Engenharia Térmica (Thermal Engineering), Vol. 15 • No. 2 • December 2016 • p. 68-75
SOLUTION OF 1D AND 2D POISSON'S EQUATION
BY USING WAVELET SCALING FUNCTIONS
R. B. Burgos
a
,
and H. F. C. Peixoto
b
a
Universidade do Estado do Rio de Janeiro
Departamento de Estruturas e Fundações
Rua S. Francisco Xavier, 524,
Rio de Janeiro, RJ, Brasil.
rburgos@eng.uerj.br
b
PUC-Rio
Departamento de Engenharia Civil
Rua Marquês de São Vicente, 225,
Rio de Janeiro, RJ, Brasil.
hfcpeixoto@gmail.com
Received: September 27, 2016
Revised: October 14, 2016
Accepted: November 07, 2016
ABSTRACT
The use of multiresolution techniques and wavelets has become
increasingly popular in the development of numerical schemes for the
solution of partial differential equations (PDEs). Therefore, the use of
wavelet scaling functions as a basis in computational analysis holds some
promise due to their compact support, orthogonality and localization
properties. Daubechies and Deslauriers-Dubuc functions have been
successfully used as basis functions in several schemes like the Wavelet-
Galerkin Method (WGM) and the Wavelet Finite Element Method
(WFEM). Another possible advantage of their use is the fact that the
calculation of integrals of inner products of wavelet scaling functions and
their derivatives can be made by solving a linear system of equations, thus
avoiding the problem of using approximations by some numerical method.
These inner products were defined as connection coefficients and they are
employed in the calculation of stiffness matrices and load vectors. In this
work, some mathematical foundations regarding wavelet scaling functions,
their derivatives and connection coefficients are reviewed. A scheme based
on the Galerkin Method is proposed for the direct solution of Poisson's
equation (potential problems) in a meshless formulation using interpolating
wavelet scaling functions (Interpolets). The applicability of the proposed
method and some convergence issues are illustrated by means of a few
examples.
Keywords: wavelets, Poisson’s equation, Wavelet-Galerkin Method
NOMENCLATURE
A matrix of filter coefficients
a
k
k
th
Daubechies filter coefficient, dimensionless
c
k
k
th
Deslauriers-Dubuc filter coefficient,
dimensionless
d interpolation coefficient
d vector of interpolation coefficients in 1D
problems
EA axial stiffness, kN
F force, kN
f load vector
FDM Finite Differences Method
g matrix of boundary conditions
I identity matrix
i wavelet translation
j wavelet translation
k stiffness matrix
k wavelet translation, integer value
M wavelet moments
N wavelet order
P matrix of filter coefficients
R reaction force, kN
u displacement, m
WGM Wavelet-Galerkin Method
Greek symbols
vector of wavelet connection coefficients
vector scaling function evaluations
vector of derivative values
vector of interpolation coefficients in 2D
problems
Kronecker delta
dimensionless coordinate in y direction
scaling function
vector of Lagrange multipliers
variance
dimensionless coordinate in x direction
wavelet function
Subscripts
i index of summation
j index of summation
k index of summation
p particular solution
Superscripts
j wavelet level of resolution
m polynomial order
n order of derivation
INTRODUCTION
The use of wavelet-based numerical schemes
has become increasingly popular in the last three
decades. Wavelet scaling functions have several
properties that are especially useful for representing
solutions of differential equations (DE’s), such as
orthogonality, compact support and a certain number