運用Deep Q Learning對教室熱舒適度與空氣品質及能耗之研究

Optimization of the Thermal Comfort, Air Quality and Energy Consumption within classroom environment using Deep Q Learning

陳弈安(Yi-An Chen)；余冠亨(Kuan-Heng Yu)；呂光欽(Kuang-Chin Lu)；廖仁忠(Jen-Chung Liao)；廖國凱(Kuo-Kai Liao)；吳武杰(Wu-Chieh Wu)；許文震(Wen-Jenn Sheu)；王啟川(Chi-Chuan Wang)

PMV ； 二氧化碳 ； Reinforcement Learning ； Deep Q-Learning ； PMV ； Air quality ； Reinforcement learning ； Deep Q-learning

冷凍空調&能源科技

115期（2019 / 05 / 10）

31 - 42

繁體中文;英文

本研究透過強化學習中的Deep Q-Learning建立了一套空調及風扇的控制演算法，能夠對室內的空氣品質、熱舒適度以及能源消耗做平衡，以實驗方式對該演算法進行測試，並探討其效能。實驗場域為交通大學工程五館132教室。研究結果顯示，未控制情況，人員使用132教室時大部分為開門關窗和空調定溫25度。教室人數超過30人時二氧化碳濃度就有機會超過1000ppm，使人不舒適。DQN Agent在維持PMV為舒適的情況下，比定溫25度節省13.7-45.0％的能源消耗。一般情況下建議冷氣定溫26度，比定溫25度節省6-18.7％的能源消耗。

This study develops a control algorithm for air conditioning and ventilation fans through Deep Q Learning, which balances indoor air quality, thermal comfort and energy consumption. The proposed DQN agent is tested and explored through experiments. The experiment was conducted in classroom 132 of Engineering building 5 in Nation Chiao Tung University. This research project is collaborated with Chunghwa Telecom and the Prius Wi-Fi system used in this study is provided by the corporation. We use carbon dioxide concentration level as indoor air quality index: concentration below 800 ppm is comfortable; 800-1000 ppm as acceptable, and over 1000 ppm is unacceptable. For thermal comfort, PMV is used as index. PMVs between -0.5 and 0.5 are comfort. For energy consumption, the sum of electricity used by the AC and fans during the 50-minute-class are included. Different experimental setup and parameters are considered to find out which control strategy has the greatest energy savings while maintaining the same comfort level. Most uncontrolled situations in room 132 are operated with closing and door opening with AC setting at 25 °C. Experimental result shows that when the number of students in classroom exceeds 30, the carbon dioxide concentration may exceed 1000 ppm, making people uncomfortable. DQN agent is able to save 13.7-45.0% energy consumption compared to constant temperature 25 °C while maintaining PMV at comfort range. For average situation it is recommended setting AC temperature at 26 °C, which saving 6-18.7% energy consumption than a fixed setting temperature of 25 °C.

1. Lin, S.,Wei, S.,Huang, C.,Chen, W. J. J.o. A.(2008).Thermal comfort study of an air conditioned presentation room in Taiwan.建築學報,65,125-138.
   連結：
2. (2014).能源產業技術白皮書.經濟部能源局.
3. Auliciems, A. J. I. J. o. B.(1981).Towards a psycho-physiological model of thermal perception.International Journal of Biometeorology,25(2),109-122.
4. Charles, K. E.(2003).,未出版
5. Dalamagkidis, K.,Kolokotsa, D.,Kalaitzakis, K.,Stavrakakis, G. S. J. B.(2007).Reinforcement learning for energy conservation and comfort in buildings.Building and Environment,42(7),2686-2698.
6. De Dear, R.,Brager, G. S. J. I. j. o. b.(2001).The adaptive model of thermal comfort and energy conservation in the built environment.International Journal of Biometeorology,45(2),100-108.
7. Fanger, P. O.(1970).Thermal comfort. Analysis and applications in environmental engineering.
8. Ferreira, P.,Ruano, A.,Silva, S.,Conceicao, E. J. E.(2012).Neural networks based predictive control for thermal comfort and energy savings in public buildings.Energy and Buildings,55,238-251.
9. Heating, A. S. o., Refrigerating, and Engineers, A.-C.(2004).Thermal environmental conditions for human occupancy.ANSI/ASHRAE Standard
10. Hwang, R.-L.,Lin, T.-P.,Kuo, N.-J. J. E.(2006).Field experiments on thermal comfort in campus classrooms in Taiwan.Energy and Buildings,38(1),53-62.
11. Intille, S. S.,Larson, K.,Tapia, E. M.,Beaudin, J. S.,Kaushik, P.,Nawyn, J.,Rockinson, R..Using a live-in laboratory for ubiquitous computing research.Proc. International Conference on Pervasive Computing
12. Kim, Y.,Schmid, T.,Charbiwala, Z. M.,Srivastava, M. B..ViridiScope: design and implementation of a fine grained power monitoring system for homes.Proc. Proceedings of the 11th international conference on Ubiquitous computing
13. Kolokotsa, D.,Gobakis, K.,Papantoniou, S.,Georgatou, C.,Kampelis, N.,Kalaitzakis, K.,Vasilakopoulou, K.,Santamouris, M. J. E.(2016).Development of a web based energy management system for University Campuses: The CAMP-IT platform.Energy and Buildings,123,119-135.
14. Kolokotsa, D.,Pouliezos, A.,Stavrakakis, G.,Lazos, C. J. B.(2009).Predictive control techniques for energy and indoor environmental quality management in buildings.Building and Environment,44(9),1850-1863.
15. Kolokotsa, D.,Stavrakakis, G.,Kalaitzakis, K.,Agoris, D. J. E. A. o. A. I.(2002).Genetic algorithms optimized fuzzy controller for the indoor environmental management in buildings implemented using PLC and local operating networks.Engineering Applications of Artificial Intelligence,15(5),417-428.
16. Ku, K.,Liaw, J.,Tsai, M.,Liu, T. J. I. T. A. S.(2015).Automatic Control System for Thermal Comfort Based on Predicted Mean Vote and Energy Saving.IEEE Transactions on Automation Science and Engineering,12(1),378-383.
17. Mnih, V.,Kavukcuoglu, K.,Silver, D.,Rusu, A. A.,Veness, J.,Bellemare, M. G.,Graves, A.,Riedmiller, M.,Fidjeland, A. K.,Ostrovski, G. J. N.(2015).Human-level control through deep reinforcement learning.Nature,518(7540),529.
18. Mui, K. W. H.,Chan, W. T. D. J. B.(2003).Adaptive comfort temperature model of air-conditioned building in Hong Kong.Building and Environment,38(6),837-852.
19. Standard, A. J. A. S. o. H., Refrigerating, and Air-Conditioning Engineers, I.(2010).Ventilation for Acceptable Indoor Air Quality.ANSI/ASHRAE Standard
20. Wei, T.,Wang, Y.,Zhu, Q.(2017).Deep reinforcement learning for building hvac control.Proc. Proceedings of the 54th Annual Design Automation Conference
21. Wong, N. H.,Khoo, S. S. J. E.(2003).Thermal comfort in classrooms in the tropics.Energy and Buildings,35(4),337-351.
22. 林貫文(2007)。國立成功大學工業設計研究所。