• Computers and Concrete
• /
• 제8권1호
• /
• pp.111-123
• /
• 2011

Severe element distortion problem is observed in finite element mesh while performing numerical simulations of high velocity steel projectiles penetration/perforation of concrete targets using finite element method (FEM). This problem of element distortion in Lagrangian formulation of FEM can be resolved by using element erosion methodology. Element erosion approach is applied in the finite element program by defining failure parameters as a condition for element elimination. In this study strain parameters for both compression and tension at failure are used as failure criteria. Since no direct method exists to determine these values, a calibration approach is used to establish suitable failure strain values while performing numerical simulations of ogive-nose steel projectile penetration/perforation into concrete target. A range of erosion parameters is suggested and adopted in concrete penetration/perforation tests to validate the suggested values. Good agreement between the numerical and field data is observed.

• 한국전산유체공학회지
• /
• 제14권2호
• /
• pp.46-51
• /
• 2009

The slip effect from the molecular interaction between fluid particles and solid surface atoms plays a key role in microscale fluid transport and heat transfer since the relative importance of surface forces increases as the size of the system decreases to the microscale. There exist two models to describe the slip effect: the Maxwell slip model in which the slip correction is made on the basis of the degree of shear stress near the wall surface and the Langmuir slip model based on a theory of adsorption of gases on solids. In this study, as the first step towards developing a general purpose numerical code of the compressible Navier-Stokes equations for computational simulations of microscale fluid flow and heat transfer, two slip models are implemented into a finite element numerical code of a simplified equation. In addition, a pressure-driven gas flow in a microchannel is investigated by the numerical code in order to validate numerical results.

• Structural Engineering and Mechanics
• /
• 제10권2호
• /
• pp.93-110
• /
• 2000

An alternative interpretation of the completeness requirements for the higher order elements is presented. Apart from the familiar condition, $\sum_iN_i=1$, some additional conditions to be satisfied by the shape functions of higher order elements are identified. Elements with their geometry in the natural form, i.e., without any geometrical distortion, satisfy most of these additional conditions inherently. However, the geometrically distorted elements satisfy only fewer conditions. The practical implications of the satisfaction or non-satisfaction of these additional conditions are investigated with respect to a 3-node bar element, and 8- and 9-node quadrilateral elements. The results suggest that non-satisfaction of these additional conditions results in poorer performance of the element when the element is geometrically distorted. Based on the new interpretation of completeness requirements, a 3-node element and an 8-node rectangular element that are insensitive to mid-node distortion under a quadratic displacement field have been developed.

• Computers and Concrete
• /
• 제1권3호
• /
• pp.261-284
• /
• 2004

Two enhancements to a recently developed plastic-damage-contact model for concrete are presented. The model itself, which uses planes of degradation that can undergo damage and separation but that can regain contact according to a contact law, is described. The first enhancement is a new damage evolution function which provides a completely smooth transition from the undamaged to the damaged state and from pre-peak to post-peak regions. The second is an improved contact function that governs the potential degree of contact with increasing opening on a crack plane. The use of a damage evolution function with a pre-peak has implications for the consistent tangent matrix/stress recovery algorithm developed for the model implementation, and amendments to this algorithm to accommodate the new function are described. A series of unpublished experimental tests on notched specimens undertaken in Cardiff in the mid 1990s are then described. These include notched beam tests as well as prismatic and cylindrical torsion tests. The tests are then considered in three dimensional finite element analyses using the modified Craft model implemented in the finite element program LUSAS. Comparisons between experimental and numerical data show reasonable agreement except that the numerical simulations do not fully describe the latter stages of the softening responses for the torsion examples. Finally, it is concluded that the torsion tests described provide useful benchmark examples for the validation of three-dimensional numerical models for concrete.

• Structural Engineering and Mechanics
• /
• 제61권6호
• /
• pp.693-700
• /
• 2017

Concrete cover cracking due to the corrosion of steel reinforcing bars is one of the main causes of deterioration in Reinforced Concrete (RC) structures. The oxidation level of the bars causes varying levels of expansion. The rebar expansions could lead to through-thickness cracking of the concrete cover, where depending on the cracking characteristics, the service life of the structures would be affected. In this paper, the effect of geometrical and material parameters, i.e., concrete cover thickness, reinforcing bar diameter, and concrete tensile strength, on the required pressure for concrete cover cracking due to corrosion has been investigated through detailed numerical simulations. ABAQUS finite element software is employed as a modeling platform where the concrete cracking is simulated by means of eXtended Finite Element Method (XFEM). The accuracy of the numerical simulations is verified by comparing the numerical results with experimental data obtained from the literature. Using a previously proposed empirical equation and the numerical model, the time from corrosion initiation to the cover cracking is predicted and then compared to the respective experimental data. Finally, a parametric study is undertaken to determine the optimum ratio of the rebar diameter to the reinforcing bars spacing in order to avoid concrete cover delamination.

• Journal of the Korean Society for Industrial and Applied Mathematics
• /
• 제7권2호
• /
• pp.13-22
• /
• 2003

Simulations on the nonlinear partial differential equation derived from Black-Scholes equation with transaction costs are performed. These numerical experiments using finite element methods are applied to KOSPI200 in 2002 and the option prices obtained with transaction costs are closer to the real prices in market than the prices used in Korea Stock Exchange.

• 한국자동차공학회논문집
• /
• 제8권3호
• /
• pp.119-129
• /
• 2000

In order to show the efficiency and accuracy of an expert drawbead model, finite element simulations of auto-body panel stampings are carried out. For numerical modeling of the drawbead of a panel die, the drawbead restraining force and bead-exit thinning calculated by the expert model are imposed to a node nearest to the drawbead position as a boundary condition, Finite element simulations show that the expert model provides the accurate solution, guarantees the stable convergence, and has the merit in the computation time.

• Steel and Composite Structures
• /
• 제10권1호
• /
• pp.1-21
• /
• 2010

As an alternative for conventional structures for tall buildings, a hybrid lateral load resisting structure has been designed, enabling the assembly of tall buildings directly from a truck. It consists of steel frames with discretely connected precast concrete infill panels provided with window openings. Besides the stiffening and strengthening effect of the infill panels on the frame structure, economical benefits may be derived from saving costs on materials and labour, and from reducing construction time. In order to develop design rules for this type of structure, the hybrid infilled frame has recently been subjected to experimental and numerical analyses. Ten full-scale tests were performed on one-storey, one-bay, 3 by 3 m infilled frame structures, having different window opening geometries. Subsequently, the response of the full-scale experiments was simulated with the finite element program DIANA. The finite element simulations were performed taking into account non-linear material characteristics and geometrical non-linearity. The experiments show that discretely connected precast concrete panels provided with a window opening, can significantly improve the performance of steel frames. A comparison between the full-scale experiments and simulations shows that the finite element models enable simulating the elastic and plastic behaviour of the hybrid infilled frame.

• Structural Engineering and Mechanics
• /
• 제32권2호
• /
• pp.337-350
• /
• 2009

Finite element simulations are increasingly used in structural analysis and design, especially in cases where complex structural and loading conditions are involved. Due to considerable progresses in computer technology as well as nonlinear finite-element analysis techniques in past years, it has become possible to pursue an accurate analysis of the complex blast-induced structural effects by means of numerical simulations. This paper aims to develop a better understanding of the behavior of steel-concrete composite beams (SCCB) under localized blast loading through a numerical parametric study. A finite element model is set up to simulate the blast-resistant features of SCCB using the transient dynamic analysis software LS-DYNA. It is demonstrated that there are three dominant failure modes for SCCB subjected to localized blast loading. The effect of loading position on the behavior of SCCB is also investigated. Finally, a simplified model is proposed for assessing the overall response of SCCB subjected to localized blast loading.

• Wind and Structures
• /
• 제8권6호
• /
• pp.407-426
• /
• 2005

A hybrid RANS/LES approach, based on the Limited Numerical Scales concept, is applied to the numerical simulation of the flow around a square cylinder. The key feature of this approach is a blending between two eddy-viscosities, one given by the $k-{\varepsilon}$ RANS model and the other by the Smagorinsky LES closure. A mixed finite-element/finite-volume formulation is used for the numerical discretization on unstructured grids. The results obtained with the hybrid approach are compared with those given by RANS and LES simulations for three different grid resolutions; comparisons with experimental data and numerical results in the literature are also provided. It is shown that, if the grid resolution is adequate for LES, the hybrid model recovers the LES accuracy. For coarser grid resolutions, the blending criterion appears to be effective to improve the accuracy of the results with respect to both LES and RANS simulations.