據台灣經濟部能源局在2016年非生產性質行業能源查核年報統計，空調設備在能源大用戶申報資料各類建築物總用電比例約為40%以上，為建築物電力消耗主因。再者，現行空調壓縮機性能評估使用設備性能係數(Coefficient of Performance，簡稱COP)或能源效率比值(Energy Efficiency Ratio，簡稱EER)兩項指標屬於靜態性能指標，僅反映空調設備在單一特定環境條件下的運轉效率，未能將空調設備實際使用情況之負載變化對壓縮機起閉的影響呈現出來。為分析在不同環境條件下對空調壓縮機的使用效率及設備異常診斷，增進電能使用效率，達節能減碳成效，本文應用溫濕度感測器及智慧電表，構築即時監測及歷史資料庫系統，以五種便捷資料探勘技術分析設備過往數據建立預測模型，使用統計方法評估設備異常狀態。結果顯示，誤差率最小之數值分析方法為支援向量機(Support Vector Machine，SVM)，相關係數(Correlation Coefficient, R)達0.898，平均絕對值誤差率(Mean Absolute Percentage Error, MAPE)值為1.643%，絕對誤差(Mean Absolute Error, MAE)為0.004，均方根誤差(Root Mean Square Error, RMSE)為0.006。另將環境因素對COP值之影響轉化成計量模型，比對不同年度之同期數據，並以統計方法有效判別設備異常狀態之界定。本研究以即時監測及歷史資料庫系統結合便捷方法人工智慧模型，評估壓縮機效能，使用者得以掌握設備性能趨勢，達撙節成本效益。本研究成果亦能將系統化建構與資料分析流程技術移轉至節能科技產業。

According to Bureau of Energy, Ministry of Economic Affairs in Taiwan, air-conditioning is more than 40% of total electricity for the non-productive industries building power consumption in 2016. In addition, currently using Coefficient of Performance (COP) or Energy Efficiency Ratio (EER) to evaluate compressor performance only reflects air conditioning in specific environmental conditions. It does not demonstrate the actual load change impact on compressor performance. In order to analyze the efficiency of air conditioning compressor and the abnormal diagnosis of equipment in different environmental conditions, this paper uses temperature sensors, humidity sensors, smart meter to build real-time monitoring and historical database system. The prediction models were created based on previous data of five data mining techniques. The abnormal status of the compressor was evaluated by statistical method. The results show that the numerical analysis method with minimum error rates is Support Vector Machine (SVM), in which Correlation Coefficient (R) is 0.898, Mean Absolute Percentage Error (MAPE) is 1.643%, Mean Absolute Error (MAE) is 0.004, and Root Mean Square Error (RMSE) is 0.006. The influence of environmental factors on COP value is transformed into the measurement model, compared with the same period in different years. The statistical method was then used to determine the abnormal status of the compressor. This study contributes to the domain knowledge by proposing an effective artificial intelligence approach for predicting the compressor performance in different environmental conditions.

[1] The Elements of Statistical Learning, Technometrics 45 (3) (2003) 267-268.
[2] T. Al-Saba, I. El-Amin, Artificial neural networks as applied to long-term demand forecasting, Artificial Intelligence in Engineering 13 (2) (1999) 189-197.
[3] E. Arcaklioğlu, A. Erişen, R. Yilmaz, Artificial neural network analysis of heat pumps using refrigerant mixtures, Energy Conversion and Management 45 (11–12) (2004) 1917-1929.
[4] M.B. Bailey, The Design and Viability of a Probabilistic Fault Detection and Diagnosis Method for Vapor Compression Cycle Equipment, University of Colorado, 1998.
[5] H. Bechtler, M.W. Browne, P.K. Bansal, V. Kecman, New approach to dynamic modelling of vapour-compression liquid chillers: artificial neural networks, Applied Thermal Engineering 21 (9) (2001) 941-953.
[6] M. Brown, C. Barrington-Leigh, Z. Brown †, Kernel regression for real-time building energy analysis, Journal of Building Performance Simulation 5 (4) (2012) 263-276.
[7] R. Chakraborty, A. Maitra, Retrieval of atmospheric properties with radiometric measurements using neural network, Atmospheric Research 181 (2016) 124-132.
[8] J.-S. Chou, Y.-C. Hsu, L.-T. Lin, Smart meter monitoring and data mining techniques for predicting refrigeration system performance, Expert Systems with Applications 41 (5) (2014) 2144-2156.
[9] J.-S. Chou, A.S. Telaga, W.K. Chong, G.E. Gibson, Early-warning application for real-time detection of energy consumption anomalies in buildings, Journal of Cleaner Production 149 (2017) 711-722.
[10] M.C. Comstock, B. Chen, J.E. Braun, R. Bernhard, P.U.S.o.M. Engineering, R.W.H. Laboratories, R. American Society of Heating, A.-C. Engineers, Literature Review for Application of Fault Detection and Diagnostic Methods to Vapor Compression Cooling Equipment, Purdue University, 1999.
[11] M. Dabiri, M. Ghafouri, H.R.R. Raftar, T. Björk, Neural network-based assessment of the stress concentration factor in a T-welded joint, Journal of Constructional Steel Research 128 (2017) 567-578.
[12] U.M. Fayyad, G. Piatetsky-Shapiro, P. Smyth, From data mining to knowledge discovery: an overview, in: M.F. Usama, P.-S. Gregory, S. Padhraic, U. Ramasamy (Eds.), Advances in knowledge discovery and data mining, American Association for Artificial Intelligence, 1996, pp. 1-34.
[13] R. Ghedamsi, N. Settou, A. Gouareh, A. Khamouli, N. Saifi, B. Recioui, B. Dokkar, Modeling and forecasting energy consumption for residential buildings in Algeria using bottom-up approach, Energy and Buildings 121 (2016) 309-317.
[14] M. Goodarzi, R. Jensen, Y. Vander Heyden, QSRR modeling for diverse drugs using different feature selection methods coupled with linear and nonlinear regressions, Journal of Chromatography B 910 (2012) 84-94.
[15] H. Hahn, S. Meyer-Nieberg, S. Pickl, Electric load forecasting methods: Tools for decision making, European Journal of Operational Research 199 (3) (2009) 902-907.
[16] G.B. Humphrey, H.R. Maier, W. Wu, N.J. Mount, G.C. Dandy, R.J. Abrahart, C.W. Dawson, Improved validation framework and R-package for artificial neural network models, Environmental Modelling & Software 92 (2017) 82-106.
[17] G.F. Hundy, A.R. Trott, T.C. Welch, Chapter 2 - The Refrigeration Cycle, Refrigeration, Air Conditioning and Heat Pumps (Fifth Edition), Butterworth-Heinemann, 2016, pp. 19-39.
[18] R. Isermann, Process fault detection based on modeling and estimation methods—A survey, Automatica 20 (4) (1984) 387-404.
[19] R.R. Kallu, E.R. Keffeler, R.J. Watters, A. Agharazi, Development of a multivariate empirical model for predicting weak rock mass modulus, International Journal of Mining Science and Technology 25 (4) (2015) 545-552.
[20] W.-K. Kao, H.-M. Chen, J.-S. Chou, Aseismic ability estimation of school building using predictive data mining models, Expert Systems with Applications 38 (8) (2011) 10252-10263.
[21] D.-S. Kapetanakis, E. Mangina, E.H. Ridouane, K. Kouramas, D. Finn, Selection of Input Variables for a Thermal Load Prediction Model, Energy Procedia 78 (2015) 3001-3006.
[22] M.-T. Ke, C.-H. Yeh, C.-J. Su, Cloud computing platform for real-time measurement and verification of energy performance, Applied Energy 188 (2017) 497-507.
[23] T.R. Kiran, S.P.S. Rajput, An effectiveness model for an indirect evaporative cooling (IEC) system: Comparison of artificial neural networks (ANN), adaptive neuro-fuzzy inference system (ANFIS) and fuzzy inference system (FIS) approach, Applied Soft Computing 11 (4) (2011) 3525-3533.
[24] D. Kolokotsa, D. Rovas, E. Kosmatopoulos, K. Kalaitzakis, A roadmap towards intelligent net zero- and positive-energy buildings, Solar Energy 85 (12) (2011) 3067-3084.
[25] E.U. Küçüksille, R. Selbaş, A. Şencan, Data mining techniques for thermophysical properties of refrigerants, Energy Conversion and Management 50 (2) (2009) 399-412.
[26] D. Liu, Y. Yuan, S. Liao, Artificial neural network vs. nonlinear regression for gold content estimation in pyrometallurgy, Expert Systems with Applications 36 (7) (2009) 10397-10400.
[27] P. Mandal, T. Senjyu, T. Funabashi, Neural networks approach to forecast several hour ahead electricity prices and loads in deregulated market, Energy Conversion and Management 47 (15–16) (2006) 2128-2142.
[28] M. Molina-Solana, M. Ros, M.D. Ruiz, J. Gómez-Romero, M.J. Martin-Bautista, Data science for building energy management: A review, Renewable and Sustainable Energy Reviews 70 (2017) 598-609.
[29] T. Orenstein, Z. Kohavi, I. Pomeranz, An optimal algorithm for cycle breaking in directed graphs, Journal of Electronic Testing 7 (1) (1995) 71-81.
[30] B. Pitt, Applications of data mining techniques to electric load profiling, Dept. Elect. Electron. Eng (2000) 1-197.
[31] M.R. Rezaee, A. Jafari, E. Kazemzadeh, Relationships between permeability, porosity and pore throat size in carbonate rocks using regression analysis and neural networks, Journal of Geophysics and Engineering 3 (4) (2006) 370.
[32] T.M. Rossi, J.E. Braun, A statistical, rule-based fault detection and diagnostic method for vapor compression air conditioners, HVAC&R Research 3 (1) (1997) 19-37.
[33] A.Ş. Şahin, Performance analysis of single-stage refrigeration system with internal heat exchanger using neural network and neuro-fuzzy, Renewable Energy 36 (10) (2011) 2747-2752.
[34] D.J. Swider, A comparison of empirically based steady-state models for vapor-compression liquid chillers, Applied Thermal Engineering 23 (5) (2003) 539-556.
[35] M.B. Swift, Comparison of Confidence Intervals for a Poisson Mean – Further Considerations, Communications in Statistics - Theory and Methods 38 (5) (2009) 748-759.
[36] R. Timofeev, Classification and regression trees (CART) theory and applications, Humboldt University, Berlin, 2004.
[37] P.M. Van Every, M. Rodriguez, C.B. Jones, A.A. Mammoli, M. Martínez-Ramón, Advanced detection of HVAC faults using unsupervised SVM novelty detection and Gaussian process models, Energy and Buildings 149 (2017) 216-224.
[38] V.N. Vapnik, V. Vapnik, Statistical learning theory, Wiley New York, 1998.
[39] I.-H. Yang, M.-S. Yeo, K.-W. Kim, Application of artificial neural network to predict the optimal start time for heating system in building, Energy Conversion and Management 44 (17) (2003) 2791-2809.
[40] A. Zendehboudi, X. Li, B. Wang, Utilization of ANN and ANFIS models to predict variable speed scroll compressor with vapor injection, International Journal of Refrigeration 74 (2017) 475-487.
[41] 唐文祥, 吳東錦, 江昭佑, 陳煥, 應用於需量反應之家庭能源管理系統, 電腦與通訊 (153) (2013) 129-137.
[42] 徐筱琪, 廖建順, 我國變頻空調機之SEER發展現況介紹, 工研院能源與環境研究所冷凍空調&熱交換 (83) (2008) 1-8.
[43] 李玉生, 林憲德, 呂罡銘, 王育忠, 黃光佑, 建築生命週期二氧化碳排放量評估之研究(二)-建築空調設備二氧化碳排放量解析, in: 內政部建築研究所 (Ed.), 2007, p. 78.
[44] 經濟部能源局, 2016年非生產性質行業能源查核年報, 2017, p. 19