Short-term load forecasting has recently become vital after the power system deregulation since the results of the forecasted value are used on determining market prices in power markets. To forecast the load is however difficult due to complex relations between electricity load and its various affected factors. Conventional regression methods are widely used to forecast loads, but it is not sufficient to deal with nonlinear relationships between the load and the variables influence the load. Forecasting load using similar day based neural networks has become popular in recent years. However, the most suitable day of the forecasted time is usually difficult to find, and leads to high forecasted errors. This thesis presents a region-based neural network using similar hour approach for short-term load forecasting that uses similar hours instead of similar days to predict next hour load, and use the summation of zonal forecasted loads in the forecasted area to improve the forecast accuracy. The proposed method is tested using the data of New England. The prediction errors (in MAPE) are ranging from 0.95% to 1.18% in all four seasons, with 1.08% in average. Besides, in order to reduce the storage space of gigantic amount of historical data, we perform data compression in each zone. The experimental results demonstrate that the prediction errors only slightly increase from 1.08% to 3.53% when the data distortion from 0 MW to 21 MW in our proposed method. That is, it is rigorous against the possible noises.

近年來，由於電力市場興起，電力變成一種能在市場上買賣的商品，作為決定電力價格的重要依據，短期電力預測顯得更為重要。然而電力預測其實相當困難，電力預測的基本概念是使用過去的歷史資訊對未來的用電量進行預估，但是影響電力的參數非常多，而且這些影響電力的參數與電量的關係相當複雜，難以進行預測。回歸方法(regression method)是傳統上常見用於預測的方法，但是在處理電力和其相關參數的非線性(nonlinear)關係上，回歸方法就顯得不足。近來採用相似天(similar day)概念的類神經網路(neural network)方法相當盛行，然而在歷史資訊中並不容易找出全天都和預測當天相似的時間，進而導致預測誤差。
本篇論文提出了以區域計算為基礎的預測方法，目的在於預測下一個小時的用電量。我們將預測地區分成較小的區域，並分區進行預測，最後再將分區的電力預測值加總作為全區的預測結果。在預測方法上，我們採用相似小時取代相似天的概念，並將結果輸入類神經網路進行預測。實驗採用新英格蘭(New England)地區八個區域每小時的電力和天氣資訊。我們的預測誤差(MAPE)在四個季節分別從0.95%到1.18%，平均而言的誤差值約為1.08%。此外為了降低需要儲存的歷史資料量，我們將資料進行壓縮，實驗證明在一定的壓縮幅度下，預測誤差成長並不顯著。

[ 1 ]
M. Negnevitsky, P. Mandal, and A.K. Srivastava, "An overview of forecasting problems and techniques in power systems," Power & Energy Society General Meeting, 2009. PES '09. IEEE , vol., no., pp.1-4, 26-30 July 2009.
[ 2 ]
T. Haida and S. Muto, "Regression based peak load forecasting using a transformation technique," Power Systems, IEEE Transactions on , vol.9, no.4, pp.1788-1794, Nov 1994.
[ 3 ]
K. -B. Song, Y. -S. Baek, D. -H. Hong, and G. Jang, "Short-term load forecasting for the holidays using fuzzy linear regression method," Power Systems, IEEE Transactions on , vol.20, no.1, pp. 96- 101, Feb. 2005.
[ 4 ]
K. -L. Ho, Y. -Y. Hsu, C. -F. Chen, T, -E. Lee, C. -C. Liang, T. -S. Lai, and K. -K. Chen, "Short term load forecasting of Taiwan power system using a knowledge-based expert system," Power Systems, IEEE Transactions on , vol.5, no.4, pp.1214-1221, Nov 1990.
[ 5 ]
Y. Chen, P.B. Luh, and S.J. Rourke, "Short-term load forecasting: Similar day-based wavelet neural networks," Intelligent Control and Automation, 2008. WCICA 2008. 7th World Congress on , vol., no., pp.3353-3358, 25-27 June 2008.
[ 6 ]
P. Mandal, T. Senjyu, N. Urasaki, and Toshihisa Funabashi, “A neural network based several-hour-ahead electric load forecasting using similar days approach,” International Journal of Electrical Power & Energy Systems, Volume 28, Issue 6, Pages 367-373, July 2006.
[ 7 ]
Z.H. Osman, M.L. Awad, and T.K. Mahmoud, "Neural network based approach for short-term load forecasting," Power Systems Conference and Exposition, 2009. PSCE '09. IEEE/PES , vol., no., pp.1-8, 15-18 March 2009.
[ 8 ]
S. Fan, K. Methaprayoon, and W.J. Lee, "Multi-area load forecasting for system with large geographical area," Industrial and Commercial Power Systems Technical Conference, 2008. ICPS 2008. IEEE/IAS , vol., no., pp.1-8, 4-8 May 2008.
[ 9 ]
R. Osczevski and B. Maurice, “The new wind chill equivalent temperature chart,” Bull. Amer. Meteorol. Soc., vol. 10, pp. 453–1458, 2005.
[ 10 ]
J. M. Masterton and F. A. Richardson, “Humidex, a Method of Quantifying Human Discomfort Due to Excessive Heat and Humidity,” Environment Canada, Atmospheric Environment Service, CLI 1-79, 1979.
[ 11 ]
T. Senjyu, H. Takara, K. Uezato, and T. Funabashi, “One-hour-ahead load forecasting using neural network,” IEEE Trans Power Syst, 17(1):113–8. 2002.
[ 12 ]
R. Lamedica, A. Prudenzi, M. Sforna, M. Caciotta, and VO. Cencellli, “A neural network based technique for short-term forecasting of anomalous load periods,” IEEE Trans Power Syst, 11(4):1749–55, 1996.
[ 13 ]
H. S. Hippert, C.E. Pedreira, and R. C. Souza, “Neural networks for short-term load forecasting: A Review and Evaluation,” IEEE Trans. Power Systems, vol. 16, no. 1, pp. 44-55 , 2001.
[ 14 ]
W. Sun, "Ant colony based feed forward NN short-term load forecasting model with input selection and DA clustering," Natural Computation, 2008. ICNC '08. Fourth International Conference on , vol.2, no., pp.300-304, 18-20 Oct. 2008.
[ 15 ]
Yong Wang; Dawu Gu; Jianping Xu; Jing Li; , "Back propagation neural network for short-term electricity load forecasting with weather features," Computational Intelligence and Natural Computing, 2009. CINC '09. International Conference on , vol.1, no., pp.58-61, 6-7 June 2009.
[ 16 ]
S. E. Papadakis, J. B. Theocharis, S. J. Kiartzis, and A. G. Bakirtzis, “A novel approach to short-term load forecasting using fuzzy neural networks,” IEEE Trans. Power Systems, vol. 13, no. 2, pp. 480–492, 1998.
[ 17 ]
J. M. Shapiro, ”Embedded image coding using zerotrees of wavelet coefficients,” IEEE Trans. Signal Processing, vol. 41, no. 12, pp. 3445-3462, Dec 1993.