Scribd
Upload
76 views3 pages

Fem Project 3

This document summarizes the θ-scheme for solving the heat equation using the finite element method. It presents: 1) The model heat equation problem and its exact solution. 2) The full discretization θ-method, which includes the forward Euler, backward Euler, and Crank-Nicolson schemes depending on the value of θ. 3) Numerical results applying the implicit backward Euler and explicit forward Euler schemes to a test problem, showing the backward Euler scheme has first-order convergence while the forward Euler scheme diverges.

Uploaded by

anon_269179442
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
76 views3 pages

Fem Project 3

This document summarizes the θ-scheme for solving the heat equation using the finite element method. It presents: 1) The model heat equation problem and its exact solution. 2) The full discretization θ-method, which includes the forward Euler, backward Euler, and Crank-Nicolson schemes depending on the value of θ. 3) Numerical results applying the implicit backward Euler and explicit forward Euler schemes to a test problem, showing the backward Euler scheme has first-order convergence while the forward Euler scheme diverges.

Uploaded by

anon_269179442
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

θ-Scheme of Finite Element Method for Heat Equation ∗

Wenqiang Feng †

Abstract
This is my MATH 574 course project report. In this report, I give some details for implement-
ing the Finite Element Method (FEM) via Matlab and Python with FEniCs. This project mainly
focuses on θ-Method for the initial boundary heat equation.

1 The Model Problem


Consider the initial boundary value problem for the heat equation:



 ut − ∆u = f (x, t), (x, t) ∈ Ω × (0, T ]

u = 0, (x, t) ∈ ∂Ω × (0, T ] (1)



= u 0 (x), x ∈ Ω,

u(·, 0)

where Ω = [0, 1] × [0, 1]. The initial data u 0 (x) and the function f (x, t) are such that the solution
is given by
u(x, y, t) = e−t sin(πx) sin(πy).

2 A Full Discretization scheme: The θ-Method


 n+1 n 
u −u 
n+1 n
  
 h τ h , vh + a θuh + (1 − θ)uh , vh = θf (t n+1 ) + (1 − θ)f (t n ), vh



 (2)
u 0 = Πu 0 .


h

So, we get
     
uhn+1 − uhn , vh + τa θuhn+1 + (1 − θ)uhn , vh = τ θf (t n+1 ) + (1 − θ)f (t n ), vh .
Since
N
X
uhn = αjn φj (x),
j=1

then
   
X N XN   XN XN 
αjn+1 φj (x) − αjn φj (x), φi (x) + τa θ αjn+1 φj (x) + (1 − θ) αjn+1 φj (x), φi (x)
   

   
j=1 j=1 j=1 j=1
 
= τ θf (t n+1 ) + (1 − θ)f (t n ), φi (x) , i = 1, 2, · · · , N .
∗ Key words: FEM, Pure Dirichlet boundary condition, Heat equation.
† Department of Mathematics,University of Tennessee, Knoxville, TN, 37909, wfeng@[Link]..edu

1
Therefore,

Mα n+1 − Mα n + θτSα n+1 + (1 − θ)τSα n = θτF n+1 + (1 − θ)τF n


where
N 
X  N 
X 
M= φj (x), φi (x) , M = ∇φj (x), ∇φi (x) , Fin = (f (t n ), φi (x)) .
j=1 j=1

Hence
(M + θτS) α n+1 = (M − (1 − θ)τS) α n + θτF n+1 + (1 − θ)τF n .
1. When θ = 0, it is first-order explicit/forward Euler Scheme.
2. When θ = 1, it is first-order implicit/backward Euler Scheme.
3. When θ = 1/2, it is second-order Crank-Nicolson/ Trapezoidal Scheme.

3 Numerical Results
We use the the following test problem which has the exact solution

u(x, y, t) = e−t sin(πx) sin(πy).

So,
u(x, y, 0) = sin(πx) sin(πy),
and
f (x, y, t) = (2π2 − 1)e−t sin(πx) sin(πy).

Figure 1: The initial data(left) and the exact solution at T = 1(right) .

2
Table 1: Errors of the computed solution (nx=ny=512)with implicit/backward Euler Scheme.

#T step ku − uh kL∞ order


q=1 20 4.9128E-04
40 2.3917E-04 1.0271
80 1.1481E-04 1.0416

Figure 2: The numerical solution at T = 1 with implicit/backward Euler Scheme.

Table 2: Errors of the computed solution (nx = ny = 64)with explicit/forward Euler Scheme.

#T step ku − uh kL∞ order


q=1 20 8.6611E+67 divergence
40 5.9872E+129 divergence
80 1.0251E+242 divergence
80 1.0251E+242 divergence
160 4.5052E-17 convergence

You might also like