基于神经元网络方法的风电场风电功率预报研究

Wind Power Forecasting for Wind Farms Based on Neural Network

风电大规模并网使风电对电网的冲击问题越来越凸显,许多地方出现了拉闸限电的情形,随着百万千瓦级风电基地、千万千瓦级风电基地的规划及建设,急需开展行之有效的风电场风电功率预报,来满足风电上网调度的实际需求,利用数值模式预报的风速、风向等预报场及风电场逐时风电功率资料,通过神经元网络方法进行了风电场风电功率预报试验,预报精度与2002—2006年欧洲风能计划中的风电场风电功率预报精度相当。

With large-scaled wind power integrated with power grids, the impacts of wind power on power grinds are more and more prominent, so much so that power cuts to limit consumption occur in many places. With 1 GW-level and 10 GW-level wind farms under planning or even under construction, effective and efficient wind power prediction is urgently needed to meet the requirements for connecting the wind power with power grids. In this paper, we try the ANN method to predict the wind power with the data of wind speed, wind direction from the numerical model and data of wind power data from the wind farm. The precision is almost equal to the one of European wind energy forecasting program.