風力機組出力預測模型之建立; A Study on the Establishment of a Forecast Model for the Power Generation of Wind Turbine(S)A Study on the Establishment of a Forecast Model for the Power Generation of Wind Turbine(S)

NCUIR > College of Earth Sciences > Department of Atmospheric Sciences and Graduate Institute of Atmospheric Physics > Research Project > Item 987654321/45046

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/45046

Title:	風力機組出力預測模型之建立;A Study on the Establishment of a Forecast Model for the Power Generation of Wind Turbine(S)A Study on the Establishment of a Forecast Model for the Power Generation of Wind Turbine(S)
Authors:	曾仁佑;林沛練
Contributors:	大氣物理研究所
Keywords:	風力發電;數值天氣預測;再生能源;Wind power generation;numerical weather prediction;renewable energy;大氣科學類;能源工程;資訊科學--軟體
Date:	2006-12-01
Issue Date:	2010-12-21 16:40:00 (UTC+8)
Publisher:	經濟部國營事業委員會
Abstract:	風力發電為乾淨的再生能源，近來隨著石化原料價格高漲以及京都協議生效而為各國所積極開發。然風力發電的最大阻礙來自其天然的變動性，使得電網整合或是機組調度產生相當大的困難。這種先天性的障礙，若能透過準確的風電預測，當可將電網運轉或是系統調度等困難減至最低。風電預測可透過數值天氣預測（numerical weather prediction）分析或統計預測方法而得，實證顯示精確的風電預測不但可有效降低電網操作的困難，同時亦可增加風電的使用效能，而使系統整體的供電成本降低。雖然電力公司對未來24小時用電量預測的誤差，大多可以控制在1.5%以內，而對一週預測的誤差也都在5%以下。但由於大型火力（煤或油）發電系統起動的時間約需8小時或更長，所以若能把因風場不穩定所造成發電量的擾動降到最低，對供電品質的維持亦有相當大的助益；特別是風力發電量已超過當地總發電量10%的區域，如澎湖等。因此，本研究計畫的主要動機是要「因應未來風力發電容量達一定程度後，可能對系統區域調度產生影響，參考國外風力發電先進國家之做法，及早做本土化之研究，以建立風力機組出力預測之模型。」國外有關風電預報之研究方式可歸納為兩大類，一為電廠直接委託氣象顧問公司或學術單位進行預報技術之研究與開發，如美、日等國；另一為多國結盟合作，共同開發相關之技術及預報模型，如歐盟。台灣的狀況則較適合採用直接委託研究的方式來進行，結合國內既有之數值天氣預報技術及風力電廠評選與發電量預測之技術，進行風力機組出力預測模型建立之研究。本研究計畫之最終目標是要「建立24小時前之風力機組出力預測模型」。研究預計18個月內完成，第一年（第一至第六個月）預計完成之項目為：(1) 蒐集並整理國外（歐美等地區）風電預測技術或模型之理論基礎（或數理推導）之探討及其應用於風場之情形；(2) 完成台灣地區風場觀測資料與中央氣象局預報資料之擷取與整合分析；(3) 完成氣象資料分析及建立初步風速統計機率模式；及(4) 利用過去之風場動力數值模式預報結果與當時實際觀測資料，建立模式與實測資料間之統計迴歸預報模式（即MOS, model output statistic），以提昇模式預報之準確度。第二年（第7至第18個月）預計完成之項目為：(1) 利用本土化之物理模式，建立一套可以進行風場預報之中尺度天氣預報模式；(2) 利用澎湖中屯及桃園大潭實際運轉之資料，分別建立一套至少預報24小時之「風力機組出力預測模型」：(3)擇取澎湖中屯及桃園大潭運轉中之風力機組進行實證分析；及(4) 完成預測模型的作業性測試至少二個月，並將所得之預測結果與機組實際出力進行差異分析，並提出後續之預測模型改善方案。 Wind power is a clean renewable energy. The demanding of wind energy is soaring in recent years due to the price increase of petroleum and the becoming effective of the Kyodo Protocol. However, the variation nature of wind hinders the development of the wind energy, and hence increase the difficulty in the management of a power system. If we can overcome the variation by an accurate wind power prediction, the difficulty in the management of power system will be reduced. Wind power generation can be predicted by numerical weather prediction or statistic methods. It has been shown that an accurate wind power prediction can reduce the difficulty in the management of a power system, and can increase the efficiency of the usage of wind energy, and also reduce the price of electricity. The electrical load can be predicted with about 1.5% accuracy for a 24-h forecast, and with about 5% accuracy for one week. This is fundamentally different from wind power forecasts. However, the typical time scales for start-up of conventional power plants are between 20 min. for gas turbines and 8 hours (or perhaps more) for large coal or oil plant. Therefore, in order to have a stable power system, we need a longer prediction horizon. This is more important to the area that the wind power generation is greater than 10%, such as Penhu. The motivation of this study is to develop a prediction model of wind power generation before the capacity of wind power reaches to a certain level and starts to affect the stability of power supply system. The main goal of this study is to establish a prediction model that can forecast wind power generation 24-hr in advance. We plan to finish this study in 18 months. In the first half year, we will finish the following four tasks: (i) to review the technical and the theoretical foundation of wind power prediction model worldwide; (ii) to collect the observed wind data and predicted wind data fron the CWB forecast model; (iii) to build up statistic forecast model; and (iv) to establish a MOS for a weather prediction model. In the second year, we will focus on the following tasks: (i) to establish a numerical weather prediction model , (ii) to build up a prediction model of wind power generation with 24-hr horizon by using the operation record from the wind turbines in Jongtun/Penhu and Dartan/Taoyung wind farms, (iii) to verify the model for at least two months, and finally (iv) to improve the prediction model by the validation result. 研究期間：9508 ~ 9512
Relation:	財團法人國家實驗研究院科技政策研究與資訊中心
Appears in Collections:	[Department of Atmospheric Sciences and Graduate Institute of Atmospheric Physics ] Research Project

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	622	View/Open

社群 sharing

Loading...