摘要
我国现行规范采用极值I型Gumbel分布作为各地区概率统计的线型,但Gumbel模型却有着样本不全等先天缺陷。根据澳大利亚某地区的实测风速记录在原样本和去掉一个高值记录的情况下分别计算出设计风速,通过对比发现:当模型样本中突然增加非常少量的高值数据的时候,所得到的设计风速结果表现出了很大的不稳定性。
The extreme value I Gumbel distribution model,used in our country′s code at present,actually has some inborn defects,such as the not all-inclusive sample.This article according to the real wind speed record in some area in Australian,calculating the design wind speed separately in the original sample and removing a high value record in the situation,discovered the contrast:when the model sample added in very few high value records,it obtains design wind speeds with obvious difference.
出处
《江西科学》
2012年第4期503-505,524,共4页
Jiangxi Science