摘要
以河北省某实际风电场为例,选取风电机组历史功率数据、风速以及数值天气预报的风速和风向作为输入因子,采用人工神经网络法对风电场超短期功率预测问题进行研究。研究结果显示,输入因子的差异性对风功率预测结果影响较大。另外,风电机组历史数据对功率预测结果的影响随时间增加而减小,进行3 h以上风电场功率预测时预测结果精度在很大程度上依赖数值天气预报数据精度。
Taking a wind farm in Hebei Province as case study, the very short-term forecasting of wind power is studied by using artificial neural network method, in which, the historical output of wind turbine, actual wind speed and the wind speed and direction obtained from Numerical Weather Prediction （NWP） are taken as input factors. The results show that the difference between input factors has greater impact on wind power prediction, the impact of historical output of wind turbine on wind power prediction is gradually decreased with the increase of operation time, and the forecasting accuracy of wind power in next 3 h and more mainly depends on the accuracy of NWP data.
出处
《水力发电》
北大核心
2013年第7期96-99,共4页
Water Power
关键词
超短期功率预测
人工神经网络法
风力发电
very short-term power forecasting
, artificial neural network
wind power
