Chapter 1

Introduction

Most of the problems arising in nature are modelled by differential equations.

Solution technique of differential equations is broadly classified as analytical,

numerical and semi-numerical methods.

Analytical methods are the solution techniques of differential equations, which

give closed form solution to the differential equations, whereas numerical and

semi-numerical method give the approximate solution. The quality of numerical

methods is depending upon the accuracy of the solution to the problem, since

most of the model problems are highly non-linear solving through numerical

method will lead to a different problem. For example, finite difference method,

if we want to calculate the derivative of the unknown function at a certain

value, then one has to develop a separate scheme and the computations in

1
 Introduction

finite difference are very sensitive to the step size. To avoid these difficulties

the researchers developed the semi-numerical techniques to solve the differential

equation.

In semi-numerical methods, the solution of the differential equation will be

assumed as some series with parameters, like power series, Dirichlets series e.tc.,

Then by substituting into the considered problem and equating the various

powers of parameter to zero gives the set of differential equations, solving this

set of differential equations and substituting into the original series will give the

approximate solution to the differential equation. Solving differential equations

through semi-numerical method will lead to the series, hence convergence of the

series will have the direct impact on the accuracy of the solution. If the series is

convergent, then few terms of the series will approximate a better solution to the

problems. If the series is divergent solution of the problem is not accurate, but

we have the technique to improve or converge the given divergent series, hence

the analysis of the series obtained in semi-numerical techniques is essential in the

study of semi-numerical techniques.

Most of the considered problems in this study are related to fluid dynamics,

in order to get a complete and fundamental understanding of the fluid dynamics

of the problems, a detail study of general governing equations of continuity and

momentum equation are taken up [1, 2] .

2
 Introduction

1.1 The Material Derivative and Acceleration

“The velocity of particles at any point in the space can be expressed in Eulerian

form as

q = f (~r, t) (1.1.1)

where ~r is position vector and also ~r = F (ζ, t), where ζ is the initial position of

the particles.

The velocity in the x-direction at time t can be reduced to from (1.1.1) in the

form

u = f1 (x, y, z, t). (1.1.2)

At the infinitesimal time interval ∆t later, this particle will move to the position

x + ∆x, y + ∆y, z + ∆z with an x-component of velocity u + ∆u. The increments

∆x, ∆y and ∆z are ∆x = u∆t, ∆y = v∆t, ∆z = w∆t.

Hence we have

∂f1 ∂f1 ∂f1 ∂f1

u+∆u = f1 (x, y, z, t)+(u ∆t+v ∆t+w ∆t+ ∆t)+Higher order terms in ∆t.
∂x ∂y ∂z ∂t
(1.1.3)

Taking the limiting form of (1.1.3) on can obtain

∆u Du ∂ ∂ ∂ ∂
lim = =( +u +v + w )u. (1.1.4)
∆t→0 ∆t Dt ∂t ∂x ∂y ∂z

3
 Introduction

The total differential with respect to time

D ∂ ∂ ∂ ∂
= +u +v +w (1.1.5)
Dt ∂t ∂x ∂y ∂z

is called as the material derivative.

1.2 The Equation of Continuity ( Conservation

of Mass)

The law of conservation of mass states that mass can be neither created nor

destroyed. Consider a moving surface A (as shown below) enclose a mass of fluid

within volume τ lying entirely in the fluid.

If ρ∆τ is an element of mass, then the mathematical expression of the law of

conservation of mass can be written as

D
(ρ∆τ ) = 0. (1.2.1)
Dt

4
 Introduction

This is one form of the equation of continuity. Equation (1.2.1) can also be

expressed in the form by dividing ρ∆τ .

Dρ/Dτ D(∆τ /Dt)

+ = 0. (1.2.2)
ρ ∆τ

From equation (1.2.2) it can be concluded in such a way that a decrease in volume

is accompanied by an increase in density and vice versa.

If we consider the total mass of the fluid enclosed in the surface A, equation

(1.2.1) becomes
X 1 Dρ D(∆τ )
lim [ ∆τ + ] = 0. (1.2.3)
∆τ →0 ρ Dt Dt

The first term in equation (1.2.3) can be expressed as

Z Z Z
 1 ∂ρ 
( + q.Oρ) dτ = 0. (1.2.4)
ρ ∂t

The second term in equation (1.2.3) represents the rate of change of volume which

is equivalent to
X D Z Z Z
lim (∆τ ) = ∆.qdτ (1.2.5)
∆τ →0 Dt

combining equations (1.2.4) and (1.2.5)

Z Z Z
1 ∂ρ
{ [ + q.Oρ] + O.q}dτ = 0
ρ ∂t

5
 Introduction

Z Z Z
1 ∂ρ
[ + O.(ρq)]dτ = 0 (1.2.6)
ρ ∂t

equation (1.2.6) is zero when

∂ρ
+ O.(ρq) = 0. (1.2.7)
∂t

or
∂ρ
+ q.Oρ + ρO.q = 0.
∂t

D
(ρ) + ρO.q = 0. (1.2.8)
Dt

In Cartesian coordinate equation (1.2.7) can be written as

∂ρ ∂(ρu) ∂(ρv) ∂(ρw)

+ + + = 0.
∂t ∂x ∂y ∂z

In the incompressible fluid the equation (1.2.8) reduces to

O.q = 0

∂u ∂v ∂w
+ + = 0.
∂x ∂y ∂z

6
 Introduction

1.3 Navier Stokes Equation ( Conservation of

Momentum)

Navier Stokes equations are derived by using the Newton’s law of motion, it states

that the total force acting on a fluid mass enclosed in an arbitrary volume fixed

in space is equal to the time rate of change of linear momentum. This is also

called the law of conservation of momentum.

Consider a differential element of fluid which is infinitely small, the three forces

are considered, the force due to gravity, the force due to the difference in pressure

and force due to the viscosity of fluid.

Consider a fluid element in three dimensions as shown in figure

Force due to gravity along x-component is mgx = ρdxdydzgx . Normal stress

acting on the right and left side is denoted by σxx (x+dx) and σxx (x), shear stress

acting on top of the cube and bottom cube is denoted byσyx (y + dy) and σyx (y),

similarly the stress acting to right in front and back of the cube is denoted by

7
 Introduction

σzx (z+dz) and σzx (z). Now the sum of the forces acting on the cube in x-direction

is

ρgx dxdydz + σxx (x + dx)dydz − σxx (x)dydz + σyx (y + dy)dxdz − σyx (y)dxdz

+σzx (z + dz)dxdy − σzx (z)dxdy = ρdxdydzax

dividing dxdydz both sides, we get

σxx (x + dx) − σxx (x) σyx (y + dy) − σyx (y) σzx (z + dz) − σzx (z)
ρgx + + + = ρax
dx dy dz

∂ ∂ ∂
ρgx + σxx + σyx + σzx = ρax
∂x ∂y ∂z

∂ ∂ ∂ ∂u ∂u ∂u ∂u
ρgx + σxx + σyx + σzx = ρ( +u +v +w )
∂x ∂y ∂z ∂t ∂x ∂y ∂z

similarly

∂ ∂ ∂ ∂v ∂v ∂v ∂v
ρgy + σxy + σyy + σzy = ρ( +u +v +w )
∂x ∂y ∂z ∂t ∂x ∂y ∂z

∂ ∂ ∂ ∂w ∂w ∂w ∂w
ρgz + σxz + σyz + σzz = ρ( +u +v +w ).
∂x ∂y ∂z ∂t ∂x ∂y ∂z

The above set of equations is called the Navier Stokes equation” [2].

8
 Introduction

1.4 Basic Idea of Homotopy Perturbation

Method

To describe the HPM [3, 4] solution to the non-linear differential equation, we

consider

D[f (η)] − f1 (η) = 0 (1.4.1)

where D denotes the operator, f (η) is unknown functions, η denote the

independent variable and f1 is known functions. D can be written as

D =L+N

where L is a simple linear part, N is remaining part of the equation (1.4.1). The

proper selection of L, N form the governing equations, one can get the homotopy

equation as follows

   
H(Φ(η, q; q)) = (1 − q) L(Φ, q) − L(v0 (η)) + q L(Φ, q) − f1 (η) = 0 (1.4.2)

where q is the embedding parameter which varies from 0 to 1 and v0 (η) is the

initial guess to the (1.4.1).

9
 Introduction

We assume the solution of (1.4.2) as follows

∞
X
Φ(η, q) = q n fn (η) (1.4.3)
n=0

The solution to the considered problems is (1.4.3) at q = 1.

1.5 Research Background

Fluid mechanics is one of the oldest branches of physics and the foundation for

understanding of many other aspects of the applied sciences and engineering. It

is concerned with the study of motion of fluids. The subject is of great interest

for two reasons. First, an understanding of fluid mechanics helps us to explain

the variety of fascinating phenomena around us. Second, an understanding of

this subject is essential to many problems encountered by an engineer. The

development of aeronautical, chemical and mechanical engineering during the

past few decades have given stimuli to study of fluid mechanics so that it now

ranks as one of the most important basic subject in engineering sciences.

Most of these problems in nature are represented by partial differential

equations. The similarity transformations reduces the equations to ordinary

differential equations. Due to nonlinearity of the equations, exact solutions

are rare. However, with assumptions, most of time solutions are obtained

10
 Introduction

by numerical methods. The schemes used in numerical methods many of

the time end up with non convergence of solution. To re-leave from such

difficulty the Homotopy perturbation method first proposed by Ji-Huan He in

1998 [4–6]. “HPM is the combination of traditional perturbation method and

homotopy in topology. In 2013 Sumit Gupta and his associates [7] applied

this method for solving nonlinear wavelike equations. This method is useful

for solving the different class of problems in the applied mathematics [4, 5, 8]

(like, functional integral equations [9], coupled system of reaction diffusion

equation [10], Helmholtz equation and fifth-order Kdv equation [11], the epidemic

model [12] etc.)” [5]. Homotopy perturbation method for MHD problem has been

successfully used by [13].

In reality fluid models involved are non-Newtonian. The complicated rheological

properties better explained by non-Newtonian fluids. One such model is known to

be Casson fluid model [14]. Obtaining exact solutions is very rare for this class of

problem. However, with assumptions, most of the time approximate solutions are

obtained by numerical methods. So to deal with these problems semi-analytical

methods are developed and more commonly used [15].

The perturbation method [16] is used earlier to solve the algebraic equations,

next it used to solve the differential equations. In 1938 Goldstein and his associate

used this technique to solve the differential equation obtained in different fields.

11
 Introduction

In 1970’s Van Dyke and his co-workers also used to solve the differential equations

along with gives the efficient techniques to improve the convergence region of the

series, this work makes the perturbation techniques to more stronger in solving

differential equations. In 1990 N. M. Bujurke and his associates successfully

apply these techniques to solve different classes of problems occur in different

engineering fields.

1.6 Motivation

A careful observation on literature, and study about the mathematical methods

in fluid dynamics of few interesting fluid models, we are motivated to investigate

the class of problems arising in fluid mechanics by semi analytical techniques.

Also the possible application of this technique to non-Newtonian fluid models,

Viz. Casson fluid with various characteristics under different conditions. Even

though, various methods of solving boundary value problems have been reported

in the literature, some attempt has been made to obtain approximations analysis

through semi analytical methods. Also there are enough inbuilt characteristics

like stability, consistency and convergence in semi analytical methods.

12