Evaluation of statistical uncertainty of cement-treated soil strength using Bayesian approach
Evaluation of statistical uncertainty of cement-treated soil strength using Bayesian approach
复制标题
使用贝叶斯方法评估水泥处理土强度的统计不确定性
DOI：
10.1016/j.sandf.2019.04.010
复制
发表时间：
2019
影响因子：
3.7
通讯作者：
Tsutomu Namikawa
中科院分区：
文献类型：
--
作者：
Atsushi Mohri;Yoshiaki Kikuchi;Shohei Noda;Kazuki Sakimoto;Yuka Sakimoto;Mitsutaka Okada;Shunsuke Moriyasu;Shin Oikawa;Tsutomu Namikawa
在深层搅拌法的质量保证过程中，通过从水泥处理桩中提取取芯试件的强度来评估水泥土强度的统计参数。样本参数包括与统计样本量相关的统计不确定性和其他因素。因此，需要对强度的统计参数进行概率表征，以量化质量保证过程中的统计不确定性。本文对水泥土搅拌桩强度估算的统计不确定度进行了定量分析。采用贝叶斯方法对由观测数据确定强度统计参数的统计不确定度进行了评定。推断是通过马尔可夫链蒙特卡罗方法进行的，其中参数的样本是从联合后验概率分布中顺序提取的。通过实例分析说明了水泥加固桩取芯试件无侧限抗压强度的统计不确定度。结果表明，从样本量约为40的数据中推断出的统计参数具有相当大的不确定性。研究发现，估计的统计参数的可变性取决于样本大小和空间相关性。在深层搅拌法的质量保证框架内，考察了强度平均值和标准差估计中所引起的统计不确定性的影响。
The statistical parameters for the strength of cement-treated soil are evaluated by the strength of cored samples retrieved from cement-treated columns for a quality assurance procedure in the deep mixing method. The sample parameters include the statistical uncertainty associated with the statistical sample size and other factors. Therefore, a probabilistic characterization of the statistical parameters of strength is required to quantify the statistical uncertainty in the quality assurance process. This paper presents a quantitative analysis of the statistical uncertainty for the estimation of the strength of cement-treated columns. The Bayesian approach is adopted to evaluate the statistical uncertainty occurring in the determination of the statistical parameters of the strength from observed data. The inference is performed via a Markov chain Monte Carlo method, in which samples of the parameters are sequentially drawn from a joint posterior probability distribution. An example analysis is performed to illustrate the statistical uncertainty of the unconfined compressive strength of cored samples retrieved from cement-treated columns. The results show that the statistical parameters, inferred from the data with the sample size of approximately 40, include considerable uncertainty. The variability of the estimated statistical parameters is found to depend on both the sample size and the spatial correlation. The influence of the statistical uncertainty, caused in the estimation of the mean and standard deviations in strength, is examined within the framework of quality assurance in the deep mixing method.
DOI：
10.1201/9781315373010-7
发表时间：
2018-04
期刊：
--
影响因子：
--
作者：
Sumio Watanabe
通讯作者：
Sumio Watanabe
基于岩土可靠性设计的挑战
DOI：
--
发表时间：
2011
期刊：
影响因子：
--
作者：
Y. Honjo
通讯作者：
Y. Honjo