基于数据驱动的盾构隧道地层可掘性识别  

Identification of formation drillability in shield tunnel based on data driven

作　　者：王文正 杨会军 林雪冰 王亚杰 何华飞 师鸿儒 宋克志[2] Wang Wenzheng;Yang Huijun;Lin Xuebing;Wang Yajie;He Huafei;Shi hong ru;Song Kezhi(Beijing Municipal Construction Group,Beijing 100044;Ludong University,Yantai,Shandong Province,264025)

机构地区：[1]北京市政建设集团,北京100044 [2]鲁东大学,山东烟台264025

出　　处：《土木工程学报》2024年第S2期142-147,共6页China Civil Engineering Journal

基　　金：国家自然科学基金(51978322)。

摘　　要：盾构掘进机与岩土体之间具有复杂的相互作用,准确预测盾构隧道地层可掘性是一个巨大的挑战。采用高斯混合模型(GMM)和K近邻算法(KNN)对盾构掘进过程中的岩体可掘性进行分类和预测。引入可掘性指数对岩体进行聚类和分类,包括贯入度(P)、推力贯入度指数(FPI)、扭矩贯入度指标(TPI)和比能(SE)。使用高斯混合模型对训练集内的TPI和FPI进行聚类分析,将岩体的可掘性分为六类。随后,基于KNN分类模型,建立了训练集刀盘速度、推进速度、盾构推力和刀盘扭矩与岩体可掘性之间的映射关系。研究表明,当K设置为4时,该模型具有较高的宏观F1值、宏观查准率P和宏观召回率R,说明了该方法的可行性和有效性。Currently,the accurate prediction of formation drillability in shield tunnel remains a significant challenge due to the complex interactions between the shield machine and the rock mass.The Gaussian mixture model(GMM)and K-nearest neighbor algorithm(KNN)are used to classify and predict the rock mass drillability in the shield machine excavation process.The drillablity indexes are introduced to cluster and classify the rock mass which contain the penetration(P),field penetration index(FPI),torque penetration index(TPI)and specific energy(SE).Clustering analysis of TPI and FPI within the training group was conducted using the Gaussian mixture model,and six drillability categories of rock mass were classified.Subsequently,the mapping relationship between cutterhead speed,advance speed,advance force,and cutterhead torque in the training group and the drillability of rock mass was established based on the KNN classification model.The results revealed that when the K is set to 4,the model has high macro-F1,macro-P and macro-R.Compared to the testing group,this method has been proven feasible and effective.

关 键 词：盾构掘进机 地层可掘性 数据挖掘 高斯混合模型 K最近邻分类 

分 类 号：U455.43[建筑科学—桥梁与隧道工程]