UNIT- II

MATHEMATICAL BEHAVIOR OF PARTIAL

DIFFERENTIAL EQUATIONS AND THEIR IMPACT
ON COMPUTATIONAL AERODYNAMICS

Department of Aerospace Engineering, SOET-SUN, Nashik 2

 Syllabus
Classification of quasilinear partial differential equations by Cramer’s
rule and eigen value method. General behaviour of different classes of
partial differential equations and their importance in understanding
physical and CFD aspects of aerodynamic problems at different Mach
numbers involving hyperbolic, parabolic and elliptic equations- domain
of dependence and range of influence for hyperbolic equations. Well-
posed problems.

Department of Aerospace Engineering, SOET-SUN, Nashik 3

 Differential Equations
 A differential equation is an equation for an unknown function of one or several
variables that relates the values of the function itself and of its derivatives of
various orders.

 Ordinary Differential Equation: Function has 1 independent variable.

 Partial Differential Equation: At least 2 independent variables.

Department of Aerospace Engineering, SOET-SUN, Nashik 4

Classification of Partial Differential Equations
Classification
of
Partial Differential Equations

Technique 1: Technique 2:
Cramer’s Rule Eigenvalue Method

Hyperbolic Parabolic Elliptic Mixed

Equations Equations Equations Types

Two types of physical problems (i.e., equilibrium and propagation problems)

Department of Aerospace Engineering, SOET-SUN, Nashik 5

 PDEs definitions
 General (implicit) form for one function u(x,y) :

 Highest derivative defines order of PDE

 Explicit PDE => We can resolve the equation to the highest derivative of u.
 Linear PDE => PDE is linear in u(x,y) and for all derivatives of u(x,y)
 Semi-linear PDEs are nonlinear PDEs, which are linear in the highest order
derivative.

Department of Aerospace Engineering, SOET-SUN, Nashik 6

 Examples of PDEs
• Examples of three simple partial differential equations having two independent
variables are

Two-dimensional Laplace equation

One-dimensional diffusion equation

One-dimensional wave equation

Department of Aerospace Engineering, SOET-SUN, Nashik 7

 Classification of quasilinear partial differential equations by
Cramer’s rule

• Consider the system of quasi linear equations

Department of Aerospace Engineering, SOET-SUN, Nashik 8

Contd…

To find characteristics of
these lines we call these
u and v as functions of x
y
and y.

x
Fig. Illustration of characteristic curve

Department of Aerospace Engineering, SOET-SUN, Nashik 9

Department of Aerospace Engineering, SOET-SUN, Nashik 10
Department of Aerospace Engineering, SOET-SUN, Nashik 11
Department of Aerospace Engineering, SOET-SUN, Nashik 12
Department of Aerospace Engineering, SOET-SUN, Nashik 13
Department of Aerospace Engineering, SOET-SUN, Nashik 14
 Classifications
 Classification of partial differential equations
 A general partial differential equation in coordinates x and y:Characterization
depends on the roots of the higher order (here Second order) terms:
➢ b2-4ac> 0 then the equation is called hyperbolic.
➢ b2-4ac = 0 then the equation is called parabolic.
➢ b2-4ac< 0 then the equation is called elliptic.
 Note: if a, b, and c themselves depend on x and y, the equations may be of
different type, depending on the location in x-y space.

Department of Aerospace Engineering, SOET-SUN, Nashik 15

 Linear PDEs of 2. Order

a(x,y)c(x,y) − b(x,y)2 / 4 > 0 Elliptic

a(x,y)c(x,y) − b(x,y)2 / 4 = 0 Parabolic
a(x,y)c(x,y) − b(x,y)2 / 4 < 0 Hyperbolic

Quasi-linear: coefficients depend on u and/or first derivative of u, but NOT on second

derivatives.

Department of Aerospace Engineering, SOET-SUN, Nashik 17

 PDEs and Quadratic Equations

 Quadratic equations in the form describe cone sections.

a(x,y)c(x,y) − b(x,y)2 / 4 > 0 Ellipse

a(x,y)c(x,y) − b(x,y)2 / 4 = 0 Parabola
a(x,y)c(x,y) − b(x,y)2 / 4 < 0 Hyperbola

Department of Aerospace Engineering, SOET-SUN, Nashik 18

 PDEs and Quadratic Equations
With coordinate transformations these equations can be
written in the standard forms:

Ellipse:

Parabola:

Hyperbola:

Coordinate transformations can be also applied to get rid of the

mixed derivatives in PDEs.
(For space dependent coefficients this is only possible locally, not
globally)

Department of Aerospace Engineering, SOET-SUN, Nashik 19

PDEs and Quadratic Equations

Department of Aerospace Engineering, SOET-SUN, Nashik 20

 Second Order PDEs with more then 2 independent variables

Classification by eigen values of the coefficient matrix:

 Elliptic: All eigen values have the same sign. [Laplace-Eq.]

 Parabolic: One eigen value is zero. [Diffusion-Eq.]
 Hyperbolic: One eigen value has opposite sign. [Wave-Eq.]
 Ultra hyperbolic: More than one positive and negative eigenvalue.

Mixed types are possible for non-constant coefficients, appear frequently in

science and are often difficult to solve.

Department of Aerospace Engineering, SOET-SUN, Nashik 21

Classification of quasilinear partial differential equations by
Eigen value method

Department of Aerospace Engineering, SOET-SUN, Nashik 22

Department of Aerospace Engineering, SOET-SUN, Nashik 23
Example

Department of Aerospace Engineering, SOET-SUN, Nashik 24

Department of Aerospace Engineering, SOET-SUN, Nashik 25
Applying Eigen Value Method

• Find [K]-1

Department of Aerospace Engineering, SOET-SUN, Nashik 26

Department of Aerospace Engineering, SOET-SUN, Nashik 27
If B2-4AC > 0 → Hyperbolic PDE , Two real distinct characteristics exist at each point in
x-y plane
B2-4AC = 0 → Parabolic PDE , one real characteristic
B2-4AC < 0 → Elliptic PDE , characteristics are imaginary

dy B  B2 − 4AC
= See Tannehill et.al. 1997, page 24 for derivation
dx 2A
Characteristic lines are related to directions in which “information” can be transmitted
in physical problems governed by PDEs.

Department of Aerospace Engineering, SOET-SUN, Nashik 28

HYPERBOLIC PROBLEMS

Department of Aerospace Engineering, SOET-SUN, Nashik 29

 Hyperbolic Problems
Hyperbolic equations are typical of marching problems with negligible dissipation.
An example is the wave equation:
This describes the transverse displacement of a string during small amplitude
vibrations. If y is the displacement, x the coordinate along the string, and a the initial
amplitude, the solution is: Note that the amplitude is independent of time, i.e. there
is no dissipation. Hyperbolic problems can have discontinuous solutions.
 Disturbances may affect only a limited region in space. This is called the zone of
influence. Disturbances propagate at the wave speed c.
 Local solutions may only depend on initial conditions in the domain of
dependence.

Department of Aerospace Engineering, SOET-SUN, Nashik 30

Characteristic lines are related to directions in which “information” can be transmitted
in physical problems governed by PDEs.
Hyperbolic PDEs with two independent variable x & y
Domain of influence: region influenced by P
y
Right-running
β2
II β3
Left-running
P
α2 α3
I III

x
a b c Initial data along the x axis
upon which P depends
Domain of dependence for P
(boundary conditions)

Department of Aerospace Engineering, SOET-SUN, Nashik 31

α = cont. & β=const. lines represent the two families of characteristics along which
signals can propagate

•Observer at point P can feel the effects of what has happened in Region I. The domain of
dependence region. Outside Region I, disturbance cannot be felt by P.
•Disturbance created at point P can be felt only in the Region II, i.e. Region II is the
domain of influence of point P.

Hyperbolic eqs. domains extend to infinity in the time like coordinate

•Solution can be obtained by “marching forward” in the distance y, starting from the given
boundary
• Spatial coordinate may or may not be bounded
• Normally associated with initial value problems
•Typically two initial conditions at t=0 are specified
• If the spatial region is bounded → boundary conditions

Department of Aerospace Engineering, SOET-SUN, Nashik 32

Steady Inviscid Supersonic Flow Problem

Fig. Illustration of initial lines for Method of Characteristics

Department of Aerospace Engineering, SOET-SUN, Nashik 35

 Three dimensional steady flows

Fig. Domain and boundary for the solution of hyperbolic equations, three dimensional steady flows

Department of Aerospace Engineering, SOET-SUN, Nashik 36

 Unsteady Inviscid Flows

• Fig. Domain and boundaries for the solution of Hyperbolic equations. One dimensional Unsteady flow

Department of Aerospace Engineering, SOET-SUN, Nashik 37

 Unsteady Inviscid Flows

Fig. Domain and boundaries for the solution of Hyperbolic equations. Two dimensional Unsteady flow

Department of Aerospace Engineering, SOET-SUN, Nashik 38

When do we encounter Unsteady Inviscid Flows?

Department of Aerospace Engineering, SOET-SUN, Nashik 39

 PARABOLIC PROBLEMS

 Parabolic equations describe marching problems. This includes time

dependent problems which involve significant amounts of dissipation.
Examples are unsteady viscous flows and unsteady heat conduction. Steady
viscous boundary layer flow is also parabolic (march along streamline, not in
time).
 An example is the transient temperature distribution in a cooling down rod:
The temperature depends on both the initial and boundary conditions. This
is also called an initial- boundary-value problem.
 Disturbances can only affect solutions at a later time.
 Dissipation ensures that the solution is always smooth.

Department of Aerospace Engineering, SOET-SUN, Nashik 41

 PARABOLIC PROBLEMS

Department of Aerospace Engineering, SOET-SUN, Nashik 42

Fig. Domain and boundaries for the solution of parabolic equations in three dimensions

Department of Aerospace Engineering, SOET-SUN, Nashik 43

What type of fluid flow types are governed by the parabolic equations?

https://www.semanticscholar.org/paper/Flow-Simulation-over-
Re-Entry-Bodies-at-Supersonic-
Prasad/989a07078e008e8d57031e7d4d8a491fb757160d

Department of Aerospace Engineering, SOET-SUN, Nashik 44

Steady Boundary Layer Flows

Fig. Schematic of a boundary Layer Flow

Department of Aerospace Engineering, SOET-SUN, Nashik 45

Parabolized Viscous Flows

What happens when the boundary layer is no thin, indeed, when the entire flow field of interest is fully
viscous?

Fig. Schematic of a totally viscous shock layer

Department of Aerospace Engineering, SOET-SUN, Nashik 46

 Parabolized Naiver Stokes Equation

Department of Aerospace Engineering, SOET-SUN, Nashik 47

Unsteady Thermal Conduction

∞
Department of Aerospace Engineering, SOET-SUN, Nashik 50
Unsteady Thermal Conduction

If there is no volumetric heat addition then 𝑞=0

ሶ

Department of Aerospace Engineering, SOET-SUN, Nashik 51

Department of Aerospace Engineering, SOET-SUN, Nashik 53
 ELLIPTIC PROBLEMS

 Elliptic equations are characteristic of equilibrium problems, this includes many

(but not all) steady state flows.

 Examples are potential flow, the steady state temperature distribution

in a rod of Solid material and equilibrium stress distributions in solid objects
under applied Loads.

Department of Aerospace Engineering, SOET-SUN, Nashik 54

 For potential flows the velocity is expressed in terms of a velocity potential:
u=∇φ. Because the flow is incompressible, ∇.u=0, which results in ∇2φ=0. This is
also known as Laplace’s equation

 The solution depends solely on the boundary conditions. This is also known as a
boundary value problem.
 A disturbance in the interior of the solution affects the solution everywhere
else. The disturbance signals travel in all directions.
 As a result, solutions are always smooth, even when boundary conditions are
Discontinuous. This makes numerical solution easier.

Department of Aerospace Engineering, SOET-SUN, Nashik 55

What type of flows are governed by the Elliptic equations?

1. Steady subsonic Viscous Flows

2. Incompressible Inviscid Flows

Department of Aerospace Engineering, SOET-SUN, Nashik 56

• Steady subsonic Viscous Flows

Department of Aerospace Engineering, SOET-SUN, Nashik 57

• Incompressible Inviscid Flows

Department of Aerospace Engineering, SOET-SUN, Nashik 58

• Classifications of Fluid flow equations

Department of Aerospace Engineering, SOET-SUN, Nashik 59

Well posed problems

A problem is well posed if:

 A solution exists.
 The solution is unique.
 The solution depends continuously on the data
(boundary and/or initial conditions).
 Problems which do not fulfill these criteria
are ill-posed.

Well posed problems have a good chance to be solved

1865-1963
numerically with a stable algorithm.

Department of Aerospace Engineering, SOET-SUN, Nashik 60

 Ill-posed problems
 Ill-posed problems play an important role in some areas, for example for inverse
problems like tomography.

 Problem needs to be reformulated for numerical treatment. Add additional

constraints, for example smoothness of the solution.

 Input data need to be regularized / preprocessed.

Department of Aerospace Engineering, SOET-SUN, Nashik 61

 Ill-conditioned problems
 Even well posed problems can be ill-conditioned.
 Small changes (errors, noise) in data lead to large errors in the solution.
 Can occur if continuous problems are solved approximately on a numerical
grid.
 PDE => algebraic equation in form Ax = b
 Condition number of matrix A:

are maximal and minimal Eigen values of A.

 Well conditioned problems have a low condition number.

Department of Aerospace Engineering, SOET-SUN, Nashik 62

 General Features of Partial Differential Equations

Type of Physical Problem

Equilibrium Propagation
Mathematical Elliptic Parabolic Hyperbolic
Classification of PDE
Characteristics Complex Real and Repeated Real and Distinct
Information Undefined Infinite Finite
Propagation speed
Domain of Dependence Entire Solution domain Present and Entire past Past solution domain
solution domain between characteristics
Range of Influence Entire Solution Present and Entire future Future solution domain
solution domain between characteristics
Type of Numerical Relaxation Marching Marching
Method

Department of Aerospace Engineering, SOET-SUN, Nashik 63

Example: The blunt-nosed body

Department of Aerospace Engineering, SOET-SUN, Nashik 64

 Example: The blunt-nosed body

• Blunt-nosed body designs are used for supersonic and hypersonic speeds (e.g.
Apollo capsules and space shuttle) because they are less susceptible to
aerodynamic heating than sharp nosed bodies.
• There is a strong, curved bow shock wave, detached from the nose by the shock
detachment distance δ.
• Calculating this flow field was a major challenge during the 1950s and 1960s
because of the difficulties involved in solving for a flow field that is elliptic in one
region and hyperbolic in others.
• Today’s CFD solvers can routinely handle such problems, provided that the flow is
calculated as being transient.

Department of Aerospace Engineering, SOET-SUN, Nashik 65

• Fluid flows can be classified in a variety of ways:
–Internal vs. external.
–Laminar vs. turbulent.
–Compressible vs. incompressible.
–Steady vs. unsteady.
–Supersonic vs. transonic vs. subsonic.
–Single-phase vs. multiphase.
–Elliptic vs. parabolic vs. hyperbolic.

Department of Aerospace Engineering, SOET-SUN, Nashik 66

 How to solve PDEs?
PDEs are solved together with appropriate Boundary conditions
and/or Initial Conditions.
Boundary value problem
Dirichlet B.C:
Specify u(x,y,...) on boundaries (say at x=0, x=Lx, y=0, y=Ly in a
rectangular box)
Von Neumann B.C:
Specify normal gradient of u(x,y,...) on boundaries.
In principle boundary can be arbitrary shaped.(but difficult to implement
in computer codes)

101
Department of Aerospace Engineering, SOET-SUN, Nashik 67
Boundary value problem

Department of Aerospace Engineering, SOET-SUN, Nashik 68

 Initial value problem
 Boundary values are usually specified
on all boundaries of the computational
domain.

 Initial conditions are specified in the

entire computational (spatial) domain,
but only for the initial time t=0.

 Initial conditions as a Cauchy problem: -

Specify initial distribution u(x,y,...,t=0)
[for parabolic problems like the Heat
equation]
- Specify u and du/dt for t=0 [for
hyperbolic problems like wave equation.]

Department of Aerospace Engineering, SOET-SUN, Nashik 69

 Boundary Conditions

Department of Aerospace Engineering, SOET-SUN, Nashik 70