• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.26 no.7
• /
• pp.1408-1415
• /
• 2002

In this paper, we investigate the effect and the importance of the accuracy of finite element analysis in the shape optimization based on the finite element method and improve the existing finite element which has inaccuracy in some cases. And then, the shape optimization is performed by using the improved finite element. One of the main stream to improve finite element is the prevention of locking phenomenon. In case of bending dominant problems, finite element solutions cannot be reliable because of shear locking phenomenon. In the process of shape optimization, the mesh distortion is large due to the change of the structure outline. So, we have to raise the accuracy of finite element analysis for the large mesh distortion. We cannot guarantee the accurate result unless the finite element itself is accurate or the finite elements are remeshed. So, we approach to more accurate shape optimization to diminish these inaccuracies by improving the existing finite element. The shape optimization using the modified finite element is applied to a two and three dimensional simple beam. Results show that the modified finite element has improved the optimization results.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.20 no.1
• /
• pp.88-99
• /
• 1996

The explicit scheme for finite element analysis of sheet metal forming problems has been widely used for providing practical solutions since it improves the convergency problem, memory size and computational time especially for the case of complicated geometry and large element number. The explicit schemes in general use are based on the elastic-plastic modeling of material requiring large computataion time. In the present work, a basic formulation for rigid-plastic explicit finite element analysis of plain strain sheet metal forming problems has been proposed. The effect of some basic parameters involved in the dynamic analysis has been studied in detail. Thus, the effective ranges of parameters have been proposed for numerical simultion by the rigid-plastic explicit finite element method. A direct trial-and-error method is introduced to treat contact and friction. In computation, sheet material is assumed to possess normal anisotropy and rigid-plastic workhardening characteristics. In order to show the validity and effectiveness of the proposed explicit scheme, computations are carried out for cylindrical punch stretching and the computational results are compared with those by the implicit scheme as well as with a commercial code. The proposed rigid-plastic exlicit finite element method can be used as a robust and efficient computational method for analysis of sheet metal forming.

• Proceedings of the Computational Structural Engineering Institute Conference
• /
• 2004.10a
• /
• pp.270-277
• /
• 2004

In this paper, path-independent values of the J-integral in the fininte element context for arbitrary two-dimensional and three-dimensional crack configurations in welds are presented. For the fracture mechanics analysis of cracks in welds, residual stress analysis and fracture analysis must be performed simultaneously. In the analysis of cracked bodies containing residual stress, the usual domain integral formulation results in path-dependent values of the J-integral. This paper discusses modifications of the conventional J-integral that yield path independence in the presence of residual stress generated by welding. The residual stress problem is treated as an initial strain problem and the J-integral modified for this class of problem is used. And a finite element program which can evaluate the J-integral for cracks in two-dimensional and three-dimensional residual stress bearing bodies is developed using the modified J-integral definition. The situation when residual stress only is present is examed as is the case when mechanical stresses are applied in conjunction with a residual stress field.

• Proceedings of the Korean Geotechical Society Conference
• /
• 2006.03a
• /
• pp.394-401
• /
• 2006

This study developed two dimensional finite element program(FE-SCP) for the analysis of a composite ground reinforced by sand compaction piles with a low area replacement ratio based on the Mohr-Coulomb elastic perfectly plastic constitutive model. Program FE-SCP give some conveniences to users such as automatic mesh generation according to the replacement ratio and the effective sand pile diameter in the post processor. Also, it contains optimum processor in calculation of In-situ stress equilibrium considering different coefficient of earth pressure between sand pile and surrounding clay. Estimated stress-strain behavior using FE-SCP and the measured one from a centrifuge test showed good agreement comparing to the result from a general finite element program.

• Proceedings of the KSME Conference
• /
• 2001.11a
• /
• pp.367-371
• /
• 2001

In the shape optimization based on the finite element method, the accuracy of finite element analysis of a given structure is important to determine the final shape. In case of a bending dominant problem, finite element solutions by the full integration scheme are not reliable because of the locking phenomenon. Furthermore, in the process of shape optimization, the mesh distortion is large due to the change of the structure outline: therefore, we cannot guarantee the accurate result unless the finite element itself is accurate. We approach to more accurate shape optimization to diminish these inaccuracies by improving the existing finite element. The shape optimization using the modified finite element is applied to a two-dimensional simple beam. Results show that the modified finite element have improved the optimization results.

• Bulletin of the Society of Naval Architects of Korea
• /
• v.23 no.4
• /
• pp.11-18
• /
• 1986

The fundamental theory and application of boundary element method for two-dimensional problem are introduced in this paper. Based on this boundary element procedure, several numerical calculations such as circular cavity problem, a thin plate with hole under tension and a long thick-walled cylinder under internal pressure are performed. The numerical results show fairly good agreement with exact solutions or results of finite element method.

• Computers and Concrete
• /
• v.16 no.5
• /
• pp.759-774
• /
• 2015

This study simulates the flexural behavior of ultra-high-performance fiber-reinforced concrete (UHPFRC) beams reinforced with steel and glass fiber-reinforced polymer (GFRP) rebars. For this, micromechanics-based modeling was first carried out on the basis of single fiber pullout models considering inclination angle. Two different tension-softening curves (TSCs) with the assumptions of 2-dimensional (2-D) and 3-dimensional (3-D) random fiber orientations were obtained from the micromechanics-based modeling, and linear elastic compressive and tensile models before the occurrence of cracks were obtained from the mechanical tests and rule of mixture. Finite element analysis incorporating smeared crack model was used due to the multiple cracking behaviors of structural UHPFRC beams, and the characteristic length of two times the element width (or two times the average crack spacing at the peak load) was suggested as a result of parametric study. Analytical results showed that the assumption of 2-D random fiber orientation is appropriate to a non-reinforced UHPFRC beam, whereas the assumption of 3-D random fiber orientation is suitable for UHPFRC beams reinforced with steel and GFRP rebars due to disorder of fiber alignment from the internal reinforcements. The micromechanics-based finite element analysis also well predicted the serviceability deflections of UHPFRC beams with GFRP rebars and hybrid reinforcements.

• Journal of Welding and Joining
• /
• v.14 no.5
• /
• pp.69-77
• /
• 1996

This study aims to analyze the S. I. F. K value upon Mode I cracks in a finite-width plate of varying thickness, which is expressed in terms of width ratio ($\omega$), thickness ratio ($\beta$) and non-dimensional crack length (λ) by using the 2-dimensional finite element method. Then, by comparing the effectiveness of the results obtained by the two finite element methods, it is seen that the 2-dimensional finite element method can be used in order to analyse the S. I. F. K values upon a various thickness model. A model is developed in order to analyze the effects of initial residual stress upon the fatigue crack growth behavior in various thickness welded specimens. In this model, crack growth rate da/dN appears to be come small as the thickness ratio with the same ΔK is increased. Also, in the initial step, as ΔK is increased with crack growth rate is decreased and then increased because the repeated compressive residual stress retards crack growth rate.

• Korean Journal of Computational Design and Engineering
• /
• v.14 no.3
• /
• pp.197-206
• /
• 2009

In a finite element analysis of the metal forming processes having large plastic deformation, largely distorted elements are unstable and hence they influence upon the result toward negative way so that adaptive remeshing is required to avoid a failure in the numerical computation. Therefore automatic mesh generation and regeneration is very important to avoid a numerical failure in a finite element analysis. In case of generating quadrilateral mesh, the automation is more difficult than that of triangular mesh because of its geometric complexity. However its demand is very high due to the precision of analysis. Thus, in this study, an automatic quadrilateral mesh generation and regeneration method using grid-based approach is developed. The developed method contains decision of grid size to generate initial mesh inside a two dimensional domain, classification of boundary angles and inner boundary nodes to improve element qualities in case of concave domains, and boundary projection to construct the final mesh.

• Transactions of the Korean Society of Mechanical Engineers
• /
• v.18 no.11
• /
• pp.2922-2931
• /
• 1994

We propose a modified 6-node element, where the sampling point of Gauss quadrature moved in the thickness direction. The modified 6-node element has been applied to static problems and forced motion analyses. In this study, this method is extended to the finite element analysis of the natural frequencies of two dimensional problems. We also propose a modified 16-node element for three dimensional problems, which behaves much like a 20-node element with smaller degree of freedom. The modified 6-node and 16-node elements have been applied to the modal analyses of beams and plates, respectively. The results agree well with the results of the 8-node or 20-node element models.