风电场超短期风速预测方法对比  

A Comparison of Very Short Term Wind Prediction by Different Methods

作　　者：伍见军[1] 王咏薇[2] 丁源[2] 祖繁[2] 高山[3] 高卓[4] 

机构地区：[1]南京信息工程大学在气物理学院 [2]南京信息工程大学大气环境中心,南京210044 [3]东南大学电气工程学院,南京211100 [4]北京四方继保自动化股份有限公司,北京100085

出　　处：《科学技术与工程》2013年第11期2965-2969,共5页Science Technology and Engineering

基　　金：江苏省科技支撑计划(BE2010200);江苏高校优势学科建设工程(PAPD);长江学者和创新团队发展计划资助

摘　　要：以沿海及山地复杂地形条件下的风电场为例,采用为期10 d的风电场测风塔实测风速资料,对比了两种预测方法在时效为4 h的超短期风速预测中的性能。其一为采用WRF数值模式预报的物理方法,其二为BP神经网络的统计方法;并探讨了两种风速预测方法的实用价值及意义。结果表明,与物理模拟方法相比,统计方法在4 h的超短期风速预测中,无论是预测准确性及计算效率都有一定的优势。Wind farms at complex terrain condition like coastal areas and mountainous areas for instances are taken to contrast. The results are two prediction methods with ten-days observations from wind-towers, the one method was WRF simulation and the other was BP neural network statistical prediction. The accuracy of wind pre- diction in 4 hours term between two methods was analyzed. The practical value and significance of the two methods were also discussed. Results show the statistical prediction had obvious advantages in accuracy and computation ef- ficiency in 4 hours term compared with the physical simulation.

关 键 词：复杂地形条件 风电场风速预测 统计预报 物理模拟 方法对比 

分 类 号：TK89[动力工程及工程热物理—流体机械及工程]