非侵入式負載監測方法及其應用

Non-intrusive Load Monitoring Approach and its Applications

章學賢(Hsueh-Hsien Chang)

指導教授 ： ARRAY(0xbd6c6a4)

中原大學/工學院/電機工程研究所/博士(2009年)

https://doi.org/10.6840/cycu200900804

摘要

非侵入式負載監測技術過去常以功率特徵為基礎，唯在負載辨識的穩定度與精確度上有待提升。本文結合類神經網路與啟動暫態能量分析技術使非侵入式負載監測系統可以更精確辨識負載需求與改善負載辨識的精確度，啟動暫態能量特徵能改善在多個負載操作下的負載辨識成效與所需計算時間。在分類器選用方面，經比較不同神經網路分類器與訓練演算法後，本文採用倒傳遞類神經網路分類器作為負載辨識的分類系統，文中利用電磁暫態程式模擬計算負載啟動暫態能量作為非侵入式負載監測的重要特徵。
本文所提非侵入式負載監測系統應用於汽電共生系統與公用電力系統的經濟調度上，透過非侵入式負載監測與管理的方法同樣地可以在電源入口端先行辨識出電力網路內所有匯流排上負載的能量需求，再將這些負載需求資料併入汽電共生系統與公用電力系統的經濟調度計算，期能有效的合理分配與管理兩者間能量供應需求，並進一步的減少空氣的汙染及電力成本。本文亦提出以基因演算法為基礎的經濟調度方法使用在汽電共生系統上，藉以獲得在環境條件的限制下取得接近全域最佳解。

關鍵字

非侵入式負載辨識 ； 非侵入式負載監測系統 ； 啟動暫態能量分析 ； 類神經網路 ； 基因演算法 ； 經濟調度

並列摘要

Non-intrusive load-monitoring (NILM) techniques were often based on power signatures in the past. These techniques are necessary to be improved for the results of reliability and accuracy of recognition. By using neural networks (NNs) in combination with turn-on transient energy analysis, this study attempts to identify load demands and improve recognition accuracy of non-intrusive load-monitoring results. The turn-on transient energy signatures can improve the efficiency of load identification and computational time under multiple operations. After comparing various training algorithms and classifiers in terms of artificial neural networks (ANN) due to various factors that determine whether a network is being used for pattern recognition, the back propagation neural network (BP) classifier is adopted in the efficiency of load identification and computational requirements. Additionally, in combination with electromagnetic transient program (EMTP) simulations, calculating the turn-on transient energy facilitate load can lead to identification and a significant feature for the turn-on transient energy.
A non-intrusive load-monitoring system is proposed in this thesis to apply to effectively manage energy demands within economic dispatch for a cogeneration system and power utility. All energy demands of loads on the buses are identified and calculated at the service entry of power by the NILM system. These data of loads will be used in economic dispatch to effectively manage and dispatch energy supplies and demands from a cogeneration system and power utility. Economic dispatch results indicate that the method can estimate accurately the energy contribution from the cogeneration system and power utility, and further reduce air pollution and power costs. Additionally, an economic dispatch method based on genetic algorithms (GAs) is used to approach the global optimum with typical environmental constraints for a cogeneration system.

並列關鍵字

Genetic Algorithms ； Non-intrusive Load Identification ； Non-intrusive Load-monitoring System ； Turn-on Transient Energy Analysis ； Economic Dispatch ； Neural Networks

參考文獻

[2] G. W. Hart, “Nonintrusive appliance load monitoring,” Proceedings of the IEEE Conference, pp. 1870–1891, 1992.

[3] J. G. Roos, I. E. Lane, E. C. Lane, and G. P. Hanche, “Using Neural networks for non-intrusive monitoring of industrial electrical loads,” Proceedings of the IEEE Instrumentation and Measurement Technology Conference, pp. 1115-1118, 1994.

[4] S. B. Leeb, A conjoint pattern recognition approach to nonintrusive load monitoring, Ph.D. Thesis, MIT, 1993.

[5] A. I. Cole and A. Albicki, “Nonintrusive identification of electrical loads in a three-phase environment based on harmonic content,” Proceedings of the IEEE Instrumentation and Measurement Technology Conference, pp. 24-29, 2000.

[6] D. C. Robertson, O. I. Camps, J. S. Mayer, and W. B. Gish, Sr., “Wavelets and electromagnetic power system transients,” IEEE Trans. on Power Delivery, Vol. 11, No. 2, pp. 1050-1056, 1996.