臺灣博碩士論文加值系統

本論文使用極速學習機(ELM)來預測臨界清除時間(CCT)，臨界清除時間是測量電力系統暫態穩定度的指標，如有一強大的臨界清除時間代表著這電力系統穩定性高，然而，通常使用一般的時域法來獲取臨界清除時間，需浪費許多的時間，為了加速計算臨界清除時間，在過去的三十年許多研究人員考慮使用類神經網路，最近，在不到十年的時間極速學習機已蛻變為類神經網路的產物，也是單層前饋式類神經網路所衍生而來的，是使用最小平方法而不是疊代梯度法，因此計算速度非常之快速，本論文研究的議題是使用極速學習機找尋臨界清除時間，本論文研究探討六個匯流排及三個發電機組的電力系統，結果顯示利用極速學習機計算出臨界清除時間是迅速的，且這結果是相當準確的。

This thesis uses extreme learning machine (ELM) to predict critical clearing time (CCT). CCT is a measurement for measuring power system transient stability. A larger CCT suggests this power system stability is stronger. However, it wastes a lot of time to obtain CCT by using the conventional time-domain method. In order to accelerate the CCT computation, many researchers have considered the usage of neural networks in the past three decades. Recently, ELM is a refined product of neural networks in less than ten years. It is the offspring of single layer feedforward network. It is very fast because of using least square method but not iterative gradient method. Therefore the calculating speed can be very fast. This thesis studies the issue of using ELM to find CCT. An example of a six-bus three-machine power system is studied in this thesis. The results show that CCT computation by ELM is fast and fairly accurate.

中文摘要 i
英文摘要 ii
致謝 iii
目錄 iv
圖表目錄 v
第一章 介紹 1
第二章 電力系統暫態穩度度 3
2-1 暫態穩定度 3
2-2 臨界清除時間 8
第三章 極速學習機 10
3-1 多層感知器 10
3-2 極速學習機 11
第四章 案例分析 15
4-1 案例分析流程 15
4-2 系統測試 16
4-3 極速學習機預測臨界清除時間 17
第五章 結論 21
參考文獻 22

[1] J.J. Wong, C. T. Su, C.S. Liu, C. L. Chang, “Study on 729 blackout in the Taiwan power system, “ International Journal of Electrical Power & Energy Systems, Vol. 29, no. 8, pp. 589-599, 2007.
[2] M. Moechtar, T. C. Cheng and L. Hu, “Transient Stability of Power System - A Survey,” Conference Record of Microelectronics Communication Technology Producing Quality Products Mobile and Portable Power Emerging Technologies, pp. 166-171, 1995.
[3] A. M. Sharaf, T. T. Lie, H. B. Gooi, “Transient stability and critical clearing time classification using neural networks,” Proceeding of 2nd IEE International Conference on Advances in Power System Control, Operation and Management, pp. 365-372, 1993.
[4] K. L. Lo, R. J. Y. Tsai, “Power system transient stability analysis by using modified Kohonen network,” Proceeding of IEEE International Conference on Neural Networks, pp. 893 – 898, 1995.
[5] J.D. Glover, M.S. Sarma, ‘Power System Analysis and Design’, 2002.
[6] Qian Ai, Danzhen Gu, Chen Chen, “New Load Modeling Approaches Based on Field Tests for Fast Transient Stability Calculations,” IEEE Trans. On Power Systems, vol. 21, no. 4, pp. 1864-1873, 2006.
[7] G. Huang, G.-B. Huang, S. Song, K. You, “Trends in extreme learning machines: A review”, Neural Networks, Vol. 61, pp. 32-48, 2015.
[8] H. Saadat, Power system analysis, McGraw Hill, 2004.
[9] 廖基宏1，蘇慶宗2，“電力系統暫態穩定度-等面積法則程式設計”，1建國科
技大學電機工程學系助理教授，2吳鳳技術學院電機工程學系教授，專題論
文，2009.
[10] 簡士翔，“靜態虛功補償器之控制器設計” ，義守大學電機工程研究所，碩
士論文，民國90年6月