簡易檢索 / 詳目顯示

回結果列表
研究生： YUDHISTIRA CHANDRA BAYU
YUDHISTIRA CHANDRA BAYU
論文名稱： A Study of Control Strategies of Water Dispensers for Energy Conservation
A Study of Control Strategies of Water Dispensers for Energy Conservation
指導教授： 周碩彥
Shuo-Yan Chou
郭伯勳
Po-Hsun Kuo
口試委員: 鄭瑞光
Ray-Guang Cheng
學位類別： 碩士
Master
系所名稱： 管理學院 - 工業管理系
Department of Industrial Management
論文出版年： 2018
畢業學年度： 106
語文別： 英文
論文頁數： 68
中文關鍵詞： Water Dispenser 、Water Consumption Prediction 、Feature Selection 、Recurrent Neural Network (RNN) 、Sleep Mode
外文關鍵詞： Water Dispenser, Water Consumption Prediction, Feature Selection, Recurrent Neural Network (RNN), Sleep Mode
相關次數： 點閱：316  下載：7 
分享至:
查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報
  •  上一筆

    • Water is crucial thing for human to live. Most important part of water is as a source for human to fulfill the needs of its body by drinking it. However, in Taiwan, water from tap is not safe. Thus many of people in Taiwan use water dispenser to take a drink. Focus on water dispenser, it also consumes a lot of energy by repeating the process such as heating and cooling on its water tank even no one uses it. Having this kind of situation, we take a chance of it by attempting to predict water consumption in water dispenser and utilizing sleep mode feature on water dispenser to save energy from it. Doing prediction with Recurrent Neural Network, we also tried to maintain the service level of water dispenser by putting sleep mode on “right time” since on this mode dispenser does not do any process either heating or cooling Those previous two statements are the main objective of this research. Focusing on water dispenser in university environment surrounded by office and labs, internal data such as water taken from water dispenser and energy usage of water dispenser is collected by attaching sensors on water dispenser that sent data either for each minute or someone takes water from dispenser. Besides internal data, external data is acquired also. We do feature selection to all attributes and Savitzky Golay filtering to water consumption data. Result of feature selection shows: Working and Not Working Hour”, “Temperature”, “Dew Point”, “Clustering Result”, “Consumption Classification”, and “Seasonal Index” are relevant attributes that correlate with water consumption data. For RNN parameters, combination of parameters produces the lowest of error on testing set is: LSTM activation is hard sigmoid, recurrent activation is hard sigmoid, and dense activation is tanH. While on this combination, dataset which utilize filtering water consumption value and related attributes gives the lowest value among other. Based on that parameters combination and dataset. dispenser could save energy about 0.86% of a whole week usage and service level decrease 1.2%

    Water is crucial thing for human to live. Most important part of water is as a source for human to fulfill the needs of its body by drinking it. However, in Taiwan, water from tap is not safe. Thus many of people in Taiwan use water dispenser to take a drink. Focus on water dispenser, it also consumes a lot of energy by repeating the process such as heating and cooling on its water tank even no one uses it. Having this kind of situation, we take a chance of it by attempting to predict water consumption in water dispenser and utilizing sleep mode feature on water dispenser to save energy from it. Doing prediction with Recurrent Neural Network, we also tried to maintain the service level of water dispenser by putting sleep mode on “right time” since on this mode dispenser does not do any process either heating or cooling Those previous two statements are the main objective of this research. Focusing on water dispenser in university environment surrounded by office and labs, internal data such as water taken from water dispenser and energy usage of water dispenser is collected by attaching sensors on water dispenser that sent data either for each minute or someone takes water from dispenser. Besides internal data, external data is acquired also. We do feature selection to all attributes and Savitzky Golay filtering to water consumption data. Result of feature selection shows: Working and Not Working Hour”, “Temperature”, “Dew Point”, “Clustering Result”, “Consumption Classification”, and “Seasonal Index” are relevant attributes that correlate with water consumption data. For RNN parameters, combination of parameters produces the lowest of error on testing set is: LSTM activation is hard sigmoid, recurrent activation is hard sigmoid, and dense activation is tanH. While on this combination, dataset which utilize filtering water consumption value and related attributes gives the lowest value among other. Based on that parameters combination and dataset. dispenser could save energy about 0.86% of a whole week usage and service level decrease 1.2%

    ABSTRACT iv ACKNOWLEDGEMENTS v TABLE OF CONTENTS vi LIST OF FIGURES viii LIST OF TABLES ix 1 CHAPTER 1 INTRODUCTION 1 1.1 Background and Motivation 1 1.2 Objective 2 1.3 Limitations 2 1.4 Organization of Thesis 2 2 CHAPTER 2 LITERATURE REVIEW 4 2.1 Water Consumption Prediction and Forecasting 4 2.2 Recurrent Neural Network (RNN) 5 2.3 Savitzki Golay Filtering Method 7 2.4 K-Means Clustering 8 2.5 Stepwise Regression 9 2.6 Research Gap 10 3 CHAPTER 3 METHODOLOGY 11 3.1 Data Collection 11 3.2 Feature Selection and Data Filtering 13 3.3 RNN Modeling 14 3.4 Sleep Mode Procedure 18 4 CHAPTER 4 RESULT AND DISCUSSION 20 4.1 Data Description 20 4.2 Feature Selection Result 22 4.3 RNN Prediction 25 4.3.1 RNN on Water Consumption Dataset 26 4.3.2 RNN on Water Consumption and Other Attributes Dataset 28 4.3.3 RNN on Filtering Water Consumption Data and Other Attributes 30 4.4 Discussion on RNN Result 33 4.5 Sleep Mode Result 34 5 CHAPTER 5 CONCLUSION AND FUTURE RESEARCH 39 REFERENCES 41 APPENDIX A 45 APPENDIX B 46 APPENDIX C 47 APPENDIX D 48 APPENDIX E 49 APPENDIX F 50 APPENDIX G 51 APPENDIX H 53 APPENDIX I 55 APPENDIX J 57 APPENDIX K 58

    1. Chen, Y.-T., The Factors Affecting Electricity Consumption and the Consumption Characteristics in the Residential Sector—A Case Example of Taiwan. Sustainability, 2017. 9(8): p. 1484.
    2. Zhongren, C. and H. Yejun. A smart power saver based on composite switch and self-learning fuzzy control for drinking water dispenser. in 2016 IEEE International Conference on Power and Renewable Energy (ICPRE). 2016.
    3. Jinxiong, X., et al., A Design of Energy-Saving Drinking Dispenser Based on Fuzzy Memory Control [J]. Journal of Inspection and Quarantine, 2010. 3: p. 011.
    4. Young, R. Watts in a Drop of Water: Savings at the Water-energy Nexus. 2014. American Council for an Energy-Efficient Economy.
    5. Ahmed, K., P. Pylsy, and J. Kurnitski, Hourly consumption profiles of domestic hot water for different occupant groups in dwellings. Solar Energy, 2016. 137: p. 516-530.
    6. Haque, M.M., A. de Souza, and A. Rahman, Water Demand Modelling Using Independent Component Regression Technique. Water Resources Management, 2017. 31(1): p. 299-312.
    7. Hucheng, L., et al. Simulation and analysis of electric water heater load regulation model based on direct load control. in 2017 IEEE Conference on Energy Internet and Energy System Integration (EI2). 2017.
    8. Mostafavi, N., F. Gándara, and S. Hoque, Predicting water consumption from energy data: Modeling the residential energy and water nexus in the integrated urban metabolism analysis tool (IUMAT). Energy and Buildings, 2018. 158: p. 1683-1693.
    9. Ferrantelli, A., et al., Analytical modelling and prediction formulas for domestic hot water consumption in residential Finnish apartments. Energy and Buildings, 2017. 143: p. 53-60.
    10. Delorme-Costil, A. and J.J. Bezian. Forecasting Domestic Hot Water Demand in Residential House Using Artificial Neural Networks. in 2017 16th IEEE International Conference on Machine Learning and Applications (ICMLA). 2017.
    11. Jithish, J. and S. Sankaran. A neuro-fuzzy approach for domestic water usage prediction. in 2017 IEEE Region 10 Symposium (TENSYMP). 2017.
    12. Walker, D., et al., Forecasting Domestic Water Consumption from Smart Meter Readings Using Statistical Methods and Artificial Neural Networks. Procedia Engineering, 2015. 119: p. 1419-1428.
    13. Candelieri, A., D. Soldi, and F. Archetti, Short-term forecasting of hourly water consumption by using automatic metering readers data. Procedia Engineering, 2015. 119: p. 844-853.
    14. Firat, M., M.E. Turan, and M.A. Yurdusev, Comparative analysis of neural network techniques for predicting water consumption time series. Journal of Hydrology, 2010. 384(1): p. 46-51.
    15. Ren, Y., et al. Based on Process Neural Network Learning Algorithm for Prediction of Urban Water Consumption. in 2010 International Conference on Computing, Control and Industrial Engineering. 2010.
    16. Jordan, M., Serial order: a parallel distributed processing approach. Technical report, June 1985-March 1986. 1986, California Univ., San Diego, La Jolla (USA). Inst. for Cognitive Science.
    17. Lipton, Z.C., J. Berkowitz, and C. Elkan, A critical review of recurrent neural networks for sequence learning. arXiv preprint arXiv:1506.00019, 2015.
    18. Rahman, A., V. Srikumar, and A.D. Smith, Predicting electricity consumption for commercial and residential buildings using deep recurrent neural networks. Applied Energy, 2018. 212: p. 372-385.
    19. Shi, Z., H. Liang, and V. Dinavahi, Direct Interval Forecast of Uncertain Wind Power Based on Recurrent Neural Networks. IEEE Transactions on Sustainable Energy, 2017: p. 1-1.
    20. Balluff, S., J. Bendfeld, and S. Krauter. Short term wind and energy prediction for offshore wind farms using neural networks. in 2015 International Conference on Renewable Energy Research and Applications (ICRERA). 2015.
    21. Panigrahi, B.K., et al. Swarm, Evolutionary, and Memetic Computing. in Third International Conference SEMCCO. 2010. Springer.
    22. Savitzky, A. and M.J. Golay, Smoothing and differentiation of data by simplified least squares procedures. Analytical chemistry, 1964. 36(8): p. 1627-1639.
    23. Acharya, D., et al., Application of adaptive Savitzky–Golay filter for EEG signal processing. Perspectives in Science, 2016. 8: p. 677-679.
    24. Baba, K., L. Bahi, and L. Ouadif, Enhancing Geophysical Signals Through the Use of Savitzky-Golay filtering method. Geofísica Internacional, 2014. 53(4): p. 399-409.
    25. Chen, J., et al., A simple method for reconstructing a high-quality NDVI time-series data set based on the Savitzky–Golay filter. Remote Sensing of Environment, 2004. 91(3): p. 332-344.
    26. Aggarwal, C.C. and C.K. Reddy, Data clustering: algorithms and applications. 2013: CRC press.
    27. Jain, A.K., Data clustering: 50 years beyond K-means. Pattern Recognition Letters, 2010. 31(8): p. 651-666.
    28. Al-Wakeel, A. and J. Wu, K-means Based Cluster Analysis of Residential Smart Meter Measurements. Energy Procedia, 2016. 88: p. 754-760.
    29. Kusrini, K., Grouping of Retail Items by Using K-Means Clustering. Procedia Computer Science, 2015. 72: p. 495-502.
    30. Software, N.S. Stepwise Regression. 2011; Available from: https://ncss-wpengine.netdna-ssl.com/wp-content/themes/ncss/pdf/Procedures/NCSS/Stepwise_Regression.pdf.
    31. Tsai, C.-F., Feature selection in bankruptcy prediction. Knowledge-Based Systems, 2009. 22(2): p. 120-127.
    32. Shin, K.-S. and Y.-J. Lee, A genetic algorithm application in bankruptcy prediction modeling. Expert Systems with Applications, 2002. 23(3): p. 321-328.
    33. Shin, K.-S., T.S. Lee, and H.-j. Kim, An application of support vector machines in bankruptcy prediction model. Expert Systems with Applications, 2005. 28(1): p. 127-135.
    34. Crone, S.F. and N. Kourentzes, Feature selection for time series prediction – A combined filter and wrapper approach for neural networks. Neurocomputing, 2010. 73(10): p. 1923-1936.
    35. Keras. Keras Documentation. 2016; Available from: https://keras.io/optimizers/.
    36. NTUST. NTUST Calender Academic 2017/2018. 2017; Available from: http://www.academic.ntust.edu.tw/ezfiles/1/1001/img/1304/AcademicCalendarfor2017-2018.pdf.
    37. CWB. Observation Data Inquiry System. 2018; Available from: http://e-service.cwb.gov.tw/HistoryDataQuery/index.jsp.

    QR CODE