In block caving projects, the rock mass fragmentation process plays a controlling role in the design and success of the operation. Poorly understood, however, is the secondary fragmentation that occurs as blocks of caved ore move through the draw column and interact with one another. The initial size of these blocks can be considerable, as can their strength and stiffness as the presence of non-persistent joints and veining will be variable. Intuitively, more planes of weakness within the blocks should facilitate early fragmentation, but this assumption could depend on additional factors such as contact type between adjacent blocks and loads imposed through varying heights of the draw column. This paper reports the findings from a laboratory investigation examining the role of block strength and planes of weakness on secondary fragmentation. Testing was carried out on small-scale concrete cubes of uniform size, with and without embedded planes of weakness. The results obtained demonstrate the significance of veining and other planes of weakness on secondary fragmentation under one-dimensional compression, and allow for the development and constraint of empirical guidelines.
The fragmentation process during block caving involves the initial (primary) fragmentation of rock in the cave back through stress-induced fracturing, release and fall of individual blocks onto the broken ore muck pile. This is followed by secondary fragmentation involving splitting and rounding of blocks as they move downwards through the draw column towards the draw point. The degree of secondary fragmentation achieved involves many variables related to the mechanical properties of the rock blocks as well as different operational factors. With respect to the block properties, the presence of planes of weakness in the form of veining and non-persistent joints is a topic of special interest given the significant influence it potentially has on the overall fragmentation process. However, the inherent heterogeneity and variable size of individual blocks generated from the cave back makes it difficult to develop any empirical relationships relating secondary fragmentation to compressive loading based on insitu observations and mine data. More preferable would be to keep constant block parameters, such as shape (aspect ratio and angularity) and size, through controlled laboratory testing. This work presents an efficient technique to fabricate concrete cubes in order to keep constant different block properties. These are then tested under one-dimensional compressive loading. The objective of these tests is to emulate a tight packing condition of single blocks such as that which would occur when the air gap between the cave back and broken ore in the draw column is small. When the air gap is small, detached blocks are unable to fall and rotate but instead maintain a very tidy arrangement with minimal voids left between adjacent blocks.
Rock fracture simulation most often presents non-linear behaviour due to loss of material integrity and strain concentration in the fracture zone. These strains generate local instabilities that propagate to the global system of equilibrium equations, generating convergence problems of the numerical solution. Often, in such cases, analyses in the post-critical regime are unfeasible. To overcome the numerical difficulties associated with problems of stiffness loss in the post-critical regime control methods, also called continuation methods, are employed. Among them are the arc-length method, the energy control method and the indirect displacement control method. These methods are employed in conjunction with a Newton- Raphson scheme for the solution of non-linear systems of equations. However, depending on the system instability, convergence in not guaranteed even with very small increments. In the present work these control techniques are investigated in combination with cohesive elements to simulate rock fracture. The constitutive model employed with the cohesive elements represents stiffness degradation through a damage law, which leads to serious convergence difficulties as reported in the literature. Here the continuation methods mentioned are applied to a Mode I rock fracture problem and to hydraulic fracture simulations. The effectiveness of the different continuation methods is compared.
The process of rock fracture induced by the fluid pressure is called hydraulic fracturing. The propagation of a hydraulic fracture is a complex process, which is basically defined by the mechanical deformation of the rock, the fluid flow within the fracture and the fracture propagation itself (Mokryakov, 2011). In the works of Bendezu et al. (2013), Chen (2012) and Carrier et al. (2012), vertical hydraulic fracture propagation is modeled by cohesive elements with traction-separation law. The softening behavior of rock in post-critical regime is demonstrated by experimental testing and is characterized mainly by high strains and low stresses (Crowder and Bawden, 2004). This nonlinear behaviour can arise due to material nonlinearity, such as micro-cracking and damage, leading to softening behaviour (Chandrakant, S.D., 2012). The cohesive fracture model considers in its formulation the softening behaviour that most rocks present in post-critical regime with propagation criteria.
A numerical method is proposed to propagate multiple discrete fractures leading to primary fragmentation in a mine that is being worked using block caving. Deformation is computed using the finite element method, fractures are represented explicitly, and the mine domain is discretized by an unstructured mesh. Bedding planes are represented by systematically varying the elastic modulus of the rock and by defining horizontal weakness planes. Fractures and matrix are represented using parametric surfaces, and tips are defined by their boundary curves. Tip advance is controlled by a failure criterion, and a criterion for propagation direction and magnitude, based on the evaluation of the modal stress intensity factors. A novel domain integral approach is applied to accurately compute stress intensity factors (K) ahead of fracture tips in three dimensions. The method does not require a structured volumetric mesh structure around the crack tip, as integration is performed over a series of virtual surface domains along the crack front. The method is efficient, as it makes direct use of automatically generated, arbitrary tetrahedral meshes, and approximates stress intensity factors (KI, KII, KIII) along each crack front using Interaction-integrals. As opposed to the J-Integral, the method does not decompose K a-posteriori, but instead uses an auxiliary field to directly compute modal K. When using this method, numerical approximations of K do not exhibit dependence on the mesh layout, and require meshes that can generally be ten times coarser than are required by displacement- and stress-based methods. Volumetric meshing requires only, on average, 17% of each computation step. Thus, cracks do not follow any pre-existing mesh structure, and the method is well suited for high-density fracture datasets. The method is demonstrated for primary fragmentation of a mine area covering 110 initial draw points, immediately beneath a 2m high undercut. Displacement is constrained at all boundaries except the surface of the undercut. The growth of the fractures is investigated at the onset of the undercut. Ninety initial disc-shaped fractures are taken into account, each having an initial radius of 10 m. Fractures grow around the undercut, intersect the bedding plane boundaries, and the domain is fragmented as a result.
Verma, A. K. (Indian Institute of Technology Kharagpur) | Rajabrahma, K. (Indian Institute of Technology Kharagpur) | Kumar, Rakesh (Indian Institute of Technology Kharagpur) | Mukesh, V. (Indian Institute of Technology Kharagpur)
In this study, laboratory investigation as well as numerical investigation is carried out to ascertain the transition from one mode to another for Semi-circular bend setup. Laboratory investigation is carried by varying the span length (s/2R); crack length (a/R); notch angle (a); and notch thickness (t). The specimen consists of cement, sand and water mixture in a ratio of 1:1:0.5 by weight and are prepared as per the ISRM standards (2014). Tests are conducted by a servo controlled universal testing machine at a strain rate of 0.2 mm/ min. The results are collected in terms of failure load, crack mouth opening displacement (CMOD) and images of crack initiation and propagation.
Image analysis of laboratory tested samples shows that crack initiates at the notch tip when the notch inclination angle is 0°. As the notch angle increases (i.e. 30° and 60°) the fracture begins to initiate behind the notch tip. The direction of crack propagation in front and back side of the sample is same for notch inclination angle is 0° whereas it differs more and more as crack inclination increases. The transition represents the change in mode of failure from mode I (0°), mode II (30°) and mixed mode I/II fracture (60°).
Finite element method used to find the transition based on MTS failure criterion. The inclination of notch is varied like 40, 42, 44, 45, and 50 deg and FEM analysis shows that the transition from mode-I to Mode-II take place at 49 deg. The same procedure can be extended to find the transition from mode-II to mode-III.
Over the years researchers have investigated the behavior of rock fracture under mode I, II or I/ II loading by different methods. ISRM have recommended four different methods for determination of fracture toughness. Out of all these method, semi-circular bend (SCB) test is the newer addition in the year 2014 because of its simplicity and ease to use. In this work, test have been performed on artificial samples by varying crack thickness (t), crack length (a/R), orientation (a), load span length (s/2R) and grain size. The Same experiment was repeated four times for more confidence in test results. The monitored data includes the load at failure, measurement of crack mouth opening displacement (CMOD) and experiments are recorded in still as well as videos pictures. The finite element models (FEM) is developed to determine the conditions SIF for Mode I, II and mixed mode sample using MTS criterion.
Mortazavi, A. (Amirkabir University of Technology) | Abbasloo, Z. (Share Babak Copper Complex, Meydook Copper Mine) | Ebrahimi, L. (Share Babak Copper Complex, Meydook Copper Mine) | Keshavarz, A. (Share Babak Copper Complex, Meydook Copper Mine) | Masoomi, A. (Share Babak Copper Complex, Meydook Copper Mine)
Investigation and 3D analysis of Heap no. 2 of Meydook copper complex was the subject of this study. In this project all geological and geotechnical investigations conducted in the mine were evaluated aiming at determination of Heap no. 2 design parameters. The main objective of the project was to evaluate the potential for construction of Heap. No. 2 on a waste dump built in a valley in the north side of the pit. Considering the space limitation for dump location this was a viable alternative to reduce the costs associated with heap no. 2 construction. Moreover, it allowed for a significant reduction in overall mining costs. Considering the project condition, a 3D analysis of heap major structures including; valley rock foundation, waste dump, and low grade ore was carried out. Five sets of analyses were conducted and the displacement and stress fields were calculated within the heap no. 2 components. In all analyses conducted the subsident profile was determined for the valley base rock and heap foundation (waste dump). Moreover, the shear interaction between heap and waste dump was evaluated at the heap interface. The analyses results showed that the maximum settlement at the heap base is about 4-5 cm and the waste dump provides a safe and reliable base for heap construction. Moreover, the analysis results showed that the selected %5 grade for the heap interface is good and does not pose any risk in terms of heap Interface shear displacements. A sensitivity analysis of heap mechanical parameters was also conducted. The obtained results show that if the friction angle of heap material is reduced below 25 degree due to acid spraying process, the heap walls become unstable and the heap geometry must be redesigned.
Meydook copper mine is one of the largest open cast copper operations in Iran. Meydook’s future expansion plan makes it one of the deepest copper mines. Meydook mine is situated 42 Km NE of Shahrebabak city in Kerman province, Iran. The ore body is located at 2842 m elevation above the see level. From a geological point of view Meydook mine is located in central Iran zone which is associated with intensive tectonic and magmatic activities. Based on exploration studies mine has 171 Mt copper ore reserve at 0.83% grade. It is planned to mine 144 Mt of ore during a 30 year life span in the form of an open pit mine.
Surfaces of rock slopes in cold regions often become frozen in winter. Therefore, understanding time-dependent behavior of frozen rock, as well as that of unfrozen rock, is important for long-term stability assessment of rock slopes. In this study, a series of uniaxial compression tests and creep tests were carried out on Shikotsu welded tuff under dry and water-saturated conditions at -20 °C to clarify time-dependent strength and deformation of the frozen rock. The influence of water content on them also was investigated. It was found that the amount of deformation of the water-saturated specimen was much greater than that of the dry specimen. It also was found that the creep behavior of the water-saturated specimen was similar to that of polycrystalline ice. These show that deformation of the water-saturated specimen is strongly affected by pore ice. The UCS of the water-saturated specimen was approximately equal to that of the dry specimen at a strain rate below 4.2×10-6/s, and its value was about 17 MPa. However, the former increased up to 26 MPa at 4.2×10-4/s with loading rate, whereas there was little increase in the latter. This means that the UCS of the water-saturated specimen was always greater than that of the dry specimen at higher strain rates. The creep life of water-saturated specimens was longer than that of the dry specimens at stress level greater than 15 MPa. These results of UCS and creep life show that inclusion effect due to presence of pore ice plays an important role on strength of the water-saturated specimen. Creep life of the non-failure specimen that was loaded for more than three days was estimated by the relationship between creep life and axial strain rate at 10 s on failure specimens. It was found that creep life of the water-saturated specimens was shorter than that of the dry specimen at stress levels less than 14 MPa. This suggests that estimation of creep life of frozen rock under water-saturated condition is important for the stability assessment of frozen rock slopes. It was clear that the loading-rate dependency of UCS was related to the stress-level dependency of creep life in both dry and wet conditions. This indicates that there are common fracture mechanism in creep and uniaxial compression tests. It also indicates that the creep life can be estimated by the relationship between the UCS and the loading rate. The time dependencies of the water-saturated specimen were in good agreement with those of polycrystalline ice. This confirms that pore ice strongly affects the strength of the water-saturated specimens as, well as the deformation behavior.
In this work, we found that contact between shale and water results in development of micro fractures. Based on results of experiments on Pierre shale, we conclude that appearance of micro fractures begin with saturation of capillaries, ionic and diffusive transport of water into the shale clays and once capillaries are saturated, the cause of micro fracture propagation is the conversion of ionic activity/exchange to excess pressure that did not exist before fracking. Based on these findings, the spread of micro fractures appear to be a time-dependent phenomenon which has not been addressed in the existing macro/micro fracture models.
Shale has often been involved as a hazard in drilling operations. This hazard can be defined as “destabilization” of shale. When contacted with water-based drilling fluids, some shales readily swell and sometimes, cause the wellbore to cave-in, slough, wash-out, close, and pack-off, impeding the drilling by sticking the drill-pipe. However, once drilling reaches the desired depth or length, the casing is set, cemented, and perforated, and then actually we wish to initiate fractures and destabilize the shale formation, using hydraulic fracturing.
Clays constitute a major portion of minerals in shale. These clays contain a large amount of free energy which is the main factor for “slick” water adsorption/absorption. In fact, the reason for using “surfactants” in hydraulic fracturing fluids is to make the penetration of the fluid into the capillaries much easier, thus water meets with less resistance to enter the small capillaries. Also, the result of high capillary suction pressure is due to small Angstrom size capillaries, smaller pores and presence of ions and hydrateable metal atoms. The free energy is thus, related to all the above mentioned and other affects. Capillary pressure, osmotic pressure and other pressures are responsible for creating the micro-fractures in shale, thus, increasing the network of micro-fractures which leads to more gas production.
The objective of this study is to evaluate the pressures of the individual ions which are released by the diffusing “slick” water into shale. These pressures would be added to the above mentioned capillary pressure, osmotic pressure, bacterially-induced pressures, chemically-induced reaction pressure, pressure due to exchangeable ion-transport, pressure due to release of free energy of solvation and eventually to the pore pressure as suggested by Terzaghi’s equation.
Developments in understanding the brittle fracture of rock over the past two decades ultimately led to the acceptance of a cohesion-weakening-friction-strengthening model as an appropriate strategy for the modelling of spalling using a continuum approach. This strength model has been used to replicate observed failure depths and geometries for a number of case studies in the literature, most notably the AECL Underground Research Laboratory in Manitoba. One thing which these case studies typically ignore, however, is the influence of rock dilatancy on the evolution of brittle failure. Physically, this phenomenon is the volumetric expansion of the rock due to fracture opening; numerically, this phenomenon is captured by the flow rule, which is often characterized by the dilation angle, Ψ. Accurately modelling rock dilatancy in conjunction with strength evolution is necessary both to predict the magnitude and distribution of ground displacements, as well as the exact depth and shape of the fracturing zone. Both of these factors can have significant implications for support design.
This paper is focussed on presenting some of the recent developments in this field in the context of a number of numerical modelling case studies. Initially, a new model for the dilatancy of brittle rock is presented, as well as some representative data and model parameters for different rock types. Next, several case studies of highly stressed tunnels and shafts are presented. The results of continuum numerical models are presented to demonstrate the ability of the back analyses to match observed in-situ measurements and obtained. The predictive capabilities of the material model are also highlighted for cases where sufficient “a priori” information (such as laboratory compression testing data) are available. The issues of model nonuniqueness and strain-localization/bifurcation are also addressed.
With respect to the numerical modelling of rockmass behaviour, there appears to be a growing acceptance of discontinuum and/or hybrid continuum/discontinuum approaches as more physically correct than conventional continuum approaches. In particular, these approaches are capable of replicating macroscopic behaviours such as complete fracture separation under gravitational loading which are inherently not possible based on the assumptions inherent in current continuum approaches; the ability to model this type of behaviour is critical in applications such as block caving where pre-existing discontinuities dominate the overall rockmass response to loading (Bobet, 2010; Elmo & Stead, 2009; Mas Ivars, Pierce, Darcel, Reyes-Montes, Potyondy, Young, & Cundall, 2011).
Successive seismic waves may cause progressive weakening of a fault zone, which has been recognized as the mechanism of earthquake aftershocks. However, it is unclear that how a fault zone slips under each wave cycle and how slip displacement accumulates under continuous wave cycles. This laboratory study provides direct evidences on the slip process of a simulated granular fault zone dynamically induced by an incident wave without the effects of late-arriving and reflected waves. The experimental observations show forward and backward slip paths of the fault zone and partially recovered slip displacement after a wave incidence. The unrecovered slip displacement after each wave cycle can be accumulated until it reaches a critical slip distance for seismic faulting, which may reinterpret delayed triggering in earthquake dynamics. The limited interseismic period restricts fault self-healing in strength. The experimental results also indicate that the dynamically induced frictional slip on the fault zone is associated with a complex frictional system, including seismic wave radiation, frictional slip initiation and normal stress vibration. Different from the dynamically induced frictional slip on the fault zone, the statically induced frictional slip causes a permanent increase of slip displacement along the shear loading direction.
An earthquake mainshock may induce aftershocks by suddenly changing external stresses applied to another fault zone in the near field (Kilb et al., 2000) and by slightly perturbing a fault zone from a critical steady state to stick-slip faulting in the far field (Gomberg and Davis, 1996). The dynamically induced frictional slip on a fault zone has been recognized as the mechanism of earthquake aftershocks (Marsan and Lengliné, 2008; van der Elst and Brodsky, 2010). Although earthquake aftershocks have been extensively studied by many seismological methods, a simplified laboratory fault zone can provide possible interpretation on real faulting and detailed analysis on a friction process (Dieterich, 1979; Nielsen at al., 2010). According to a few laboratory experiments related to the dynamically induced frictional slip on a simulated fault zone (e.g., Uenishi et al., 1999; Johnson et al., 2012), the slip process induced by successive seismic waves exhibits continuous motions. However, it is unclear that fault slip response under each wave cycle and slip displacement accumulation under continuous wave cycles.
The Thyssen Mining Group of Companies has successfully applied ground freezing technology to numerous projects throughout Europe and North America for over 100 years. Recently, the aspect of data acquisition and data verification has received greater focus.
Thyssen Mining has been engaged in several relatively deep shaft sinking projects over the past five (5) years, all of which have required ground freezing to successfully sink through water bearing rock. To safely and economically sink through a quantity of water similar to Saskatchewan’s Potash Mines, it is imperative to understand freeze wall thickness and growth rate.
Critical to understanding freeze wall thickness and growth rate, is the collection of accurate data as well as the interpretation of those data. The data can then be utilized to gain a better understanding of the freeze wall growth rate in active projects as well as to provide data to be utilized in modeling for future projects.
This paper elaborates on the installation and operation of the data acquisition system utilized at the Newmont Leeville project, Nevada’s first frozen shaft project.
In addition, this paper elaborates on the methodology developed by Thyssen Mining to manipulate the freeze wall data derived through the data acquisition system.
The successful incorporation of the new system of collection and analysis of freeze wall data for the Leeville Project has given Thyssen Mining a high degree of confidence in the understanding, as well as prediction, of the ground conditions during shaft sinking.
This paper will include:
• A brief description of ground freezing and its objective;
• An overview of previous data acquisition methods employed;
• Current data acquisition methods;
• Installation requirements; and,
• Data manipulation and presentation methods.
Thyssen Mining Construction of Canada (“TMCC”) was chosen by Newmont to sink a ventilation shaft (Leeville No.3 Ventilation Shaft) at its Leeville underground mining operation located on the Carlin Trend approximately 40 kilometers North of Carlin, Nevada. A major contributing factor in the selection of TMCC was the company’s extensive experience in sinking shafts through frozen ground.