C CA AL LC CU UL LA AT TI IN NG G L LI IQ QU UI ID D R RE EG GI IM ME ES S

Blazo Ljubicic, Brown Fintube-Koch Company, USA, describes the flow

of non-Newtonian fluids in heat exchangers.

It is taken for granted that reference to liquids and solids leaves no
ambiguity about what these terms actually mean. A liquid is
material that flows under its own weight, whilst a solid suffers
from a chronic desire to retain its original shape, and when
deformed, returns to its initial configuration. A liquid finds its own
level when it fills a bottle, while a solid bounces when dropped on
the floor. These would be acceptable explanations and most would
agree with these definitions.
For many years, Hook's law for solids and Newton's law for
liquids were considered universal. In classical mechanics, material
properties were considered known when either viscosity or
modulus of elasticity were given; the basic 'constitutive equations'
governing the behavior of each small material element were known.
Practical analysis often involved coupling these equations with the
momentum equation and applying them to the increasingly
complicated flows or deformations. However, dealing with solids
and liquids seemed deceptively simple. As nature is never so
straightforward, this was a gross oversimplification of a complex
situation. In virtually all industries including petroleum, foods,
pharmaceuticals, minerals, etc, many materials cannot be strictly
classified as liquids or solids. The science that deals with these
complex materials is called 'Theology', the study of the deformation
and flow of matter.

Background

To determine the details of the flow behavior of a complex
materials classification on the basis of phase structure alone is not
adequate. The way in which the fluid responds to the application
of stress must also be considered. The application of shearing
stress to fluids is by far the most important factor to be taken into
account in the characterization of fluids.
If the shearing stress imposed on the fluid is T= F/A, then fluid is
subjected to strain at the rate y= dV/dr, where y is the velocity
gradient or the rate of shear. It is reasonable to expect a
relationship between the imposed shearing stress, T, and the
resulting rate of shear, y. Experience confirms these quantities are
indeed related, but the relationship is complex. It is different not
only for different fluids, but also for the same fluid under different
temperature and/or pressure conditions.
The relationship

is known as a constitutive or rheological equation of a given fluid at
any given pressure and temperature.
Although extensive research has been made into an exact
relationship among the variables which characterize the theological
behavior of complex fluids, practical design
guidelines derived from these studies have been limited. Exact
methods exist only for the simplest case, the Newtonian fluid.
One of the major problems is the difficulty of separating,
either theoretically or experimentally, the independent effect
of the many influencing factors. The challenges in this field
are to relate the motion of the bulk fluid to the dynamics of
individual fluid particles, and to be able to predict how these
fluids will flow in complicated geometries.

Classification
Several types of mechanical responses are possible when
shearing stress is applied to a material. The strain may be
small and completely recoverable upon removal of stress.
This response is typical of elastic solids. The material may
undergo permanent deformation or strain, which is not
recovered or dissipated upon removal of the stress. This is
typical of plastic deformation of solids. The deformation may
be only temporary and disappear immediately upon removal
of the stress, allowing the material to flow readily, and to
assume the shape of its container. This response is typical of
true liquids. Some materials may exhibit a combination of
these types of response, sometimes appearing to behave as
elastic solid, sometimes as a liquid. Those materials are
classified as visco-elastic liquids, or as elasti-viscous semi-
solids.
Fluids are classified as various rheological types on the
basis of the form of their rheological equations or rheograms.
The major division is between fluids that are described as
purely viscous and those that exhibit both viscous and elastic
properties. Purely viscous fluids are those, which on removal
of the shearing force do not completely recover from any
deformation they may have undergone under its action. These
fluids may be divided further into the categories of time
independent and time dependent. Fluids in the first category
exhibit development of structure or orientation or particles,
attaining an equilibrium condition immediately. The time
dependent category includes fluids that have lower rates of
structure or orientation development; the time of structure
decay to attain equilibrium condition is measurable. Thus,
their behavior depends both upon the rate constants and
duration of imposed shear.
No general equation is available, relating the shear rate,
shear stress, thermodynamic variables, composition variables
and time. Simplified empirical flow models are used instead to
fit experimental observation and to relate the shear stress and
shear rate data. For fluids that exhibit non-Newtonian or
shear dependent flow behavior, the shear stress is not
directly proportional to the shear rate . Shear stress is
defined as the ratio of the shearing force to the area on which
( ) ( ) 1 f
CALCULATING LIQUID REGIMES
HYDROCARBON ENGINEERING APRIL 1999 Palladian Publications 1999
( ) 8
L 128
p D
Q
4



it is applied. The apparent viscosity is defined as the ratio of wall
shear stress to the rate of shear applied.

Viscometers
Viscometers are used to measure the deformation or strain that
occurs to the fluid when a force or a shearing action is applied to it
and thus determine flow behavior. Determination of Theological
properties at the lower shear range is best accomplished using a
rotational viscometer. Measurement of the torque required to
impart an angular velocity to the inner cylinder, corresponding to
some rate of shear, gives the shear stress required. Measurement
of flow properties at high shear rate with rotational viscometers
such as concentric cylinder viscometer become very cumbersome
as the frictional heat buildup becomes significant.
Capillary tube viscometers, such as extrusion rheometers,
avoid this problem by once through flow of the fluid, which carries
out the built up heat. While weight-driven capillary flow testers
(better known as melt indexes) provide quick control for molten
thermoplastic, extrusion rheometers provide viscosity vs. shear
rate information over a range of shear rates. Extrusion rheometers
use a pressurized vessel to force the fluid through a long, smooth
cylindrical capillary tube of known dimensions. The frictional
pressure drop associated with the laminar flow of the fluid and the
corresponding flow-rate is measured. To determine the flow
curve, several measurements are necessary at different
pressures and length-to-diameter ratios of the capillary tube.
Flow curves are normally used for the graphical description of
flow behavior. They are either manually drawn or produced by the
software coming with the measuring instrument.
For a fluid in steady laminar flow within a tube, viscous force
tending to retard flow will be exactly balanced by force resulting
from the pressure difference between the two ends. Therefore,
the shear stress at any point in the capillary tube is directly
proportional to the distance from the center and to the pressure
gradient. At the
wall, the shear stress is,


The shear rate varies along the tube radius, and its nonlinearity
depends on the nature of the fluid. Hence, direct experimental
measurement is not possible as in the case of rotational
viscometers.

Power law fluids
A power law model, where the rate of shear is proportional to
the power of the shearing stress, describes flow behavior of many
non-Newtonian fluids and particularly of many slurries. When the
rate of shear increases more than in proportion to the shearing
stress, the fluid is called pseudo-plastic or shear thinning. For the
case of a dillatant or shear thickening fluid, the rate of shear
increases less than in proportion to the shearing stress.
The main reason for the wide use of the power law model is that
the Theological behavior of a fluid, in a log shear-rate/shear-stress
graph, is represented by a straight line. The basic equations for
calculating the pressure drop and shear rate in pipes for power law
non Newtonian fluids appear as follows:


where,
n - is the flow factor
k - is the consistency index
The flow factor and consistency index are not true constants,
but normally remain unchanged over a range of shear rate of
practical interest.
The power law relation can also be defined in terms of the
mean flow shear rate .

where,

The parameter n' is the slope of the log-log plot of
w

versus 8V/D and must not be confused with the flow factor,
n, in equation 3. In the most general case, k' and n' are not
constants, but vary with 8V/D.
It has been shown that over the range of shear rates for
which the power law applies with constant values of k and
n, n' and n are identical. Under these conditions, a plot of
log(
w
) vs. log () is linear, and n' is constant over this
range.
Mean shear rate and shear stress are related by the
coefficient of proportionality known as the apparent
viscosity. Unlike Newtonian viscosity, this apparent
viscosity is not a
constant and is a function of the shear rate.
Apparent viscosity is determined from measurement of the
pressure drop in the capillary tube and the associated flow
rate; shear stress and shear rate are calculated at the same
point in the capillary i.e. at the wall.
Integrating the velocity distribution over the cross section of
the tube, the volumetric flowrate becomes:

For n=1 (Newtonian fluid), this equation reduces to the
Hagen-Poisseuille relation.


The power law parameters, k and n can be determined
by fitting experimental measurement of P and Q to a log
form of equation 7, or

Similarly, parameters k' and n' are also estimated by
fitting same experimental data to a log form of equation 4
( ) 7
2
D
k L 2
p
1 n 3
n
Q
n
1 n 3
n
1 +

,
_

,
_

,
_

+

( ) 2
L 4
D
p
w

( ) 3 n k
( ) 4 ' k
'
n
w

( ) 5
32 8
3
D
Q
D
V
dr
du

,
_

( ) 6
app

( ) ( ) 9
L
p
log
n
1
2
D
1 n 3
n
k 2
1
log Q log
n
1 n 3
n
1

,
_

+
1
]
1

,
_

+

,
_

+
CALCULATING LIQUID REGIMES
HYDROCARBON ENGINEERING APRIL 1999 Palladian Publications 1999


Parameters n and n' obtained from equations 9 and 10
will be identical for the range of shear rate over which
power law parameters are unchanged. However, k and k'
are not equal and their relation is given by:

Apparent viscosity at the wall is calculated from the
obtained wall shear stress and wall shear rates as:

where the wall and mean shear rate relation is given by:

( ) 11
' n 4
1 ' n 3
k ' k
n

,
_

+

( ) ( ) ( ) ( ) 10 log ' n ' k log log
w
+
( ) 12 w k
w
1 n
w
app

( ) 13
' n
1 ' n 3
w

+

CALCULATING LIQUID REGIMES
HYDROCARBON ENGINEERING APRIL 1999 Palladian Publications 1999

For the laminar flow, it is not necessary to calculate the
slope of the experimentally determined log-logplot of
t
w
. versus 8V/D. In a case where the process conditions are
within the range of experimental data, the plot can be used
directly. If the flow is not laminar or turbulent, or if the
experimental data were not determined over the range of
process shear rat es, then the parameters n' and k' are not
valid.

Friction factor
Laminar flow regime
Substituting equation 4 into the friction factor relation as
defined for Newtonian fluids gives:

where,

is a generalized Reynolds number valid for all time
independent fluids. In the special case, for n'=1 and
k'= , equation 15 reduces to the well known Reynolds
number for a Newtonian fluid.

Transition flow regime
For Newtonian fluids, the beginning of the transition flow
regime is determined by calculating the Reynolds number
and comparing it to the empirically determined value of
Below the Reynolds number of 2100, the flow is considered
laminar. Metzner and Reed, and Dodge and Metzner used
the same approach to establish the criteria for the beginning
of the transition zone for non-Newtonian fluids. They
found that at n' values of 1.0, 0.726, and 0.380, the
transition region begins at R'
e
of 2100, 2700, and 3100
respectively. Others suggested similar values for the
beginning of turbulence in non-Newtonian fluids, but the
conservative recommendation still remains R'
e
= 2100. The
friction factor in the transition flow regime is normally
prorated between the limiting values for laminar and
turbulent zones.

Turbulent flow regime
The friction factor for the time independent, inelastic, non
Newtonian fluids in the turbulent flow can be calculated
from:

The constants in equation 16 were determined
experimentally. For n'=l, friction factor from Eq. (16)
reduces to the von Karman equation for Newtonian flow.

If the power law parameters n' and k' vary with 8V/D,
then to calculate f, iterative procedure is required. The
procedure convergence is fast, and for a given flow rate, pipe
diameter, and tube length is as follows.

If the pressure drop p' calculated from the step 7 does
not agree with the assumed value p, another iteration is
needed.
Conclusion
Knowledge of the Theological properties of a fluid is
essential for designing industrial plants and processes in
which non-Newtonian fluids are used. These industrially
important fluids are Theologically complex giving rise to
processing problems, especially in the heat transfer and
mixing equipment. Engineers often use simplified
Newtonian fluid models to predict its behavior. Simple
models are convenient and in some cases their use can be
justified. However, this is not always possible. Simple
models may be theoretically correct, but they are not real.
Engineers dealing with non-Newtonian fluids should always
be aware that the model is a framework and that in general
the simplicity of the model is adversely proportional to its
analytical depth.

References
1 DODGE D.W, METZNER A.B., AlChE, 5,189,1955.
2. METZNER A.B., Reed J.C., AlChE, 1, 434,1957.
3 COGSWELL RN, Polymer Melt Theology - A Guide for Industrial
Practice, George Goodwin Ltd. In association with The Plastic and
Rubber Institute., Wiley & Sons, New York, 1981.
4 DOI M., EDWARDS S. F., The Theory of Polymer Dynamics, Clarendon
Press, Oxford, 1986.
5. CASSON W., A flow equation for Pigment-Oil Suspensions of the
printing ink type, Theology of Dispersed Systems, ed. C.C. Mill,
Pergamon, pp 84, 1959.
6 TSAI S.C. and KNELL E.W., Theology and its Effect on Atomization of
Coal Water Slurry, Proceedings of First Annual Pittsburgh Coal
Conference, Sept. 17-21, Pittsburgh, PA, 1984.
7 CHEREMISINOFF N.P. and GUPTA R., Handbook of Fluids in
Motion, Ann Arbor Science Publishers, Ann Arbor, Ml, p. 135-77,
1983.
8 SKELLAND A.H.P., Non-Newtonian Flow and Heat Transfer; Wiley,
( ) 14
' R
16
f
e

( ) 15
8 ' k
V D
e ' R
1 ' n
'
n 2 ' n

( ) 16
' n
4 . 0
f ' R log
' n
0 . 4
f
1
2 . 1
2
' n
1
e
75 . 0

,
_

( ) ( )
( ) ( )

,
_

,
_

,
_

2
V
D
L
f 4 ' p : p drop pressure Calculate
' R
16
f : f factor friction Calculate
8 ' k
V ' D
e ' R : ' R number ynolds Re Calculate
D
V 8
n 1 ' n n 1 ' k n 1 : from ' k Calculate
D
V 8
n 1 d
n 1 d
' n : from ' n Find
L 4
p D
: as stress shear wall Calculate
p of value a Assume
2
e
1 ' n
' n 2 n
e
w
w
w
CALCULATING LIQUID REGIMES
HYDROCARBON ENGINEERING APRIL 1999 Palladian Publications 1999
New York, 1967.
9. VAN WAZER J.R., LYONS,J.W., KIM K.Y., and COLWELL R.E.,
Viscosity and Flow Measurement-A laboratory Handbook;
Interscience, 1973.
10. BOX G.E.P., HUNTER W.G. and HUNTER J.S., Statistics for
Experimenters; Wiley, 1978.