长期极值统计理论及其在海洋环境参数统计分析中的应用

Long-term Extreme Value Statistics Theory and Its Application tothe Marine Environmental Parameter Statistics and Analysis

海洋环境极值参数(如风速、流速、波高、周期等)在海洋工程设计中具有重要意义。利用次序统计和极值理论方面的较新研究成果,从理论上证明了多种统计分布中Weibull分布是最优的,使长期极值统计建立在一个更坚实的基础上;同时引入基于序列统计的最大似然估计方法。利用大量数据,对最小二乘估计方法和最大似然估计法进行对比分析,指出最大似然估计法是精确估计,而最小二乘估计方法是保守估计。

The marine environmental extreme value parameters, such as wind speed, current speed, wave height and wave period play an important role in marine engineering design. In this paper, the new developments in extreme value statistics theory are used to theoretically prove that the Weibull distribution is the best among many statistical distributions and makes the long-term extreme value statistics more effective. Furthermore, a number of data are used to make a comparison between the least square estimation method and the maximum likelihood estimation method, and it is shown from the comparison results that the latter is a precise estimation, and the former is a conservative estimation.