| Autor: |
Feng, Wang, ErKeng, Yu, Bo, Dong, Xiao, Liu |
| Zdroj: |
IFAC-PapersOnLine; August 1997, Vol. 30 Issue: 17 p541-544, 4p |
| Abstrakt: |
This paper presents a method for ultrashort-tenn load forecasting (USTLF) based on artificial neural network, which is training and forecasting on-line by selecting typical sample sets and thereby improves the quality of sample sets. Extensive studies have been performed about the effect of various factors such as learning rate, moméntum, etc. on the efficiency and accuracy of the backpropagation algorithm which are employed in the training of neural networks. The proposed method has been implemented in Central China Electric Power Network. It is found that the forecasting result is very accurate. |
| Databáze: |
Supplemental Index |
| Externí odkaz: |
|
