跳到主要內容

臺灣博碩士論文加值系統

(44.221.73.157) 您好!臺灣時間:2024/06/20 10:49
字體大小: 字級放大   字級縮小   預設字形  
回查詢結果 :::

詳目顯示

我願授權國圖
: 
twitterline
研究生:蔣貴君
研究生(外文):KUEI CHUN CHIANG
論文名稱:基於非監督式學習策略之住宅用戶異常用電行為診斷模型
論文名稱(外文):Anomaly Detection for Power Usage Behavior of Domestic Household Based on Unsupervised Learning Strategy
指導教授:張耀中、蔡家緯
學位類別:碩士
校院名稱:國立臺東大學
系所名稱:資訊工程學系
學門:工程學門
學類:電資工程學類
論文種類:學術論文
論文出版年:2023
畢業學年度:111
語文別:中文
論文頁數:56
中文關鍵詞:住宅用戶異常用電行為異常檢測自動編碼器集成式學習單類分類器
外文關鍵詞:resident anomaly power consumption behavioranomaly detectionautoencoderensemble learningand one-class classifier
相關次數:
  • 被引用被引用:0
  • 點閱點閱:83
  • 評分評分:
  • 下載下載:0
  • 收藏至我的研究室書目清單書目收藏:0
全球暖化成為事實,節能減碳風氣正熾,如何提升節能的資訊技術蔚為風潮,而高投資成本與時效都緩不濟急。有鑑於全球智慧電表(Smart meter)日趨普及,除了管理方便,更能深入分析能源電力消費,達到節能效果。因此,在獲取住宅用戶用電資料的成本便能快速下降,透過機器學習技術分析用電量小且用電行為複雜的住宅用戶,用電行為技術則成節能減碳措施中重要研究議題,而辨識用戶用電行為異常,為其中一項重要課題。目前針對住宅用戶異常用電行為辨識多直接使用由感測器或電表所直接蒐集的低階特徵值來進行模型建置,然而,因低階特徵值維度較大且可能含較多冗餘訊息,容易降低模型訓練效能。故本研究使用深度學習技術-自動編碼器(Autoencoder)技術來萃取住宅用戶高階用電資訊,以提升異常診斷模型所找出異常事件的效能。除此之外,為了提升模型泛化能力,本研究採取非監督學習(Unsupervised learning)方法,提出集成式學習(Ensemble learning)架構與策略,整合二個單類分類器(One-class classifier)模型,並透過5戶實際住宅用戶一年度用電資料進行模型建置與實驗,具以證明提出技術與架構之可行性。
As global warming intensifies, energy conservation and carbon reduction have become major priorities in many areas. The widespread deployment of advanced metering infrastructure and smart meters, which aid in energy conservation through electricity consumption behavior analysis, has significantly lowered the cost of obtaining data on residential electricity usage. This change promotes the analysis of residential electricity consumption, which features both small and complicated consumption behaviors, using machine learning to become an important research topic among various energy saving and carbon reduction measures. A main subtopic of this subject is the identification of abnormal electricity consumption behaviors. At present, anomaly detection is typically realized using models based on low-level features directly collected by sensors and electricity meters. However, owing to the significant number of dimensions and large amount of redundant information in these low-level features, the training efficiency of the model is often low. To resolve this, this study adopts an autoencoder, which is a deep learning technology, to extract the high-level electricity consumption information of residential users to improve the performance of anomaly detection models. In addition, for improving the generalization ability of the anomaly detection model, this thesis proposes an ensemble learning architecture and manner under the unsupervised learning strategy to integrate two one-class classifier algorithms. This thesis proves the feasibility of the proposed anomaly detection of resident power consumption behaviors by using the experiments, in which the annual electricity consumption data of 5 actual residential users are used to analyzed and built the models.
中文摘要 i
英文摘要 ii
表目錄 iv
圖目錄 v
第壹章 緒論 1
第一節研究機動 1
第二節研究方向與貢獻 4
第三節章節提要 4
第貳章 背景介紹與文獻探討 6
第一節行為科學理論與其於節能上應用 6
第二節異常診斷簡介 9
第三節用電行為之異常診斷技術 17
第參章 研究方法 20
第一節總體架構 20
第二節模型架構與訓練流程 21
第三節集成學習策略 25
第肆章 實驗結果分析與評估 28
第一節系統開發環境與實驗資料 28
第二節高階特徵值實驗與效能分析 29
第三節異常診斷模型實驗與效能分析 43
第伍章 結論與未來研究 51
第一節結論 51
第二節未來展望 52
參考文獻 54
[1]Sanstad, A. H. and R. B. Howarth, 1994. ‘Normal’ markets, market imperfections and energy efficiency. Energy Policy, 22(10), 811–818.
[2]Net Zero by 2050 - A Roadmap for the Global Energy Sector, https://iea.blob.core.windows.net/assets/4719e321-6d3d-41a2-bd6b-461ad2f850a8/NetZeroby2050-ARoadmapfortheGlobalEnergySector.pdf
[3]經濟部能源局「主動式節能技術與示範應用計畫」
[4]蔡家緯、洪永杰、蔣貴君、趙寬居, "用戶電能資料驅動服務之國際趨勢與發展方向",臺灣能源期刊第五卷,第四期,107年12月。
[5]"Empowering Cities for a Net Zero Future: Unlocking resilient, smart, sustainable urban energy systems", International Energy Agency (IEA), 2021年7月。
[6]劉書安,蔣貴君,洪永杰,趙寬居, 基於行為科學之節能系統設計與驗證,2022年5月。
[7]A.Sudarshan, “Nudges in the marketplace: The response of household electricity consumption to information and monetary incentives,” Journal of Economic Behavior & Organization, Vol. 134, Feb 2017
[8]K. Koroleva, and M. Melenhorst, “Designing an integrated socio-technical behaviour change system for energy saving,” Energy Informatics, 2019
[9]Edgeworth, F. Y. 1887. On discordant observations. Philosophical Magazine 23, 5, 364-375.
[10]Varun Chandola, Arindam Banerjee, Vireshwar Kumar, “Anomaly detection: A survey”, ACM Computing Surveys, September 2009.
[11]Wolpert, D.H., Macready, W.G. (1995), No Free Lunch Theorems for Search, Technical Report SFI-TR-95-02-010 (Santa Fe Institute).
[12]Wolpert, D.H., Macready, W.G. (1997), "No Free Lunch Theorems for Optimization," IEEE Transactions on Evolutionary Computation 1, 67.
[13]Pimentel, Marco AF, et al. "A review of novelty detection." Signal processing 99 (2014): 215-249.
[14]H. Solberg, A. Lahti, Detection of outliers in reference distributions: performance of Horn0s algorithm, Clin. Chem. 51 (12) (2005) 2326–2332.
[15]Wang, Xinlin, and Sung-Hoon Ahn. "Real-time prediction and anomaly detection of electrical load in a residential community." Applied Energy 259 (2020): 114145.
[16]Zhang Y, Chen WW, Black J. Anomaly detection in premise energy consumption data. In: 2011 IEEE power and energy society general meeting; 2011.
[17]Chou JS, Telaga AS. Real-time detection of anomalous power consumption. Renew Sustain Energy Rev 2014;33:400–11.
[18]Luo J, Hong T, Yue M. Real-time anomaly detection for very short-term load forecasting. Clean Energy 2018;6:235–43.
[19]Fenza G, Gallo M, Loia Vi. Drift-aware methodology for anomaly detection in smart grid. IEEE Access 2019;7:9645–57.
[20]Jokar P, Arianpoo N, Leung VCM. Electricity theft detection in AMI using customers’ consumption patterns. IEEE Trans Smart Grid 2016;7:216–26.
[21]Pinceti A, Sankar L, Kosut O. Load redistribution attack detection using machine learning: a data-driven approach. 2018 IEEE power & energy society general meeting. 2018.
[22]Iwayemi A, Zhou C. SARAA: semi-supervised learning for automated residential appliance annotation. IEEE Trans Smart Grid 2017;8:779–86.
[23]Fan C, Xiao F, Zhao Y, Wang JY. Analytical investigation of autoencoder-based methods for unsupervised anomaly detection in building energy data. Apply Energy 2018;211:1123–35.
[24]Pereira J, Margarida S. Unsupervised anomaly detection in energy time series data using vibrational recurrent autoencoders with attention. In: 2018 17th IEEE international conference on machine learning and applications; 2018.
[25]David Weinberg. Our machines now have knowledge we will never understand; 2017. https://www.wired.com/story/our-machines-now-have-knowledge-wellnever-understand/.
[26]Janetzko, H., Stoffel, F., Mittelstädt, S., & Keim, D. A. (2014). Anomaly detection for visual analytics of power consumption data. Computers & Graphics, 38, 27-37.
[27]Naganathan, H., Chong, W. K., Huang, Z., & Cheng, Y. (2016). A Non-Stationary analysis using Ensemble Empirical Mode Decomposition to detect anomalies in building energy consumption. Procedia Engineering, 145, 1059-1065.
[28]Cabrera, David F. Motta, and Hamidreza Zareipour. "Data association mining for identifying lighting energy waste patterns in educational institutes." Energy and Buildings 62 (2013): 210-216.
[29]Li, Zhang. "Abnormal Energy Consumption Analysis Based on Big Data Mining Technology." 2020 Asia Energy and Electrical Engineering Symposium (AEEES). IEEE, 2020.
[30]Rumelhart, D.E.; Hinton, G.E.; Williams, R.J. Parallel Distributed Processing: Explorations in the Microstructure of Cognition, Chap. Learning Internal Representations by Error Propagation; MIT Press: Cambridge, MA, USA, 1986; Volume 1, pp. 318–362. Available online: http://dl.acm.org/citation.cfm?id=104279.104293 (accessed on 25 March 2022).
[31]Baldi, P. Autoencoders, unsupervised learning, and deep architectures. In Proceedings of the ICML Workshop on Unsupervised and Transfer Learning, JMLR Workshop and Conference Proceedings, Edinburgh, Scotland, 26 June–1 July 2012.
[32]Schölkopf, B.; Williamson, R.; Smola, A.; Shawe-Taylor, J.; Platt, J. Support Vector Method for Novelty Detection. In Proceedings of the 12th International Conference on Neural Information Processing Systems, Denver, CO, USA, 29 November–4 December 1999.
[33]Li, K.L.; Huang, H.K.; Tian, S.F.; Xu, W. Improving one-class SVM for anomaly detection. In Proceedings of the 2003 International Conference on Machine Learning and Cybernetics (IEEE Cat. No. 03EX693), Xi’an, China, 2–5 November 2003; Volume 5,pp. 3077–3081.
[34]Wang, Y.; Johnny, W.; Andrew, M. Anomaly intrusion detection using one class SVM. In Proceedings of the Fifth Annual IEEE SMC Information Assurance Workshop, IEEE, New York, NY, USA, 10–11 June 2004
QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top
無相關期刊