A finite difference scheme with arbitrary mesh systems for solving high-order partial differential equations

doi:10.1016/0045-7949(89)90379-9

Computers & Structures

Volume 31, Issue 3, 1989, Pages 319-328

https://doi.org/10.1016/0045-7949(89)90379-9 Get rights and content

Abstract

A finite difference technique for arbitrary irregular mesh systems is developed for numerical solutions of high-order boundary value problems. A successive differencing scheme is used to formulate the higher-order differencing equations. The details of the computational procedure for plate bending problems which are governed by a biharmonic equation are presented. In order to demonstrate the validity of the formulation procedure, numerical results obtained from the present technique were first compared with the analytical solutions for problems having regular and curved boundaries. For the regular boundary problem, the present results were also compared with that from the conventional finite difference method which is only suitable for treating regular meshes. Computations were then conducted to evaluate the accuracy of the solutions with different mesh arrangements. These studies have shown that, if well implemented, this method not only maintains the inherent simplicity of the finite difference method, but also puts it into a more programmable and flexible formulation so that the geometric generality, such as that in the finite element method, can be achieved.

References (16)

T. Liszka et al.
The finite difference method at arbitrary irregular grids and its application in applied mechanics
Comput. Struct.
(1980)
P.S. Jensen
Finite difference techniques for variable grids
Comput. Struct.
(1972)
A.A. Tseng
A generalized finite difference scheme for convection-dominated metalforming problems
Int. J. Numer. Meth. Engng
(1984)
R. von Mises
On network methods in conformal mapping and in related problems
Nat. Bur. Stand. Appl. Math. Ser.
(1952)
R.H. MacNeal
An asymmetrical finite difference network
Q. J. appl. Math.
(1953)
L. Collatz
The Numerical Treatment of Differential Equations
(1960)
G.E. Forsythe et al.
Finite-Difference Methods for Partial Differential Equations
(1960)
G. Davies et al.
Application of an irregular mesh finite difference approximation to the plate buckling problem

There are more references available in the full text version of this article.

Cited by (16)

A new approach for solving heat conduction under zero and non-zero initial conditions
2022, Engineering Analysis with Boundary Elements
Citation Excerpt :
Several mesh-based approaches are alternatives if the domain is discretized and allows the simulation of non-homogeneous media. For instance, Zhang et al. [16] and Mitsoulis et al. [17] used the finite element method (FEM), Cai et al. [18] applied the finite volume method (FVM) and Ozisik [19], Tseng et al. [20] used the finite difference method (FDM) in large models with more complex shapes but finite dimensions. Cascio et al. [21] presented a hybrid formulation coupling the virtual element method (VEM) [22–24] with the boundary element method formulation for heterogeneous materials.
This paper provides a new approach for the analysis of transient heat diffusion arising from conduction. The diffusion is subject to zero and non-zero initial conditions in a two-dimensional domain, containing confined heterogeneous subdomains. This system can be subjected to a heat source. A new formulation is presented, in the frequency domain. It allows coupling between the boundary element method (BEM) and other methods (such as the method of fundamental solutions (MFS), the finite element method (FEM), the finite difference method (FDM), the meshless Petrov-Galerkin (MLGP) method, the virtual element method (VEM), and analytical formulations). The BEM is used for the homogeneous exterior domain and the other method is used to evaluate the solutions for the non-homogeneous confined subdomains, to overcome the specific limitations of each of these methods.
An inverse fast Fourier transform was used to provide the time domain responses and the results were then compared with those of reference solutions. It was then possible to demonstrate that the proposed algorithm could be used in the frequency and time domains by performing simulation analyses to study heat diffusion in the neighbourhood of heterogeneous inclusions within a solid medium.
The generalized finite difference method with third- and fourth-order approximations and treatment of ill-conditioned stars
2021, Engineering Analysis with Boundary Elements
Citation Excerpt :
However, higher-order approximations have been rarely used in the GFDM, due to two main reasons: the need to increase the minimum number of points per star, which often causes problems of ill-conditioning, and the high computational cost. To avoid the use of the fourth-order approximation when the differential equation is of fourth-order, Orkisz [4] and Tseng and Gu [24] suggested decomposing the fourth-order operators by successive approximations using second-order operators. Nowadays, the power of computers allows using the fourth-order approximation with many more points per star than the minimum necessary, thus avoiding ill-conditioned stars.
In this paper, we solve 2D and 3D second-order partial differential equations considering the Generalized Finite Difference Method with third- and fourth-order approximations. We analyze the influence of the number of points per star and establish some values as references. We propose a new strategy to deal with ill-conditioned stars, which are frequent in higher-order approximations. This strategy uses a few points per star and presents excellent results both for detecting ill-conditioned stars and for increasing the accuracy of the numerical approximation. We apply parallel processing to the formation of stars and derivatives, and show the speedup achieved.
Multilayer perceptrons and radial basis function neural network methods for the solution of differential equations: A survey
2011, Computers and Mathematics with Applications
Citation Excerpt :
A series of problems in many scientific fields can be modeled with the use of differential equations such as problems in physics [1–3], chemistry [4–6], biology [7,8], economics [9], etc. Due to the importance of differential equations many methods have been proposed in the relevant literature for their solution such as Runge–Kutta methods [10,11], Predictor–Corrector methods [12,13], Finite Difference methods [14–16], Finite Element methods [17–19], Splines [20–23] and other methods [24–41] etc. These methods require the discretization of domain into the number of finite elements where the functions are approximated locally.
Since neural networks have universal approximation capabilities, therefore it is possible to postulate them as solutions for given differential equations that define unsupervised errors. In this paper, we present a wide survey and classification of different Multilayer Perceptron (MLP) and Radial Basis Function (RBF) neural network techniques, which are used for solving differential equations of various kinds. Our main purpose is to provide a synthesis of the published research works in this area and stimulate further research interest and effort in the identified topics. Here, we describe the crux of various research articles published by numerous researchers, mostly within the last 10 years to get a better knowledge about the present scenario.
A generalized finite difference method using Coatmèlec lattices
2009, Computer Physics Communications
Citation Excerpt :
The list of methods referred to as meshless is very large and it is continuously growing. For example, smooth particle hydrodynamics [2], generalized finite differences [3,4], moving least squares techniques [5], diffuse elements [6], element-free Galerkin [7], to name only a few (a more comprehensive list can be found in [1]). Meshless methods are useless without an evaluation of the nodal connectivity bounded in time and a computational cost which grows linearly with the total number of nodes in the domain [1].
Generalized finite difference methods require that a properly posed set of nodes exists around each node in the mesh, so that the solution for the corresponding multivariate interpolation problem be unique. In this paper we first show that the construction of these meshes can be computerized using a relatively simple algorithm based on the concept of a Coatmèlec lattice. Then, we present a generalized finite difference method which provides a numerical solution of a partial differential equation over an arbitrary domain, using the generated meshes. The accuracy and mesh adaptivity of the method is evaluated using elliptical equations in several domains.
Numerical solution of quenching problems using mesh-dependent variable temporal steps
2007, Applied Numerical Mathematics
In this paper, we introduce a new adaptive method for computing the numerical solutions of a class of quenching parabolic equations which exhibit a solution with one singularity. Our method systematically generates an irregular mesh with mesh-dependent temporal increments based on the solution behavior from which an implicit finite difference scheme associated with the irregular mesh is constructed. The convergence and stability of the finite difference scheme is analyzed for the solution before quenching. An equivalent linearized model is used to justify the stability of the method near quenching as well. A numerical example is provided to demonstrate the viability of the proposed method.
A finite difference scheme with variable rectilinear mesh for solving multi-harmonic partial differential equations
1992, Computers and Structures
A finite difference technique for variable rectilinear meshes is developed for numerical solutions of harmonic, biharmonic and multi-harmonic boundary value problems. A successive operator scheme is used to formulate a biharmonic equation from a harmonic one. The details of the computational procedure are included for torsion and plate bending problems which are governed by harmonic and biharmonic equations, respectively. The numerical results obtained from the present technique are compared with analytic and generalized finite difference solutions of Tseng and Gu for problems having regular and curved boundaries. Computations are then conducted to evaluate the accuracy of the solutions with different mesh arrangements.

View all citing articles on Scopus

^†: Visiting Scholar, originally with the Department of Mathematics and Mechanics, Lanzhou University, People's Republic of China.

View full text

A finite difference scheme with arbitrary mesh systems for solving high-order partial differential equations

Abstract

Comput. Struct.

Comput. Struct.

A generalized finite difference scheme for convection-dominated metalforming problems

Int. J. Numer. Meth. Engng

On network methods in conformal mapping and in related problems

Nat. Bur. Stand. Appl. Math. Ser.

An asymmetrical finite difference network

Q. J. appl. Math.

The Numerical Treatment of Differential Equations

Finite-Difference Methods for Partial Differential Equations

Application of an irregular mesh finite difference approximation to the plate buckling problem