Search

The mean stress, as a fundamental statistical property of a group of stress data, is essential for stress variability characterisation. However, currently in rock mechanics, the mean stress is customarily and erroneously calculated by separately averaging the principal stress magnitudes and orientations. In order to draw the attention of our community to the appropriate approach for stress variability characterisation, here we compare the customary scalar/vector mean with that obtained by the mathematically rigorous tensorial approach—the Euclidean mean. Calculation of mean stress using both a small group of actual in situ stress measurement results and a large group of simulated stress data (obtained using the combined finite-discrete element method, FEMDEM) demonstrates that the two approaches yield different results. Further investigation of these results shows that the scalar/vector approach may yield non-unique and non-orthogonal mean principal stresses, and these may deviate significantly from those of the Euclidean mean. Our calculations and comparisons reveal that the scalar/vector approach is deficient because it processes the principal stress magnitudes and orientations separately as independent quantities and ignores the connection between them. Conversely, the tensorial approach appropriately averages the tensors that simultaneously carry not only the information of stress magnitudes and orientations, but also the inherent relations between them. Therefore, arbitrarily using scalar/vector mean stress of in situ stress measurements as input in further rock engineering analyses may yield significantly erroneous results. We advise that stress data should be statistically processed in a tensorial manner using tensors referred to a common Cartesian coordinate system. [ABSTRACT FROM AUTHOR]

This paper presents a novel approach for hydromechanical modelling of fractured rocks by linking a finite-discrete element solid model with a control volume-finite element fluid model based on an immersed-body approach. The adaptive meshing capability permits flow within/near fractures to be accurately captured by locally-refined mesh. The model is validated against analytical solutions for single-phase flow through a smooth/rough fracture and reported numerical solutions for multi-phase flow through intersecting fractures. Examples of modelling single- and multi-phase flows through fracture networks under in situ stresses are further presented, illustrating the important geomechanical effects on the hydrological behaviour of fractured porous media. [ABSTRACT FROM AUTHOR]

Copyright of Hydrogeology Journal is the property of Springer Nature and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

In situ stress is an important parameter in rock mechanics, thus robust estimation of far–field stress to be used as boundary loadings for further rock engineering analysis based on the local in situ stress data seems indispensable. Here, as part of a preliminary investigation into this problem, we use the combined finite–discrete element method to examine how the mean of local stress tensors is related to the far–field stress. We have conducted a series of stress simulations on a model of a fractured rock mass subjected to various boundary loadings, and calculated the Euclidean mean of the stress data and compared them with the boundary loadings. The results shows that the Euclidean mean and boundary loadings are approximately equal, which gives us an indication that the Euclidean mean of the stress data can be a reasonable estimation of the far–field stress. [ABSTRACT FROM AUTHOR]

The construction of Core-Loc armour layers requires that unit centroids are systematically placed according to a grid pattern. To give the type of interlock required by this unit’s developers, neither a fully random nor a strict identically orientated neighbour-to-neighbour pattern of placement is successful. This means that a certain degree of irregularity in unit orientations is required. To guide the contractors, armour layers deemed to have been placed with acceptable standards, units must satisfy special placement criteria given by the Core-Loc developer. There is a lack of instruction on how to achieve such ‘acceptable irregular orientations and placements’. The construction success seems more a function of experience and skill of the crane operator than any repeatable or potentially machine-programmable methodology. For this reason, it is also challenging to design a repeatable numerical method that reliably positions units into armour layers that satisfy the developer’s special placement criteria. For the numerical creation of full-scale Core-Loc armour layers, a placement method was briefly introduced in previous work by the authors, based on the multi-body FEMDEM code, Y3D-R combined with a versatile deposition code, POSITIT. While satisfactory layers were previously demonstrated, here we elaborate on the sensitivity of the placed armour layer attributes to the numerical construction process parameters. In this work, new capabilities of POSITIT/Y3D-R are harnessed in a computationally efficient row-by-row numerical placement of units. These include vibration and artificially low friction in order to achieve acceptable irregular packs of the appropriate packing density. To examine the role that initial armour layer build quality has on unit immovability or the percentage of units that do not displace significantly under a standard oscillatory disturbing force, a sinusoidal velocity vibration is applied to the foundations in a series of tests on differently constructed full-scale armour layers. What emerges from this sensitivity study is an implicit verification of the importance of the technical requirements set by the developer of Core-Loc units. [ABSTRACT FROM AUTHOR]

Abstract: Rubble mound breakwaters armoured with concrete units rely on collective behaviour between adjacent concrete armour units but existing largely empirical approaches have been unable to provide a detailed understanding of how these gigantic granular systems work. The problem has been that current methods cannot investigate the interdependence of hydraulic and structural stability at the scale of individual units. Numerical methods have the potential to provide such answers but there are many challenges to overcome. We present a solution to the first major bottleneck concerning the solids modelling: the numerical creation of a breakwater trunk section of single layer concrete units with geometrical and mechanical properties that conform to realistic prototype structure placements. Positioning of units is achieved with a new versatile software tool, POSITIT, which incorporates user-defined deposition variables and the initial positioning grid necessary to achieve the required design packing densities. The code Y3D, based on the combined finite-discrete element method, FEMDEM, solves the multi-body mechanics of the problem. First, we show numerically constructed breakwater sections with armour layers of 8m3 CORE-LOC™ units placed on rock underlayers. The numerically-generated packs are deemed acceptable when examined according to a range of criteria indicative of acceptably placed armour layers, as set by concrete unit designers. Breakwater sections with packing densities ranging from 0.59 to 0.63 are then created. Using a set of analysis tools, local variation in packing density as an indicator of heterogeneity, centroid spacing, unit contacts and orientation of unit axes are presented, together with mechanical information showing the variation in contact forces. For these five packs examined, an increasingly tighter pack was associated with a steady increase in coordination number and a more steeply and accelerating increase in average maximum contact force per unit. The force distribution results and discussion presented illustrate the potential of discrete unit FEMDEM modelling methods to address initial placement quality, armour layer design and future innovation in rubble mound structures. [Copyright &y& Elsevier]