Int. J. Exergy, Vol. 19, No.

1, 2016 15

Computational study of entropy generation

minimisation in continuous stirred tank reactor

Akhilesh Khapre*
and Basudeb Munshi
Department of Chemical Engineering,
National Institute of Technology,
Rourkela, Odisha, 769008, India
Email: akhilesh_khapre@yahoo.co.in
Email: basudeb@nitrkl.ac.in
*Corresponding author

Abstract: Entropy generation minimisation (EGM) for optimising stirred

tank reactor (CSTR) is performed by considering two cases: constant wall
temperature and heat flux thermal boundary conditions. Entropy generation is
computed from the temperature and velocity gradients which are predicted
using CFD simulation. The selected parameters for EGM are inlet Reynolds
number, impeller rotation, impeller clearance and impeller blade width.
The optimum inlet Reynolds number increases with the wall temperature and
heat flux. The effect of impeller rotation is also found on entropy generation.
The optimum impeller clearance varies with the wall temperature and is
invariant with the wall heat flux. The optimum impeller blade width depends
on the magnitude of the wall temperature and heat flux. Bejan number is also
determined, and it shows that the entropy generation is mostly dominated by
heat transfer over the fluid friction irreversibility in the present study.

Keywords: entropy generation minimisation; CSTR; CFD; optimisation;

temperature; heat flux; Bejan number.

Reference to this paper should be made as follows: Khapre, A. and Munshi, B.

(2016) Computational study of entropy generation minimisation in continuous
stirred tank reactor, Int. J. Exergy, Vol. 19, No. 1, pp.1540.

Biographical notes: Akhilesh Khapre received BE in Chemical Engineering

from Pune University, India in 2007. He received his MTech in Computational
Fluid Dynamics in 2009 from UPES, Dehradun, India. He is now pursing his
PhD at National Institute of Technology, Rourkela, India. His research interests
are CDF and heat transfer.

Basudeb Munshi is an Associate Professor in the Department of Chemical

Engineering at National Institute of Technology, Rourkela, India. He received
his PhD in Chemical Engineering from Birla Institute of Technology & Science
(BITS), Pilani, India. His major interests are transport phenomena, CFD,
separation technology and process control.

Copyright 2016 Inderscience Enterprises Ltd.

16 A. Khapre and B. Munshi

1 Introduction

Reactor is the heart of chemical processes. Among all kind of reactors, stirred tank
reactor (STR) is widely used. The efficiency of heat and mass transfer in STR depends
largely on the extent of a mixing. The optimum designing of STR is a challenging task to
achieve the required mixing of the reactants. Many researchers carried out it by solving
the standard process model equations based on the first law of thermodynamics along
with the conventional or latest optimisation techniques (Powell and Hill, 2008; Petrov
and Tzonkov, 2009; Chaibakhsh et al., 2010; Chen and Wang, 2010; Song et al., 2010;
Vaneshani and Jazayeri-Rad, 2011; Sivapathasekaran and Sen, 2013; Aghbolaghy and
Karimi, 2014). The objectives of all these optimisation techniques are either to minimise
the cost or to obtain the optimal operating condition to control the product quality. The
application of the second law of thermodynamics with entropy generation minimisation
(EGM) approach is a useful technique to achieve the optimum conditions (Alebrahim and
Bejan, 1999; Zimparov and Penchev, 2003; Balaji et al., 2007; Jankowski, 2009;
Guo et al., 2010). In general, the mathematical models of stirred tank include all three
transport equations: momentum, mass and energy. The finite temperature gradient in a
heat transfer process and a frictional loss in the flow process increase the entropy
generation at the cost of work potential of the system. Thus, the destruction of available
work decreases with decreasing entropy generation and it is minimal at the minimum
entropy generation point. Nowadays, the energy is one of the most concerned issues.
At the minimum entropy generation, the use of supplied energy is a maximum for
particular system.
Many fluid flows and heat transfer devices are optimised using EGM technique. It
was first proposed by McClintock (1951) who did optimum design of fluid passages of a
heat exchanger. Later, an abundant work in the area of EGM was carried out and most of
the works were analytical in nature. Bejan (1979) examined the entropy generation in the
convective heat transfer process. He further carried out entropy generation analysis of
numbers of fundamental applications and showed that the minimisation of entropy
generation increased the efficiency of the system (Bejan, 1995). Thereafter, the EGM was
used by many researchers for the purpose of analysis and optimisation of systems
such as fluid flow system, heat exchanger design, fins design and power plant design.
The entropy generation analysis of fluid flow system included pipe flow (Mahmud
and Fraser, 2002; Sahin and Mansour, 2003; Al-Zaharnah and Yilbas, 2004; Guo et al.,
2011), channel flow (Haddad et al., 2004; Koo and Kleinstreuer, 2004; Basha et al., 2007;
Adesanya and Makinde, 2014) and annular flow (Demirel and Kahraman, 2000;
Yilbas et al., 2004; Kahraman and Yurusoy, 2008; Kucuk, 2011). Sciubba (2010)
performed the multi-variable optimisation of the bifurcation of the branch of a pipe using
the EGM. The selected parameters were the aspect ratio of domain, diameter ratio
of branches and the length of secondary branch. Both the aspect ratio and diameter
ratio influenced strongly the entropy generation, while the patterns of branching had
negligible effect.
Alebrahim and Bejan (1999) optimised a cross-flow heat exchanger. They studied the
EGM of the optimised geometry and found that the optimised geometry is insensitive to
additional thermodynamic irreversibility. Cimpean and Pop (2011) used EGM approach
to optimise parameters like Peclet number, mixed convection parameter and inclination
angle of an inclined channel where fluid flows through a porous media. Thiel (2012)
proposed an improved design of the humidificationdehumidification and desalination
 Computational study of entropy generation minimisation 17

system to maximise its efficiency using EGM. Zimparov and Penchev (2003) replaced
smooth tubes with three-dimensional dimple and pyramidal surface roughness tubes in
the heat exchanger. They optimised it thermodynamically using the EGM method. Singh
et al. (2008) optimised an air-cooled tube fin heat exchanger design using the principle of
EGM. Asadi (2013) proposed the cuckoo algorithm to optimise a heat exchanger.
The algorithm was based on the EGM concept. Similarly, Ahmadi et al. (2011) used a
multi-objective genetic algorithm to minimise both the entropy generation and total
annual cost of a compact heat exchanger. They considered six geometrical variables to
study the optimisation and observed that any change in geometrical parameters which
decrease the entropy generation, leads to increasing the total annual cost. Culham and
Muzychka (2001) designed plate fin type heat sink using the EGM technique.
The viscosity of fluid accounts the presence frictional loss, i.e. viscous dissipation in
the flow system. Thus, entropy generation is significantly influenced by the viscosity of
fluid (Sahin, 2014). Many researchers studied the effect of non-Newtonian parameters
(Peters and Baaijens, 1997; Mahmud and Fraser, 2002; Ayinde and Yilbas, 2004;
Pakdemirli and Yilbas, 2006; Galanis and Rashidi, 2012; Shojaeian and Kosar, 2014)
and temperature-dependent viscosity (Sahin, 1999; 2002; Tasnim and Mahmud, 2002;
Al-Zaharnah and Yilbas, 2004) of fluid on the entropy generation. Sahin (2002) used
temperature-dependent viscosity of fluid to determine the heat transfer coefficient and
the fluid friction factor in the pipe at constant heat flux. He found a strong effect of the
variation of viscosity on entropy generation. Pakdemirli and Yilbas (2006) studied the
entropy generation in a pipe due to non-Newtonian fluid flow. They examined the effect
of non-Newtonian parameter and Brinkman number on the entropy generation. The non-
Newtonian parameters affected inversely the entropy generation, while Brinkman number
changed proportionately with the entropy generation. Galanis and Rashidi (2012) used
power law fluid to study the heat and mass transfer in circular and parallel plate ducts.
Analytical solution of velocity, temperature and concentration were used to evaluate the
local entropy generation rate. They observed that the local entropy generation increases
with the flow behaviour index. Khan and Gorla (2012) analysed the entropy generation
rate over a horizontal plate enclosed in a porous medium due to free convection of
non-Newtonian fluids. Hung (2008) studied the entropy generation of non-Newtonian
fluid flow through a circular micro-channel with and without viscous dissipation term.
He showed that neglecting viscous dissipation term can lead to significant deviation in
the entropy generation.
Naterer and Adeyinka (2009) studied experimentally the entropy generation of
laminar fluid flow in a tank induced by a magnetic stirrer. They calculated entropy
production rates using the measured hydrodynamics data and found the highest entropy
production rates near the stirrer. Manzi and Carrazzoni (2008) optimised a stirred tank
reactor (CSTR) in the presence of a typical irreversible, exothermic and first-order
chemical reaction. They found that entropy generation is minima when inlet temperature
of reactant is equal to the reaction temperature. Manzi et al. (2009) worked on direct
method of entropy minimisation to produce propylene glycol inside a CSTR. They used
chemical process model equations in their work. Driss et al. (2012) studied numerically
the hydrodynamics and mixing performance of a stirred tank to show the effect of
multiple Rushton turbine configurations. They found the highest viscous dissipation rate
near the blade tip. But they did not study the EGM of the stirred tank system.
Several applications of the EGM were found in the literature in designing many
heat transfer devices. Such applications are of two types, deterministic approach and
18 A. Khapre and B. Munshi

heuristic approach. Most of the application used deterministic approach, which needs the
development of analytical expressions of entropy generations as a function of system
parameter. But for complicated system, it is difficult to formulate the analytical
expression of the entropy generation. Beside this, an analytical expression was developed
as function of average heat transfer flux and using the fluid friction expression available
in the open literature. Therefore, the deterministic approach fails for the complicated
problems, which can be solved better by the heuristic approach. In heuristic approach, the
detail solution of the system is obtained let say using CFD approach and then the
calculated local entropy generation is used to calculate the overall entropy generation.
The heuristic approach was already applied on few systems as discussed above to design
heat exchangers and fluid flow systems using EGM method. It was also found that there
were few works carried out in the area of EGM of CSTR. Those work used only
analytical equations for mass, energy and entropy. It is expected that the spatial
distributions of temperature and velocity have a definite effect on the entropy
distribution. Therefore, the EGM analysis is carried out in the present study with heuristic
approach to optimise the performance of the CSTR. A commercial CFD code, ANSYS
Fluent, is used to simulate the models of CSTR. The identification of optima of the CSTR
is found solving the initial process numerically followed by the improvement of the
process performance by EGM technique with respect to system parameters such as inlet
Reynolds number, impeller speed, impeller clearance and impeller blade width.

2 Methodology

2.1 Physical system description

The batch stirred tank (Figure 1(a)) used by Venneker et al. (2010) in their experimental
work is considered for the validation of CFD models. Both the diameter (D) and height
(H) of the tank are 0.627 m. Four baffles having width (Wb) (1/10)th of the diameter of
the tank are mounted on the tank wall at 90 apart from each other. The diameter (Di) of
the Rushton turbine is (1/3)rd of the tank diameter. It is mounted at a clearance height,
Ci equal to (1/3)rd of the tank diameter from the bottom wall of the tank. The impeller
blade width, b, is (1/5)th of the impeller diameter. To analyse entropy generation of
CSTR, a pair of inlet and outlet pipe of diameter (d) 0.025 m is added to the batch
stirred tank. They are placed on the side wall of the tank and in between two baffles
(Figure 1(b)). carboxymethyl cellulose (CMC) is taken as a working fluid with
consistency index, K, 0.0132 and flow index, n, 0.85 (Venneker et al., 2010).
For non-isothermal reactor, a temperature-dependent viscosity is considered. The thermal
conductivity () of the fluid is 0.7 W/(m K) (Denys et al., 2007).

2.2 Model equations

The governing equation of continuity, momentum and temperature are as follows (Ansys
Fluent 13: Theory guide, 2011):
G
+ ( ) = 0 (1)
t
 Computational study of entropy generation minimisation 19

G GG G
( ) + ( ) = P + ( ) + g (2)
t
T G
CP + CP ( T ) = K eff 2T v (3)
t
G
where is the velocity vector, T is the temperature, P is the static pressure, is the
G
stress tensor, g is the gravitational body force, K eff is the effective thermal
conductivity, CP is the heat capacity of the liquid at constant pressure and v is the
viscous dissipation function. The stress tensor is expressed as
G G 2 G
= ( + T ) . I (4)
3
where is the apparent viscosity and I is the unit tensor.

Figure 1 (a) Batch stirred tank and (b) Continuous stirred tank with inlet and outlet (see online
version for colours)

The turbulent flow, induced by the Rushton turbine, is modelled using realisable k
turbulence model. The transport equations of turbulence kinetic energy, k, and its rate of
dissipation, , are (Ansys Fluent 13: Theory guide, 2011) given as follows:

G
( k ) + ( kv ) = + t k + Gk (5)
t k
and

G 2
( ) + ( v ) = + t + C 2 (6)
t k +

where C2 is constant. k and are the turbulent Prandtl numbers for k and , respectively.
The values used for the constants are C2 = 1.92, k = 1.0, = 1.2.
20 A. Khapre and B. Munshi

In equation (5), Gk represents the generation of turbulence kinetic energy due to mean
velocity gradients and calculated as
u j
Gk = ui' u 'j (7)
xi

The turbulent or eddy viscosity, t, is computed by combining k and as follows (Ansys

Fluent 13: Theory guide, 2011):
k2
t = C = t S 2 (8)

where S is the modulus of the mean rate of strain tensor and expressed as

1 u j ui
S = 2 Sij Sij and Sij = + (9)
2 xi x j
The variable C is calculated as (Ansys Fluent 13: Theory guide, 2011):
1
C = (10)
kU
A0 + As

where U = Sij Sij + ij ij and ij = ij 2 ijk k

ij is the mean rate of rotation of tensor viewed in a moving reference frame with
angular velocity, k. The model constants A0 and As are given as:

A0 = 4.04, As = 6 cos

where
1 Sij S jk S ki
= cos 1
3
( )
6W , W =
S3
S = Sij Sij .

For isothermal non-Newtonian fluid, the power law is used to model viscosity. It is
given as:
= K  n 1 (11)

where K is the consistency index and n is the flow behaviour index.

For non-isothermal condition, the temperature and shear rate dependent viscosity is
given as (Ansys Fluent 13: Users Guide, 2011):
= K  n 1 H (T ) (12)

H(T) is the temperature-dependent part in the viscosity expression and is given in

equation (13) (Ansys Fluent 13: Users Guide, 2011):

1 1
H (T ) = exp (13)
T T0 T T0
 Computational study of entropy generation minimisation 21

where is the ratio of molar activation energy to universal gas constant. The value of is
2324.5 K (Converti et al., 1999). The reference temperature T is taken as 288.15 K. T0 is
the minimum temperature shift and taken as zero in the present work.
The transport equation for entropy is (Kock and Herwig, 2004) given as:
G

( s ) + ( sv ) = div + + 2
G q
(14)
t T T T
where s is the specific entropy and T is the temperature. The last two terms on the right
side are related to entropy generation: first term represents the entropy generation due to
viscous dissipation and second term describes the entropy generation due to heat transfer
by finite temperature gradient. The above equation is applicable for both laminar and
turbulent flow state. In turbulent state, according to Reynolds averaged NavierStokes
approach, the instantaneous entropy is decomposed into mean and fluctuation term.
The time-averaged equation of entropy is written as
G
t
( )
( s ) + ( sv ) = div ( q / T ) u s + / T + / T 2 (15)

/ T describes the entropy generation due to viscous dissipation and it has two parts:
mean viscous dissipation ( Sgen,VD ) and turbulent viscous dissipation ( Sgen,VD ). Thus,

/ T = Sgen,VD = Sgen,VD + Sgen,VD (16)

In terms of mean velocity and velocity fluctuation gradients, these can be written as

eff u v 1 w 2
2 2 2
1 v w
Sgen,VD = 2 + +u + + +
T r r 2 z z r (17)
2 2
w u 1 u v
+ + + + r
r z r r r

eff u v 1 w 2
2 2 2
1 v w
Sgen,VD = 2 + 2 + u + + +
T r r z z r
(18)
2 2
w u 1 u v
+ + + + r
r z r r r

where eff is the effective viscosity of fluid.

The term / T 2 in equation (15) is the entropy generation due to heat transfer
irreversibility. It also has two parts: entropy generation due to gradient of mean
temperature ( Sgen, HT ) and gradient of fluctuating temperature ( Sgen, HT ). Thus

/ T 2 = Sgen, HT = Sgen, HT + Sgen, HT (19)

In terms of mean temperature and temperature fluctuation gradients, these can be

written as
22 A. Khapre and B. Munshi

K T 2 1 T 2 T 2
Sgen, HT = eff2 + + (20)
T r r z

K eff T 2 1 T 2 T 2
Sgen, HT = + + (21)
T2 r r z

where Keff is the effective thermal conductivity.
The solution of turbulent momentum, k model equation and energy equation makes
the mean velocities and temperatures available. Therefore, both Sgen,VD and Sgen,HT can be
determined directly. But the remaining two terms are unknown and need additional
equations. The equations are given below (Kock and Herwig, 2004, 2005; Herwig and
Kock, 2007).

Sgen,VD = (22)
T
t
Sgen, HT = S (23)
gen, HT
where t is the turbulent thermal diffusivity and is thermal diffusivity of fluid.
From equations (15), (16) and (19), the total entropy generation due to both heat
transfer irreversibility and fluid viscous dissipation can be written as:
Sgen,Total = Sgen, HT + Sgen,VD = Sgen, HT + Sgen, HT + Sgen,VD + Sgen,VD (24)

Further, the total entropy generation can be written in dimensionless form as (Yilbas
et al., 1999; Basha et al., 2007):
Sgen,Total D 2
S* = (25)

To know the relative contribution of entropy generation due to viscous dissipation and
heat transfer irreversibility on the total entropy generation, the Bejan number, Be, is
defined as follows (Ko and Wu, 2009):
Sgen, HT
Be = (26)
Sgen,Total

Bejan number ranges from 0 to 1. When Be = 1, the irreversibility is because of the heat
transfer only, and for Be = 0, the irreversibility occurs only by the fluid viscous
dissipation.

2.3 Numerical methodology and boundary condition

The tank geometry is meshed approximately with 9000001500000 unstructured
tetrahedral cells. A very refined grid near impeller region and tank wall is developed. The
grid independence test was carried out comparing the present numerical data with the
experimental data of Venneker et al. (2010). The computational work finds that further
 Computational study of entropy generation minimisation 23

increase of meshing beyond 1200000 cells does not alter the computational results.
Hence, all the results shown in present study use 1200000 unstructured tetrahedral cells.
Multiple reference frame commonly known as MRF is used to model rotating
impeller. In this approach, the tank is divided into two domains: one is rotating domain
with impeller and the other is stationary domain with baffles and remaining part of the
tank. The rotating domain rotates with rotational speed of the impeller but impeller itself
remains stationary. Both the domains exchange values of flow parameters through a
common face known as interface. The impeller shaft outside MRF zone is considered as
moving wall. No-slip boundary condition is used on the walls, baffles and impellers.
The velocity at the inlet and pressure at the outlet of the tank are specified for the CSTR.
As the thermal boundary condition either a uniform surface temperature or constant heat
flux is imposed on the tank walls. At the inlet, the temperature of the working fluid is
kept at 298 K.
The finite volume method is used to discretise the governing equations. All transport
equations are discretised using second-order upwind difference scheme. A pressure-
correction method, SIMPLE (semi-implicate method for pressure linked equations),
is used to couple pressure and velocity equations. The turbulent flow is modelled
using realisable k turbulence model. Power law model is used for the viscosity of
the non-Newtonian fluid. In order to verify the convergence of the simulations, the
residuals as well as additional parameter such as the torque on the shaft are monitored.
The convergence criterion for each transport equation is set to 105.

Figure 2 Comparison of predicted velocity profile with experimental data

Source: Venneker et al. (2010)

2.4 Validation
It is necessary to validate CFD models to check the reliability and accuracy of numerical
methods. The CFD model validation for a stirred batch tank is done by comparing
24 A. Khapre and B. Munshi

velocity flow field with the experimental data of Venneker et al. (2010). The CMC is
used as a working fluid with the flow index n = 0.85. The specified rotation of the
impeller is 180 rpm. Figure 2 shows the comparison of velocities. All the velocities are
taken at the impeller disk level and made dimensionless using the tip velocity of
the impeller blade, Vtip = NDi . The dimensionless velocities are plotted against the
dimensionless radial distance r/Ri, where r is the radial position and Ri is the radius of the
impeller. Thus, r/Ri = 1 is the position of the impeller blade tip. The CFD predicted
velocity follows the trend of experimental data and shows an overall good agreement
with it except very near impeller region.

3 Results and discussion

To reach the objective of the present study, i.e. to design a CSTR using the EGM method,
the previous batch stirred tank is converted to CSTR by adding an inlet and outlet
flow pipes to the tank. The diameter and length of the inlet and outlet pipes are 0.025 m
and 0.03 m.
The numerical study consists of minimising the total entropy generation rate in the
stirred tank by varying tank parameters and keeping either constant wall temperature or
constant wall heat flux as the heating source for tanks liquid. For each thermal boundary
condition of the wall, the optimisation is involved with four systematic steps where the
entropy generation is minimised in each step with respect to the inlet Reynolds number
(Re), impeller rotation, impeller clearance and blade width.
Step 1: Entropy generation is minimised with respect to inlet fluid flow Reynolds
number, Re, using 180 rpm speed of the impeller, 0.33D impeller clearance and 0.042 m
impeller blade width.
Step 2: Entropy generation is minimised with respect to impeller rotation using optimised
inlet Re obtained in the step 1, 0.33D impeller clearance and 0.042 m impeller blade
width.
Step 3: Entropy generation is minimised with respect to impeller clearance using
optimised inlet Re, optimised impeller rotation and 0.042 m impeller blade width.
Step 4: Entropy generation is minimised with respect to impeller blade width using
optimised inlet Re, optimised impeller rotation, and optimised bottom clearance obtained
in step 3.

3.1 Constant wall temperature thermal boundary condition

3.1.1 Optimisation with respect to inlet Reynolds number
The dimensionless entropy generation in stirred tank at different boundary wall
temperature is shown in Figure 3. The optimum inlet Reynolds number is determined as a
function of wall temperature. Due to large temperature gradient near the wall, the entropy
generation is found high at very low inlet Re. The convective energy transport increases
with inlet Re, and it reduces the temperature gradient in the CSTR. Therefore, the entropy
generation due to heat transfer irreversibility decreases with increase in inlet Re.
However, the entropy generation due to frictional loss increases with increase in inlet Re.
The total entropy generation, therefore, passes through a minimal. On the left side of the
 Computational study of entropy generation minimisation 25

minimal point, the contribution of thermal entropy generation and on the right side of it
the contribution of viscous entropy generation to the total entropy generation is more.
The entropy generation is found higher at higher wall temperature. The figure also shows
that the magnitude of minimum entropy generation is higher for higher wall temperature.
At high temperature, the viscosity of the liquid is less and hence, the frictional loss is also
less. Therefore, the optimised inlet Re increases with the wall temperature. The optimised
inlet Reynolds numbers are 3600, 3400, 3200 and 3000 for the wall temperatures 353,
343, 333 and 323 K, respectively.
Figure 4 shows the Bejan number, Be, profile with inlet Re. Be is approximately
equal to 1.0 till inlet Re 3000, and it becomes 0.5 for inlet Re between 8000 and 10,000.
Thus, over a wide range of inlet Re, the thermal entropy generation is more than the
entropy generation due to viscous dissipation. Comparison of Figures 3 and 4 shows that
around the entropy minimisation point, the heat transfer irreversibility is much larger than
the frictional irreversibility. The linear and sharp variation of Be at higher inlet Re occurs
due to domination of the convective energy transport over the conductive heat transfer.
The higher wall temperature leads to larger temperature gradients, and hence higher
Bejan number is observed.

Figure 3 Dimensionless entropy generation vs. inlet Reynolds number at constant wall
temperature with 180 rpm of the impeller rotations, 0.33D impeller clearance and
0.042 m impeller blade width

3.1.2 Optimisation with respect to impeller rotations

The comparison of Figures 3 and 5 reveals that the entropy generation varies more with
the inlet Re than impeller rotations. It occurs due to larger variation of temperature
gradient of fluid inside the reactor with inlet Re. The less dependency of entropy
generation on the impeller rotations proves that the entropy generation shown in Figure 5
is affected mainly by the heat transfer irreversibility. The same is also observed in
26 A. Khapre and B. Munshi

Figure 6 where Be is closer to 1.0 at all impeller rotations. The optimum impeller
rotations are 104, 100, 91 and 85 rpm for the wall temperature 353, 343, 333 and 323 K,
respectively. Hence, a little increase in optimum impeller rotations is observed with the
temperature of the wall of the CSTR.

Figure 4 Bejan number vs. inlet Reynolds number at constant wall temperature with 180 rpm of
the impeller rotations, 0.33D impeller clearance and 0.042 m impeller blade width

Figure 5 Dimensionless entropy generation vs. impeller rotations at constant wall temperature
with optimised inlet Re, 0.33D impeller clearance and 0.042 m impeller blade width
 Computational study of entropy generation minimisation 27

Figure 6 Bejan number vs. impeller rotations at constant wall temperature with optimised
inlet Re, 0.33D impeller clearance and 0.042 m impeller blade width

3.1.3 Optimisation with respect to impeller clearance

The impeller clearance affects the flow and temperature distribution inside the CSTR.
The effect causes the variation of the entropy generation and it is shown in Figure 7.
The figure depicts that for wall temperature 353 K, the entropy generation is minimum at
impeller clearance, Ci = 0.33D. Whereas, for other wall temperatures, the minimum
entropy generation occurs at impeller clearance, Ci = 0.2D. The Bejan number
distribution is shown in Figure 8. It shows that the entropy generation at all the impeller
clearance is affected predominantly by the heat transfer irreversibility. At lower clearance
than 0.33D, the flow around the impeller forms single loop and it results in axially up
movement of fluid, whereas two loops are formed around the impeller and it gives better
mixing for Ci = 0.33D. Due to this, the temperature gradient for the latter case is higher
and hence, more entropy is generated due to the thermal irreversibility. Therefore, Be is
found maximum at Ci = 0.33D.

3.1.4 Optimisation with respect to impeller blade width

The variation of entropy generation with the impeller blade width is shown in Figure 9.
In case of Ci = 0.2D, the entropy generation is observed to decrease up to the width of
0.042 m and thereafter it becomes invariant with the width of the impeller. For the
smaller impeller blade width, the mixing of liquid inside the CSTR is low and hence,
there is more thermal irreversibility due to the presence of larger thickness of thermal
boundary layer. It can also be seen in the Bejan number distribution in Figure 10.
The convective energy transport reaches a saturated state if the impeller width is more
than 0.042 m. Hence, the entropy generation becomes constant at higher impeller blade
width. For the wall temperature 353 K, the impeller is mounted at higher position with
28 A. Khapre and B. Munshi

clearance Ci = 0.33D. The higher position makes sure a little dependency of heat transfer
irreversibility on the impeller blade width. Therefore, the entropy generation is found
almost invariant with the blade width of the impeller for Ci = 0.33D. Figure 9 also shows
that there is a specific optimum value of impeller blade width that corresponds to
minimum entropy generation and the optimum impeller blade width is 0.042 m.

Figure 7 Dimensionless entropy generation vs. impeller clearance at constant wall temperature
with optimised inlet Re, optimised impeller rotations and 0.042 m impeller blade width

Figure 8 Bejan number vs. impeller clearance at constant wall temperature with optimised
inlet Re, optimised impeller rotations and 0.042 m impeller blade width
 Computational study of entropy generation minimisation 29

Figure 9 Dimensionless entropy generation vs. impeller blade width at constant wall temperature
with optimised inlet Re, impeller rotations and impeller clearance

Figure 10 Bejan number vs. impeller blade width at constant wall temperature with optimised inlet
Re, impeller rotations and impeller clearance

Table 1 denotes the optimal tank parameters for the specified wall temperature boundary
condition.
30 A. Khapre and B. Munshi

Table 1 Optimal tank parameters for the specified wall temperature

Wall Inlet Reynolds Impeller Impeller Impeller blade

temperature (K) number rotations (rpm) clearance (m) width (m)
353 3600 104 0.33D 0.042
343 3400 100 0.20D 0.042
333 3200 91 0.20D 0.042
323 3000 85 0.20D 0.042

3.2 Constant wall heat flux thermal boundary condition

3.2.1 Optimisation with respect to inlet Reynolds number
The entropy minimisation study is also carried out at the specified constant wall
heat flux. The effect of inlet Re on the total entropy generation at specified wall
heat fluxes is shown in Figure 11. Due to higher temperature gradient, the entropy
generation is found more at higher wall heat flux. The entropy generation decreases,
reaches at minimum and further increases with the inlet Re. Thus, the minimum entropy
generation point is clearly observed on each curve. If inlet Re is lower than the optimum,
the entropy generation is significantly affected by the wall heat flux magnitude.
If inlet Re is more than the optimum, then the effect of heat flux on the entropy
generation is decreased and the entropy production becomes independent of the heat flux
at higher inlet Re. The explanation can be drawn from the corresponding Bejan number
distribution plot given in Figure 12. It shows that Be approaches zero with higher inlet
Re. At higher inlet Re, the convective energy transport dominates over thermal
conduction. Hence, the effect of wall heat flux on entropy generation becomes negligible.
The calculated optimum inlet Re is 2800, 2500, 2100 and 1700 for wall heat flux 3000,
2000, 1000 and 500 W/m2, respectively. The figure finds that the optimum inlet Re
increases linearly with the wall heat flux. It also shows that around the minimum entropy
generation point, Be is greater than 0.5 in all the cases, and hence, the effect of heat
transfer irreversibility on the total entropy generation is more than the frictional
irreversibility.

3.2.2 Optimisation with respect to impeller rotations

Figure 13 shows the effect of the rotational speed of the impeller on the entropy
generation. The minimum entropy is obtained at impeller rotations 79, 67, 55 and 45 rpm
for the respective wall heat fluxes of 3000, 2000, 1000 and 500 W/m2. The entropy
generation increases with heat flux and it leads to increase of the optimum impeller
rotations with the wall heat flux. The figure shows that the total entropy generation is
almost independent of the impeller rotations. The distribution of Be in Figure 14 depicts
that the contribution of thermal entropy generation decreases with the impeller speed and
the decline rate increases above 180 rpm. It happens due to substantial decrease in
thermal entropy generation and increase in frictional entropy generation. The figure also
finds that the decrease of entropy generation occurs at higher impeller rotations with the
increase in wall heat flux.
 Computational study of entropy generation minimisation 31

Figure 11 Dimensionless entropy generation vs. inlet Reynolds number at constant wall heat flux
with 180 rpm of the impeller rotations, 0.33D impeller clearance and 0.042 m impeller
blade width

Figure 12 Bejan number vs. inlet Reynolds number at constant wall heat flux with 180 rpm of the
impeller rotations, 0.33D impeller clearance and 0.042 m impeller blade width
32 A. Khapre and B. Munshi

Figure 13 Dimensionless entropy generation vs impeller rotation at constant wall heat flux with
optimised inlet Re, 0.33D impeller clearance and 0.042 m impeller blade width

Figure 14 Bejan number vs. impeller rotation at constant wall heat flux with optimised inlet Re,
0.33D impeller clearance and 0.042 m impeller blade width

3.2.3 Optimisation with respect to impeller clearance

The effect of impeller clearance on the total entropy generation is shown in Figure 15.
The minimum entropy generation occurs at impeller clearance, Ci = 0.33D, for all the
 Computational study of entropy generation minimisation 33

cases except for 500 W/m2 wall heat transfer flux where entropy generation is found
almost constant. The optima become more distinct at higher wall heat flux. The
corresponding Be distribution as shown in Figure 16 finds that the Be passes through the
maximum at impeller clearance of 0.33D for all the wall heat fluxes except the lowest
one. The physical explanations follow the same for the specified wall temperature case.

Figure 15 Dimensionless entropy generation vs. impeller clearance at constant wall heat flux with
optimised inlet Re, optimised impeller rotations and 0.042 m impeller blade width

Figure 16 Bejan number vs. impeller clearance at constant wall heat flux with optimised inlet Re,
optimised impeller rotations and 0.042 m impeller blade width
34 A. Khapre and B. Munshi

Figure 17 Dimensionless entropy generation vs. impeller blade width at constant wall heat flux
with optimised inlet Re, impeller rotations and impeller clearance

Figure 18 Bejan number vs. impeller blade width at constant wall heat flux with optimised
inlet Re, impeller rotations and impeller clearance

3.2.4 Optimisation with respect to impeller blade width

The effect of impeller blade width on the entropy generation at the specified heat flux is
shown in Figure 17. The minimum entropy generation point becomes prominent at higher
 Computational study of entropy generation minimisation 35

wall heat flux and it is obtained at blade width 0.042 m. The Bejan number distribution
with respect to blade width is shown in Figure 18, and the nature of its distribution is
found similar to Figure 16.
Table 2 denotes the optimal tank parameters for the specified wall heat flux boundary
condition.

Table 2 Optimal tank parameters for the specified wall heat flux

Wall heat flux Inlet Reynolds Impeller Impeller Impeller blade

(W/m2) number rotations (rpm) clearance (m) width (m)
3000 2800 79 0.33D 0.042
2000 2500 67 0.33D 0.042
1000 2100 55 0.33D 0.042
500 1700 45 0.33D 0.042

4 Conclusions

In this paper, the entropy generation of CSTR with non-Newtonian fluid has been
calculated, and their performances have been compared with respect to entropy
generation. Entropy generations in all the simulation work have been evaluated on the
satisfactorily meshed geometry and it has been computed as a function of temperature
and velocity gradients which were calculated using commercial CFD software, ANSYS
Fluent. Two cases have been demonstrated:
optimal design for the specified wall temperature
optimal design for the specified wall heat flux.
The heuristic technique has been used where basic design of the CSTR has been
improved systemically with respect to the system parameters, such as Reynolds number,
impeller rotations, impeller clearance and blade width. In each of the optimal parameter
findings, the optimal value of the parameter which produces minimum entropy generation
has been selected.
The total entropy generation is the sum of thermal and frictional entropy generations.
The thermal and frictional entropy generations have been found to decrease and increase,
respectively, with the inlet Reynolds number. The total entropy generation, therefore,
passed through minima. The entropy generations have been increased quite quickly, if the
inlet Reynolds number is less than the optimum values. But, it has been increased
gradually for inlet Re greater than the optimum value. The calculated optimal inlet
Reynolds numbers were 3600, 3400, 3200 and 3000 for the wall temperatures 353, 343,
333 and 323 K, respectively. Due to higher temperature gradient, the entropy generation
has been found more at higher wall heat flux. The calculated optimum inlet Re were
2800, 2500, 2100 and 1700 for wall heat flux 3000, 2000, 1000 and 500 W/m2,
respectively. The optimal inlet Reynolds number has been found to vary linearly with
both the specified wall temperature and wall heat flux. Around the entropy minimisation
point, the heat transfer irreversibility has been found much higher than the frictional
irreversibility. The Bejan number, Be, was found greater than 0.5 at optimal point.
36 A. Khapre and B. Munshi

Though the EGM point has been identified, the effect of impeller rotations on
the entropy generation has been found less significant as compared to the inlet Re. The
optimum impeller rotations were 104, 100, 91 and 85 rpm for the wall temperature 353,
343, 333 and 323 K, respectively. These were 79, 67, 55 and 45 rpm for the respective
wall heat fluxes of 3000, 2000, 1000 and 500 W/m2. For impeller rotations lower than the
optimum value, the variation of the entropy generation with impeller rotations has been
found almost negligible. However, for higher impeller rotations than the optimum,
the variation of the entropy generation has been found more for constant wall temperature
than constant wall heat flux boundary problem. At high impeller rotations, the Bejan
number has been affected more by a frictional pressure drop in case of constant wall heat
flux problem.
The optimal impeller clearance position has been found to be dependent on the wall
temperature. It has been found as 0.33D for the specified wall temperature 353 K,
whereas for others it has been found as 0.2D. In the case of constant wall heat flux, the
optimum impeller clearance has been found 0.33D except for wall heat flux 500 W/m2
where entropy generation has been found invariant with impeller clearance. Although the
trend of Bejan number distribution has been found similar in both the cases,
the contribution of viscous dissipation to entropy generation is relatively more for the
constant wall heat flux boundary problem.
The optimum impeller blade width depends on the magnitude of wall temperature and
wall heat flux. In case of constant wall temperature, a large variation of the entropy
generation occurs, if the blade width of the impeller is less than the optimum. But,
it becomes almost insensitive on the width, if the used impeller blade width is more than
the optimum. At low heat flux, the entropy generation does not vary much with the
impeller blade width. However, at high wall heat flux, a clear minimum entropy
generation point appears. In case of constant wall temperature boundary condition, the
thermal entropy generation dominates over the entropy generation due to viscous energy
dissipation as Be is closer to 1.0 for the entire range of the impeller blade width.
Whereas, for constant wall heat flux condition, Be appears close to 0.5 for certain range
of impeller blade width where both the thermal and viscous dissipation entropy
generation have equal contribution to the total entropy generation.

References
Adesanya, S.O. and Makinde, O.D. (2014) Entropy generation in couple stress fluid flow
through porous channel with fluid slippage, International Journal of Exergy, Vol. 15, No. 3,
pp.344362.
Aghbolaghy, M. and Karimi, A. (2014) Simulation and optimization of enzymatic hydrogen
peroxide production in a continuous stirred tank reactor using CFDRSM combined method,
Journal of the Taiwan Institute of Chemical Engineers, Vol. 45, pp.101107.
Ahmadi, P., Hajabdollahi, H. and Dincer, I. (2011) Cost and entropy generation minimization of a
cross-flow plate fin heat exchanger using multi-objective genetic algorithm, Journal of Heat
Transfer, Vol. 133, No. 2, pp.021801/110.
Alebrahim, A. and Bejan, A. (1999) Entropy generation minimization in a ram-air cross-flow heat
exchanger, International Journal of Applied Thermodynamics, Vol. 2, No. 4, pp.145157.
Al-Zaharnah, I.T. and Yilbas, B.S. (2004) Thermal analysis in pipe flow: influence of variable
viscosity on entropy generation, Entropy, Vol. 6, pp.344363.
Ansys Fluent 13: Theory Guide (2011) Ansys Inc., USA.
 Computational study of entropy generation minimisation 37

Asadi, M. (2013) Entropy minimization in plate-fin heat exchanger using Cuckoo algorithm,
Wyno Journal of Engineering & Technology Research, Vol. 1, No. 2, pp.2129.
Ayinde, T.F. and Yilbas, B.S. (2004) Entropy generation due to non-Newtonian flow around a
spherical particle at low Reynolds number, International Journal of Exergy, Vol. 1, No. 2,
pp.256267.
Balaji, C., Hlling, M. and Herwig, H. (2007) Entropy generation minimization in turbulent mixed
convection flows, International Communications in Heat and Mass Transfer, Vol. 34, No. 5,
pp.544552.
Basha, M., Al-Qahtani, M. and Yilbas, B.S. (2007) Entropy generation in a rotating channel,
Proceedings of the Institution of Mechanical Engineers Part A: Journal of Power and Energy,
Vol. 221, pp.291299.
Bejan, A. (1979) A study of entropy generation in fundamental convective heat transfer, Journal
of Heat Transfer, Vol. 101, No. 4, pp.718725.
Bejan, A. (1995) Entropy Generation Minimization, CRC Press, New York.
Chaibakhsh, N., Rahman, M.B.A., Vahabzadeh, F., Abd-Aziz, S., Basri, M. and Salleh, A.B.
(2010) Optimization of operational conditions for adipate ester synthesis in a stirred tank
reactor, Biotechnology and Bioprocess Engineering, Vol. 15, pp.846853.
Chen, M.L. and Wang, F.S. (2010) Optimization of a fed-batch simultaneous saccharification and
cofermentation process from lignocellulose to ethanol, Industrial & Engineering Chemistry
Research, Vol. 49, pp.57755785.
Cimpean, D.S. and Pop, I. (2011) A study of entropy generation minimization in an inclined
channel, WSEAS Transactions on Heat and Mass Transfer, Vol. 6, No. 2, pp.3140.
Converti, A., Zilli, M., Arni, S., Felice, R.D. and Borghi, M.D. (1999) Estimation of viscosity of
highly viscous fermentation media containing one or more solutes, Biochemical Engineering
Journal, Vol. 4, pp.8185.
Culham, J.R. and Muzychka, Y.S. (2001) Optimization of plate fin heat sinks using entropy
generation minimization, IEEE Transactions on Components and Packaging Technologies,
Vol. 24, No. 2, pp.159165.
Demirel, Y. and Kahraman, R. (2000) Thermodynamic analysis of convective heat transfer in an
annular packed bed, International Journal of Heat and Fluid Flow, Vol. 21, pp.442448.
Denys, S., Pieters, J. and Dewettinck, K. (2007) CFD analysis of thermal processing of eggs, in
Sun D.W. (Ed.), Computational fluid dynamics in food processing, CRC press, Taylor and
Francis group, pp.347380.
Driss, Z., Karray, S., Chtourou, W., Kchaou, H. and Abid, M.S. (2012) A study of mixing
structure in stirred tanks equipped with multiple four-blade Rushton impellers, The Archive of
Mechanical Engineering, Vol. 59, No. 1, pp.5372.
Galanis, N. and Rashidi, M.M. (2012) Entropy generation in non-Newtonian fluids due to heat
and mass transfer in the entrance region of ducts, Heat and Mass Transfer, Vol. 48,
pp.16471662.
Guo, J., Cheng, L. and Xu, M. (2010) The entropy generation minimisation based on the revised
entropy generation number, International Journal of Exergy, Vol. 7, No. 5, pp.607626.
Guo, J., Xu, M., Cai, J. and Huai, X. (2011) Viscous dissipation effect on entropy generation in
curved square microchannels, Energy, Vol. 36, pp.54165423.
Haddad, O., Abuzaid, M. and Al-Nimr, M. (2004) Entropy generation due to laminar
incompressible forced convection flow through parallel-plates microchannel, Entropy, Vol. 6,
No. 5, pp.413426.
Herwig, H. and Kock, F. (2007) Direct and indirect methods of calculating entropy generation
rates in turbulent convective heat transfer problems, Heat and Mass Transfer, Vol. 43,
pp.207215.
38 A. Khapre and B. Munshi

Hung, Y.M. (2008) Viscous dissipation effect on entropy generation for non-Newtonian fluids
in micro channels, International Communications in Heat and Mass Transfer, Vol. 35,
pp.11251129.
Jankowski, T.A. (2009) Minimizing entropy generation in internal flows by adjusting the shape of
the cross-section, International Journal of Heat and Mass Transfer, Vol. 52, Nos. 1516,
pp.34393445.
Kahraman, A. and Yurusoy, M. (2008) Entropy generation due to non-Newtonian fluid flow in
annular pipe with relative rotation: constant viscosity case, Journal of Theoretical and
Applied Mechanics, Vol. 46, No. 1, pp.6983.
Khan, W.A. and Gorla, R.S.R. (2012) Second law analysis for free convection in non-Newtonian
fluids over a horizontal plate embedded in a porous medium: (prescribed heat flux), Brazilian
Journal of Chemical Engineering, Vol. 29, No. 3, pp.511518.
Ko, T.H. and Wu, C.P. (2009) A numerical study on entropy generation induced by turbulent
forced convection in curved rectangular ducts with various aspect ratios, International
Communications in Heat and Mass Transfer, Vol. 36, No. 1, pp.2531.
Kock, F. and Herwig, H. (2004) Local entropy production in turbulent shear flows: a high-
Reynolds number model with wall functions, International Journal of Heat and Mass
Transfer, Vol. 47, pp.22052215.
Kock, F. and Herwig, H. (2005) Entropy production calculation for turbulent shear flows and
their implementation in CFD codes, International Journal of Heat and Fluid Flow, Vol. 26,
pp.672680.
Koo, J. and Kleinstreuer, C. (2004) Viscous dissipation effect in microtubes and microchannels,
International Journal of Heat and Mass Transfer, Vol. 47, pp.31593169.
Kucuk, H. (2011) Numerical analysis of entropy generation and minimisation in eccentric curved
annular ducts, International Journal of Exergy, Vol. 9, No. 1, pp.4065.
Mahmud, S. and Fraser, R.A. (2002) Inherent irreversibility of channel and pipe flows for
non-Newtonian fluids, International Communications in Heat Mass Transfer, Vol. 29,
pp.577587.
Manzi, J. and Carrazzoni, E. (2008) Analysis and optimization of a CSTR by direct entropy
minimization, Journal of Chemical Engineering of Japan, Vol. 41, No. 3, pp.194199.
Manzi, J., Vianna, R. and Bispo, H. (2009) Direct entropy minimization applied to the production
of propylene glycol, Chemical Engineering and Processing: Process Intensification, Vol. 48,
No. 1, pp.470475.
McClintock, F.A. (1951) The design of heat exchangers for minimum irreversibility, 1951 ASME
Annual Meeting, Paper No. 51-A-108, Atlantic City, New Jersey, USA.
Naterer, G.F. and Adeyinka, O.B. (2009) Imaging velocimetry measurements for entropy
production in a rotational magnetic stirring tank and parallel channel flow, Entropy, Vol. 11,
pp.334350.
Pakdemirli, M. and Yilbas, B.S. (2006) Entropy generation in a pipe due to non-Newtonian fluid
flow: Constant viscosity case, Sadhana, Vol. 31, No. 1, pp.2129.
Peters, G.W.M. and Baaijens, F.P.T. (1997) Modelling of non-isothermal viscoelastic flows,
Journal of Non-Newtonian Fluid Mechanics, Vol. 68, Nos. 23, pp.205224.
Petrov, M. and Tzonkov, St. (2009) Multiple objective optimization of a mass transfer in stirred
tank bioreactors, Bioautomation, Vol. 13, No. 4, pp.173184.
Powell, E.E. and Hill, G.A. (2008) Optimization of continuously stirrer tank bioreactor design for
cost minimization: effect of microbial species and operating conditions, International Journal
of Chemical Reactor Engineering, Vol. 6, No. 1, Article A20, DOI: 10.2202/1542-6580.1642.
Sahin, A.Z. (1999) The effect of variable viscosity on the entropy generation and pumping
power in a laminar fluid flow through a duct subjected to constant heat flux, Heat and Mass
Transfer, Vol. 35, pp.499506.
 Computational study of entropy generation minimisation 39

Sahin, A.Z. (2002) Entropy generation and pumping power in a turbulent fluid flow through a
smooth pipe subjected to constant heat flux, Exergy, Vol. 2, pp.314321.
Sahin, A.Z. (2014) A simple method of determining entropy generation rate in viscous fluid flow
through ducts, Arabian Journal for Science and Engineering, Vol. 39, pp.12411249.
Sahin, Z.A. and Mansour, B.R. (2003) Entropy generation in laminar fluid flow through a circular
pipe, Entropy, Vol. 5, pp.404416.
Sciubba, E. (2010) Entropy generation minima in different configurations of the branching of a
fluid-carrying pipe in laminar isothermal flow, Entropy, Vol. 12, pp.18551866.
Shojaeian, M. and Kosar, A. (2014) Convective heat transfer and entropy generation analysis on
Newtonian and non-Newtonian fluid flows between parallel-plates under slip boundary
conditions, International Journal of Heat and Mass Transfer, Vol. 70, pp.664673.
Singh, V., Aute, V. and Radermacher, R. (2008) Usefulness of entropy generation minimization
through a heat exchanger modeling tool, Paper presented at International Refrigeration and
Air Conditioning Conference, Paper 958, 1417 July, Purdue, http://docs.lib.purdue.edu/
iracc/958
Sivapathasekaran, C. and Sen, R. (2013) Performance evaluation of an ANNGA aided
experimental modeling and optimization procedure for enhanced synthesis of marine
biosurfactant in a stirred tank reactor, Journal of Chemical Technology and Biotechnology,
Vol. 88, pp.794799.
Song, X., Zhang, M., Wang, J., Li, P. and Yu, J. (2010) Optimization design for DTB industrial
crystallizer of potassium chloride, Industrial & Engineering Chemistry Research, Vol. 49,
pp.1029710302.
Tasnim, S.H. and Mahmud, S. (2002) Entropy generation in a vertical concentric channel with
temperature dependent viscosity, International Communications in Heat and Mass Transfer,
Vol. 29, No. 7, pp.907918.
Thiel, G.P. (2012) Entropy Generation Minimization of a Heat and Mass Exchanger for Use in a
Humidification-dehumidification Desalination System, Masters thesis, Department of
Mechanical Engineering, Massachusetts Institute of Technology, USA.
Vaneshani, S. and Jazayeri-Rad, H. (2011) Optimized fuzzy control by particle swarm
optimization technique for control of CSTR, World Academy of Science, Engineering and
Technology, Vol. 5, No. 11, pp.1126.
Venneker, B., Derksen, J. and Van den Akker, H.E.A. (2010) Turbulent flow of shear-thinning
liquids in stirred tanks-The effects of Reynolds number and flow index, Chemical
Engineering Research and Design, Vol. 88, No. 7A, pp.827843.
Yilbas, B.S, Shuja, S. and Budair, M. (1999) Second law analysis of a swirling flow in a
circular duct with restriction, International Journal of Heat and Mass Transfer, Vol. 42,
pp.40274041.
Yilbas, B.S., Yurusoy, M. and Pakdemirli, M. (2004) Entropy analysis for non-Newtonian fluid
flow in annular pipe: constant viscosity case, Entropy, Vol. 6, pp.304315.
Zimparov, V. and Penchev, P. (2003) Entropy generation minimization of tubes with
threedimensional surface roughness for shell-and-tube heat exchangers, Proceedings of the
First International Exergy, Energy and Environment Symposium, 1317 July, Izmir, Turkey,
http://yops-bg.com/files/pdf/13.pdf

Nomenclature
b Impeller blade width (m)
Be Bejan number
Ci Impeller clearance (m)
d Inlet and outlet diameter of pipe (m)
40 A. Khapre and B. Munshi

D Tank diameter (m)

Di Impeller diameter (m)
H Tank height (m)
K Consistency index (kg sn1/ m)
Keff Effective thermal conductivity (W/(m2 K))
n Flow behaviour index
N Impeller rotational speed (rps)
Rei Inlet Reynolds number
Sgen,Total Total entropy generation rate (W/K m3)
Sgen,HT Entropy generation rate due to temperature gradient (W/K m3)
Sgen,VD Entropy generation rate due to fluid viscous dissipation (W/K m3)
S* Dimensionless Entropy generation
T Temperature (K)
s Specific entropy (J/kg K)
Vtip Impeller tip velocity (m/s)
Wb Baffles width (m)
Greek letters
Viscosity (Pas)
Apparent viscosity (Pas)
eff Effective viscosity (Pas)
Density (kg/m3)
Thermal conductivity (W/(mK))
Dissipation due to turbulence kinetic energy (m2/s3)
Thermal diffusivity (m2/s)
t Turbulent thermal diffusivity (m2/s)
v Viscous dissipation function (s2)