摘要：The geological condition is complicated andchangeable, and difficult to predict accurately, so the shieldtunneling construction will face great risks, even lead to serioussafety accidents. Real-time and accurate geologicalidentification and prediction is an important guarantee for thesafe construction of shield machine, which is also an importantproblem of shield technology. Based on the real-time tunnelingcontrol parameters of shield machine construction, anunsupervised data tree K-Means cluster method is proposed inthis paper. Firstly, the correlation between shield tunnelingparameters and geological types is analyzed to determine themain tunneling parameters which are greatly affected bygeological changes. Secondly, the unsupervised data tree is usedto optimize the K value of K-Means cluster algorithm, and thena geological recognition cluster algorithm based on maintunneling parameters is constructed. Finally, a simulationexperiment is carried out based on the field construction data,and the accuracy of geological identification is 100%. Theresults show that the method can accurately identify geologicalcategories and provide decision support for effective parameterregulation of shield machine.
关键词：Shield machine; tunneling parameter; geological identification; K-Means cluster; data tree
