| 作成者 |
|
|
|
|
|
| 本文言語 |
|
| 出版者 |
|
|
|
| 発行日 |
|
| 収録物名 |
|
| 巻 |
|
| 号 |
|
| 開始ページ |
|
| 終了ページ |
|
| 出版タイプ |
|
| アクセス権 |
|
| JaLC DOI |
|
| 関連DOI |
|
|
|
|
|
|
|
| 関連URI |
|
|
|
|
|
|
|
| 関連情報 |
|
|
|
|
|
|
|
| 概要 |
To make weekly operational plan for power stations, we have to forecast daily peak load till one week ahea., In case of one-week-ahead load forecasting, we can obtain only insufficient and unnumerical... information with respect to load such as holiday information and trend of temperature. Therefore the forecasting using a time series model is necessary. However we must pay sufficient attention in modelling to seasonal and weekly variations of load. We propose a method for forecasting with suitable modelling and removing of these effects on the load. First step is to get numerical expected temperature based on weekly weather forecast expressed in "words", and to construct two models that represent the relations between the temperature and the load. Second step involves dividing the time series of load data into weekly variation caused by holidays, seasonal variation owing to temperature and residual variation due to unknown factors, and forecasting each of them using digital filters or an autoregressive model. The forecasting examples show the ability of the method in forecasting with practically good accuracy without suffering from the effect of seasons and holidays. Prediction errors are around 100-300 MWh.続きを見る
|
Mendeley出力
