
页码:230 / 239
页数:10
相关论文
共 37 条
[1]  
Akkaya I, 2019, Arxiv, DOI arXiv:1910.07113
[2]  
Barrett Enda, 2015, Autonomous HVAC Control, A Reinforce- ment Learning Approach
[3]   Transfer learning with deep neural networks for model predictive control of HVAC and natural ventilation in smart buildings [J].
Chen, Yujiao ;
Tong, Zheming ;
Zheng, Yang ;
Samuelson, Holly ;
Norford, Leslie .
JOURNAL OF CLEANER PRODUCTION, 2020, 254
[4]   Experimental analysis of data-driven control for a building heating system [J].
Costanzo, G. T. ;
Iacovella, S. ;
Ruelens, F. ;
Leurs, T. ;
Claessens, B. J. .
SUSTAINABLE ENERGY GRIDS & NETWORKS, 2016, 6 :81-90
[5]  
Crawley DB, 2000, ASHRAE J, V42, P49
[6]  
Da Silva FL, 2019, J ARTIF INTELL RES, V64, P645
[7]   Using reinforcement learning to optimize occupant comfort and energy usage in HVAC systems [J].
Fazenda, Pedro ;
Veeramachaneni, Kalyan ;
Lima, Pedro ;
O'Reilly, Una-May .
JOURNAL OF AMBIENT INTELLIGENCE AND SMART ENVIRONMENTS, 2014, 6 (06) :675-690
[8]  
Gao GY, 2019, Arxiv, DOI arXiv:1901.04693
[9]   DeepComfort: Energy-Efficient Thermal Comfort Control in Buildings Via Reinforcement Learning [J].
Gao, Guanyu ;
Li, Jie ;
Wen, Yonggang .
IEEE INTERNET OF THINGS JOURNAL, 2020, 7 (09) :8472-8484
[10]  
Goodfellow I, 2016, ADAPT COMPUT MACH LE, P29