大跨度公路隧道围岩动态分级的Bayes优化  

Optimization of Rock Mass Dynamic Classification of Large Span Road Tunnel Based on Bayes

作　　者：马超锋[1] 李晓[1] 成国文[1] 赫建明[1] 

机构地区：[1]中国科学院地质与地球物理研究所工程地质力学重点实验室,北京100029

出　　处：《地下空间与工程学报》2011年第2期322-328,共7页Chinese Journal of Underground Space and Engineering

基　　金：国家重点基础研究发展计划(973)项目(2009CB724605;2010CB731501);国家自然科学基金资助项目(40702052);中国科学院知识创新工程重要方向项目(KJCX3-SW-L1;KZCX2-YW-109)

摘　　要：基于贝叶斯统计理论,充分重视隧道掌子面相同或相似工程岩组已有的围岩分级信息,建立围岩质量指标[BQ]的先验分布和后验分布的概率模型,推导出围岩质量指标[BQ]的后验分布超参数的计算表达式,获得围岩质量指标[BQ]的可信概率为1-α的可信区间,并据此对公路隧道施工阶段的围岩动态分级进行优化。通过某大跨度公路隧道的若干个掌子面的围岩动态分级实践和应用,验证了贝叶斯优化方法的可行性。结果表明:基于贝叶斯优化方法进行的隧道掌子面围岩动态分级有效降低了施工阶段围岩分级的主观差异,据此进行的围岩级别划分更加合理。Based on the Bayesian statistics,the dynamic classification message of the same or similar engineering rock mass is fully considered.The probability models of the prior distribution and the posterior distribution of are founded,the expressions of the super-parameters of the posterior distribution are deduced,and further the confidence interval of at 1-α confidence level is obtained.Then the dynamic classification results of rock mass are optimized in order to weaken the subjective difference caused by human factors.Taking a road tunnel as example,the feasibility of this approach is validated.It is shown that the subjective difference is weakened by the optimized dynamic classification effectively.Hence,the classification results based on Bayesian statistics are more rational.

关 键 词：公路隧道 围岩动态分级 贝叶斯统计 优化 先验分布 

分 类 号：U452.12[建筑科学—桥梁与隧道工程]