Search results

The purpose of this paper is to present an investigation of the effect of different gravity conditions on the penetration mechanism using the two-dimensional Distinct Element Method (DEM), which ranges from high gravity used in centrifuge model tests to low gravity incurred by serial parabolic flight, with the aim of efficiently analyzing cone penetration tests on the lunar surface.

Seven penetration tests were numerically simulated on loose granular ground under different gravity conditions, i.e. one-sixth, one-half, one, five, ten, 15 and 20 terrestrial gravities. The effect of gravity on the mechanisms is examined with aspect to the tip resistance, deformation pattern, displacement paths, stress fields, stress paths, strain and rotation paths, and velocity fields during the penetration process.

First, under both low and high gravities, the penetration leads to high gradients of the value and direction of stresses in addition to high gradients in the velocity field near the penetrometer. In addition, the soil near the penetrometer undergoes large rotations of the principal stresses. Second, high gravity leads to a larger rotation of principal stresses and more downward particle motions than low gravity. Third, the tip resistance increases with penetration depth and gravity. Both the maximum (steady) normalized cone tip resistance and the maximum normalized mean (deviatoric) stress can be uniquely expressed by a linear equation in terms of the reciprocal of gravity.

This study investigates the effect of different gravity conditions on penetration mechanisms by using DEM.

The purpose of this paper is to study the distribution of active earth pressure in retaining walls with narrow cohesion less backfill considering arching effects.

To this end, the approach of principal stresses rotation was used to consider the arching effects.

According to the presented formulation, the active soil pressure distribution is nonlinear with zero value at the wall base. The proposed formulation implies that by increasing the frictional forces at both sides of the backfill, the arching effect is increased and so, the lateral earth pressure on the retaining wall is decreased. Also, by narrowing the backfill space, the lateral earth pressure is extremely decreased.

A comprehensive analytical solution for the active earth pressure of narrow backfills is presented, such that the effects of the surcharge and the characteristics of the stable back surface are considered. The magnitude and height of the application of lateral active force are also derived.

A new model for handling non‐orthogonal cracks within the smeared crack concept is described. It is based on a decomposition of the total strain increment into a concrete and into a crack strain increment. This decomposition also permits a proper combination of crack formation with other non‐linear phenomena such as plasticity and creep and with thermal effects and shrinkage. Relations are elaborated with some other crack models that are currently used for the analysis of concrete structures. The model is applied to some problems involving shear failures of reinforced concrete structures such as a moderately deep beam and an axisymmetric slab. The latter example is also of interest in that it confirms statements that ‘reduced integration’ is not reliable for problems involving crack formation and in that it supports the assertion that identifying numerical divergence with structural failure may be highly misleading.

In the finite element method two options are currently available for dealing with problems involving an axisymmetric geometry but in which loads, displacements or other boundary conditions do not have rotational symmetry. The first involves a full three‐dimensional solution whereas the second is to use a two‐dimensional (2D) axisymmetric formulation and to express the non‐symmetric loads/displacements as Fourier series in the circumferential direction. There are some cases, however, that, in spite of not being truly axisymmetric, it can be shown that the non‐zero components of the full strain tensor number four or less. In this paper it is shown that such problems may be solved using simple 2D finite element formulations and two alternative solution methods are presented. One of these involves a modification to the matrix relating strains to displacements and the second employs conventional 2D formulations with tied degrees of freedom. The solution procedures are applied to three examples which have some geotechnical interest, namely the behaviour of a long rigid pile under either torsional or vertical loading and the behaviour of a hollow cylinder sample subjected to torsion. In all three cases the soil is modelled by means of an elastoplastic constitutive law of the Cam‐clay type.

THE importance of the aircraft engine supercharger is being emphasized by the increasing demands for high altitude performance in the present war. Centrifugal stresses of considerable magnitude are induced in the supercharger impeller by reason of the high rotative speeds necessary to obtain the desired pumping effect. A speed of 20,000 r.p.m. is not uncommon for an impeller of 12 in. outside diameter and over. Consequently, a knowledge of the centrifugal stresses constitutes a basic design consideration. Unfortunately, a direct determination of these stresses is not an easy matter.

A nonlinear finite element (FE) beam-column model for the analysis of reinforced concrete (RC) frames with due account of shear is presented in this paper. The model is an expansion of the traditional flexural fibre beam formulations to cases where multiaxial behaviour exists, being an alternative to plane and solid FE models for the nonlinear analysis of entire frame structures. The paper aims to discuss these issues.

Shear is taken into account at different levels of the numerical model: at the material level RC is simulated through a smeared cracked approach with rotating cracks; at the fibre level, an iterative procedure guarantees equilibrium between concrete and transversal reinforcement, allowing to compute the biaxial stress-strain state of each fibre; at the section level, a uniform shear stress pattern is assumed in order to estimate the internal shear stress-strain distribution; and at the element level, the Timoshenko beam theory takes into account an average rotation due to shear.

The proposed model is validated through experimental tests available in the literature, as well as through an experimental campaign carried out by the authors. The results on the response of RC elements critical to shear include displacements, strains and crack patterns and show the capabilities of the model to efficiently deal with shear effects in beam elements.

A formulation for the nonlinear shear-bending interaction based on the fixed stress approach is implemented in a fibre beam model. Shear effects are accurately accounted during all the nonlinear path of the structure in a computationally efficient manner.

The purpose of this paper is to present several methods on how to deal with yield surface discontinuities. The explicit formulations, first presented by Koiter (1953), result in multisingular constitutive matrices which can cause numerical problems in elasto-plastic finite element calculations. These problems, however, are not documented in previous literature. In this paper an amendment to the Koiter formulation of the constitutive matrices for stress points located on discontinuities is proposed.

First, a review of existing methods of handling yield surface discontinuities is given. Examples of the numerical problems of the methods are presented. Next, an augmentation of the existing methods is proposed and its robustness is demonstrated through footing bearing capacity calculations that are usually considered “hard”.

Previous studies documented in the literature all present “easy” calculation examples, e.g. low friction angles and few elements. The amendments presented in this paper result in robust elasto-plastic computations, making the solution of “hard” problems possible without introducing approximations in the yield surfaces. Examples of “hard” problems are highly frictional soils and/or three-dimensional geometries.

The proposed method makes finite element calculations using yield criteria with corners and apices, e.g. Mohr-Coulomb and Hoek-Brown, much more robust and stable.

The objective of this paper is to quantitatively assess shear band evolution by using two-dimensional discrete element method (DEM).

The DEM model was first calibrated by retrospectively modelling existing triaxial tests. A series of DEM analyses was then conducted with the focus on the particle rotation during loading. An approach based on particle rotation was developed to precisely identify the shear band region from the surrounding. In this approach, a threshold rotation angle ω₀ was defined to distinguish the potential particles inside and outside the shear band and an index g(ω₀) was introduced to assess the discrepancy between the rotation response inside and outside shear band. The most distinct shear band region can be determined by the ω₀ corresponding to the peak g(ω₀). By using the proposed approach, the shear band development of two computational cases with different typical localised failure patterns were successfully examined by quantitatively measuring the inclination angle and thickness of shear band, as well as the microscopic quantities.

The results show that the shear band formation is stress-dependent, transiting from conjugated double shear bands to single shear band with confining stress increasing. The shear band evolution of two typical localised failure modes exhibits opposite trends with increasing strain level, both in inclination angle and thickness. Shear band featured a larger volumetric dilatancy and a lower coordination number than the surrounding. The shear band also significantly disturbs the induced anisotropy of soil.

This paper proposed an approach to quantitatively assess shear band evolution based on the result of two-dimensional DEM modelling.

Large eddy simulation (LES) is widely used in prediction of turbulent flow. The purpose of this paper is to propose a new dynamic mixed nonlinear subgrid‐scale (SGS) model (DMNM), in order to improve LES precision of complex turbulent flow, such as flow including separation or rotation.

The SGS stress in DMNM consists of scale‐similarity part and eddy‐viscosity part. The scale‐similarity part is used to describe the energy transfer of scales that are close to the cut‐off explicitly. The eddy‐viscosity part represents energy transfer of the other scales between smaller than grid‐filter size and larger than grid‐filter size. The model is demonstrated through two examples; one is channel flow and another is surface‐mounted cube flow. The computed results are compared with prior experimental data, and the behavior of DMNM is analyzed.

The proposed model has the following characteristics. First, DMNM exhibits significant flexibility in self‐calibration of the model coefficients. Second, it does not require alignment of the principal axes of the SGS stress tensor and the resolved strain rate tensor. Third, since both the rotating part and scale‐similarity part are considered in the new model, flow with rotation and separation is easily simulated. Compared with the prior experimental data, DMNM gives more accurate results in both examples.

The SGS model DMNM proposed in the paper could capture the detail vortex characteristics more accurately. It has the advantage in simulation of complex flow, including more separations.

The purpose of this paper is to present a new eddy‐viscosity formulation designed to exhibit a correct response to streamline curvature and flow rotation. The formulation is implemented into a linear k‐ ε turbulence model with a two‐layer near‐wall treatment in a commercial computational fluid dynamics (CFD) solver.

A simple, robust formula is developed for the eddy‐viscosity that is curvature/rotation sensitive and also satisfies realizability and invariance principles. The new model is tested on several two‐ and three‐dimensional problems, including rotating channel flow, U‐bend flow and internally cooled turbine airfoil conjugate heat transfer. Predictions are compared to those with popular eddy‐viscosity models.

Converged solutions to a variety of turbulent flow problems are obtained with no additional computational expense over existing two‐equation models. In all cases, results with the new model are superior to two other popular k‐ ε model variants, especially for regions in which rapid rotation or strong streamline curvature exists.

The approach adopted here for linear eddy‐viscosity models may be extended in a straightforward manner to non‐linear eddy‐viscosity or explicit algebraic stress models.

The new model is a simple “plug‐in” formula that contains important physics not included in most linear eddy‐viscosity models and is easy to implement in most flow solvers.

The present model for curved and rotating flows is developed without the need for second derivatives of velocity in the formulation, which are known to present difficulties with unstructured meshes.