• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.40 no.11
• /
• pp.972-978
• /
• 2012

This paper discusses the parallelization of multifrontal solution method, widely used for finite element structural analyses, for a shared memory architecture. Multifrontal method is easier than other linear solution methods because the solution procedure implies that unknowns can be eliminated simultaneously. Two innovative ideas are introduced to achieve optimal solver performance on a shared memory computer. Those are pairing two frontal matrices and splitting the frontal matrix in order to reduce the temporal memory space required by independent computing tasks. Performance comparisons between original algorithm and proposed one prove that proposed method is more computationally efficient on current multicore machines.

• Proceedings of the Korean Society for Technology of Plasticity Conference
• /
• 2001.10a
• /
• pp.113-116
• /
• 2001

Finite element analysis programs have been for metal forming process design They will become more and more important in understanding forming process For large-scale forging analysis problems, the performance of a linear equation solver is very important for the overall efficiency of the analysis code. With problem size increased, the computation time needs to be reduced, which is spent on setting the system of algebraic equations associated with finite element model Many matrix solvers have been developed and used usefully in finite element program for this purpose.

• Journal of the Korean Society for Industrial and Applied Mathematics
• /
• v.13 no.1
• /
• pp.13-20
• /
• 2009

The IPSAP which is a finite element analysis program has been developed for high parallel performance computing. This program consists of various analysis modules - stress, vibration and thermal analysis module, etc. The M orthogonal block Lanczos algorithm with shiftinvert transformation is used for solving eigenvalue problems in the vibration module. And the multifrontal algorithm which is one of the most efficient direct linear equation solvers is applied to factorization and triangular system solving phases in this block Lanczos iteration routine. In this study, the performance enhancement procedures of the IPSAP are composed of the following stages: 1) communication volume minimization of the factorization phase by modifying parallel matrix subroutines. 2) idling time minimization in triangular system solving phase by partial inverse of the frontal matrix and the LCM (least common multiple) concept.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.21 no.12
• /
• pp.2096-2105
• /
• 1997

A parallel approach using a network of engineering workstations is presented for the efficient computation in the elastoplastic analysis of strip rolling process. The domain decomposition method coupled with the frontal solver for elimination of internal degrees of freedom in each subdomain is used. PVM is used for message passing and synchronization between processors. A 2-D plane strain problem and the strip rolling process are analyzed to demonstrate the performance of the algorithm and factors that have a great effect on efficiency are discussed. In spite of much communication time on the network the result illustrates the advantages of this parallel algorithm over its corresponding sequential algorithm.

• Computational Structural Engineering
• /
• v.2 no.2
• /
• pp.73-83
• /
• 1989

A computer program, MSSLL, was developed for the analysis of the "Multiple Shell Subjected to Lateral Loads" by utilizing 2-dimensional finite elements in a 3-dimensional global assemblage with 6 DOF at each nodal point. In this program, substructuring procedure with frontal solver was introduced in the solution procedure to save both human and computer resource when the whole structure consists of repeated identical substructures. Some of the results obtained by MSSLL were compared with the existing solutions by other methods. The effect of rise to span-length ratio was investigated for the behavior of the multiple conical shell with 8 substructures subjected to seismic loads by performing a parametric study.

• The Magazine of the Society of Air-Conditioning and Refrigerating Engineers of Korea
• /
• v.6 no.4
• /
• pp.262-266
• /
• 1977

A new computer system for the stress analysis and design of piping network has been devlo-ped in this study. For the stress analysis, the system utilizes the finite element technique in which the frontal technique is used as the equation solver. The element library of the system has (1) Pipe Element (2) Beam Element, (3) Hanger Element and (4) Spring Element which should be sufficient to model the entire piping system including flexible supports, joints, piping rack and hangers. Based on the element stresses, code check has been performed and the safety factor for each element is calculated.

• Journal of Mechanical Science and Technology
• /
• v.14 no.6
• /
• pp.633-644
• /
• 2000

Recently, in proportion to the increase of earthquake occurrence-frequency and its strength in the countries within the circum-pan Pacific earthquake belt, a concept of earthquake-proof design for huge structures containing liquid has been growing up. This study deals with the refinement of classical numerical approaches for the free vibration analysis of separated structure and liquid motions. According to the liquid-structure interaction, LNG-storage tanks exhibit two distinguished eigenmodes, the sloshing mode and the bulging mode. For the sloshing -mode analysis, we refine the classical rigid-tank model by reflecting the container flexibility. While, for the bulging-mode analysis, we refine the classical uncoupled structural vibration system by taking the liquid free-surface fluctuation into consideration. We first construct the refined dynamic models for both problems, and present the refined numerical procedures. Furthermore, in order for the efficient treatment of large-scale matrices, we employ the Lanczos iteration scheme and the frontal-solver for our test FEM program. With the developed program we carry out numerical experiments illustrating the theoretical results.

• Transactions of the Korean Society of Automotive Engineers
• /
• v.10 no.5
• /
• pp.214-222
• /
• 2002

Computer simulations are widely used to analyze passenger safety in simulated traffic accidents. ATB, Articulated Total Body, is a computer simulation model developed to predict gross human body response to such dynamic environments as vehicle crashes and pilot ejections. ATB, whose code is open, has high flexibility and application capability that users can easily insert defined modules and functions. ATB is, however, inconvenient as it was coded in FORTRAN and it needs a formated input file. Moreover, it takes much time to make input files and to modify coding errors. This study aims to increase user friendliness by adding a preprocessor program, WINATB(WINdows ATB), to the conventional ATB. WINATB, programmed in Visual C＋＋ and OpenGL, uses ATB IV as a dynamic solver. The preprocessor helps users prepare input files through graphic interface and dialog box. An additional postprocessor makes the graphical presentation of simulated results. In these case of the frontal crash, the rear impact and the side impact, the simulation results obtained by WINATB and MADYMO(MAthematical Dynamic Model) are compared to validate the effectiveness of WINAIB.

• Journal of the Korean Institute of Telematics and Electronics D
• /
• v.36D no.11
• /
• pp.45-55
• /
• 1999

The small signal analyzer for the stationary drift-diffusion equation is developed. The slotboom variables of the potential, electron and hole concentrations for the response of applied small signal are defined and the stationary drift-diffusion equation is linearlized on DC operation point by $S^3A$ method. Frontal solver, which is used to solve the global matrix, progresses the accuracy of the solution in high frequency and minimizes the requirement of the memory. The simulations are executed on the structure of 3 dimensional N'P junction diode and 2 dimensional n-MOSFET to verify the proposed algorithm. The average relative errors of the conductance and the capacitance compared with MEDICI are about 26% and 0.67 for N'P junction diode and 7.75% and 2.24% for n-MOSFET. The simulation by the proposed algorithm can analyze the stationary drift-diffusion equation for applied small signal in high frequency region about 100GHz.

• Composites Research
• /
• v.26 no.6
• /
• pp.363-372
• /
• 2013

Numerical simulation for Resin Transfer Molding manufacturing process is attempted by using the eXtended Finite Element Method (XFEM) combined with the level set method. XFEM allows to obtaining a good numerical precision of the pressure near the resin flow front, where its gradient is discontinuous. The enriched shape functions of XFEM are derived by using the level set values so as to correctly describe the interpolation with the resin flow front. In addition, the level set method is used to transport the resin flow front at each time step during the mold filling. The level set values are calculated by an implicit characteristic Galerkin FEM. The multi-frontal solver of IPSAP is adopted to solve the system. This work is validated by comparing the obtained results with analytic solutions. Moreover, a localization method of XFEM and level set method is proposed to increase the computing efficiency. The computation domain is reduced to the small region near the resin flow front. Therefore, the total computing time is strongly reduced by it. The efficiency test is made with a simple channel flow model. Several application examples are analyzed to demonstrate ability of this method.