The present work describes investigations on the anisotropic strength behavior of rocks in the splitting tensile test (Brazilian test). Three transversely isotropic rocks (gneiss, slate and sandstone) were studied in the Lab. A total of more than 550 indirect tensile strength tests were conducted, with emphasis was placed on the investigation of the influence of the spatial position of anisotropic weakness plane to the direction of the load on the fracture strength and fracture or fracture mode. In parallel, analytical solutions were evaluated for stress distribution and developed 3D numerical models to study the stress distribution and the fracture mode at the transversely isotropic disc. There were new findings on the fracture mode of crack propagation, the influence of the disc thickness, the influence of the applying loading angle and angle of the loading-foliation for transversely isotropic material.:ACKNOWLEDGMENTS 5 ABSTRACT 7 TABLE OF CONTENTS 9 LIST OF FIGURES 13 LIST OF TABLES 19 I. INTRODUCTION 21 Objective of this work 22 Scope of work 23 Research procedure 23 Significance of the work 24 Layout 24 1 STATE OF THE ART 27 1.1 Review of the Brazilian tensile strength test 27 1.1.1 General overview 27 1.1.2 Development of the Brazilian tensile strength test 29 1.1.3 The Brazilian tensile strength test on anisotropic rocks 31 1.1.4 Summary 32 1.2 Analytical aspects 33 1.2.1 Hypotheses for the conventional Brazilian test 34 1.2.2 Failure criteria 36 1.2.3 Crack initiation and propagation 39 1.2.4 Summary 41 1.3 Numerical considerations 41 1.3.1 Numerical methods 42 1.3.2 Summary 42 1.4 Conclusion 43 2 DIAMETRAL COMPRESSION IN A SOLID DISC – COMPILATION OF ANALYTICAL AND SEMI-ANALYTICAL SOLUTIONS 45 2.1 Introduction 45 2.2 Diametral compressive stress distribution in an isotropic elastic disc 45 2.2.1 Elastic theory of line load 46 2.2.2 2D analytical solutions 47 2.2.3 3D disc under line and diametral compressive distributed loads 55 2.2.4 3D solution under diametral compressive distributed load 56 2.3 Stress and strain in an isotropic solid disc 59 2.4 Stress and strain in anisotropic rocks 61 2.5 Conclusion 65 3 LABORATORY TESTS 69 3.1 Introduction 69 3.2 Laboratory test program 70 3.3 Sample preparation 71 3.4 Ultrasonic measurements 72 3.5 Uniaxial and triaxial compression tests 73 3.5.1 Uniaxial compression test 73 3.5.2 Triaxial compression tests 74 3.6 Brazilian tensile strength tests 76 3.6.1 Test apparatus 76 3.6.2 Laboratory test results 77 3.6.3 Interpretation of the test results 89 3.7 Conclusion 96 4 NUMERICAL SIMULATION OF ISOTROPIC MATERIALS - COMPARISON WITH ANALYTICAL SOLUTIONS 97 4.1 Introduction 97 4.2 Numerical simulation of isotropic materials 97 4.2.1 FLAC3D simulation program 97 4.2.2 Simulation procedure 98 4.2.3 Numerical model setup 98 4.2.4 Influence of mesh type 99 4.2.5 Influence of specimen thickness 100 4.2.6 Influence of Poisson’s ratio 102 4.2.7 Influence of loading angle (2) 106 4.2.8 Comparison of 3D analytical and numerical results 110 4.2.9 Influence of stress concentration at the loading jaws 112 4.3 Comparison with experimental results of Postaer Sandstone (FG.Ss) 112 4.4 Conclusion 114 5 NUMERICAL SIMULATION OF ANISOTROPIC MATERIALS - COMPARISON WITH LABORATORY TESTS 117 5.1 Introduction 117 5.2 General procedure for simulating the Brazilian test using FLAC3D 117 5.2.1 Conceptual model 119 5.2.2 Boundary Conditions 119 5.2.3 Numerical model set-up 120 5.3 Constitutive model 121 5.3.1 Choice of constitutive model 121 5.3.2 Bilinear Strain-Hardening/Softening Ubiquitous-Joint Model [98] 121 5.4 Parameter calibration 124 5.4.1 Material parameters used 124 5.4.2 Contact between disc and loading jaws 126 5.4.3 Post-failure deformation properties 128 5.4.4 Tension cut-off 129 5.5 Numerical simulation results 131 5.5.1 Introduction 131 5.5.2 Stress distribution and failure state 133 5.5.3 Stress state in an isotropic elastic medium with arbitrary orientation planes 136 5.5.4 Plasticity states 139 5.5.5 Damage and fracture process 141 5.5.6 Fracture patterns – Comparison of lab results and numerical simulations 148 5.6 Tensile strength – Comparison of lab results and numerical simulations 149 5.6.1 Tensile strength of Le.Gs Gneiss 150 5.6.2 Tensile strength of My.Sc Slate 155 5.7 Summary and Review 159 5.7.1 Potential failure state deduced from pure elastic considerations 159 5.7.2 Tensile strength distribution 160 5.7.3 Tensile strength – determining the anisotropy factor 161 5.7.4 Tensile strength – different procedures - different results 163 6 CONCLUSION AND RECOMMENDATIONS 165 APPENDICES 171 Appendix 3.1 - Fracture patterns in FG.Ss samples 171 Appendix 3.2 - Fracture patterns in FG.Gs samples 177 Appendix 3.3 - Fracture patterns in Le.Gs samples 183 Appendix 3.4 - Fracture patterns in My.Sc samples 190 Appendix 4.1 - Influence of loading angle 197 Appendix 4.2 - Influence of material properties 203 Appendix 5.1 - Failure zone state in Le.Gs Gneiss 209 Appendix 5.2: Failure zone state in My.Sc Slate 216 REFERENCES 223 / Inhalt der Arbeit sind Untersuchungen zum anisotropen Festigkeitsverhalten von Gesteinen beim Spaltzugversuch (Brazilian Test). Laborativ wurden drei transversalisotrope Gesteine (Granit, Schiefer und Sandstein) untersucht. Insgesamt wurden mehr als 550 Spaltzugversuche durchgeführt, wobei der Schwerpunkt auf die Untersuchung des Einflusses der räumlichen Lage der Anisotropieebene zur Richtung des Lasteintrages auf die Bruchfestigkeit und das Bruchbild bzw. den Bruchmodus gelegt wurde. Parallel dazu wurden analytische Lösungen zur Spannungsverteilung ausgewertet sowie numerische 3D-Modelle entwickelt, um die Spannungsverteilung sowie den Bruchmodus bei einer transversalisotropen Scheibe zu untersuchen. Es wurden neue Erkenntnisse zum Bruchmodus, der Rissausbreitung, des Einflusses der Scheibendicke, dem Einfluss des Lasteinleitungswinkel sowie des Winkels Lasteintrag - Anisotropieebene für transversalisotropes Material gewonnen.:ACKNOWLEDGMENTS 5 ABSTRACT 7 TABLE OF CONTENTS 9 LIST OF FIGURES 13 LIST OF TABLES 19 I. INTRODUCTION 21 Objective of this work 22 Scope of work 23 Research procedure 23 Significance of the work 24 Layout 24 1 STATE OF THE ART 27 1.1 Review of the Brazilian tensile strength test 27 1.1.1 General overview 27 1.1.2 Development of the Brazilian tensile strength test 29 1.1.3 The Brazilian tensile strength test on anisotropic rocks 31 1.1.4 Summary 32 1.2 Analytical aspects 33 1.2.1 Hypotheses for the conventional Brazilian test 34 1.2.2 Failure criteria 36 1.2.3 Crack initiation and propagation 39 1.2.4 Summary 41 1.3 Numerical considerations 41 1.3.1 Numerical methods 42 1.3.2 Summary 42 1.4 Conclusion 43 2 DIAMETRAL COMPRESSION IN A SOLID DISC – COMPILATION OF ANALYTICAL AND SEMI-ANALYTICAL SOLUTIONS 45 2.1 Introduction 45 2.2 Diametral compressive stress distribution in an isotropic elastic disc 45 2.2.1 Elastic theory of line load 46 2.2.2 2D analytical solutions 47 2.2.3 3D disc under line and diametral compressive distributed loads 55 2.2.4 3D solution under diametral compressive distributed load 56 2.3 Stress and strain in an isotropic solid disc 59 2.4 Stress and strain in anisotropic rocks 61 2.5 Conclusion 65 3 LABORATORY TESTS 69 3.1 Introduction 69 3.2 Laboratory test program 70 3.3 Sample preparation 71 3.4 Ultrasonic measurements 72 3.5 Uniaxial and triaxial compression tests 73 3.5.1 Uniaxial compression test 73 3.5.2 Triaxial compression tests 74 3.6 Brazilian tensile strength tests 76 3.6.1 Test apparatus 76 3.6.2 Laboratory test results 77 3.6.3 Interpretation of the test results 89 3.7 Conclusion 96 4 NUMERICAL SIMULATION OF ISOTROPIC MATERIALS - COMPARISON WITH ANALYTICAL SOLUTIONS 97 4.1 Introduction 97 4.2 Numerical simulation of isotropic materials 97 4.2.1 FLAC3D simulation program 97 4.2.2 Simulation procedure 98 4.2.3 Numerical model setup 98 4.2.4 Influence of mesh type 99 4.2.5 Influence of specimen thickness 100 4.2.6 Influence of Poisson’s ratio 102 4.2.7 Influence of loading angle (2) 106 4.2.8 Comparison of 3D analytical and numerical results 110 4.2.9 Influence of stress concentration at the loading jaws 112 4.3 Comparison with experimental results of Postaer Sandstone (FG.Ss) 112 4.4 Conclusion 114 5 NUMERICAL SIMULATION OF ANISOTROPIC MATERIALS - COMPARISON WITH LABORATORY TESTS 117 5.1 Introduction 117 5.2 General procedure for simulating the Brazilian test using FLAC3D 117 5.2.1 Conceptual model 119 5.2.2 Boundary Conditions 119 5.2.3 Numerical model set-up 120 5.3 Constitutive model 121 5.3.1 Choice of constitutive model 121 5.3.2 Bilinear Strain-Hardening/Softening Ubiquitous-Joint Model [98] 121 5.4 Parameter calibration 124 5.4.1 Material parameters used 124 5.4.2 Contact between disc and loading jaws 126 5.4.3 Post-failure deformation properties 128 5.4.4 Tension cut-off 129 5.5 Numerical simulation results 131 5.5.1 Introduction 131 5.5.2 Stress distribution and failure state 133 5.5.3 Stress state in an isotropic elastic medium with arbitrary orientation planes 136 5.5.4 Plasticity states 139 5.5.5 Damage and fracture process 141 5.5.6 Fracture patterns – Comparison of lab results and numerical simulations 148 5.6 Tensile strength – Comparison of lab results and numerical simulations 149 5.6.1 Tensile strength of Le.Gs Gneiss 150 5.6.2 Tensile strength of My.Sc Slate 155 5.7 Summary and Review 159 5.7.1 Potential failure state deduced from pure elastic considerations 159 5.7.2 Tensile strength distribution 160 5.7.3 Tensile strength – determining the anisotropy factor 161 5.7.4 Tensile strength – different procedures - different results 163 6 CONCLUSION AND RECOMMENDATIONS 165 APPENDICES 171 Appendix 3.1 - Fracture patterns in FG.Ss samples 171 Appendix 3.2 - Fracture patterns in FG.Gs samples 177 Appendix 3.3 - Fracture patterns in Le.Gs samples 183 Appendix 3.4 - Fracture patterns in My.Sc samples 190 Appendix 4.1 - Influence of loading angle 197 Appendix 4.2 - Influence of material properties 203 Appendix 5.1 - Failure zone state in Le.Gs Gneiss 209 Appendix 5.2: Failure zone state in My.Sc Slate 216 REFERENCES 223

The time-independent and -dependent damage characteristics of brittle rocks inside the pre-failure range have been investigated using numerical simulations and lab testing. Grain-based discrete element models have been developed to simulate both, time-independent and -dependent damage evolution leading to ultimate failure of sandstone and granite, respectively. The models take into account elastic grain and elasto-plastic contact deformation, inter- and intra-granular fracturing and lifetime prediction on the basis of subcritical crack growth. The time-independent mechanical behavior of Coconino sandstone and Lac du Bonnet granite during uniaxial compression tests, Brazilian splitting tests and fracture toughness tests was simulated. Triaxial compression tests and fracture toughness tests for Kirchberg II granite and fracture patterns tests for Eibenstock II granite were carried out in laboratory to perform time-independent damage and failure criterion analysis. The corresponding simulations showed reasonable damage phenomena compared with experimental results. Damage indices were deduced and were applied for different time-independent simulations. Based on calibrations of the time-independent damage simulations of selected brittle rocks, Charles equation and Hillig-Charles equation, which are generally used to describe subcritical crack growth, were implemented into the numerical code to simulate time-dependent damage. One-edged crack growth in Coconino sandstone specimen due to stress corrosion has been analyzed theoretically and numerically. Uniaxial compressive creep tests for Lac du Bonnet granite were simulated and time-dependent behavior in terms of the damage process during primary, secondary and tertiary creep until final failure characterized by macroscopic fracturing was discussed in detail. Subsequent to this, the time-dependent Mode-I crack growth tests and uniaxial compressive creep tests for Kirchberg II granite were carried out and the corresponding simulations were performed. Simulation results are in good agreement with experimental observations. In addition, damage indices and time-dependent fracture development were monitored and illustrated. The developed approach was applied to two potential practical applications: the damage analysis of a sandstone landscape arch and a tunnel. Finally, the results are summarized and recommendations for future work are proposed.:1 Introduction 2 State of the art 3 Time-independent damage analysis 4 Time-dependent damage analysis 5 Applications of numerical models . 6 Conclusions and outlook References

The presented work documents the influence of Voronoi block size and shape as well as internal mesh size on the calibrated fracture toughness KIC. It is documented that Voronoi based procedures have an inevitable error of up to ± 30%. On the other hand, this approach is able to reproduce complex fracture pattern in a realistic manner with reasonable computational power. The work propose a KIC calibration procedure and documents based on the comparison with lab tests, that crack propagation, fracture pattern as well as stress-strain behavior of brittle solids can be duplicated by calibrated Voronoi based DEM simulations. The thesis also documents a swelling law for the DEM code UDEC including parameter determination and validation on lab tests with swelling cement. Finally, calibrated concrete models with one or two holes under different boundary conditions are used to predict swelling induced cracking. Numerical predictions were compared with corresponding lab tests and showed satisfying results.

‘Coupled process’ implies that one process affects the initiation and progress of the others and vice versa. The deformation and damage behaviors of rock under loading process change the fluid flow field within it, and lead to altering in permeable characteristics; on the other side inner fluid flow leads to altering in pore pressure and effective stress of rock matrix and flow by influencing stress strain behavior of rock. Therefore, responses of rock to natural or man-made perturbations cannot be predicted with confidence by considering each process independently. As far as hydro-mechanical behavior of rock is concerned, the researchers have always been making efforts to develop the model which can represent the permeable characteristics as well as stress-strain behaviors during the entire damage process. A brittle low porous granite was chosen as the study object in this thesis, the aim is to establish a corresponding constitutive law including the relation between permeability evolution and mechanical deformation as well as the rock failure behavior under hydro-mechanical coupled conditions based on own hydro-mechanical coupled lab tests. The main research works of this thesis are as follows: 1. The fluid flow and mechanical theoretical models have been reviewed and the theoretical methods to solve hydro-mechanical coupled problems of porous medium such as flow equations, elasto-plastic constitutive law, and Biot coupled control equations have been summarized. 2. A series of laboratory tests have been conducted on the granite from Erzgebirge–Vogtland region within the Saxothuringian segment of Central Europe, including: permeability measurements, ultrasonic wave speed measurements, Brazilian tests, uniaxial and triaxial compression tests. A hydro-mechanical coupled testing system has been designed and used to conduct drained, undrained triaxial compression tests and permeability evolution measurements during complete loading process. A set of physical and mechanical parameters were obtained. 3. Based on analyzing the complete stress-strain curves obtained from triaxial compression tests and Hoek-Brown failure criterion, a modified elemental elasto-plastic constitutive law was developed which can represent strength degradation and volume dilation considering the influence of confining pressure. 4. The mechanism of HM-coupled behavior according to the Biot theory of elastic porous medium is summarized. A trilinear evolution rule for Biot’s coefficient based on the laboratory observations was deduced to eliminate the error in predicting rock strength caused by constant Biot’s coefficient. 5. The permeability evolution of low porous rock during the failure process was described based on literature data and own measurements, a general rule for the permeability evolution was developed for the laboratory scale, a strong linear relation between permeability and volumetrical strain was observed and a linear function was extracted to predict permeability evolution during loading process based on own measurements. 6. By combining modified constitutive law, the trilinear Biot’s coefficient evolution model and the linear relationship between permeability and volumetrical strain, a fully hydro-mechanical coupled numerical simulation scheme was developed and implemented in FLAC3D. A series of numerical simulations of triaxial compression test considering the hydro-mechanical coupling were performed with FLAC3D. And a good agreement was found between the numerical simulation results and the laboratory measurements under 20 MPa confining pressure and 10 MPa fluid pressure, the feasibility of this fully hydro-mechanical coupled model was proven.

The realistic approximation of structural behavior in a post fracture state by the phase-field method requires information about the spatial orientation of the crack surface at the material point level. For the directional phase-field split, this orientation is specified by the crack orientation vector, that is defined perpendicular to the crack surface. An alternative approach to the determination of the orientation based on standard fracture mechanical arguments, i.e. in alignment with the direction of the largest principle tensile strain or stress, is investigated by considering the amount of dissipated strain energy density during crack evolution. In contrast to the application of gradient methods, the analytical approach enables the determination of all local maxima of strain energy density dissipation and, in consequence, the identification of the global maximum, that is assumed to govern the orientation of an evolving crack. Furthermore, the evaluation of the local maxima provides a novel aspect in the discussion of the phenomenon of crack branching. As the directional split differentiates into crack driving contributions of tension and shear stresses on the crack surface, a consistent relation to Mode I and Mode II fracture is available and a mode dependent fracture toughness can be considered. Consequently, the realistic simulation of rock-like fracture is demonstrated. In addition, a numerical investigation of Ƭ-convergence for an AT-2 type crack surface density is presented in a two-dimensional setup. For the directional split, also the issues internal locking as well as lateral phase-field evolution are addressed.