基于长期实测资料的风速测量相关推测方法对比
Comparison of Measure-correlation-predict Algorithms in Offshore Wind Power Assessment with Multi-year Observation of Automatic Weather Stations
-
摘要:
针对海上风电的风能资源评估，利用4个沿海自动气象站2003—2010年的逐时测风数据与同期MERRA再分析资料，通过6组数值试验，从平均风速、风速标准差、平均风功率密度和风力发电机组的年发电量等4个评价指标全面对比分析了8种风资源测量-相关-推测方法。结果表明：简单正交回归法、简化主轴回归法、分位数回归法、Weibull尺度拟合法和矩阵时间序列法对平均风速的推测结果更优，简单正交回归法、简化主轴回归法和Weibull尺度拟合法对风速标准差和平均风功率密度的推测结果更优，联合概率分布法对风向-风速联合频率分布和风机发电量的推测结果明显优于其他方法。参证站自身的数据质量和空间代表性对测量-相关-推测方法的表现具有潜在影响，在实际工作中应对多种测量-相关-推测方法进行集成运用和综合判断，以符合海上风电多层次、多目标的风资源评估需求。 Abstract:For offshore wind power assessment, eight kinds of Measure-Correlation-Predict (MCP) algorithms have been employed to compare mean wind speed, standard deviation of wind speed, mean power density and annual energy output of wind turbine in six numerical experiments with wind observation of four automatic weather stations in Pearl River Delta and reanalysis dataset MERRA from 2003 to 2010. The results show that Simple Orthogonal Regression (SOR), Reduced Major Axis regression (RMA), Quantile Regression (QR), Weibull Scale Fitting (WS) predict better mean wind speed. Besides, SOR, RMA and WS predict better standard deviation of wind speed. Joint Probability Distribution algorithm (JPD) predicts better energy output of wind turbine.
下载:
