风能资源估算中的不确定性分析

UNCERTAINTY ANALYSIS OF WIND ENERGY RESOURCE ESTIMATION

针对已投运风电场的利用小时数大多达不到设计水平的现状,该文从风能资源评估的角度寻找原因。通过对决定平均有效风能密度估算值误差大小的不完全伽玛函数的数值实验,发现计算误差不稳定,有时会超出可接受范围,造成估值偏高。利用改进FAST和Sobol两种基于方差的灵敏度分析方法,分别计算出近似平均有效风能密度对Weibull分布参数及风力机切入、切出风速4个不确定量的一阶灵敏度和全影响灵敏度。一致的灵敏度排序结果说明尺度系数是影响相对误差的主要原因。通过分析数值实验结果提出了近似算法的适用条件。算例显示,所提约束条件可以很好地控制近似算法的误差范围。

Based on the current situation that the operating hours of the most of running wind power farms can not reach the design level, the causes was investigated from the perspective of the estimation of wind energy resources. According to the numerical experiment of incomplete gamma function which determines the estimated value error of average effective wind energy density, it is found that the error of approximate calculation is unstable and could be far away from reasonable bound sometimes, and results in the overestimation of average effective wind energy densi- ty. The extended FAST method and Sobol method are respectively applied to calculate one order sensitivity and total sensitivity of uncertainty quantities, including the effect of Weibull distribution parameters, cut in speed and cut out speed on estimation variance. The main factor contributing to variation of error was locked by the same sensitivities rank. Based on the results of numerical experiments analysis, the suitable constraint set of the application calculation is also presented. The test cases show that the error of approximate calculation can be limited by constraints.