半導體產業為高度能源密集產業，晶圓製造需要仰賴電力，製程機台與廠務端用電約各占全廠55%、45%，而冰水系統約占廠務用電50%，因而為節能重點。但現有節能研究多以分析歷史資料，最佳化冰水系統負載，受限於資料提供限制，無法統合致平台中統一管理，因此少有讓管理者判斷當下之能源決策。而物聯網(Internet of Things, IoT)透過智慧感測器(Smart sensor)和智慧電表(Smart meter)收集資料，提高機器或生產線能源消耗的可視性。
本研究提出應用物聯網架構於能源管理，使用物聯網平台Thingworx模擬廠房冰水系統運作，透過收集四項設備冰水主機、冷卻水塔、冰水泵、冷卻水泵運轉時的實時數據(Real time data)，將案例廠房冰水系統導入，於物聯網平台中建構智慧監控系統模式，釐清何為模式實體關係、實體溝通、使用者介面。建立智慧監控系統的三個功能，一為檢視冰水系統及各設備之用電量，二為檢視各設備運轉參數的歷史數據，三為運轉異常所提供之警戒機制，作為記錄的部落格功能。
通過收集運轉參數，應用冰水系統能源轉換因素(Energy conversion factor, ECF)，利用演算法分析設備與冰水系統耗電量相關性。分析結果指出，在不同的運轉條件下，泵浦的耗電量相關性上升，因而推測為影響冰水系統耗電量主因，而非耗電量最高的冰水主機。2016群組分析指出設備組合(冷卻水塔、冷卻水泵及冰水泵)影響冰水系統耗電量，而2017年並無特定設備組合影響系統耗電量。
建立冰水系統能源消耗模型後，比較5種演算法的模型誤差，選擇誤差最小的整體學習演算法建立能源消耗模型。首先使用2016年10~12月數據建立預測模型預測2017同時段數據，誤差值約3%，後用2017年10~12月數據預測2018年1~3月數據，誤差值為9%，而此二模型決定係數都大於0.95，顯示足以預測大多數情形。未來若有感測數據需分析，可以此模型預測耗電量。

The chilled water system is further an energy saving hot spot because it consumes 50% of the facility electricity. However, the energy performance of the chilled water system is lacking in evaluating the overall energy efficiency. Using Internet of Things (IoT) technology can increase the visibility and awareness of energy consumption, also optimize the process of energy-awareness and decision-making.

This study proposes a IoT framework to use for energy management, by using the IoT platform Thingworx to simulate the operation of a chilled water system through the real time data of four kinds of equipment: chiller, cooling tower, chilled water pump and cooling water pump. It collected operational parameter data and used algorithms to analyze the correlation between the power consumptions of equipment energy conversion factor (ECF) and chilled water system.

The results indicate that under different operational choices, the correlation of the pump's power consumption increases, indicating that the pump, not the chiller, is the main cause of the power consumption of the chilled water system.

The ensemble learning algorithm was finally chosen to establish the energy consumption model. The data from October to December 2016 was used first to build a prediction model to predict 2017 values. The deviation is about 3%. Then using the data from October to December of 2017 to predict the values from January to March of 2018, the deviation is about 9%. Both models’ determination coefficients are greater than 0.95, which is sufficient to predict most situations. The derived model can predict power consumption and help make energy management decisions.

英文文獻
1. Bitondo, M. J. and Tozzi, M. J. (1999). Chiller Plant Controls: Multiple Chiller Controls. Carrier Corporation Syracuse, New York.
2. Crane, L. (2007). S23-The Guideline Basics. Presented at the SEMICON WEST, San Francisco Marriott.
3. Fadi, S., and Giovanni, M. (2015) “Energy management based on Internet of Things: practices and framework for adoption in production management.” Journal of Cleaner Production, Vol. 100, pp. 235~246.
4. Haller, S., Karnouskos, S., and Schroth, C., (2009). “The Internet of Things in an enterprise context.” In: Domingue, J., Fensel, D., Traverso, P. (Eds.), First Future Internet Symposium, FIS 2008 Vienna, Austria, September 29-30, Springer-Verlag, Berlin, Heidelberg, pp. 14~28.
5. Jovanović, B. and Filipović, J. (2016). “ISO 50001 standard-based energy management maturity model - proposal and validation in industry.” Journal of Cleaner Production, Vol. 112, pp. 2744~2755.
6. Lee, I. and Lee, K.(2015). “The Internet of Things (IoT)：Applications, investments, and challenges for enterprises.” Business Horizons, Vol. 58, pp. 431~440.
7. Lidia, P., Dimitrios, A., Georgios, M., and Ioannis, C. (2017). “Open source IoT meter devices for smart and energy-efficient school buildings.” HardwareX, Vol. 1, pp. 54~67.
8. Lin, Y. B., Lin Y. W., Hsiao, C. Y., and Wang, S. Y. (2017). “Location-based IoT applications on campus: The IoT talk approach”, Pervasive and Mobile Computing, vol. 40 ,pp. 660-673.
9. Lu, Y. (2017). “Industry 4.0: A survey on technologies, applications and open research issues”, Journal of Industrial Information Integration, vol. 6 , pp. 1~10.
10. Luigi, A., Antonio, I., and Giacomo, M., (2010), “The Internet of Things: A survey.” Computer Networks, Vol. 54, No. 15, pp. 2787~2805.
11. Ma, H. T., Zhou, W. Y., Lu, X. Y., Ding, Z., and Cao, Y., (2016). “Application of Low Cost Active and Passive Energy Saving Technologies in an Ultra-low Energy Consumption Building.” Journal of Energy Procedia, Vol. 88, pp. 807~813.
12. Min, W., Hong, S., and Musharraf, A. (2016), “An IoT-based energy-management platform for industrial facilities”, Applied Energy, Vol. 164, pp 607~619.
13. Miorandi, D., Sicari, S., Pellegrini, F.D., and Chlamtac, I. (2012). “Internet of Things：Vision, applications and research challenges”, Ad hoc networks, Vol. 10, issue 7, pp. 1497~1516.
14. Song, M. L., and Wang, S. H. (2017) “Measuring Environment-Biased Technological Progress Considering Energy Saving and Emission Reduction.” Process Safety and Environmental Protection, Vol. 116, pp. 745-753.
15. Taylor, S. T. (2002). “Primary-Only vs. Primary-Secondary Variable Flow Systems” ASHRAE Journal, No. 44, Vol. 2, pp. 25~29.
16. Vazan, P., Janikova, D., Tanuska, P., Kebisek, M., and Cervenanska, Z. (2017). “Using data mining methods for manufacturing process control”, IFAC-PapersOnLine, Vol. 50, No. 1, pp. 6178~6183.
17. Weber, H. (2011). “Accountability in the Internetof Things”, Computer Law & Security Review, Vol. 27, No. 2, pp. 133~138.
18. Weinert, N., Chiotellis, S., and Seliger, G. (2011). “Methodology for planning and operating energy-efficient production systems.” CIRP Annal. Vol. 60, No. 1, pp. 41~44.

中文文獻
1. 朱允杰(2014)，「應用回歸分析於冰水主機與冷卻水塔系統之最佳化控制策略」，國立台北科技大學能源與冷凍空調工程系碩士論文。
2. 周碩彥(2015)，物聯網發展趨勢展示內容，國立科學工藝博物館委託研究計畫。
3. 林暐倫(2016)，「高科技廠房水系統用電量與用水量之關係」，國立成功大學土木工程研究所碩士論文。
4. 河村雅人(2015)，圖解物聯網：感測器的架構與運用，碁峰出版社，台灣。
5. 孫鴻德(2010)，「台灣高科技廠節能技術之分析」，中原大學電機工程研究所碩士論文。
6. 張峻銓(2007)，「冰水主機與冷卻水塔群組最佳化運轉策略研究」，國立臺灣大學工學院機械工程學系碩士論文。
7. 許家榮(2008)，「降載下冷卻水塔之節水分析」，國立成功大學機械工程學系專班碩士論文。
8. 陳南蓁(2012)，語料庫建構技術，國家教育研究院研究報告。
9. 黃乃容(2017)，「半導體廠房冰水系統之能源績效指標」，國立成功大學土木工程研究所碩士論文。
10. 黃仲翊(2017)，「外部輸入非線性自動迴歸模型應用於冰水主機耗能分析」，國立台北科技大學能源與冷凍空調工程系碩士學位論文。
11. 黃學亮(2015)，統計學，五南出版社，台灣。
12. 楊禮源(2012)，「我國產業導入ISO 50001之探討」，國立台北科技大學環境工程與管理研究所碩士論文。
13. 經濟部能源局(2015)，第四次全國能源會議具體行動計畫，經濟部能源局，台北。
14. 經濟部能源局(2016)，中華民國104年能源統計手冊，經濟部能源局，台北。
15. 葉怡成(1999)，類神經網路模式應用與實作，第六版，儒林圖書股份有限公司，台北。
16. 蔡瑞益(2004)，「廠務空調系統冰水主機水側溫度節能實務分析」，中原大學機械工程學系碩士學位論文。
17. 蔡耀文(2014)，「廠務系統能源轉換係數(ECF)之建立及應用」，國立台北科技大學能源與冷凍空調工程系碩士學位論文。

網站資料
1. Dnuggets, K. (2014). “Where analytics, data mining, data science is applied.” http://www.kdnuggets.com/2014/12/where-analyticsdata- mining-data-science-applied.html. , accessed on 2018/5/18.
2. EE Times Taiwan (2017)，台灣IC產業2016年產值表現優於全球，http://www.eettaiwan.com/news/article/20170221NT21-Taiwan-IC-industry-revenue- 2016, acceesed on 2017/7/18.
3. PTC (2017a), “Thingworx Developer Builder Help Document”, https://developer.thingworx.com/, accessed on 2017/10/5.
4. PTC (2017b), “PTC university precision lms”, https://precisionlms.ptc.com/app/home?1, accessed on 2017/10/8.
5. PTC (2017c), “ThingWorx Platform Overview”, http//www.thingworx.com/platform, acceesed on 2017/8/14.
6. PTC (2018), “Thingworx Analytics Builder Help Document”, https://support.ptc.com/help/thingworx_hc/thingworx_analytics_7/index.html#page/thingworx_analytics_7%2Fanalytics-builder-overview.html%23, accessed on 2018/2/2.
7. RPubs (2018a)，R關聯式規則；決策樹，http://www.rpubs.com/skydome20/R-Note6-Apriori-DecisionTree, accessed on 2018/3/5.
8. RPubs (2018b)，R, Ensemble Learning，http://rpubs.com/skydome20/R-Note16-Ensemble_Learning, accessed on 2018/3/5.
9. RPubs (2018c)，R使用與資料預處理，http://rpubs.com/skydome20/Table, accessed on 2018/3/5.
10. RPubs (2018d)，遺漏值處理，http://www.rpubs.com/skydome20/R-Note10-Missing_Value, accessed on 2018/3/6.
11. 台灣區冷凍空調工程工業同業公會(2017)，冰水主機，http://www.hvac-net.org.tw/archive/files/250b1a.doc, accessed on 2018/4/15.
12. 台灣電力公司103年度空調運用技術研討會(2014)，「空調系統節能技術」， http://www.taipower.com.tw/UpFile/_userfiles/file/ 電力生活館/ 空調系統節能技術.pdf, accessed on 2017/6/12.
13. 台灣積體電路製造股份有限公司(2017) ，空調系統節能實務分享，https://www.tsmc-csr.com/upload_files/course/pdf/energy-saving-4.pdf, accessed on 2017/5/26.
14. 林尚鋒(2014)，趨勢眺望:四大組成要件到位，物聯網應用市場加速起飛，新通訊元件165期，http://www.2cm.com.tw/markettrend_content.asp?sn=1411050024, accessed on 2017/8/14.
15. 研華科技(2017)，研華與微軟共建物聯網開放平台，http://www2.advantech.tw/digital-logistics-digital-services/news/6b173618-d3a4-4a12-b993-8736f16b4f10/, accessed on 2017/9/11.
16. 經濟部能源局(2014)，全國能源會議，http://2014energy.tw/, accessed on 2018/4/5.
17. 經濟部能源局(2017)，新能源政策，http://www.moeaboe.gov.tw/ECW/populace/home/Home.aspx, accessed on 2018/4/2.