风电场风速数值预报的误差分析及订正

Error analysis and correction of wind speed numerical forecast at wind farm

使用WRF模式对内蒙古某风电场区域内的2011年1—6月,50 m高度的风速进行了模拟,并结合实测风速对模拟结果进行了评估。在此基础上再利用自回归模型(AR模型)和持续法对WRF模式模拟结果进行了订正预报,订正结果表明:AR模型和持续法都能有效地减小WRF模式风速的模拟误差,AR模型订正效果优于持续法。为能对订正预报时效进行延长,提出了"假设观测值"概念。在AR模型的基础上建立一种新的订正模型称之为New AR模型。其订正预报结果表明:新模型能在12 h时效内,改善WRF模式风速模拟精度,其中6 h的改进效果较好。

The wind speed at 50 m at a wind farm of Inner Mongolia from January to June of 2011 was simulated by WRF（Weather Research and Forecasting） model, then the simulation results were e-valuated, combined with the observed wind speed. Furthermore, the wind speed was corrected by Auto-Regression（AR model） and Persistence. The results show that the mentioned above two methods can botheffectively reduce the simulation error, while the AR model performs better than the Persistence. In orderto extend the prediction period, a new revised model（New AR model） was established, based on theconcept of hypothetical observation. The correction forecasting results of the new model show that New ARmodel can improve the wind speed accuracy of WRF within 12 hours, especially within 6 hours.