极值风速的最优概率模型

THE OPTIMAL PROBABILISTIC DISTRIBUTION FOR EXTREME WIND SPEED

风速过程作为平稳高斯随机过程 ,风速母体服从指数型分布。基于时段最大取样和跨阈取样法 ,研究了极值风速理论分布的特征 ;研究了威布尔分布、极值Ⅰ型分布和广义Pareto分布 (GPD)等概型的尾部特性及对极值风速的估计精度 ,得出威布尔分布为年最大风速普遍最优概率模型 ,极值Ⅰ型分布次之 ,GPD估计的结果偏差最大且极值风速估计的变异性最大。采用MonteCarlo抽样试验和小长山站实测风资料统计分析验证了这一结论。

The wind parent is conformed to the exponential distribution; it is based on that the wind process is assumed as a stationary Gaussian process. The distribution of the extreme wind speed is studied by using different sampling methods. It is focused to study different distributions about both the tail characteristics and the capabilities to estimates of the extreme wind speed. The Weibull distribution is evaluated as the optimal one. It is verified by the analysis of the Monte Carlo samples and the wind speed observations at the Xiaochangshan station.