【作者】 秦胜伍；

【导师】 陈剑平；

【作者基本信息】 吉林大学 ， 地质工程， 2009， 博士

【摘要】 本文以“浙江诸永高速公路台州段超前地质预报与监控量测”为依托,在国家自然科学基金“裂隙岩体隧道施工地质灾害三维信息化超前预报研究”(40272117)基础上,结合白鹤隧道施工阶段的实际情况,构建了隧道三维工程地质数据库。基于COM的GIS组件ArcObjects9.2进行二次开发,以Visual Basic作为开发平台,进行系统集成,开发隧道施工超前地质预报系统。在开发的平台上,对隧道实体、地层、节理裂隙、钻孔数据进行了三维实体建模,进行三维可视化模拟。将不连续面三维网络模拟与GIS结合,在3DGIS环境下进行不连续面三维网络模拟。将数据挖掘理论引入到超前地质预报系统中,针对三维工程地质数据库进行知识发现和数据挖掘,从空间和时间两个角度,利用开挖揭露的地质信息对掌子前方可能出现的塌方等灾害进行超前地质预报。空间角度上,利用粗糙集理论的属性约简,判断围岩稳定性影响因素,并进行权系数计算。利用人工神经网络、可拓学理论进行对隧道围岩的稳定性、可能出现的塌方等灾害进行预测,开拓了根据开挖揭露的地质信息进行非线性超前地质预报的方法。时间角度上,将BP神经网络与ARMA模型相结合,从趋势性和随机性两方面,对围岩压力等监测数据进行趋势提取和预测。在GIS平台上将超前地质预报结果三维可视化输出,可以利用GIS强大的查询和空间分析功能,便于结果的综合查询分析。更多还原

【Abstract】 With the rapid development of national economic, projects of hydraulic engineering and traffic engineering is flourishing, and so is the tunnel construction, which is always confined mainly by various kinds of geological disasters. In order to ensure the normal operation of tunnel construction, it plays a decisive role to do geological prediction in the process of tunnel information-based construction.Although more and more attention are paid to geological prediction ahead of working face during tunnel construction at home and broad, the prediction level is not high and the method is not mature. The key reason of this situation is that the information of most prediction method is from investigations before construction, or simply depends on geophysics explorer means and ignores the geological information disclosed during construction. How to predict geological situation ahead of working face using the geological information disclosed during construction, is the core of information construction. In this thesis, this thought is through out the whole research.Relying on Geological Prediction Ahead of Working Face during Tunnel Construction of Zhuyong Highway (Taizhou Part) in Zhejiang Province and National Natural Science Foundation of P. R. China (40272117), some research is carried on in this thesis. Baihe tunnel, a part of Zhuyong highway in Zhejiang province, locates in Xianju County, the physiognomy features of which are mid-low hill controlled by cathaysian and neocathaysian structural system. The dominant rock mass is weakly or slightly weathered tuff. Joints and cracks are well developed, mainly trending N.W and N.N.E, widely opened, weak cementation, with a result of collapse and water inrushing.In this thesis, it is introduced the standard of layer division and property structure of all layers in the 3D geological information database based on GeoDataBase. 3D geological information database of Baihe tunnel is built, which consists of graphics library including contour, drilling hole, stratum information, joints and cracks, and monitoring instruments position, etc. and property databases, such as database of joints and cracks, monitoring database. And geological information disclosed in the process of tunnel excavation is spatial informationized, which provides a database support to do data mining and geological prediction ahead of working face.Through a deep understanding of the 3D GIS model, the 3D modeling approach working for geological prediction ahead of working face during tunnel construction is developed. The way of building 3D tunnel model based on GIS, and conversion between geodetic coordinates and construction stake number are introduced as well as 3D model of Baihe tunnel is successfully built. Besides, the way to express and construct drill data is suggested, and then the 3D stratum model of Baihe tunnel is built based on the drill data and Kriging interpolation. Finally, 3D display of Baihe tunnel as well as simulation and analysis come to fulfillment, after the construction of 3D physical model and joints and cracks model.Based on the GIS platform, mining useful knowledge from the previous 3D GeoDataBase, sufficiently analyzing the engineering information and geological data disclosed during construction, the intellectualized prediction model of 3D information in advance of geological disaster during construction is built. The general technical route of the study is informationized construction that is forecasting the condition of undisclosed part and modified the original design in real time according to the information of disclosed part. Thus, it provides the basis to modify support scheme, excavation method and working sequence, is a technical support and protection to safe construction reducing geological disaster.Introduce data mining theory to the system of geological prediction in advance to do knowledge finding and data mining in 3D engineering geological database and predict in advance the disaster possibly existing ahead of tunnel working face in the view of time and space using the disclosed geological information. From the standpoint of space, draw information of dominant instruction surface after analyzing the joint data from excavation process using 3D net work numerical modeling technique for random discontinuities of rock mass and fuzzy clustering analysis; build decision information table of tunnel geo-disaster using rough set theory; do data mining and knowledge finding in 3D engineering geological database adopting the function of attribution reduction in rough set theory; judge the influence factors to rock mass stability as well as the weight coefficient computing; and predict the possible disaster and stability of rock mass applying artificial net work and extenics theory. From the standpoint of time, combining the BP net work and ARMA model, do trend abstract and prediction of the monitoring data of rock mass pressure in both aspects of trendy and random. 3D visual output of prediction result on the platform of GIS, which is powerful in query and spatial analysis, is convenient to query and analysis of result.The main algorithms adopted in this thesis are rough set theory, fuzzy clustering analysis, artificial neural network, and Extenics in data mining theory. Based on the Arc Objects 9.2—GIS component of COM, with the development platform of Visual Basic, the system of tunnel construction forecasts in advance is developed on the basis of GIS. The system comprises some modules such as the management of spatial database, the construction of 3D model, the simulation of 3D net, spatial overlay analysis, spatial query, geological forecast, etc. what the system can do are data inputting, inquiring and managing of drilling data, joints statistic data and monitoring data; 3D model building of drilling hole, stratum and tunnel; 3D net simulation of discontinuities based on GIS, as well as computing RQD, abstracting dominant surface; spatial analysis and inquiring characterized by click inquiry and location query by special property; geological prediction ahead of working face by both approach of ANN and extenics; 3D rock classification, and analyzing the influence basin of shuttered zone applying the 3D analysis of the buffer zone in GIS; and forecasting time series of monitoring data by both approaches of BP network and BP-ARMA.Adopting the way and the system mentioned in this thesis, geological prediction ahead of working face is pursued in typical part of Baihe tunnel, and some conclusions can be obtained as below:1. The main origins of collapse and water inrushing in Baihe tunnel are its considerable length, developed joints and cracks, widely opened structure surface and its weak cohesion.2. By the attribute reduction in rough set, to analyze the affect factor of the stability of rock mass in Baihe tunnel, the result suggests that RQD, the angel between the tunnel axes and the controlling structural plane, the type of structure plane, the length of structure plane, etc. influenced the stability by different level, especially the type of structure plane as well as the character of groundwater.3. Comparing the results of stability of rock mass predicting from extenics and BP network, it can be found that these two prediction results are similar and accordant actual situation, which indicate that nonlinear theory is effect in dealing with geological prediction ahead in tunnel construction4. Time series analysis of rock mass pressure using BP-ARMA model indicate that this model can be used to analysis and predict monitoring data of rock mass pressure. Trend term extracted by BP reflects the overall trend of rock mass pressure, and ARMA reflects the random induced by rock mass structure and excavation progress. Compared with BP prediction, this method has higher accuracy.5. The prediction result of rock mass stability indicate that geological prediction ahead in tunnel construction using geological information disclosed in excavation is scientific and feasible. The research method is innovative and has application and promotion value.更多还原