J. Sladek¹, V. Sladek¹, S.N. Atluri²

CMES-Computer Modeling in Engineering & Sciences, Vol.6, No.3, pp. 309-318, 2004, DOI:10.3970/cmes.2004.006.309

Abstract Meshless methods based on the local Petrov-Galerkin approach are proposed for solution of steady and transient heat conduction problem in a continuously nonhomogeneous anisotropic medium. Fundamental solution of the governing partial differential equations and the Heaviside step function are used as the test functions in the local weak form. It is leading to derive local boundary integral equations which are given in the Laplace transform domain. The analyzed domain is covered by small subdomains with a simple geometry. To eliminate the number of unknowns on artificial boundaries of subdomains the modified fundamental solution and/or the More >

T. Tsangaris¹, Y.–S. Smyrlis^{1, 2}, A. Karageorghis^{1, 2}

CMC-Computers, Materials & Continua, Vol.1, No.3, pp. 245-258, 2004, DOI:10.3970/cmc.2004.001.245

Abstract The Method of Fundamental Solutions (MFS) is a boundary-type method for the solution of certain elliptic boundary value problems. In this work, we develop an efficient matrix decomposition MFS algorithm for the solution of biharmonic problems in annular domains. The circulant structure of the matrices involved in the MFS discretization is exploited by using Fast Fourier Transforms. The algorithm is tested numerically on several examples. More >

Jan Sladek¹, Vladimir Sladek¹, Chuanzeng Zhang²

CMES-Computer Modeling in Engineering & Sciences, Vol.4, No.6, pp. 637-648, 2003, DOI:10.3970/cmes.2003.004.637

Abstract A new computational method for solving transient elastodynamic initial-boundary value problems in continuously non-homogeneous solids, based on the meshless local Petrov-Galerkin (MLPG) method, is proposed in the present paper. The moving least squares (MLS) is used for interpolation and the modified fundamental solution as the test function. The local Petrov-Galerkin method for unsymmetric weak form in such a way is transformed to the local boundary integral equations (LBIE). The analyzed domain is divided into small subdomains, in which a weak solution is assumed to exist. Nodal points are randomly spread in the analyzed domain and More >

E. J. Sellountos¹, D. Polyzos²

CMES-Computer Modeling in Engineering & Sciences, Vol.4, No.6, pp. 619-636, 2003, DOI:10.3970/cmes.2003.004.619

Abstract A new meshless local Petrov-Galerkin (MLPG) method for solving two dimensional frequency domain elastodynamic problems is proposed. Since the method utilizes, in its weak formulation, either the elastostatic or the frequency domain elastodynamic fundamental solution as test function, it is equivalent to the local boundary integral equation (LBIE) method. Nodal points spread over the analyzed domain are considered and the moving least squares (MLS) interpolation scheme for the approximation of the interior and boundary variables is employed. Two integral equations suitable for the integral representation of the displacement fields in the local sub- domains are… More >

Q. Li¹, S. Shen¹, Z. D. Han¹, S. N. Atluri¹

CMES-Computer Modeling in Engineering & Sciences, Vol.4, No.5, pp. 571-586, 2003, DOI:10.3970/cmes.2003.004.571

Abstract In this paper, a truly meshless method, the Meshless Local Petrov-Galerkin (MLPG) Method, is developed for three-dimensional elasto-statics. The two simplest members of MLPG family of methods, the MLPG type 5 and MLPG type 2, are combined, in order to reduce the computational requirements and to obtain high efficiency. The MLPG5 method is applied at the nodes inside the 3-D domain, so that any domain integration is eliminated altogether, if no body forces are involved. The MLPG 2 method is applied at the nodes that are on the boundaries, and on the interfaces of material More >

Yiorgos-Sokratis Smyrlis¹, Andreas Karageorghis¹

CMES-Computer Modeling in Engineering & Sciences, Vol.4, No.5, pp. 535-550, 2003, DOI:10.3970/cmes.2003.004.535

Abstract In this study, we investigate the application of the Method of Fundamental Solutions for the solution of biharmonic Dirichlet problems on a disk. Modifications of the method for overcoming sources of inaccuracy are suggested. We also propose an efficient algorithm for the solution of the resulting systems which exploits the symmetries of the matrices involved. The techniques described in the paper are applied to standard test problems. More >

I. S. Raju¹, D. R. Phillips²

CMES-Computer Modeling in Engineering & Sciences, Vol.4, No.1, pp. 141-160, 2003, DOI:10.3970/cmes.2003.004.141

Abstract An accurate and yet simple Meshless Local Petrov-Galerkin (MLPG) formulation for analyzing beam problems is presented. In the formulation, simple weight functions are chosen as test functions as in the conventional MLPG method. Linear test functions are also chosen, leading to a variation of the MLPG method that is computationally efficient compared to the conventional implementation. The MLPG method is evaluated by applying the formulation to a variety of patch tests, thin beam problems, and problems with load discontinuities. The formulation successfully reproduces exact solutions to machine accuracy when higher order power and spline functions More >

Jin Yeon Cho¹, Seung Jo Kim²

CMES-Computer Modeling in Engineering & Sciences, Vol.3, No.6, pp. 687-698, 2002, DOI:10.3970/cmes.2002.003.687

Abstract In this work, an explicit solution procedure for the recently developed discontinuous time integration method is proposed in order to reduce the computational cost while maintaining the desirable numerical characteristics of the discontinuous time integration method. In the present explicit solution procedure, a two-stage correction algorithm is devised to obtain the solution at the next time step without any matrix factorization. To observe the numerical characteristics of the proposed explicit solution procedure, stability and convergence analyses are performed. From the stability analysis, it is observed that the proposed algorithm gives a larger critical time step More >

Olivier Ricou¹, Michel Bercovier²

CMES-Computer Modeling in Engineering & Sciences, Vol.3, No.1, pp. 87-102, 2002, DOI:10.3970/cmes.2002.003.087

Abstract In this article, we present a method of reduction of the dimension of the Stokes equations by one in a quasi-cylindrical domain. It takes the special shape of the domain into account by the use of a projection onto a space of polynomials defined over the thickness. The polynomials are defined to fit as well as possible with the variables they approximate. Hence, this method restricted to the first polynomial, recovers the Hele-Shaw approximation.
The convergence of the approximate solution to the continuous one is shown. Under a regularity hypothesis, we also obtain error estimates. More >

Weimin Han, Xueping Meng¹

CMES-Computer Modeling in Engineering & Sciences, Vol.3, No.1, pp. 65-76, 2002, DOI:10.3970/cmes.2002.003.065

Abstract Interests in meshfree (or meshless) methods have grown rapidly in the recent years in solving boundary value problems arising in mechanics, especially in dealing with difficult problems involving large deformation, moving discontinuities, etc. Rigorous error estimates of a meshfree method, the reproducing kernel particle method, for smooth solutions have been theoretically derived and experimentally tested in Han, Meng (2001). In this paper, we provide an error analysis of the meshfree method for solving problems with singular solutions. The results are presented in the context of one-dimensional problems. The error estimates are of optimal order and More >