G. Gambolati¹, M. Ferronato¹, C. Janna¹

The International Conference on Computational & Experimental Engineering and Sciences, Vol.9, No.2, pp. 99-106, 2009, DOI:10.3970/icces.2009.009.099

Abstract The Finite Element (FE) solution to consolidation equations in large geological settings raises a few numerical issues depending on the actual process addressed by the analysis. There are two basic problems where the solver efficiency plays a crucial role: 1- fully coupled consolidation, and 2- non-linear faulted (uncoupled) consolidation. Using a proper nodal numbering the FE matrices exhibit a block (or multilevel) structure. Krylov subspace solvers are attracting a growing attention, provided that a relatively inexpensive and effective preconditioner is available. For both problems possible preconditioners include the Diagonal Scaling (DS), the Incomplete Triangular Factorization (ILU), the Mixed Constraint Preconditioning… More >

E-N.G. Grylonakis¹, C.K. Filelis-Papadopoulos¹, G.A. Gravvanis¹

CMES-Computer Modeling in Engineering & Sciences, Vol.109-110, No.6, pp. 505-517, 2015, DOI:10.3970/cmes.2015.109.505

Abstract A new transform method for solving boundary value problems in two dimensions was proposed by A.S. Fokas, namely the unified transform. This approach seeks a solution to the unknown boundary values by solving a global relation, using the known boundary data. This relation can be used to characterize the Dirichlet to Neumann map. For the numerical solution of the global relation, a collocation-type method was recently introduced. Hence, the considered method is used for solving the 2D Laplace equation in several irregular convex polygons. The linear system, resulting from the collocation-type method, was solved by the Explicit Preconditioned Generalized Minimum… More >

Chein-Shan Liu¹

CMES-Computer Modeling in Engineering & Sciences, Vol.92, No.3, pp. 241-269, 2013, DOI:10.32604/cmes.2013.092.241

Abstract In order to solve an ill-posed linear problem, we propose an innovative Jacobian type iterative method by presetting a conditioner before the steepest descent direction. The preconditioner is derived from an invariant manifold approach, which includes two parameters α and γ to be determined. When the weighting parameter α is optimized by minimizing a properly defined objective function, the relaxation parameter γ can be derived to accelerate the convergence speed under a switching criterion. When the switch is turned-on, by using the derived value of γ it can pull back the iterative orbit to the fast manifold. It is the… More >

Christos K. Filelis-Papadopoulos¹, George A. Gravvanis¹

CMES-Computer Modeling in Engineering & Sciences, Vol.90, No.3, pp. 233-253, 2013, DOI:10.3970/cmes.2013.090.232

Abstract During the last decades, multigrid methods have been extensively used in order to solve large scale linear systems derived from the discretization of partial differential equations using the finite difference method. Approximate Inverses in conjunction with Richardon’s iterative method could be used as smoothers in the multigrid method. Thus, a new class of smoothers based on approximate inverses could be derived. Effectiveness of explicit approximate inverses relies in the fact that they are close approximants to the inverse of the coefficient matrix and are fast to compute in parallel. Furthermore, the class of finite difference approximate inverses proposed in conjunction… More >

G.A. Gravvanis¹, K.M. Giannoutakis¹

CMES-Computer Modeling in Engineering & Sciences, Vol.16, No.2, pp. 69-82, 2006, DOI:10.3970/cmes.2006.016.069

Abstract A new class of normalized explicit optimized approximate inverse finite element matrix techniques, based on normalized finite element approximate factorization procedures, for solving sparse linear systems resulting from the finite element discretization of partial differential equations in three space variables are introduced. A new parallel normalized explicit preconditioned conjugate gradient square method in conjunction with normalized approximate inverse finite element matrix techniques for solving efficiently sparse finite element linear systems on distributed memory systems is also presented along with theoretical estimates on speedups and efficiency. The performance on a distributed memory machine, using Message Passing Interface (MPI) communication library, is… More >

John L. Volakis¹, Kubilay Sertel¹, Erik Jørgensen², Rick W. Kindt¹

CMES-Computer Modeling in Engineering & Sciences, Vol.5, No.5, pp. 463-476, 2004, DOI:10.3970/cmes.2004.005.463

Abstract In this paper we address several topics relating to the development and implementation of volume integral and hybrid finite element methods for electromagnetic modeling. Comparisons of volume integral equation formulations with the finite element-boundary integral method are given in terms of accuracy and computing resources. We also discuss preconditioning and parallelization of the multilevel fast multipole method, and propose higher-order basis functions for curvilinear quadrilaterals and volumetric basis functions for curvilinear hexahedra. The latter have the desirable property of vanishing divergence within the element but non-zero curl. In addition, a new domain decomposition is introduced for solving array problems involving… More >

G.A. Gravvanis¹, K.M. Giannoutakis²

CMES-Computer Modeling in Engineering & Sciences, Vol.71, No.4, pp. 305-330, 2011, DOI:10.3970/cmes.2011.071.305

Abstract In this paper we present parallel explicit approximate inverse matrix techniques for solving sparse linear systems on shared memory systems, which are derived using the finite element method for biharmonic equations in three space variables. Our approach for solving such equations is by considering the biharmonic equation as a coupled equation approach (pair of Poisson equation), using a FE approximation scheme, yielding an inner-outer iteration method. Additionally, parallel approximate inverse matrix algorithms are introduced for the efficient solution of sparse linear systems, based on an anti-diagonal computational approach that eliminates the data dependencies. Parallel explicit preconditioned conjugate gradient-type schemes in… More >

R. Djebali¹, M. El Ganaoui²

CMES-Computer Modeling in Engineering & Sciences, Vol.71, No.3, pp. 179-202, 2011, DOI:10.3970/cmes.2011.071.179

Abstract The Lattice Boltzmann method (LBM) became, today, a powerful tool for simulating fluid flows. Its improvements for different applications and configurations offers more flexibility and results in several schemes such as in presence of external/internal forcing term. However, we look for the suitable model that gives correct informations, matches the hydrodynamic equations and preserves some features like coding easily, preserving computational cost, stability and accuracy. In the present work, high order incompressible models and equilibrium distribution functions for the advection-diffusion equations are analyzed. Boundary conditions, acceleration, stability and preconditioning with initial fields are underlined which permit to rigorously selecting two… More >

Peter Lucas¹, Alexander H. van Zuijlen¹, Hester Bijl¹

CMES-Computer Modeling in Engineering & Sciences, Vol.59, No.1, pp. 79-106, 2010, DOI:10.3970/cmes.2010.059.079

Abstract Despite the advances in computer power and numerical algorithms over the last decades, solutions to unsteady flow problems remain computing time intensive.
In previous work [Lucas, P.,Bijl, H., and Zuijlen, A.H. van(2010)], we have shown that a Jacobian-free Newton-Krylov (JFNK) algorithm, preconditioned with an approximate factorization of the Jacobian which approximately matches the target residual operator, enables a speed up of a factor of 10 compared to nonlinear multigrid (NMG) for two-dimensional, large Reynolds number, unsteady flow computations. Furthermore, in [Lucas, P., Zuijlen, A.H. van, and Bijl, H. (2010)] we show that this algorithm also greatly outperforms NMG for parameter… More >

Luca Bergamaschi^1,2, ,Ángeles Martínez², Giorgio Pini²

CMES-Computer Modeling in Engineering & Sciences, Vol.49, No.3, pp. 191-216, 2009, DOI:10.3970/cmes.2009.049.191

Abstract A meshless model, based on the Meshless Local Petrov-Galerkin (MLPG) approach, is developed and implemented in parallel for the solution of axi-symmetric poroelastic problems. The parallel code is based on a concurrent construction of the stiffness matrix by the processors and on a parallel preconditioned iterative method of Krylov type for the solution of the resulting linear system. The performance of the code is investigated on a realistic application concerning the prediction of land subsidence above a deep compacting reservoir. The overall code is shown to obtain a very high parallel efficiency (larger than 78% for the solution phase) and… More >