采矿与安全工程学报2025，Vol.42Issue(2)：440-451,12.DOI:10.13545／j.cnki.jmse.2023.0416
基于数字钻进的层状岩体界面及岩石强度识别试验研究
Experimental study on identification of layered rock mass interface and rock strength based on digital drilling
摘要
Abstract
Accurate prediction of rock properties is one of the major guarantees of safe tunnel construc-tion.In recent years,the rapidly developing digital drilling technology has provided more convenient methods for predicting the mechanical properties of rock.To reveal the machine-rock interaction during rotary cutting drilling of layered rock mass,a method for predicting rock strength and rock mass interface based on machine learning and the digital drilling technology was proposed here.Digital drilling tests were carried out on layered rock mass combinations with different strength grades,and three ma-chine learning algorithms,i.e.,back propagation(BP)neural network,support vector machine(SVM),and CatBoost,were employed to identify rock mass interface and rock strength.The results demonstrate that the machine learning methods can accurately identify the interface position,stratum thickness,and strength of layered rock masses.The balanced accuracy and F1-score of interface recogni-tion both exceed 95％,and the prediction error is below 5％,indicating high prediction accuracy.The overall error range of the predicted rock strength parameters and the measured results is within 10％.Among these three algorithms,the BP neural network is highly accurate in predicting the strengths of all types of rock,boasting a superior prediction effect.The research results are expected to provide reference for realizing dynamic in-situ detection of rock mass in coal mine roadways.关键词
数字钻进/机器学习/层状岩体/强度预测/界面识别Key words
digital drilling/machine learning/layered rock mass/strength prediction/interface recog-nition分类
建筑与水利引用本文复制引用
岳小磊,岳中文,马文彪,李杨,闫逸飞..基于数字钻进的层状岩体界面及岩石强度识别试验研究[J].采矿与安全工程学报,2025,42(2):440-451,12.基金项目
国家重点研发计划项目(2021YFC2902103) （2021YFC2902103）
