摘要
为克服实际工程中测试数据不足情况下,采用经典统计学方法推断可变作用的代表值时,由统计不确定性导致推断结果偏小的缺点,利用可变作用服从极大值I型分布,与极小值I型分布可相互转换的特性,根据极小值I型分布分位值的线性回归推断方法,提出了可变作用标准值、频遇值的线性回归推断方法,可在不同的置信度下考虑统计不确定性的影响,适用于小样本统计。分析结果表明:小样本条件下,按点估计法推断可变作用的标准值和频遇值时,结果偏小;按线性回归推断方法推断可变作用的标准值、频遇值和设计值时,可取置信度为0.75;可变作用的准永久值受统计不确定性的影响较小,可取点估计法的推断结果。
It is needed sometimes in engineering to infer the representative values of variable actions on the conditions of small samples( the test data are inadequate),but the classical statistics methods adopted now do not take into account the succedent influences of statistical uncertainty,and the inferring results are always small. In order to overcome the above shortcomings,using the feature that variable actions obey type I maximum distribution and can be converted to type I minimum distribution and applying the current linear regression estimation for inferring the fractiles of type I minimum distribution,the linear regression estimation of characteristic and frequent values of variable actions were proposed,which could take into account the influences of statistical uncertainty with different degree of confidence. According to the results of theoretical analysis and examples,when using the indirect point estimation methods to infer the characteristic value and the frequent value of variable actions on the conditions of small samples,the inferring results are always on the aggressive side. When using the linear regression estimation to infer the characteristic value,frequent value and design value of variable actions,it is desirable for the 0. 75 confidence degree. Because the statistical uncertainty has little influence on inferring the quasi-permanent value of variable actions,the current point estimate method can still be used.
出处
《建筑结构学报》
EI
CAS
CSCD
北大核心
2014年第10期98-103,共6页
Journal of Building Structures
基金
国家自然科学基金项目(50678143
51278401)
陕西教育厅专项科研项目(14JK1412)
关键词
可变作用
代表值
小样本
统计推断
线性回归
variable actions
representative value
small sample
statistical inference
linear regression