摘要
风的间歇性和时变性制约电力系统能量平衡,准确的风电功率预测有助于电网减小旋转备用、合理制定检修计划。为减小预测误差,提出一种基于多计算流体力学(computational fluid dynamics,CFD)模型的新型风电场组合功率预测方法。首先,利用小波混合神经网络对数值天气预报降尺度;其次,提出了考虑多重尾流的风电场物理CFD模型,并建立了根据测风塔风速外推各台风电机组风速的加速比相关系数;最后,提出了仅考虑自由流场和带有激盘模型的变权重组合流场模型。实际算例仿真证明,所提出的预测方法更准确地反映了风电场实际运行状态,有效提高了预测准确性。
Time-dependency and intermittence of wind power impacts energy balance between wind power output and load in power system in real-time. Accurate wind power forecast is helpful to reduce spinning reserve capacity and make reasonable maintenance plan. For impacts of wind power uncertainties on power system, a new combination method of wind power forecasting based on improved computational fluid dynamics（CFD） model was proposed. Firstly, wavelet-hybrid neural network（HNN） was used to downscale NWP for wind measurement mast. Secondly, CFD speedup correlation coefficient was established to reflect wind flow distribution accurately. Finally, a new variable-weight CFD combination method of wind power forecast for each wind turbine was put forward considering rotor thrust coefficient of free flow field and wakes of actuator disc. Engineering applications show that the proposed method can effectively reflect wind farm operation state and improve wind power forecasting accuracy.
出处
《电网技术》
EI
CSCD
北大核心
2017年第1期79-85,共7页
Power System Technology
基金
国家电网公司科技项目(DKYKJ[2012]001-1)~~
关键词
功率预测
组合方法
计算流体力学
小波混合神经网络
尾流模型
power forecast
combination method
computational fluid dynamics
wavelet-hybrid neural network
wake model
