臺灣博碩士論文加值系統

本論文探討利用類神經網路之預測能力與靜態無效率補償器(SVC)結合，同時達成改善電力系統三相負載不平衡與功率因數修正。考慮到當負載變動較大，使SVC的補償能力受到時間延遲的影響，故本論文應用類神經網路的預測能力，先預測出往後數個週期的有效功率與無效功率，再對電流進行補償，此方法的平均絕對值誤差率(MAPE)只有10%以下，符合高預測準確度，藉以改善延遲的情形。本論文以實際鋼鐵廠爐變一次側三相電壓、電流、有效功率、無效功率資料模擬。為得到SVC所需要的基頻數值以及準確的預測結果與加快運算速度，在訓練類神經網路之前，對這些電力資料進行快速傅立葉轉換(FFT)以取得基頻值，之後使用自定義函數L2範數法規一化法。因為資料型態的關係，此法比Matlab內建規一化指令更適合使用在本文。

This thesis examines the combination of artifical neural network(ANN) forecasting ability and the static var compensator(SVC). It can improve three-phase unbalance of load on the power system and correct the power factor at the same time. The compensation ability of an SVC is limited by delays in reactive power measurements and thysitor ignition. In order to improve the SVC performance, this thesis presents a technique based on the predicting ability of ANN. At first, it is to predict three phase active power and reactive power several cycles ahead, then it is to compensate current components. The MAPE is only 10% or less and has high accuracy of prediction to improve delay situations. This thesis used huge field data, collected from the primary side of an electric arc furnace transformer in a real steel plant. The data included three phase voltage, line current, active power, and reactive power. It wants to obtain the basedband value of SVC, get accurate results of prediction, and accelerate program speed of operation. Before traning the ANN, the program needs to evaluate the baseband of electrical data by the fast fourier transform(FFT). Then it is to employ L2 normalization to get normalized data. Because of data type, this is more suitable than the normalized instruction of Matlab. This thesis finds a better way in reactive power of compensation. It can also improve reactive power variation of the electric arc furnace load and power quality problem.

目錄
頁次
摘要...........................................................................i
ABSTRACT......................................................................ii
誌謝.........................................................................iii
圖索引........................................................................vi
表索引........................................................................ix
第一章 緒論....................................................................1
1.1 研究動機...................................................................1
1.2 文獻回顧...................................................................1
1.3 研究內容...................................................................3
1.4 章節敘述...................................................................4
第二章 靜態無效功率補償器......................................................5
2.1 前言.......................................................................5
2.2靜態無效功率補償器組成與動作原理............................................5
2.3 靜態無效功率補償器應用於電力品質..........................................10
2.3.1 電力量的定義............................................................10
2.3.2 功率因數修正方法........................................................12
2.3.3三相負載不平衡定義與改善方法.............................................15
第三章 類神經網路.............................................................23
3.1 前言......................................................................23
3.2類神經網路介紹.............................................................23
3.2.1類神經網路原理與功能.....................................................23
3.2.2 類神經網路學習規則......................................................30
3.2.3類神經網路應用...........................................................33
3.3倒傳遞類神經網路...........................................................35
3.3.1倒傳遞法則簡介...........................................................35
3.3.2倒傳遞演算法.............................................................37
3.3.3有彈性的倒傳遞演算法.....................................................39
3.4預測效果之評估方式.........................................................40
3.5傅立葉轉換.................................................................42
3.5.1傅立葉轉換定義...........................................................42
3.5.2 離散傅立葉轉換..........................................................43
3.5.3 快速傅立葉轉換..........................................................44
第四章 SVC補償結果............................................................46
4.1實際資料計算分析...........................................................46
4.2資料規一化.................................................................59
4.3模擬方法與結果.............................................................62
4.3.1未使用類神經網路進行補償.................................................62
4.3.2使用類神經網路進行補償...................................................73
第五章 結論...................................................................93
5.1 研究成果..................................................................93
5.2 未來研究方向..............................................................94
參考文獻......................................................................95
符號說明......................................................................99

參考文獻
[1]江榮城，「電力品質實務(一)」，全華科技圖書股份有限公司，2000。
[2]H. Samet, M. R. Farhadi, and M. R. B. Mofrad, "Employing Artifical Neuron Networks for prediction of electrical arc furnace reactive power to improve compensator performance,” Proceeding of the IEEE International Conference on Energy and Exhibition, Florence, Italy, pp. 249-253, September, 2012.
[3]IEEE Working Group on Nonsinusoidal Situations, “A Survey of North American Electric Utility Concerns Regarding Nonsinusoidal Waveforms,” IEEE Transaction on Power Delivery, Vol. 11, No. 1, PP. 73-78, 1996.
[4]李尚懿，「三相不平衡配電饋線無效功率補償方法之研究」，博士學位論文，國立台灣科技大學，台北，民國84年。
[5]L. Gyugyi, R. A. Otto, and T. H. Putman, “Principles and Application of Static, Thyristor-Controlled Shunt Compensators,” IEEE Transation on Power Apparatus and Systems, Vol. 97, No. 5, pp. 1935-1945.
[6]劉鈺韋，黃舜鴻和洪晨軒，「電弧爐變動性虛功率與先進改善技術」，行政院國家科學委員會補助專題研究計畫期末報告，台北，民國102年。
[7]A. G. Exposito, F. G. Vazques, C. I. Michell, T. G. Garcia, and F. P. Madronal, “Microprocessor-based control of an SVC for optimal load compensation,” IEEE Transations on Power Delivery, Vol. 7, No. 2, pp. 706-712, 1992.
[8]羅華強，「類神經網路:Matlab的應用」，高立圖書，2011。

[9]黃瞬鴻，「利用小波之電力品質與電壓閃爍計算及分析」，碩士學位論文，國立台灣科技大學，台北，民國102年。
[10]羅文毅，「基於氣象資料的配電饋線負載預測之研究」，碩士學位論文，國立台灣科技大學，台北，民國98年。
[11]張延亮，「Matlab神經網路30個案例分析」，北京航空航天大學出版社，2010。
[12]新浪博客，http://blog.sina.com.cn/s/blog_d76227260101dlxn.html
[13]洪芳婷，「應用混合型負載預測模型於最佳化之需量規劃」，碩士學位論文，國立成功大學，台南，2007。
[14]台電營業規則，台灣電力公司，民國102年。
[15]黃鎮平，「負載特性對功率因數計算值的影響」，博士學位論文，國立台灣科技大學，台北，2010。
[16]IEEE Recommended Practice and Requirements for Harmonic Control in Electric Power Systems, IEEE Standard 519-1992, 1993.
[17]G. R. Harley, E. B. Makram, “Starting Transients of Induction Motors Connected to Unbalanced Networks,” Electric Power System Research, pp. 189-197, Vol.17 , 1989.
[18]陳在相，「輸電系統不平衡之量測與分析」，台灣電力公司研究發展專題，民國81年。
[19]陳政宏，「運用氣溫輪廓與類神經網路的台電短期負載預測」，碩士學位論文，國立台灣科技大學，台北，民國94年。
[20]古瓊景，EMC類神經網路IC原理及應用，全華科技圖書股份有限公司，1999。
[21]A. E. Emanuel, “Summary of IEEE standard 1459:definitions for the measurement of electric power quantities under sinusoidal, nonsinusoidal, balanced, or unbalanced conditions,”IEEE Transaction on Industry Application, Vol. 40, No. 3, pp. 869-876, 2004.
[22]H. Samet, M. Parniani, “Predictive method for improving SVC speed in electric arc furnace compensation,” IEEE Transaction on Power Delivery, Vol. 22, No. 1, pp. 732-734, Jan. ,2007.
[23]IEEE Standard for Shunt Power Capacitors, ANSI/IEEE Standard 18-1992, 1993.
[24]L. Gyuhyi and E. R. Taylor, Jr., “Characteristic of static, thyristor-conrolled shunt compensators for power transmission system applications,” IEEE Transactions on Power Apparatus and Systems, Vol. 99, No. 5, pp. 1795-1894, 1980.
[25]S. L. Marple, “Computing the discrete-time analytic signal via FFT,” IEEE Transaction on Signal Processing, Vol. 47, pp. 2600-2603, September, 1999.
[26]T. W. S. Chow and Y. F. Yam, “Measurement and evaluation of inatananeous reactive power using neural networks,” IEEE Transaction on Power Delivery, Vol. 9, No. 3, July, 1994.
[27]D. Asber, S. Lefebvre, M. Saad, and C. Desbiens, “Modeling of distribution loads for short and medium-term load forecasting,” Proceedings of the IEEE Internation Conference on Power Engineering Society General Meeting, Tampa, FL, USA, pp. 1-5, June, 2007.
[28]J. H. Mathews and K. K. Fink, Numerical Methods Using Matlab, 4th Edition, New Jersey, USA:Prentice-Hall, 2004.
[29]C. John, O.Salaan, and N. R. Estoperez, “An Artifical Neural Network Based Real-time Reactive Power Controller,” Proceedings of the World Congress on Engineering and Computer Science, San Francisco, USA, Vol. 1, October, 2011.
[30]柳松青，「MATLAB神經網路BP網路研究與應用」，計算機工程與設計，第24卷，第11期，第81-88頁，2003。
[31]S. H. Friedberg, A. J. Insel, and L. E. Spence, LINEAR ALGEBRA, 4th, Edition, New Jersey, USA:Prentice-Hall, 2003.
[32]T. J. E. Miller, Reactive Power Control in Electric System, New York, USA:John Wiley and Sons, 1982.
[33]T. Senjyu, H. Takara, K.Uezato and T. Funabashi, “One-Hour-Ahead Load Forecasting Using Neural Network,” IEEE Transactions on Power Systems, Vol. 17, Issue 1, pp. 113-118, Feb. , 2002.
[34]D. Srinivasan, S. S. Tan, C. S. Cheng and E. K. Chan, “Parallel neural network-fuzzy expert system strategy for short-term load forecasting:system implementation and performance evaluation,” IEEE Transactions on Power Systems, Vol. 14, Issue3, pp. 1100-1106, Aug. ,1999.