短期风速-风电功率预测方法.pdf

短期风速一风电功率预测方法 张友鹏，叶爱贤， 高锋阳，董唯光 (兰州交通大学自动化与电气T程学院，甘肃鼍州730070) 摘要：传统灰色风速预测模型累加处理时不能预测突变风速。使风电功率预测误差过大。采用数值逼近算法对传统灰色 GM(1，1)预测模型进行优化改进．以优化的灰色GM《1，1)预测模型对未来时段风速进行预测，突变风速预测误差降低了 34．3％。再将优化风速预测模型和时间序列动态神经网络相结合．构建出风电功率预测模型。应用该模型对酒泉地区某 风电场现场数据进行仿真测试。预测效果可信度大于93％。 关键词：灰色优化模型；短期风速预测；时间序列动态神经网络；风电功率预测 中图分类号：TM73 文献标识码：A 文章编号：1002—087 X(2013)04—0614—03 Predictionmethodofshorttermwind speed—windpower ZHANG Ai—xian，GAO You-peng,YE Feng-yang，DONGWei—guang ofAutomationandElectrical Gansu fSchool Engineering，Lanzhou Zj0070．China) JiaotongUniversity,Lanzhou Abstract：Thetraditional modelcannot mutationofwind wind graypredictor predict speed．1eadingpowerforecasting toexcessive errors．The wasusedto model approximationalgorithm improvegraypredictorGM(1。1)for predicting fhewind Offuture simulationresultsshowfhatforecastmutationOfwind errorsdecrease speed periods．The speed 34．3％．Thenewwind forecastmodeIiscombinedwithtimeseriesbasedon neura}networkto by speed dynamic wind model at wastested windfarm，andforecast was than93％． predictorpower．This Jiuquan credibility greater wind series Key model；short-term neuralnetwork；wind words：grayoptimiz