Z. Chen^1,2, Y. Gan¹, J.K. Chen²

CMES-Computer Modeling in Engineering & Sciences, Vol.26, No.2, pp. 123-138, 2008, DOI:10.3970/cmes.2008.026.123

Abstract A coupled thermo-mechanical constitutive model with decohesion is proposed to simulate the material failure evolution due to localized heating. A discontinuous bifurcation analysis is performed based on a thermoviscoplasticity relation to identify the transition from continuous to discontinuous failure modes as well as the orientation of the discontinuous failure. The thermo-mechanical model is then implemented within the framework of the Material Point Method (MPM) so that the different gradient and divergence operators in the governing differential equations could be discretized in a single computational domain and that continuous remeshing is not required with the evolution More >

Amit Shaw¹, B. Banerjee¹, D Roy^1,2

CMES-Computer Modeling in Engineering & Sciences, Vol.26, No.1, pp. 31-60, 2008, DOI:10.3970/cmes.2008.026.031

Abstract A generalization of a NURBS based parametric mesh-free method (NPMM), recently proposed by Shaw and Roy (2008), is considered. A key feature of this parametric formulation is a geometric map that provides a local bijection between the physical domain and a rectangular parametric domain. This enables constructions of shape functions and their derivatives over the parametric domain whilst satisfying polynomial reproduction and interpolation properties over the (non-rectangular) physical domain. Hence the NPMM enables higher-dimensional B-spline based functional approximations over non-rectangular domains even as the NURBS basis functions are constructed via the usual tensor products of… More >

L.K. Keppas¹, G.I. Giannopoulos¹, N.K. Anifantis¹

CMES-Computer Modeling in Engineering & Sciences, Vol.25, No.3, pp. 181-196, 2008, DOI:10.3970/cmes.2008.025.181

Abstract In the present paper a boundary element procedure is formulated to treat two-dimensional time dependent thermo-elastic contact problems incorporating thermal resistance along the contacting surfaces. The existence of pressure-dependent thermal contact leads to coupling of temperature and stress fields. Therefore, the inherent non-linearity of the problem demands simultaneous treating of both thermal and mechanical boundary integral equations while iterative procedures are introduced to ensure equilibrium of mechanical and thermal contact conditions at each step of the process. The transient behavior of interfacial cracks in bimaterial solids when undergo thermal shock in the presence of partial More >

C. Moulinec¹, R. Issa², J.-C. Marongiu³, D. Violeau⁴

CMES-Computer Modeling in Engineering & Sciences, Vol.25, No.3, pp. 133-148, 2008, DOI:10.3970/cmes.2008.025.133

Abstract The gridless Smoothed Particle Hydrodynamics (SPH) numerical method is preferably used in Computational Fluid Dynamics (CFD) to simulate complex flows with one or several convoluted free surfaces. This type of flows requires distorted meshes with classical Eulerian mesh-based methods or very fine meshes with Volume of Fluid method. Few 3-D SPH simulations have been carried out to our knowlegde so far, mainly due to prohibitive computational investment since the number of particles required in 3-D is usually too large to be handled by a single processor. In this paper, a parallel 3-D SPH code is More >

Atsuya Oishi¹, Shinobu Yoshimura²

CMES-Computer Modeling in Engineering & Sciences, Vol.25, No.2, pp. 115-132, 2008, DOI:10.3970/cmes.2008.025.115

Abstract This paper describes the finite element analyses of dynamic problems using graphics hardware. The graphics hardware, known as GPU that is an acronym of Graphics Processing Unit, was first developed only for processing 3D computer graphics. However it has obtained both flexible programmability using a high-level shader programming language such as OpenGL Shading Language (GLSL), and has also obtained fast numerical processing ability of over 160 GFLOPS that is much faster than CPU. In this paper, GPU is utilized for the finite element analyses of dynamic problems. Two different computational tasks in the dynamic finite More >

J.R. Jimenez-Octavio¹, O. Lopez-Garcia², E. Pilo¹, A. Carnicero²

CMES-Computer Modeling in Engineering & Sciences, Vol.25, No.2, pp. 81-98, 2008, DOI:10.3970/cmes.2008.025.081

Abstract This paper presents a multidisciplinary design and optimization method of power transmission lines. This optimization method solves both mechanical and electrical problems by a new strongly coupled method that also optimizes the potential designs using a genetic algorithm. A multi-objective function is formulated to simplify a constrained typical optimization problem into an unconstrained one. The scope of this work is the sizing and configuration optimization problem with fixed topology. The method is applied to a railway overhead transmission line. The genetic algorithm is applied to mechanical, electrical and electromechanical optimization problems obtaining good results. Finally, More >

Liming Xiong¹, Youping Chen¹

CMES-Computer Modeling in Engineering & Sciences, Vol.24, No.2&3, pp. 203-214, 2008, DOI:10.3970/cmes.2008.024.203

Abstract This paper presents the application of an atomistic field theory (AFT) in modeling and simulation of boron- , boron/nitrogen and silicon/nitrogen-doped nanocrystalline silicon carbide (B-, BN-, SiN-SiC). Intergranular glassy films (IGFs) and nano-sized pores have been obtained in triple junctions of the grains in nanocrystalline SiC (nc-SiC). Residual tensile stress in the SiC grains and compressive stress in the grain boundaries (GBs) are observed. Under uniaxial tension, the constitutive responses of nanocrystalline SiC were reproduced from the simulations. It is found that the mechanical properties of nanocrystalline SiC are strongly dependent on the compositions of More >

Shardool U. Chirputkar¹, Dong Qian²

CMES-Computer Modeling in Engineering & Sciences, Vol.24, No.2&3, pp. 185-202, 2008, DOI:10.3970/cmes.2008.024.185

Abstract A multiscale method based on the extended space-time finite element method is developed for the coupled atomistic/continuum simulation of nanoscale material systems. Existing single scale approach such as the finite element method has limited capability of representing the fine scale physics in both the spatial and temporal domains. This is a major disadvantage for directly incorporating FEM in coupled atomistic/continuum simulations as it results in errors such as spurious wave reflections at the atomistic/continuum interface. While numerous efforts have been devoted to eliminating the interfacial mismatch effects, less attention has been paid to developing fine More >

G. Fitzgerald¹, G. Goldbeck-Wood², P. Kung¹, M. Petersen¹, L. Subramanian¹, J. Wescott²

CMES-Computer Modeling in Engineering & Sciences, Vol.24, No.2&3, pp. 169-184, 2008, DOI:10.3970/cmes.2008.024.169

Abstract Molecular modeling has established itself as an important component of applied research in areas such as drug discovery, catalysis, and polymers. Algorithmic improvements to these methods coupled with the increasing speed of computational hardware are making it possible to perform predictive modeling on ever larger systems. Methods are now available that are capable of modeling hundreds of thousands of atoms, and the results can have a significant impact on real-world engineering problems. The article reviews some of the modeling methods currently in use; provides illustrative examples of applications to challenges in sensors, fuel cells, and More >

F.C. Araújo¹, L.J. Gray²

CMES-Computer Modeling in Engineering & Sciences, Vol.24, No.2&3, pp. 103-122, 2008, DOI:10.3970/cmes.2008.024.103

Abstract In recent years, carbon nanotubes (CNTs) have been widely employed to build advanced composites. In this work, a Boundary Element Method (BEM) is applied to 3D representative volume elements (RVEs) to estimate mechanical properties of CNT-based composites. To model the thin-walled nanotubes, special integration procedures for calculating nearly-strongly-singular integrals have been developed. The generic BE substructuring algorithm allows modeling complex CNT-reinforced polymers, containing any number of nanotubes of any shape (straight or curved). The subregion-by-subregion strategy, based on Krylov solvers, makes the independent generation, assembly, and storage of the many parts of the complete BE More >