Institute of Aeronautical

Engineering
Department of Information Technology

PARTIAL DIFFERENTIAL
EQUATIONS

By Veda Sandeep
23951A12G9
 1. Differential Equation
• Let x be independent variable. Derivatives of any function
(say z1), depend upon x is the rate of change z1 as x changes
slightly.
• The equation that contains one or more unknown functions • =2
of single independent variable and their derivatives is
differential equation.
• , (zx , zy)
• When no. of independent variable is 2 or more(say x,y)
such that z(x,y) is function of x and y, then partial
derivative of dependent variable with any one independent
variable keeping other independent variable as constant. • + + z2 = xy
• Ordinary Differential Equations (ODEs) deal with functions
of one independent variable, while Partial Differential
Equations (PDEs) deal with functions of multiple
independent variables.
REMEMBER

Taking z(x,y), as a dependent and x,y as an independent variable,

we will adopt the following notations through out the presentation.

z x= zy = zxx = zxy= zyy=

 Introduction To PDE
• A partial differential equation (PDE) is a mathematical equation that contains
an unknown function (z) of two or more independent variables(x,y), as well as
the partial derivatives of that unknown function(zx, zy, zxx) with respect to those
variables.
• That unknown Function (z) is called dependent variables.
• PDEs can be classified into various types based on their order, linearity, and
coefficients.
Examples:
Order of PDE is order of highest order partial derivative
involved in that PDE

(1st order)

(3rd order)
Degree of PDE is degree of highest order partial derivative involved in that PDE

(1 order)
(1st degree)
st
(2nd degree)

(2nd order) (1st degree) (1st degree)

The concept of degree cannot be attributed to all PDE. For example, the given PDE doesnot have any degree.
sin zx + e^(zy) = 1
 Linearity: Linear and Non Linear PDE:
• LINEAR :- If the dependent variable (z) and its partial derivatives (zx, zy, zxx, zxy, zyy)
occurs in the first power only and are not multiplied.
• NON LINEAR :- Else

EXAMPLES
Linear: Non-Linear:
• zx +zy =0 • zx2+zy2=0
• x.zx+y.zy=z • z.zx+z.zy=z
• zx+zy+zxx+zxy+zyy • zx+zy+zxx+zxy+zyy
=z =z2
First Order PDE (L)
• The general form of first order PDE is of type f(x,y,z,p,q)=0
• It can also be written as:
A.zx+B.zy+C.z = D or A
where is the function of independent
variables and constants

• If D=0, the PDE above becomes Homogeneous.

 Origin Of First Order Partial
By the elimination Differential Equation
of the arbitrary constants By the elimination of arbitrary
from a relation between x, y and z. functions of these variables.

g(x,y,z,a,b)=0 f(u, v) = 0,

Differentiating g wrt. x and y partially, and where u and v are function of x,y,z
from f, fx, and fy Differentiating f wrt. x and y, taking z as dependent
variable
We get equation of the form
We get equation of the form
f(x,y,z,p,q)=0 pP + qQ = R
is required PDE
is required PDE Where P,Q,R are Lagrange’s linear equation.
x2+y2 = (z-c)2 tan2α Algebric z = xy + f(x2+y2) Algebric
xq-yp=0 PDE py-y2 = qx-x2 PDE
 Solution: Linear PDE of the First Order

• The PDE of the type pP + qQ = R, Rules for solving pP + qQ = R

where P, Q, R are functions of (x, y,
• Put the given PDE in the standard form pP + qQ = R.
z), is called a linear PDE of the
first order or Lagrange’s linear • Write down Lagrange’s auxiliary equations = =
equation. • Solve these equations
• Its solution is in the form of F(u, v) • Let u(x, y, z) = c1 and v(x, y, z) = c2 are two
independent solutions.
= 0, where F is arbitrary function
and u(x, y, z) = c1 and v(x, y, z) = c2 • 4. The general solution is then written one of the
equivalent form F(u, v) = 0

• PDE : y2p − xyq = x(z-2y) • Step2: x2+y2 = c1, zy−y2 = c2

• Step1: P= y2, Q= xy, R= x(z-2y) • Step3: F(c1,c2) = F(x2+y2, zy−y2)=0
Second Order PDE (L)
A second order PDE involves second-order partial derivatives of an unknown
function(z) with respect to one or more independent variables.

General Second order Linear PDE (with 1D/2I variable) is:

Azxx + Bzxy + Czyy + Dzx + Ezy + Fz = G ------(i)
when, G=0, equation (i) is homogeneous.

Depending on the coefficients, second-order Solutions (BCs and ICs): Second-order PDEs often require

PDEs can be classified as: boundary conditions for elliptic and hyperbolic equations, while
parabolic equations typically require initial conditions along
Parabolic if B2-4AC = 0
with boundary conditions.
Elliptic if B2-4AC < 0
Hyperbolic if B2-4AC > 0
 Solution method for second order L. PDE
(Direct Integration)
• Direct integration (PDEs) is a method where both sides of a PDE are integrated
with respect to one of the independent variables to eliminate one derivative from
the equation. And the process continues till all Partial derivatives are removed.

• Example:

where f(t) and g(t) are unknown function

using ICs,
Final Sol.n
When the equation is more complex or doesn't lend itself well to direct integration. In
such cases, separation of variables becomes a valuable alternative.

Why Separation of Variable?

• Reduction to ODEs: Separate solution into functions of single variables, simplifying PDE
into manageable ODEs.
• Homogeneous Boundary Conditions: Effective for PDEs with homogeneous boundary
conditions, facilitating solution combination.
• Orthogonality of Solutions: Solutions may form orthogonal sets, easing coefficient
determination for boundary value problems.
• Versatility: Applicable to various linear second-order PDEs, including heat, wave, and
Laplace's equations.
• Physical Interpretation: Provides clear physical interpretations, aiding understanding in
fields like mathematical physics and engineering.
 Solution method for second order L. PDE
(Separation of Variable)
• If we have a second order PDE
ut=αuxx -------------(i)
(where u is dependent variable depend upon x and t)
• To get solution we assume product 2 different function X(x) and T(t) of x and t respectively,
be the solution of PDE above. ie’.
u(x,t)=X(x).T(t) -----(ii)
• Solving (i) and (ii), we get = k (Separation Constant)
• According to the value of k (k <,=,> 0) we can found 3 pair of distinct solution for X and T
and final Solution will be u(x,t)=X(x).T(t), where, 3 different solutions can be found.
• Finally, using initial and boundary condition, we can found one out of 3 as a non trivial
solution.
 Three Different solution for above equation(ut=αuxx )
is found as:
1. u(x,t)=(A.e+B.e)eαλ^2.t , when k=λ2 for λ >0
2. u(x,t)=(C.cos+B.sin)e-αλ^2.t , when k=-λ2 for λ >0
3. u(x,t)= E.x+F , when k=0

Example 2:
Example:
PDE: ut=a.uxx , 0<x<L, t>0
PDE: ut=3.uxx , 0<x<2, t>0
BCs: u(0,t)=0, ux(L,t)=0, t>0
BCs: u(0,t)=0, u(2,t)=0, t>0
IC: u(x,0)=x
IC: u(x,0)=x
We get the non-trivial solution from
We get the non-trivial solution from
2nd solution
2nd solution
u(x,t)=. exp{}
u(x,t)=. exp{}
 Other Solution Methods:

• Fourier and Laplace Transforms: Transforming the PDE into a different

domain (frequency or Laplace space) can simplify the problem, allowing for
easier solution.
• Numerical Methods: When analytical solutions are not feasible, numerical
methods such as finite difference, finite element, or spectral methods are
employed to approximate solutions.
 One Dimension Heat Equation:
• We need to find the temperature distribution 𝑢( 𝑥, 𝑡)
along the medium over time. This involves determining
how the temperature varies with both position 𝑥 along
the medium and time 𝑡.
• It is a classic example of a parabolic PDE and is used to
model heat conduction processes.

1. Homogeneous Medium (ρ,s:const)

2. Heat flows in direction of decreasing Temperature
ASSUMPTIONS 3. Heat flow rate (Q) across an area (A) is proportional to A
and temperature gradient. (ρ as proportionality Constant)
4. Quantity of heat gained and lost by body is proportional
to mass of body ‘m’ and change in temperature ‘dT’ ((‘s’
as proportionality Constant)
• Q1= -kAx
• Q2= -kAx+Δx

• ΔQ= Q2-Q1
=kAx+Δx -kAx
• ΔQ= (A. Δx.ρ) . S .

•
• ut=a2uxx