Research On Theories And Methods Of Geological Structure Modeling In Hydropower Projects Based On Uncertainty Analysis

With the development of hydropower projects in China,more and more hydropower projects are constructed in areas with complex geological condition.Since geological structure will exert great impact on the stability as well as permeability of rocks,it is indispensable to investigate geological structure in detail during construction of hydropower projects.However,the uncertainty of geological structure is widely spread because of the uncertainty from itself,and the limitation of geological investigation technology as well as current cognitive level.This uncertainty will influence the accuracy of geological model,the design of scheme,and finally affect the schedule,cost and quality of the projects.Therefore,it is one of critical problems to analyze the position,attributive classification and spatial relationships of geological structure with uncertainty theory,and build up accurate geological structure model in the research on the geological structure of hydropower projects.In-depth research and analysis on the problems mentioned above are shown in this thesis,and the achievements are listed as follows.1.Considering the high uncertainty of geological structure in hydropower projects,theory framework and mathematical model of geological structure modeling on the basis of uncertainty analysis are proposed.First,uncertain factors in the modeling of geological structure are analyzed.Then,according to the analysis results,modeling framework and mathematical model of geological structure on the basis of uncertainty analysis are proposed.Finally,with decomposition analysis of uncertainty problems in the modeling of geological structure,mathematical model for analyzing uncertainty of geological structure data in hydropower projects,mathematical model for modeling 3 dimensional(3D)large scale geological structure and that for simulating 3D fracture network are proposed.This work can provide the theoretical basis for uncertainty analysis and modeling of geological structure in hydropower projects.2.Based on classification study of the uncertainty existing in modeling data of hydropower engineering geological structure,uncertainty analysis method for location,attributes classification and spatial relationship of geological structure hydropower projects is proposed.Geological structure position,attribute and spatial relationship are basis data for the modeling of geologic structure in hydropower project,which are also featured with uncertainty.The uncertainty will directly affect the accuracy of geological structure model.Therefore,it is necessary to make a deep analysis of the uncertainty of geological structure position,attribute and spatial relationship in hydropower projects.This thesis proposes the method that using point error ellipse model and line error ellipse model to analyze position uncertainty of geological structure in hydropower projects.The approach can overcome the deficiency that current geological section drawing adopts a series of line to approximate curves,and ignores the intermediate point error of line element.Then,the cubic spline curve is used to fit the data of geological structure point.When compared with random polyline error band,stratigraphic position error resulting from cubic spline curve in fitting stratigraphic demarcation point is smaller.By comparing with the random line error band,it is proved that the cubic spline curve is less in the uncertainty of the stratigraphic position when fitting the geological boundary points.In terms of uncertainty analysis of attribute classification of geological structure,attribute classification method based on rough set and conditional entropy is proposed.This approach can solve the problem that current attribute classification method is difficult to calculate the attribute weight and cannot evaluate the classification accuracy,and avoid subjectivity.In addition,a classification accuracy evaluation method based on error matrix is proposed,which comprehensively evaluates classification accuracy with indictors of user accuracy,producer accuracy,commission Errors,omission error and Kappa coefficient.For the uncertainty analysis of spacial relations of geological structure,a unified model of geological structure spacial relations is proposed.In this model,Region connection calculus theory(RCC),3D27 direction relation model and extended Euclidean distance are adopted to express topological relations,direction relations and distance relations,which overcomes the insufficient that current research lacks description of spacial relation of geological structure objects and the capacity to describe 3D spacial relation is inadequate.3.For the uncertainty problems of data discreteness,inhomogeneity and geometry complexity in large-scale geological structure of hydropower projects,a refined modeling method of 3D geological structure based on T-splines is proposed.There is difficulty in subdivision modeling of engineering key areas in current studies,and the ability of describing complex geometry of geological structure is also insufficient.Based on NURBS(Non-Uniform Rational B-Splines)modeling,a new 3D geological modeling data structure named T-splines–NURBS–BREP and a refined modeling method of large-scale geological structure are proposed.There are three advantages of this method:(1)it inherits the accuracy that NURBS uses parameter analytical form to express geologic body surface,and can realize satisfactory fitting of geological discrete data;(2)since T-splines has a feature of locally refinement,it can realize local subdivision modeling of project key areas and build geological structure with fractal complexity.(3)T-splines subdivision surface has the flexibility of presenting any topological structure,so that it can construct complex geological structure model with high genus.In general,the refined modeling method of 3D geological structure based on T-splines improves the modeling precision of large-scale geological structure and the ability of describing complex geological structure,which provides an effective technology for geological modeling and analysis in hydraulic engineering.4.Considering the uncertainty problems of data incompleteness,uncertain geometry and random distribution in rock fractures of hydropower projects,simulation method of 3D fracture network based on Latin Hypercube Sampling(LHS)is put forward.Current research always ignores the effect that fracture data below exposed surface exerts on simulation results,and Monte Carlo method(MC)has become the main sampling method which can lead to poor stability and low precision,and sample collapse problem is easy to occur.In light of this,this paper proposes simulation method of 3-D fracture network in fractured rock based on LHS.The fracture data used from exposed surface is supplemented and corrected by borehole camera data,and fracture parameters such as location,size,quantity,and orientation are analyzed.Then LHS method is adopted for stochastic simulation of fracture parameters.Finally,3-D fracture network model is established using 3-D modeling software VisualGeo.When compared with MC method,LHS can realize better fitting degree of random sample to distribution and space filling capacity,which provides accurate 3-D fracture network model for dam foundation stability analysis of hydraulic engineering.The theory and method of geological structure modeling based on uncertainty analysis is applied to a hydraulic project in southwest of China.It is shown that the theory and method proposed in this thesis are able to analyze the uncertainty of geological structure data reasonably,accurately depict geometric complexity and topological complexity of terrain body,cover layer bod and,lens body,establish accurate 3-D fracture network model,which provides accurate model and decisionmaking basis for the analysis of hydraulic engineering geological structure.