1. Post-doctoral Research Station of Safety Science and Engineering,Anhui University of Science and Technology,Huainan　 232001,China; 2. School of Ar-chitectural Engineering,Nantong Vocational College,Nantong　 226001,China; 3. School of Resources and Environmental Engineering,Anhui University,Hefei　 230022,China; 4. School of Civil Engineering and Architecture,Anhui University of Science and Technology,Huainan　 232001,China

In order to resist complex and changeable loading of underground structure engineering effectively,and im- prove the safety and reliability of the shaft lining,regarding shafts in the deep alluvium of Huainan and Huaibei mining area as the prototype and pouring reinforced concrete shaft lining model according to the similarity principle,a lot of ultimate bearing capacity tests of reinforced concrete lining models are conducted. The result shows that the main fac- tors affecting load bearing capacity are concrete compression strength,ratio of lining thickness to inner radius and reinforcement ratio. Among them,the impact of concrete compressive strength on shaft lining bearing capacity is obvious, and the impact of reinforcement ratio is weak. However,various influencing factors are accompanied by varying degrees of uncertainty in practical engineering. Aiming at the fuzzy random of ultimate bearing capacity in deep alluvium,based on the sample big data set of shaft lining structure parameters and tests of high strength reinforced concrete in Huainan and Huaibei mining area,the uncertainty distribution of structural materials,geometric parameters and calculation mod- el are analyzed to obtain the fuzzy random distributive rules of concrete compression strength,ratio of lining thickness to inner radius and reinforcement ratio. The traditional data mining HMM model is improved by using the algorithm of maximum expected (EM). The maximum likelihood estimate value is calculated in step E and the parameter expecta- tion estimate is calculated in step M respectively. The improved model has gone through two fuzzy random processes. Compared with the original algorithm,it has the advantages of small error,high efficiency and fast convergence,thus can better suit the uncertain characteristics of actual underground engineering. Based on the improved data mining al- gorithm,the integrated fuzzy random distribution of structural materials,and the geometric parameters and calculation model under big data environment, an ultimate bearing capacity fuzzy random model with big data mining of high strength reinforced concrete shaft lining has been set up,and proved to be more reasonable and practical for engineer- ing,thus providing reliable references for the design of reinforced concrete shaft lining structural parameters in deep al- luvium in the future.

国家自然科学基金资助项目(51874005，51374010，51474004);江苏省建设系统科技计划资助项目(2017ZD062);南通市级科技发展计划资助项目(MS12018054)

YAO Yafeng,CHENG Hua,RONG Chuanxin,et al. Fuzzy random analysis on ultimate bearing capacity based on big data mining in deep alluvium[ J]. Journal of China Coal Society,2020,45 (3):1089 -1098.