电力系统保护与控制
電力繫統保護與控製
전력계통보호여공제
POWER SYSTM PROTECTION AND CONTROL
2013年
12期
79-85
,共7页
风速预测%风电功率特性曲线%灰色理论%FIR-MA 模型%辨识%最小二乘迭代
風速預測%風電功率特性麯線%灰色理論%FIR-MA 模型%辨識%最小二乘迭代
풍속예측%풍전공솔특성곡선%회색이론%FIR-MA 모형%변식%최소이승질대
wind speed prediction%wind power characteristic curve%grey theory%FIR-MA model%identification method%least squares iteration
为了准确预测风电机组的输出功率,针对实际风场,给出一种基于灰色 GM(1,1)模型和辨识模型的风电功率预测建模方法,采用残差修正的方法对风速进行预测,得出准确的风速预测序列。同时为了提高风电功率预测的精度,引入 FIR-MA迭代辨识模型,从分段函数的角度对风电场实际风速-风电功率曲线进行拟合,取得合适的 FIR-MA 模型。利用该模型对额定容量为850 kW 的风电机组进行建模,采用平均绝对误差和均方根误差,以及单点误差作为评价指标,与风电场的实测数据进行比较分析。仿真结果表明,基于灰色-辨识模型的风电机组输出功率预测方法是有效和实用的,该模型能够很好地预测风电机组的实时输出功率,从而提高风电场输出功率预测的精确性。
為瞭準確預測風電機組的輸齣功率,針對實際風場,給齣一種基于灰色 GM(1,1)模型和辨識模型的風電功率預測建模方法,採用殘差脩正的方法對風速進行預測,得齣準確的風速預測序列。同時為瞭提高風電功率預測的精度,引入 FIR-MA迭代辨識模型,從分段函數的角度對風電場實際風速-風電功率麯線進行擬閤,取得閤適的 FIR-MA 模型。利用該模型對額定容量為850 kW 的風電機組進行建模,採用平均絕對誤差和均方根誤差,以及單點誤差作為評價指標,與風電場的實測數據進行比較分析。倣真結果錶明,基于灰色-辨識模型的風電機組輸齣功率預測方法是有效和實用的,該模型能夠很好地預測風電機組的實時輸齣功率,從而提高風電場輸齣功率預測的精確性。
위료준학예측풍전궤조적수출공솔,침대실제풍장,급출일충기우회색 GM(1,1)모형화변식모형적풍전공솔예측건모방법,채용잔차수정적방법대풍속진행예측,득출준학적풍속예측서렬。동시위료제고풍전공솔예측적정도,인입 FIR-MA질대변식모형,종분단함수적각도대풍전장실제풍속-풍전공솔곡선진행의합,취득합괄적 FIR-MA 모형。이용해모형대액정용량위850 kW 적풍전궤조진행건모,채용평균절대오차화균방근오차,이급단점오차작위평개지표,여풍전장적실측수거진행비교분석。방진결과표명,기우회색-변식모형적풍전궤조수출공솔예측방법시유효화실용적,해모형능구흔호지예측풍전궤조적실시수출공솔,종이제고풍전장수출공솔예측적정학성。
To predict the output power of wind turbine accurately, based on the GM (1, 1) model and the identification method, a wind power generation short-term prediction method is presented for the real wind farm. The revision of residual error is applied to forecast the wind speed and get the accurate predicted wind speed series. Then, in order to increase the prediction precision of wind power, the FIR-MA iterative identification model is adopted to fit the real relationship between sequential wind speed and wind power and get the proper FIR-MA model. By modeling the wind turbine whose rated capacity is 850 kW, this paper compares the predicted wind generation power with the observed data using mean absolute percentage error, root mean square error and single point error as its evaluation indexes. The simulation shows the effectiveness and the practical applicability of the presented method, which can predict the real time generation power of wind turbineness and raise the accuracy of the wind power prediction. Finally, the simulation using the actual data from wind farm in China proves the efficiency of the proposed grey-identification model. <br> This work is supported by National Natural Science Foundation of China (No. 61174032).