CHAPTER 3

SOLUTION METHODOLOGY

The chapter deals with the solution methodology of the commercial software, ANSYS
FLUENT that has been used in the study. Details about the computational approach
and numerical schemes chosen in the study are briefly explained. Brief information
about the widely known k- turbulence model is also mentioned. Also, the boundary
conditions that have been used for simplifying the study are provided here. The
chapter concludes with basic information about the dimensionless numbers that
frequently come up in problems related to fluid mechanics and heat transfer.

Computational fluid dynamics is the analysis of systems involving fluid flow,

heat transfer and associated phenomenon such as chemical reactions by means of
computer based simulation. The technique is very powerful and spans a wide range
of industrial and non-industrial application areas. CFD codes are structured around
the numerical algorithms that can tackle fluid flow problems. In order to provide
easy access to their solving power all commercial CFD packages include sophisticated
user interfaces to input problem parameters and to examine the results. Hence all
codes contain three main elements: (i) a pre-processor, (ii) a solver and (iii) a post-
processor. Based on the solver there are three distinct streams of numerical solution
techniques: finite difference, finite element and spectral methods.

The finite volume method was originally developed as a special finite

difference formulation. It is central to the main commercially available CFD codes
namely PHOENICS, FLUENT, FLOW3D and STAR-CD. In this present analysis of
convection properties of nanofluids all simulations have been carried out in
FLUENT. This chapter gives the details of the solver settings that have been used
through-out the analysis.

3.1 COMMERCIAL CFD SOFTWARE – ANSYS FLUENT

The FLUENT code solves the partial differential equations namely the Navier-
Stokes equations i.e., continuity, momentum, energy, turbulent kinetic energy and
turbulent dissipation rate by discretizing them into algebraic equations and solves
them with appropriate boundary conditions by Gauss Seidel matrix method.

15
Chapter 3_______________________________________________________________

The differential and integral forms of the Navier-Stokes equations are as follows. The
scalar variable (𝜑), Diffusion coefficient (𝜑 ) and Source term (𝑆𝜑 ) of the equations
are given below.

Differential form

𝜕𝜑 𝜕 𝜕 𝜕
+ ⃗ 𝜑) =
(𝑉 (𝜑 ) + 𝑆𝜑
𝜕𝑡 𝜕𝑥𝑖 𝜕𝑥𝑖 𝜕𝑥𝑖

Integral form

𝜕
∭ 𝜑𝑑 + ∬ 𝑉
⃗ 𝜑. 𝑛⃗𝑑𝑆 = ∬ 𝜑 ∇𝜑. 𝑛⃗𝑑𝑆 + ∭ 𝑆𝜑 𝑑
𝜕𝑡 
 𝑆 𝑆

Scalar Diffusion Source term

Equation variable coefficient
(𝜑) (𝜑 ) (𝑆𝜑 )
Continuity 1 0 0
𝜕𝑝 𝜕 𝜕𝑢 𝜕 𝜕𝑣
− + (𝑒𝑓𝑓 ) + (𝑒𝑓𝑓 )
x 𝜕𝑥 𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑥
u 𝑒𝑓𝑓
momentum 𝜕 𝜕𝑤
+ (𝑒𝑓𝑓 )
𝜕𝑧 𝜕𝑥
𝜕𝑝 𝜕 𝜕𝑢 𝜕 𝜕𝑣
− + (𝑒𝑓𝑓 ) + (𝑒𝑓𝑓 )
y 𝜕𝑦 𝜕𝑥 𝜕𝑦 𝜕𝑦 𝜕𝑦
v 𝑒𝑓𝑓
momentum 𝜕 𝜕𝑤
+ (𝑒𝑓𝑓 )
𝜕𝑧 𝜕𝑦
𝜕𝑝 𝜕 𝜕𝑢 𝜕 𝜕𝑣
− + (𝑒𝑓𝑓 ) + (𝑒𝑓𝑓 )
z 𝜕𝑥 𝜕𝑥 𝜕𝑧 𝜕𝑦 𝜕𝑧
w 𝑒𝑓𝑓
momentum 𝜕 𝜕𝑤
+ (𝑒𝑓𝑓 )
𝜕𝑧 𝜕𝑧
Turbulent
kinetic 𝑡 𝑃 − 
k  +
energy 𝑘
Turbulent
𝑡  2
dissipation   + 𝐶1 𝑃 − 𝐶2
 𝑘 𝑘
rate

16
______________________________________________________ Solution Methodology

FLUENT has two flow solvers – pressure based and density based. In the
pressure based approach, the pressure field is obtained from by manipulating
continuity and momentum equations. FLUENT also solves the mass and momentum,
and when required energy, turbulence and chemical species. Of the two pressure-
based solver algorithms available, pressure based coupled algorithm has been used in
this study. In this method, the coupled system of equations comprising the
momentum equations and the continuity equations are solved together. Semi-
Implicit Method for Pressure-Linked Equations (SIMPLE) algorithm by Patankar
and Spalding (1972) has been used in this study for coupling pressure and velocity.

Fluent stores discrete values of the scalar 𝜑 at the cell centres. However face
values 𝜑𝑓 are required for the convection terms and must be interpolated from the
cell centre values. This is done using the upwind scheme. Upwinding means that the
face value 𝜑𝑓 is derived from quantities in the cell upstream relative to the direction
of normal velocity. In the present analysis second order upwind scheme has been
used. In the second order upwind scheme, the quantities at the cell faces are
computed using a multidimensional linear reconstruction approach. In this approach
higher order accuracy is achieved at cell faces through a Taylor series expansion of
the cell centred solution about the cell centroid. Gradients are needed not only for
constructing values of a scalar at the cell faces, but also for computing secondary
diffusion terms and velocity derivatives. Least square cell based method has been
used in the present work. In this method the solution is assumed to vary linearly.
Residual values have been used for determining the convergence of the discretized
equations. For continuity and momentum equations the convergence criteria has
been fixed at 10-6 and for turbulence and energy equations 10-9. Figure 3.1 shows the
flowchart of the methodology followed.

3.2 MATHEMATICAL FORMULATION

A. SINGLE PHASE MODEL (pseudo-homogeneous model)

In the single phase model, the basic hypothesis is that the mixture behaves
like a single phase fluid. In this model the nanofluid is treated as a normal fluid, but
with enhanced properties due to the addition of nanoparticles. The governing
equations for this model are as follows:

17
Chapter 3_______________________________________________________________

Conservation of mass:
∇. (𝑚 𝑉
⃗)=0 (3.1)

Conservation of momentum:
∇. (𝑚 𝑉
⃗𝑉⃗ ) = −∇𝑃 + ∇. ( − 𝑡 ) (3.2)

Conservation of energy:
∇. (𝑉
⃗ 𝐶𝑃 𝑇) = ∇. (𝜆∇𝑇 − 𝐶𝑃  𝑣𝑡
̅) (3.3)
𝑚

In the energy equation (3.3), compression work and the viscous dissipation are
assumed to be negligible.

B. MIXTURE MODEL

In the mixture model, the fluid is considered to be a single fluid with two
phases and the coupling between them is strong. Each phase has its own velocity
vector and within a given control volume there is a certain fraction of each phase.
The governing equations for this model are as follows:

Conservation of mass:
∇. (𝑚 𝑉
⃗𝑚 ) = 0 (3.4)

Conservation of momentum:
𝑛
(3.5)
∇. (𝑚 𝑉 ⃗𝑚 ) = −∇𝑃𝑚 + ∇. ( − 𝑡 ) + ∇. (∑   𝑉
⃗𝑚 𝑉 ⃗ ⃗
𝑉
𝑘 𝑘 𝑑𝑟,𝑘 𝑑𝑟,𝑘
)
𝑘=1

⃗ 𝑑𝑟,𝑘 is the drift velocity of the kth phase.

where 𝑉
Volume fraction:
(3.6)
∇. (𝑃  𝑉
⃗𝑚 ) = −∇. (  𝑉
𝑝
⃗ 𝑑𝑟,𝑝 )
𝑃 𝑝

Conservation of energy:
𝑛
(3.7)
∇. (∑ 𝑘 𝑉
⃗ 𝑘 ( 𝐻𝑘 + 𝑃)) = ∇. (𝜆∇𝑇 − 𝐶𝑃  𝑣𝑡
𝑘 𝑚
̅)
𝑘=1

In the energy equation (3.7), compression work and the viscous dissipation are
⃗ 𝑑𝑟,𝑘 is the
assumed to be negligible. In the conservation of momentum equation (3.5) 𝑉
drift velocity for secondary phase k (i.e., the nanoparticles in the present study),
defined as:

18
______________________________________________________ Solution Methodology

⃗ 𝑑𝑟,𝑘 = 𝑉
𝑉 ⃗𝑘 − 𝑉
⃗𝑚

The shear relation is given by:

 = 𝑚 ∇𝑉
⃗𝑚
𝑛

𝑡 = ∑ 𝑘 𝑘 ̅̅̅̅̅̅
𝑣𝑘 𝑣𝑘
𝑘=1

The slip velocity (relative velocity) is defined as the velocity of secondary phase (p)
relative to the velocity of the primary phase (bf):
⃗ 𝑝𝑏𝑓 = 𝑉
𝑉 ⃗𝑝 − 𝑉
⃗ 𝑏𝑓

The drift velocity is related to the relative velocity:

𝑛
𝑘 𝑘
⃗ 𝑑𝑟,𝑝 = 𝑉
𝑉 ⃗ 𝑝𝑏𝑓 − ∑ ⃗
𝑉
𝑚 𝑏𝑓𝑘
𝑘=1

The relative velocity is determined by expressions proposed by Manninen et al.

(1996).

3.3 DRAG FUNCTION

The slip velocity in case of multiphase flows for fluid-fluid interaction can be
calculated by the following formulations:

 Schiller and Naumann (1935) correlation: Acceptable for general use in all
fluid-fluid multiphase calculations.
 Morsi and Alexander (1972) model: The most complete, adjusting the function
definition frequently over a large range of Reynolds numbers, but calculations
with this model may be less stable than with the other models.
 Symmetric: The symmetric model is recommended for flows in which the
secondary (dispersed) phase in one region of the domain becomes the primary
(continuous) phase in another.

In the present study, the drag function has been calculated by Schiller and Naumann
(1935) correlation.

𝑝 𝑑𝑝2 (𝑝 − 𝑚 ) (3.8)

⃗ 𝑝𝑏𝑓 =
𝑉 𝑎
18𝑏𝑓 𝑓𝑑𝑟𝑎𝑔 𝑝
⃗𝑚 . ∇)𝑉
𝑎 = 𝑔 − (𝑉 ⃗𝑚

19
Chapter 3_______________________________________________________________

1 + 0.15𝑅𝑒𝑝0.687 Re𝑃  1000 (3.9)

𝑓𝑑𝑟𝑎𝑔 ={
0.0183𝑅𝑒𝑝 Re𝑃 > 1000

3.4 k- TURBULENCE MODEL

The Navier–Stokes equations govern the velocity and pressure of a fluid flow.
In a turbulent flow, each of these quantities may be decomposed into a mean part
and a fluctuating part. Averaging the equations gives the Reynolds averaged Navier–
Stokes (RANS) equations, which govern the mean flow. However, the nonlinearity of
the Navier–Stokes equations means that the velocity fluctuations still appear in the
RANS equations, in the nonlinear term from the convective acceleration. Hence for
closing system of mean governing equations two new equations are required. This
can be achieved in various ways.

In the present study, the standard k- turbulence model (Launder and
Spalding (1974)) has been used for closing the system of mean flow equations. The
model introduces two new equations, one for turbulent kinetic energy and the other
for the rate of dissipation. The two equations are

𝜇𝑇 (3.10)
∇. (𝜌𝑉̅ 𝑘) = ∇. ( ∇𝑘) + 𝐺𝑘 − 𝜌𝜖
𝜎𝑘

𝜇𝑇 𝜖 (3.11)
∇. (𝜌𝑉̅ 𝜖) = ∇. ( ∇𝜖) + (𝐶1 𝐺𝑘 − 𝐶2 𝜌𝜖)
𝜎𝑘 𝑘

𝐶1 = 1.44, 𝐶2 = 0.09, 𝐶𝜇 = 0.09, 𝜎𝑘 = 1.0, 𝜎𝜖 = 1.3

The values of the constants have been arrived at by numerous iterations of data
fitting for a wide range of turbulent flows. Complete derivation of equations 3.10 and
3.11, are given in APPENDIX A for reference. Detailed analysis of the k- turbulence
model is given in Mohammadi and Pironneau (1994). Figure 3.1 shows the
methodology followed in the present study.

20
______________________________________________________ Solution Methodology

Figure 3.1: Methodology flowchart

3.5 BOUNDARY CONDITIONS

In the present study the following boundary conditions has been used:

(a.) Velocity inlet

The velocity vector and scalar properties of the flow at the inlet boundary are
provided here. Uniform profile has been used in the present study.

(b.) Pressure outlet

The gauge pressure at the outlet boundary is the input parameter. This is interpreted
as the static pressure of the environment into which the flow exhausts.

21
Chapter 3_______________________________________________________________

(c.) Wall

At the wall the no slip condition has been used (Dirichlet conditions). Also the input
for constant heat flux condition is to be provided here.

(d.) Axisymmetric

Used for simulating 2D axisymmetric flow conditions.

(e.) Symmetry

Used for simulating 3D symmetric flow conditions. This condition considers zero
normal gradients of all the variables.

3.6 DIMENSIONLESS NUMBERS

For better understanding dimensionless numbers are used to characterize results in

fluid and heat transfer problems. Following non-dimensionless numbers are used in
this study.

Reynolds number, Re:

Reynolds number is a dimensionless quantity that is used to help predict similar flow
patterns in different fluid flow situations. Re is used as the input parameter in the
present study.

𝜌𝑣𝐷𝐻
𝑅𝑒 =
𝜇

Prandtl number, Pr:

The Prandtl number is defined as the ratio of momentum diffusivity to

thermal diffusivity. Literature shows that 𝑁𝑢 = 𝑓(𝑅𝑒, 𝑃𝑟)

𝜇𝐶𝑃
𝑃𝑟 =
𝑘

Nusselt number, Nu:

Nusselt number is the ratio of convective to conductive heat transfer across

the boundary. A Nusselt number close to one, namely convection and conduction of
similar magnitude, is characteristic of laminar flow. A larger Nusselt number

22
______________________________________________________ Solution Methodology

corresponds to more active convection, with turbulent flow typically in the 100 to
1000 range.

ℎ𝐿
𝑁𝑢 =
𝑘

Stanton number, St:

Stanton number is a dimensionless number that measures the ratio of heat

transferred into a fluid to the thermal capacity of the fluid.

ℎ 𝑁𝑢
𝑆𝑡 = =
𝜌𝑣𝐶𝑃 𝑅𝑒 𝑃𝑟

Dean number, De:

Dean number occurs in the study of flow in curved pipes and channels. It is
typically denoted by De and defined as

𝜌𝑣𝐷𝐻 𝐷𝐻 0.5 √𝑐𝑒𝑛𝑡𝑟𝑖𝑝𝑒𝑡𝑎𝑙 𝑓𝑜𝑟𝑐𝑒𝑠 × 𝑖𝑛𝑒𝑟𝑡𝑖𝑎𝑙 𝑓𝑜𝑟𝑐𝑒𝑠

𝐷𝑒 = ( ) =
𝜇 2𝑅𝐶 𝑣𝑖𝑠𝑐𝑜𝑢𝑠 𝑓𝑜𝑟𝑐𝑒𝑠

Using the above described solution methodology, simulations have been carried out.
The simulation results are presented in the subsequent chapters.

23