Partial Differential Equations

Slides:



Advertisements
Similar presentations
Courant and all that Consistency, Convergence Stability Numerical Dispersion Computational grids and numerical anisotropy The goal of this lecture is to.
Advertisements

One-phase Solidification Problem: FEM approach via Parabolic Variational Inequalities Ali Etaati May 14 th 2008.
Computational Modeling for Engineering MECN 6040
Chapter 8 Elliptic Equation.
Parabolic Partial Differential Equations
Numerical Computation
CSCI 317 Mike Heroux1 Sparse Matrix Computations CSCI 317 Mike Heroux.
PART 7 Ordinary Differential Equations ODEs
CS 584. Review n Systems of equations and finite element methods are related.
PARTIAL DIFFERENTIAL EQUATIONS
CHE/ME 109 Heat Transfer in Electronics LECTURE 11 – ONE DIMENSIONAL NUMERICAL MODELS.
NAE C_S2001 FINAL PROJECT PRESENTATION LASIC ISMAR 04/12/01 INSTRUCTOR: PROF. GUTIERREZ.
Lecture 9 Interpolation and Splines. Lingo Interpolation – filling in gaps in data Find a function f(x) that 1) goes through all your data points 2) does.
Ordinary Differential Equations (ODEs) 1Daniel Baur / Numerical Methods for Chemical Engineers / Implicit ODE Solvers Daniel Baur ETH Zurich, Institut.
Boundary-value Problems and Finite-difference Equations Douglas Wilhelm Harder, M.Math. LEL Department of Electrical and Computer Engineering University.
1 Chapter 9 NUMERICAL SOLUTION OF PARTIAL DIFFERENTIAL EQUATIONS.
Partial differential equations Function depends on two or more independent variables This is a very simple one - there are many more complicated ones.
1 Systems of Linear Equations Error Analysis and System Condition.
CISE301: Numerical Methods Topic 9 Partial Differential Equations (PDEs) Lectures KFUPM (Term 101) Section 04 Read & CISE301_Topic9.
SE301: Numerical Methods Topic 9 Partial Differential Equations (PDEs) Lectures KFUPM Read & CISE301_Topic9 KFUPM.
Numerical Solutions to ODEs Nancy Griffeth January 14, 2014 Funding for this workshop was provided by the program “Computational Modeling and Analysis.
Scientific Computing Partial Differential Equations Explicit Solution of Wave Equation.
COMPUTATIONAL MODELING FOR ENGINEERING MECN 6040 Professor: Dr. Omar E. Meza Castillo Department.
Scientific Computing Partial Differential Equations Introduction and
Hyperbolic PDEs Numerical Methods for PDEs Spring 2007 Jim E. Jones.
Simpson Rule For Integration.
Scientific Computing Partial Differential Equations Poisson Equation.
CS 219: Sparse Matrix Algorithms
1 EEE 431 Computational Methods in Electrodynamics Lecture 4 By Dr. Rasime Uyguroglu
1 ELEC 3105 Basic EM and Power Engineering Start Solutions to Poisson’s and/or Laplace’s.
Scientific Computing Partial Differential Equations Implicit Solution of Heat Equation.
To Dream the Impossible Scheme Part 1 Approximating Derivatives on Non-Uniform, Skewed and Random Grid Schemes Part 2 Applying Rectangular Finite Difference.
Programming assignment #2 Solving a parabolic PDE using finite differences Numerical Methods for PDEs Spring 2007 Jim E. Jones.
Solution of a System of ODEs with POLYMATH and MATLAB, Boundary Value Iterations with MATLAB For a system of n simultaneous first-order ODEs: where x is.
+ Numerical Integration Techniques A Brief Introduction By Kai Zhao January, 2011.
Elliptic PDEs and the Finite Difference Method
Solving an elliptic PDE using finite differences Numerical Methods for PDEs Spring 2007 Jim E. Jones.
Boundary Value Problems l Up to this point we have solved differential equations that have all of their initial conditions specified. l There is another.
Numerical Methods.
Engineering Analysis – Computational Fluid Dynamics –
Engineering Analysis – Computational Fluid Dynamics –
Colorado Center for Astrodynamics Research The University of Colorado 1 STATISTICAL ORBIT DETERMINATION The Minimum Variance Estimate ASEN 5070 LECTURE.
Parallel and Domain Decomposed Marching Method for Poisson Equation 大氣五 b 簡睦樺.
MECN 3500 Inter - Bayamon Lecture 9 Numerical Methods for Engineering MECN 3500 Professor: Dr. Omar E. Meza Castillo
Lecture 21 MA471 Fall 03. Recall Jacobi Smoothing We recall that the relaxed Jacobi scheme: Smooths out the highest frequency modes fastest.
Copyright © 2006 The McGraw-Hill Companies, Inc. Permission required for reproduction or display. by Lale Yurttas, Texas A&M University Chapter 271 Boundary-Value.
Copyright © The McGraw-Hill Companies, Inc. Permission required for reproduction or display. 1 Chapter 27.
Island Recharge Problem
Clicker Question 1 Suppose a population of gerbils starts with 20 individuals and grows at an initial rate of 6% per month. If the maximum capacity is.
Approximating Derivatives Using Taylor Series and Vandermonde Matrices.
Marching Solver for Poisson Equation 大氣四 簡睦樺. Outline A brief review for Poisson equation and marching method Parallel algorithm and consideration for.
Solving Partial Differential Equation Numerically Pertemuan 13 Matakuliah: S0262-Analisis Numerik Tahun: 2010.
MA2213 Lecture 9 Nonlinear Systems. Midterm Test Results.
EEE 431 Computational Methods in Electrodynamics
Finite Difference Methods
Chapter 30.
Scientific Computing Lab
Scientific Computing Lab
Lecture 19 MA471 Fall 2003.
ME/AE 339 Computational Fluid Dynamics K. M. Isaac Topic2_PDE
Chapter 27.
Numerical Solutions of Partial Differential Equations
Scientific Computing Partial Differential Equations Implicit Solution of Heat Equation.
CS6068 Applications: Numerical Methods
6.5 Taylor Series Linearization
topic13_grid_generation
Discrete Least Squares Approximation
Numerical Methods on Partial Differential Equation
Programming assignment #1 Solving an elliptic PDE using finite differences Numerical Methods for PDEs Spring 2007 Jim E. Jones.
Derivatives and Gradients
Presentation transcript:

Partial Differential Equations Scientific Computing Partial Differential Equations Explicit Solution of Heat Equation

One Dimensional Heat Equation The one-dimensional Heat Equation for the temperature u(x,t) in a rod at time t is given by: where c >0, T>0 are constants

One Dimensional Heat Equation We will solve this equation for x and t values on a grid in x-t space: Approximate Solution uij=u(xi, tj) at grid points

Finite Difference Approximation To approximate the solution we use the finite difference approximation for the two derivatives in the heat equation. We use the forward-difference formula for the time derivative and the centered-difference formula for the space derivative:

Finite Difference Approximation Then the heat equation (ut =cuxx ) can be approximated as Or, Let r = (ck/h2) Solving for ui,j+1 we get:

One Dimensional Heat Equation Note how this formula uses info from the j-th time step to approximate the (j+1)-st time step:

One Dimensional Heat Equation On the left edge of this grid, u0,j = g0,j = g0(tj). On the right edge of this grid, un,j = g1,j = g1(tj). On the bottom row of the grid, ui,0 = fi = f(xi). Thus, the algorithm for finding the (explicit) solution to the heat equation is:

Matlab Implementation function z = simpleHeat(f, g0, g1, T, n, m, c) %Simple Explicit solution of heat equation h = 1/n; k = T/m; r = c*k/h^2; % Constants x = 0:h:1; t = 0:k:T; % x and t vectors % Boundary conditions u(1:n+1, 1) = f(x)'; % Transpose, since it’s a row vector u(1, 1:m+1) = g0(t); u(n+1, 1:m+1) = g1(t); % compute solution forward in time for j = 1:m u(2:n,j+1) = r*u(1:n-1,j) + (1-2*r)*u(2:n,j) + r*u(3:n+1,j); end z=u'; mesh(x,t,z); % plot solution in 3-d

Matlab Implementation Usage: f = inline(‘x.^4’); g0 = inline(‘0*t’); g1 = inline(‘t.^0’); n=5; m=5; c=1; T=0.1; z = simpleHeat(f, g0, g1, T, n, m, c);

Matlab Implementation Calculated solution appears correct:

Matlab Implementation Try different T value: T=0.5 Values seem chaotic

Matlab Implementation Why this chaotic behavior?

Matrix Form of Solution

Matrix Form of Solution This is of the form Start: u(0)=u(:,0)=f(:) Iterate:

Matrix Form of Solution Now, suppose that we had an error in the initial value for u(0), say the initial u value was u(0)+e, where e is a small error. Then, under the iteration, the error will grow like Ame. For the error to stay bounded, we must have Thus, the largest eigenvalue of A must be <= 1.

Gerschgorin Theorem Let A be a square nxn matrix. Around every element aii on the diagonal of the matrix draw a circle with radius Such circles are known as Gerschgorin disks. Theorem: Every eigenvalue of A lies in one of these Gerschgorin disks.

Gerschgorin Theorem Example: The circles that bound the Eigenvalues are: C1: Center point (4,0) with radius r1 = |2|+|3|=5 C2: Center point (-5,0) with radius r2=|-2|+|8|=10 C3: Center Point (3,0) with radius r3=|1|+|0|=1

Gerschgorin Theorem Example: Actual eigenvalues in red

Gerschgorin Theorem Example: C1: Center point (1,0) radius r1 = |0|+|7|=7 C2: Center point (-5,0) radius r2=|2|+|0|=2 C3: Center Point (-3,0) radius r3=|4|+|4|=8

Matrix Form of Solution Circles: center=(1-2r), radius = r or 2r. Take largest radius 2r. Then, if λ is an eigenvalue of A, we have 1-2r

Matrix Form of Solution So, the eigenvalue satisfies: For the error to stay bounded, we need Thus, we need For our first example, we had T=0.1 and so, we expect the solution to be stable. For the second example, T=0.5, we have r =2.5, so the error in the iterates will grow in an unbounded fashion, which we could see in the numerical solution.