智慧型節能技術：以感測網路自動偵測異常空調狀態之研究

現今許多建築物皆使用中央空調系統以達到冷卻效果。然而，許多使用者常常不瞭解自己使用空調之行為是否正確，亦不知自己的行為所造成的影響為何。在本篇論文，我們在台灣大學資訊系館建立一個無線感測網路，使用普及運算的技術收集各區域的環境資料。接著，我們對數據集進行特徵萃取、缺失資料與正規化，並使用機器學習演算法SVM對各區域建立空調模式之模型，用以辨識空調模式為關閉、送風或冷氣。實驗結果顯示各區域建立之分類器可達85%之準確率。收集資料方面，我們亦使用紅外線感測器偵測環境之有無人使用狀態。另外，為了得到資訊系館的熱舒適溫度範圍，我們使用ASHRAE發展的熱感尺度對環境使用者進行問卷調查。從問卷統計結果我們得到資訊系館的熱舒適溫度範圍為19:32◦C 至24:67◦C，中性溫度為21:99◦C。因此，由上述資訊可分析出各房間之空調使用情形屬正常或異常。其中，異常包括了「該環境無人使用，且存在區域之空調模式為開啟」(狀態0)、「該環境有人使用、存在區域之空調模式為冷氣且其室內溫度過冷」(狀態1) 與「該環境有人使用、存在區域之空調模式為冷氣且其室內溫度過熱」(狀態2)。我們從歷史資料(2011/01/01至2011/05/31)分析各種空調使用狀態，發現R336實驗室與R204電腦教室之異常使用較為嚴重，R336約佔40% 至50%，主要為狀態0與狀態2，且狀態2發生比例隨天氣變熱而增加。R204異常使用則高達80%至90%，主要為狀態0與狀態1。而其餘房間之異常使用比例較小，約為10%。此外，我們以時間分布觀點分析各種空調使用狀態，分成一周分布、平日分布與假日分布，藉以找出異常空調使用狀態之樣式。因此，若能妥善將這些分析結果回饋給使用者，使其修正異常使用空調的行為，便能達到節能減碳之目的。

關鍵字

感測網路；熱舒適溫度；空調模式；異常空調狀態；節能

並列摘要

Many buildings employ central air conditioning system for cooling. However, people in such buildings usually do not know the usage pattern and the impact of their behavior on usage of air conditioning. In this thesis, we establish a wireless sensor network in the NTU CSIE building to collect the environmental data. In air conditioning mode recognition, we preprocess the dataset, including feature extraction, missing data treatment and feature normalization. Then, we build the SVM model to recognize the air conditioning mode. The result shows the accuracy of each zone’s model is higher than 85%. We also use infrared motion sensors to get the occupancy state of the environment. In thermal comfort calculation, we conduct a questionnaire survey using ASHRAE thermal sensation scale to derive the thermal comfort range. The statistical result shows the thermal comfort range in NTU CSIE building is between 19:32 ◦C and 24:67 ◦C, and the neutral temperature is 21:99 ◦C. From all above information, we can determine whether the air conditioning state is abnormal or not. We define three abnormal air conditioning states. The first is that people is absent but the air conditioning is turned on. The second is that when people is in the room and the air conditioning is in cooling mode, the indoor temperature is below the lower bound of the thermal comfort zone. The third is like the second state but the indoor temperature is above the upper bound. We analyze the air conditioning states from January 2011 to May 2011. We find that R336 and R204 have a high percentage of abnormal state. The abnormal states of R336 are mainly the first state and the third state, and the third state happens more frequently when the weather bacomes warmer. The abnormal states of R204 are mainly the first state and the second state. Abnormal states in other experiment environments take 10% only. Furthermore, we also demonstrate the time distribution of each air conditioning state in weekdays and weekends. As a result, we believe that these analysis information can help people to be aware of their incorrect behavior in air conditioning and to achieve the goal of energy conservation.

並列關鍵字

sensor network ； thermal comfort ； A/C mode ； abnormal A/C state ； enery conservation

參考文獻

[24] S.-L. Lin, S.-M. Wei, C.-H. Huang, and W.-K. Chen. Thermal comfort study of an air-conditioned presentation room in taiwan. Journal of Architecture, 65:125–138, Sep. 2008.

[2] ASHRAE Standard 55: Thermal environmental conditions for human occupancy, American Society of Heating, Refrigeration and Air-conditioning Engineers, Atlanta, Georgia, USA, 2004.

[4] C.-C. Chang and C.-J. Lin. LIBSVM: a library for support vector machines, 2001. Software available at http://www.csie.ntu.edu.tw/ cjlin/libsvm.

[5] C.-S. Chang. Hybrid decision fusion system for hvac and lighting control. Master’s thesis, National Taiwan University, 2004.

[10] R. de Dear and G. S. Brager. The adaptive model of thermal comfort and energy conservation in the built environment. International Journal of Biometeorology, 45:100–108, 2001.

國際替代計量

智慧型節能技術：以感測網路自動偵測異常空調狀態之研究

全文下載

主題瀏覽