【作者】 孟陆波；

【导师】 李天斌；

【作者基本信息】 成都理工大学 ， 岩土工程， 2009， 博士

【摘要】 隧道施工过程中,各类不同成因的不良地质和施工地质灾害常常成为制约隧道修建的最主要因素。因此,对掌子面前方的地质条件和可能的地质灾害开展超前地质预报,对确保隧道施工安全具有举足轻重的作用。本文采用工程地质、岩体力学、地球物理探测、数字图像识别、非线性科学理论、现代信息技术相结合的方法,从隧道不良地质超前预报、隧道施工地质灾害预测和隧道超前地质综合预报三大方面对隧道超前地质预报技术进行深入研究,并最终开发了一套较为完善的隧道超前地质预报计算机辅助软件系统。论文取得的主要成果如下:(1)从TSP(Tunnel Seismic Prediction)、地质雷达(Ground Penetrating Radar)、瞬变电磁法(Transient Electromagnetic Methods)、BEAM(Bore Tunnelling Electrical Ahead Monitoring)的探测原理出发,通过分析其探测隧道常见不良地质的典型案例,总结并提炼了TSP对空洞、断层破碎带、含水溶隙等不良地质的波形特征;地质雷达对完整岩体、破碎岩体、空洞、含水溶洞等不良地质的波形特征;瞬变电磁法对富水断层、干燥断层、富水溶洞、干燥溶洞、含水溶隙裂隙等不良地质以及不良地质组合和后方低阻体的响应特征;Beam对岩溶含水体、含水裂隙、充水断层带、无水断层等不良地质的响应特征。这些成果为建立超前地质预报解译标志、提高隧道超前地质预报精度奠定了良好基础。(2)探讨了数字图像模式识别技术在超前地质预报中的应用。通过地质雷达图像预处理,采用神经网络模式识别技术,探索地质雷达溶洞图像的智能识别,建立了智能识别二维地质雷达图像中由于含水小溶洞造成双曲线特征的神经网络模型,为识别地质雷达图像中的其它不良地质体提供了新途径。(3)探讨了岩爆与洞室形态和尺寸的关系。通过分析同一尺寸不同洞室形状、同一洞室形状不同尺寸下的岩爆强度,得到在其它影响因素相同条件下岩爆存在洞室尺寸和洞形效应。椭圆形洞(谐洞)对减少岩爆最有利,岩爆将随洞室开挖尺寸的增大而增强,但最终趋于一个稳定值。将这一成果应用于锦屏二级水电站引水隧洞和辅助洞的岩爆预测中,结果表明,在相同桩号、岩爆其它发生条件相似的条件下,引水隧洞中段的岩爆强度将比辅助洞提高半个等级。(4)在围岩大变形预测方面,提出了根据洞壁位移-时间观测曲线预测围岩失稳时间的方法。认为洞壁位移-时间观测曲线的拐点即为围岩失稳的时间,由此建立了多阶非线性回归预报模型,求解了围岩失稳时间的数学表达式,并利用最小二乘法,对监测数据进行拟合,求得预报模型的参数,即可预测围岩失稳的时间。该方法一般适用于围岩失稳的临短预报,其可靠程度取决于位移-时间观测曲线的连续性。(5)从预测隧道掌子面前方涌突水危险性角度出发,建立了隧道涌突水危险性的经验公式:α=Log 10 ((A1×A2)/(B1×B2)),其中a为一无量纲的涌突水危险性综合系数,A1为岩墙厚度,A2为岩墙强度,B1为富水情况,B2为水压条件,并从地层岩性、地质构造等地质指标和TSP、地质雷达、瞬变电磁等物探相关参数指标对地下水富集条件进行综合判断。通过实例验证表明,隧道涌突水危险性综合系数a基本能反映掌子面前方涌突水危险程度的高低。(6)提出了“以地质分析为核心,综合物探分析,洞内外结合、长短预测结合,物性参数互补”的综合预报原则,建立了隧道超前地质综合预报的工作体系。在此基础上,构建了隧道常见不良地质的综合预报方法和工作流程,并选取相关地质参数和物探成果参数,采用模糊神经网络技术,建立了隧道常见不良地质的综合预报模型。该综合预报体系在铜锣山隧道得到较好应用,为确保隧道施工安全提供了保障。(7)以工程地质、岩体力学、岩土工程监测、地球物理探测技术、图像识别技术为依据,采用人工神经网络技术、模糊-层次综合评判、模糊神经网络等非线性智能科学技术,利用数据库和软件工程技术,集成开发了隧道超前地质预报计算机辅助软件系统(TGP-CAS)。主要功能包括隧道基本地质条件预报、隧道施工地质灾害预报、地球物理超前探测预报、综合预报和查询等。国内几条长大深埋隧道的应用表明,该软件具有管理方便、预报精度较高的优点。更多还原

【Abstract】 Various geological problems and hazards are the uppermost factors which often restricted tunnel construction; therefore geological forecast in tunnels are very important for the safety of tunnel construction. This paper researches the forecast of geological problems, the prediction of geological hazards during construction and synthetically forecast of geological condition ahead the workface with the disciplines of engineering geology, rock mass mechanics, geophysical exploration, digital image identification, nonlinear science theory and modern information technology. Finally, a computer assistant prediction software system for tunnel geological forecast has been obtained . The main results are as follows:(1) Base on the principles of TSP(Tunnel Seismic Prediction), GPR(Ground Penetrating Radar), TEM(Transient Electromagnetic Methods) and BEAM(Bore Tunneling Electrical Ahead Monitoring), typical examples on geological problems forecast are collected and analyzed. Then the seismic wave characters for empty cave, fault, hydrous crack by TSP, the electromagnetic wave characters for intact rock mass, rock with cracks, empty cave, hydrous cave by GPR, the response characters for hydrous fault, dry fault, hydrous cavity, dry cavity, hydrous crack,and low electric resistance by TEM, the response characters of hydrous cavity, hydrous crack, hydrous fault,dry fault by BEAM are analyzed and summarized. These results are helpful for creating elucidative criterion of tunnel geological problems forecast and improving prediction accuracy.(2) The application of digital image mode identify technology for tunnel geological problems forecast has been discussed. The intelligent identification methods of GPR image for cave has been explored by pre-deposal of the image and the artificial nervous network identify technology. A nervus network model which can intelligently identify hyperbola which means hydrous cave in GPR image has been gained. This method offers a new way for identifying other unfavorable geological problems in GPR image.(3) The relationships of rock burst and chamber physical dimension has been discussed. By analyzing the grade of rock burst with the condition of same dimension but different shape as well as the same shape but different dimension. Several conclusions have been gotten as fellows. Rock burst has the chamber shape effect. Ellipse is the best section shape of tunnel for reducing rock burst on condition of same chamber dimension. Rock burst has chamber dimensional effect. The grade of rock burst will enhance with the chamber dimension enlargement and finally stay at a constant on condition of same chamber shape. The methods were used to predict rock burst of a certain hydroelectric power station’s and traffic tunnels. The result indicated that the grade of rock burst in the middle section of hydropower tunnels will increase 0.5 grade than traffic tunnel.(4) On the field of large deformation prediction, a method of predicting collapse time of by sidewall displacement-time monitoring data was expounded. The inflection point of displacement-time curve is regarded as the collapse time . the model of multistage nonlinear regression analysis has been created. The mathematic representation of surrounding rock collapse time has been solved. The parameters of the model were gained by fitting the data with least-squares procedure. The method is suit for critical prediction of surrounding rock collapse. The degree of reliability depends on the continuity of displacement-time monitoring data.(5) Advancing a new method for prediction of water gushing risk in front of the workface. A empirical equation about tunnel water gushing risk coefficient has been constructed. That isα=Log 10 ((A1×A2)/(B1×B2)), a is water gushing risk coefficient with no unit, A1 is thickness of wall rock, A2 is intensity of wall rock, B1 is degree of containing water, B2 is condition of water pressure. It judges degree of water contain from two factors: one is geologic factors such as rock character and geologic structure, the other is geophysical exploration parameters such as TSP, GPR and TEM. It is proved that the coefficient a can really show the water gushing risk degree in front of workface by some tunnel examples.(6) A principle of comprehensive prediction that“Take the geological analysis as the core, combine geological analysis and geophysical exploration, combine the condition inside and outside of the tunnel, combine long distance prediction methods with short distance prediction methods and fetch up mutually geophysical parameter”has been concluded. The system of geological comprehensive prediction in tunnel has been established. Then, the comprehensive prediction methods and workflows of common unfavorable geological condition in tunnel have been created. Through extracting geological and geophysical parameters, a comprehensive prediction model of geological problems has been constructed with fuzzy neural network. The comprehensive prediction system was used in TongLuoShan tunnel which leads to pretty good effect and ensure the tunnel’s safety.(7) According to the theory of engineering geology, rock mass mechanics, geotechnical engineering monitoring, geophysical exploration and digital image identification, using nonlinear intelligent science technology such as artificial nervus network, fuzzy arrangement synthesis evaluate, fuzzy neural network, utilizing data-base and software engineering technique, a software about tunnel geological prediction computer assistant system (which named TGP-CAS) has been developed. This software has four main functions. The first, it can manage and predict the basic geology conditions of tunnel. The second, it can forecast tunnel’s geological hazards such as rock burst ,l arge deformation, rock slip and gushing water. The third, it can manage the four geophysical explorations such as TSP, GPR, TEM and BEAM. Finally, it can synthetically predict tunnel geological problems and synthetically query all kinds of information. The software was applied in some over-length and deep tunnels, the results indicate that it has many advantages such as easy management, high prediction precision.更多还原