摘要
通过对实测样本进行直方图分析和假设检验,得到样本的概率分布,结合先验分布,基于贝叶斯方法得到沉降修正系数的后验分布,对传统的沉降修正系数进行优化。实例分析结果表明:当先验分布为无信息分布时,其后验分布受样本信息的影响较大;当样本信息服从高斯分布时,其后验分布也服从参数为μ和σ的高斯分布,μ值随外荷载增大而减小,σ值呈无明显规律变化;通过贝叶斯方法的优化,修正经验系数考虑了荷载的影响,并且取值的区间明显缩小。
The probability distributions of samples measured are obtained through histogram analysis and hypothesis testing. The posterior distribution coefficients can be gotten based on the Bayesian approach and combined with priori distributions of the settlement modification coefficients. The analysis results show that when the prior distributions are the non-information, the posterior distributions of coefficients are greatly influenced by samples information. When the distributions of samples are approved to obey Gaussian distributions, the posterior distributions of coefficients also obey these distributions, whose parameters are μ and σ, the values of μ are decreased with external loads increasing, and the values of σ have no significant change law. Through the optimization by Bayesian methods, the range of settlement modification coefficients is narrowed, which the influence of external loads are taken into account. 3 tabs, 4 figs, 10 refs.
出处
《长安大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2009年第4期30-33,81,共5页
Journal of Chang’an University(Natural Science Edition)
基金
国家自然科学基金项目(50778180)
关键词
岩土工程
贝叶斯理论
沉降修正系数
先验信息
后验分布
均匀分布
高斯分布
geotechnical engineering
Bayesian theory
settlement modification coefficients
prior knowledge
posterior distribution
uniform distribution
Gaussian distribution
