發(fā)布時間：2018-02-09 11:55

  本文關鍵詞： 風速預測 空間相關性 動態(tài)特征 離線分類建模 在線特征匹配　出處：《電力系統(tǒng)自動化》2017年12期 　論文類型：期刊論文

【摘要】：風速的空間相關性有助于提高其預測質(zhì)量,特別是在風速突變的情況下。將"離線分類建模,在線匹配模型"的預測思路應用到利用空間相關性的超短期風速預測之中:通過歷史數(shù)據(jù)的時序分析,識別其中各風電場風速存在空間相關性的時段;按其時序特征及其他的條件特征,將觀察時窗內(nèi)的風速序列劃分為不同演化形態(tài)的樣本子集;在離線環(huán)境下,分別根據(jù)各類形態(tài)的訓練樣本子集優(yōu)化其專用的預測模型及參數(shù);在線應用時,則根據(jù)當下窗口內(nèi)風速序列的演化形態(tài)及相關的條件特征,按匹配所得模型及參數(shù),根據(jù)參考風電場的實測數(shù)據(jù)預測目標風電場的風速。以實際的歷史數(shù)據(jù)驗證了所述方法的有效性。
[Abstract]:Spatial correlation of wind speed helps to improve the quality of its prediction, especially in the case of sudden changes in wind speed. The on-line matching model is applied to the ultra-short-term wind speed prediction based on spatial correlation. Through the time series analysis of historical data, the time period in which the wind speed of each wind farm is spatially correlated is identified. According to its time series and other conditional features, the wind velocity series in the observation window is divided into subsets of samples with different evolution patterns, and its special prediction model and parameters are optimized according to the training sample subsets of various forms in off-line environment. In online application, according to the evolution of wind velocity series in the current window and related conditional features, according to the matching model and parameters, The wind speed of the target wind farm is predicted according to the measured data of the reference wind farm. The effectiveness of the method is verified by the actual historical data.
【作者單位】： 東南大學電氣工程學院;新能源與儲能運行控制國家重點實驗室(中國電力科學研究院);南瑞集團公司(國網(wǎng)電力科學研究院);智能電網(wǎng)保護和運行控制國家重點實驗室;神華集團有限責任公司;國網(wǎng)甘肅省電力公司電力科學研究院;國網(wǎng)甘肅省電力公司風電技術中心;
【基金】：國家自然科學基金重點項目(61533010) NSFC-NRCT(中泰)合作研究項目(51561145011) 國家電網(wǎng)公司科技項目~~
【分類號】：TM614
，

本文編號：1497871