Research On Fine Descriptive Method Of Geometric Characteristics Of Complex Rock Mass Structures

Fine characterization of geometric characteristics of complicated rock mass structures is an important part of fine description of engineering rock masses.The geometric description of rock mass structures is mainly the quantitative description of the geometric characteristic parameters of fractrures and the spatial distribution and connectivity of fracture networks in rock masses.The fracture network in rock masses usually plays a key role in controlling the mechanical properties,hydraulic properties and engineering stability of rock masses.However,how the fracture network and its related geometric characteristics really affect the rock mass characteristics is still a problem worthy exploring continuously.In such case,the complete three-dimensional characterization of fracture networks is very important for the fine description of engineering rock masses,and it is also one of the essential basic tasks in rock mass classification and geotechnical engineering design.This thesis takes Songta hydropower station on Nu River and Xulong hydropower station on Jinsha River in Southwest China as engineering backgrounds.Following the guiding ideology of“comprehensive research→disaggregate research→comprehensive research”on the description of the geometric characteristics of complex rock mass structures,some new methods and theories of fine characterization of the geometric characteristics of complex rock mass structures were introduced,aiming at quantifying and systematizing the characterization of rock mass structures.In this thesis,the disaggregate research mainly focused on fracture orientation,fracture trace length and fracture abundance in the geometric characteristics of rock mass structures,while the comprehensive research focused on the generation of comprehensive homogeneous domains of rock masses,the three-dimensional characterization simulation of rock mass structures and the connectivity of fracture networks of rock masses.The main research contents and achievements obtained in this thesis are summarized as follows:（1）For the complicated fractured rock masses at Xulong dam site,a quantitative,logical and reasonable progressive framework was proposed to demarcate comprehensive homogeneous domains by rationally utilizing and integrating all kinds of available data.Since the proposed framework comprehensively considers geological,geotechnical and structural aspects,the obtained comprehensive homogeneous domains of rock masses have practical engineering significance.In addition,to better quantify the spatial distribution characteristics of rock fracture networks,the concept of fracture tensors was first introduced to determine the structural similarity of rock masses.（2）The comprehensive characterization of orientation data in fractured rock masses is studied and a framework combining the idea of fractal geometry,data visualization and Schmidt upper-hemisphere equal-area projection plot was proposed to quantitatively describe the geometric distribution and statistical characteristics of orientation data.The fractal indicator（i.e.,monofractal dimension）was proposed to quantitatively describe the dispersion degree of orientation data distribution,and the multifractal spectrum was introduced to describe the nonlinear distribution characteristics of orientation data.（3）In this thesis,a finite mixture model composed of multiple Fisher distribution componets was proposed to simulate the observed complex orientation data set,so as to realized the automatic clustering and simulation of fracture orientation data in one step.In the proposed workflow,the spherical K-means algorithm was applied to select the initial cluster centers to avoid the defect that the clustering result is sensitive to the initial cluster centers,so as to ensure that the global optimal solution can be obtained.The component-wise expectation-maximization algorithm using the minimum message length criterion was used to automatically determine the optimal number of fracture sets.An additional advantage of the proposed method is the representation of orientation data using a full sphere,instead of the conventional hemispherical characterization.The use of a full spherical representation effectively solves the issue of clustering for fractures with high dip angles.The results of artificial simulation and measured fracture orientation data sets show that the proposed method based on finite mixture model has excellent performance.（4）To better describe the characteristics of rock fracture length,a new distribution-free method using L-moment theory and improved multi-scanline method was proposed to infer the fracture trace length distributions in long narrow sampling windows,which are common in large-scale hydropower project and deep rock mass engineering.The proposed method emphasizes the use of contained trace lengths to replace the traditional observed trace lengths for distribution inference and parameter estimation.In addition,a correction method for orientation bias of curved sampling windows was proposed,which makes it possible to use curved sampling window to infer trace length distribution.（5）How to simulate and characterize the heterogeneity of fracture parameter distribution in large-scale rock mass on the scale of engineering was introduced in detail in this thesis.By using the biased one-dimensional and two-dimensional fracture abundance indexes collected along the tunnel in the abutment rock mass of Songta dam site,the three-dimensional fracture abundance indexes P₃₀ and P₃₂ in the rock mass near the tunnels were deduced.Then,the spatial distribution of P₃₀ and P₃₂of abutment rock masses was simulated by sequential gaussian simulation method of geostatistics,and then the spatial distribution characteristics of fractures in abutment rock mass were reproduced.In this way,we can fully capture the heterogeneity of entire rock masses at engineering scale.（6）According to the two-dimensional fracture information obtained from the tunnel namely PD231 at Songta hydropower station,a discrete fracture network model was generated using the theory of probability and statistics to synthetically characterize the fracture network with the rock mass.A stepwise approach based on bounding volume technology and separating axis theorem was proposed to accelerate the fracture intersection analysis in the discrete fracture network model.On this basis,the connectivity of the discrete fracture network was analyzed.The main practical conclusion is that the sub-vertical fracture set 3 and the shallow fracture set 1constitute the main flow pathways in the fractured rock mass at the dam site.Besides,the results suggest that the sub-vertical fracture sets（set 2 and set 3）are more connected than the shallow fracture set（set 1）.Moreover,it is found that the self-intersection behavior of fracrues within a fracture set is related to the orientation dispersion,fracture intensity and fracture size.