以我国多个地区实际运行风电场为研究对象,以实现风电功率预测结果不确定性较优估计为目的,提出了基于风过程方法的风电功率预测结果不确定性估计方法,并给出了各风过程的数学模型。采用风过程方法和功率水平划分方法划分预测误差数据,可有效识别不同特性预测误差。采用非参数回归方法拟合误差概率密度分布,获得了较优的拟合结果。根据实际风电场数据验证文中方法的先进性,其与基于正态分布优化拟合的估计方法的比较结果表明,文中方法的各项评价指标均优于后者。
Taking wind farms actually operated in various areas of China as the object and the implementation of optimal uncertainty estimation of wind power prediction results as the goal, a wind process method-based approach to estimate the uncertainty of wind power prediction results is proposed and mathematical models for different wind processes are given. To effectively identify prediction errors of different characteristics, the prediction error data is divided by wind process method and power level division method. Adopting nonparametric regression method to fit probability density distribution of the error, a better fitting result is attained. The advancement of the proposed method is verified by actual wind farm data; and comparing the proposed estimation method with the estimation method based on normal distribution optimized fitting, the results show that the evaluation indices obtained by the proposed method are better than those from the latter.