10th INDO-GERMAN WINTER ACADEMY 2011

Classification of Partial Differential Equations and their solution characteristics

Presented by: Akhilesh Kumawat Indian Institute of Technology Roorkee

Outline
Introduction Classification of PDEs Hyperbolic PDE Parabolic PDE Elliptic PDE Numerical Methods References

Differential Equations
What are Differential Equations?

CLASSIFICATION Dimension of unknown: Ordinary differential equation (ODE) Partial differential equation (PDE) Number of equations: single differential equation system of differential equations (coupled) Order : nth order DE has nth derivative, and no higher

Differential Equations
Classification based on Linearity :

Differential Equations

Linear

Non-linear

Homogeneous

Nonhomogeneous

Significance in Engineering

Laplace s Equation: 2u = uxx + uyy + uzz = 0 unknown: u(x,y,z) gravitational / electrostatic potential Heat Equation: utt = a22u unknown: u(t,x,y,z) heat conduction Wave Equation: utt= a22u unknown: u(t,x,y,z) wave propagation

Significance in Engineering

Schrdinger Wave Equation quantum mechanics (electron probability densities)

Navier-Stokes Equation fluid flow (fluid velocity & pressure)

Introduction

What are Partial Differential Equations? General form of Partial Differential Equation: F(x, y, , u, ux, uy, .., uxx, uyy, uxy,

)=0

Classification

Elliptic Partial Differential Equations

Hyperbolic Partial Differential Equations

Parabolic Partial Differential Equations

Classification
System of coupled equations for several variables: Time : first-derivative (second-derivative for wave equation) Space: first- and second-derivatives General Formula Auxx + Buxy + Cuyy + Dux +Euy + Fu + G = 0 The PDE is Elliptic if B2-4AC <0 Parabolic if B2-4AC=0 Hyperbolic if B2-4AC >0

Well Posedness

A problem is called well posed, if it satisfies all the following criteria : Existence Uniqueness Stability

Boundary and initial conditions

A boundary condition expresses the behavior of a function on the boundary (border) of its area of definition. An initial condition is like a boundary condition, but only for one direction Three kinds of boundary conditions: Dirichlet conditions with u = f on R Neumann conditions with d/dn(u) = f or d/ds(u) = g on R Mixed (Robin) conditions d/dn(u) + ku = f, k > 0, on R.

Classification

Case 1: b2 > ac

hyperbolic

There are two characteristic directions at each point Generally not too hard to solve Solutions to the Cauchy problem tend to be well behaved Prototype model: the wave equation uxx= c2utt + F

Classification

Case 2: b2 < ac

elliptic

There are no real characteristics Solving the Cauchy problem can be much more difficult Solutions can be extremely poorly behaved Prototype model: Laplace s equation uxx+ uyy= 0

Classification

Case 3: b2 = ac

parabolic

There is one (repeated) real characteristic direction Can have features of both hyperbolic and elliptic problems Prototype model: the heat equation ut = uxx + Q

Physical interpretation

Propagation problems lead to parabolic or hyperbolic PDE s. Equilibrium equations lead to elliptic PDE. Most fluid equations with an explicit time dependence are Hyperbolic PDEs For dissipation problem, Parabolic PDEs

Physical problem

Elliptic Partial Differential Equation Consider a thin metal square plate with dimensions 0.5 meters by 0.5 meters. Two adjacent boundaries are held at a constant 0 deg C. The heat on the other two boundaries increases linearly from 0 deg C to 100 deg C. We want to know what the temperature is at each point when the temperature of the metal has reached steady-state.

Physical problem

Parabolic Partial Differential Equation Consider a rod of length l that is perfectly insulated. We want to examine the flow of heat along the rod over time. The two ends of the rod are held at 0 deg C. We denote heat as temperature, and seek the solution to the equation u(x, t): the temperature of x at time t > 0.

Physical problem

Hyperbolic Partial Differential Equation An elastic string is stretched between two horizontal supports that are separated by length l. The PDE defines vertical displacement of the string at time t. (We are assuming that when we pluck the string, it only vibrates in one direction.)

Hyperbolic PDEs

Hyperbolic equations retain any discontinuities of functions or derivatives in the initial data If a disturbance is made in the initial data of a hyperbolic differential equation, then not every point of space feels the disturbance at once. Relative to a fixed time coordinate, disturbances have a finite propagation speed. They travel along the characteristics of the equation

Hyperbolic PDEs
Region of influence: Part of domain, between the characteristic curves, from point P to away from the initial data line Region of dependence: Part of domain, between the characteristic curves, from the initial data line to the point P

Hyperbolic PDEs
Examples: i. uxx =a2uxx wave equation (linear wave equation)

ii. u =a2u + f(x, t) non-homogeneous wave equation iii. utt=a2uxx bu Klein-Gordon equation iv. utt=a2uxx bu + f(x, t) non-homogeneous Klein-Gordon equation

Hyperbolic PDEs

v. utt =a2 (urr + (1/r) ur) + g(r, t) nonhomogeneous wave equation with axial symmetry vi. u =a2 (u + (2/r)u ) + g(r, t) non- nonhomogeneous wave equation with central symmetry vii. utt + kut=a2uxx + bw Telegraph equation

Hyperbolic PDEs

DAlemberts solution Introduce new independent variables: y=x + at, z=x at Substituting these in (4), we get uyz=0 (5) Integrating (5) w.r.t. z, we get uy=f(y) (6) Integrating (6) w.r.t. y, we obtainu = (y) + (z), where (y) = f(y)dy

Hyperbolic PDEs

Thus, u(x, t)= (x + at) + (x at) (7) is the general solution of (4) Now suppose, u(x, 0)=g(x) and ut(x, 0)=0, then (7) takes the form : u(x, t)=g(x + at) + g(x - at) which is the d Alembert s solution of the wave equation (4)

Hyperbolic PDEs

Hyperbolic PDEs

Elliptic PDEs

Solutions of elliptic PDEs are as smooth as the coefficients allow, within the interior of the region where the equation and solutions are defined. For example, solutions of Laplace's equation are analytic within the domain where they are defined, but solutions may assume boundary values that are not smooth. Region of Influence: Entire domain Region of Dependence: Entire domain

Elliptic PDEs

1. Laplace equation 2 u = 0. This is a very common and important equation that occurs in the studies of (a) electromagnetism including electrostatics, dielectrics, steady currents, and magneto statics; (b) hydrodynamics (irrotatinal flow of perfect fluid and surface waves); (c) heat flow; (d) gravitation.

Elliptic PDEs

2 . Poisson Equation: u + =0 The two dimensional Poisson equation has the following form: uxx + uyy + f(x, y) = 0 (Cartesian coordinate system) (1/r)(rur) r +(1/r2) uvv + g(r, v) = 0 (polar coordinate system) Poisson s equation is a partial differential equation with broad utility in electrostatics, mechanical engineering and theoretical physics. E.g. In electrostatics: V=- /

Elliptic PDEs
3.Helmholtz Equation: The wave (Helmholtz) and time-independent diffusion equation is of the form uxx + uyy + u = -f(x, y) (Cartesian form) (1/r)(rur) t+(1/r2) uvv + u=-g(r, v) (Cylindrical form) These equations appear in such diverse phenomena as (a) elastic waves in solids including vibrating strings, bars, membranes; (b) sound or acoustics; (c) electromagnetics waves; (d) nuclear reactors.

Elliptic PDEs
Analytic functions: The real and imaginary parts of a complex analytic function both satisfy the Laplace equation. If f(x + i y)=u(x, y) + i v(x, y) is an analytic function, then uxx + uyy =0, vxx + vyy =0 The close connection between the Laplace equation and analytic functions implies that any solution of the Laplace equation has derivatives of all orders, and can be expanded in a power series, at least inside a circle that does not enclose a singularity.

Parabolic PDEs

Equations that are parabolic at every point can be transformed into a form analogous to the heat equation by a change of independent variables. Solutions smooth out as the transformed time variable increases Region of influence: Part of domain away from initial data line from the characteristic curve Region of dependence: Part of domain from the initial data line to the characteristic curve

Parabolic PDEs
Examples: i. ut=auxx heat equation (linear heat equation) ii. ut=auxx + f(x, t) non-homogeneous heat Equation iii. ut=auxx + bux+ cu + f(x, t) convective heat equation with a source iv. ut=a(urr + (1/r) ur) heat equation with axial symmetry

Parabolic PDEs

v. ut=a(urr + (1/r) ur) + g(r, t) heat equation with axial symmetry (with a source) vi. ut=a(urr + (2/r) ur) heat equation with central Symmetry vii. ut=a(urr + (2/r) ur) + g(r, t) heat equation with central symmetry (with a source) viii.i ut=-( 2/2m) uxx + h(x)u Schrodinger equation (linear schrodinger equation)

Parabolic PDEs
Heat equation: ut=a u The maximum value of u is either earlier in time than the region of concern or on the edge of the region of concern. even if u has a discontinuity at an initial time t = t , the temperature becomes smooth as soon as t > t0. For example, if a bar of metal has temperature 0 and another has temperature 100 and they are stuck together end to end, then very quickly the temperature at the point of connection is 50 and the graph of the temperature is smoothly running from 0 to 100.

Parabolic PDEs
Solving by separation of variables: Let u(x, t)=X(x)T(t) Substituting this in heat equation, we get X /X = T /aT = k i. k=c2: X=c1epx + c2e-px, T=c2ea2pt ii. k=-p2: X=c4cos px + c5sin px, T=c6e-a2tp iii. k=0: X=c7x + c8, T=c9 Thus, various possible solutions are: u=(c1epx + c2e-px)(c3eap2t) u=(c4cos px + c5sin px)(c6e-a2tp) u=(c7x + c8)c9

Numerical Methods

Objective: Speed, Accuracy at minimum cost Numerical Accuracy (error analysis) Numerical Stability (stability analysis) Numerical Efficiency (minimize cost) Validation (model/prototype data, field data, analytic solution, theory, asymptotic solution Reliability and Flexibility (reduce preparation and debugging time) Flow Visualization (graphics and animations)

Numerical Methods

Governing Equations ICS/BCS

Discretization

System of Algebraic Equations

Equation (Matrix) Solver

Approximate Solution

Continuous Solutions

Finite Difference Finite-Volume Finite-Element Spectral Boundary Element Hybrid

Discrete Nodal Values

Tridiagonal ADI SOR GaussSeidel Krylov Multigrid DAE

Ui (x,y,z,t) p (x,y,z,t) T (x,y,z,t)

Discretization
 Time derivatives Almost exclusively by finite-difference methods  Spatial derivatives Finite-difference: Taylor-series expansion Finite-element: low-order shape function and interpolation function, continuous within each element Finite-volume: integral form of PDE in each control volume There are also other methods, e.g. collocation, Spectral method, spectral element, panel Method, boundary element method

References References

Numerical Methods for Elliptic and Parabolic Partial Differential Equations by Peter Knabner Lutz Angermann An Introduction to partial Differential Equations by Yehuda Pinchover and Jacob Rubinstein Last years lectures Partial Differential Equations by Kreyzig