
126 APPLICATIONS IN FLUID DYNAMICS
(f) What is the criterion for determining whether the scaling in part (c) or in
part (e) is appropriate for the describing equations?
(g) Use the results of your scaling analysis in part (e) to estimate the velocity
gradients in both liquids 1 and 2; that is, develop appropriate scales for the
velocity gradients in each liquid.
(h) For the scaling analysis done in part (e), determine the dimensionless groups
that are necessary to correlate the total rate of entropy production
˙
S,which
is given by
˙
S =−
H
1
0
μ
1
T
du
x
d
y
2
WLdy −
H
2
H
1
μ
2
T
du
x
d
y
2
WLdy (3.P.21-2)
where T is the absolute temperature and W and L denote the width and
length, respectively, of the two flat plates.
(i) One means for reducing the viscous drag at the walls experienced in pumping
viscous liquids such as petroleum is to inject a less viscous immiscible liquid
such as water, which will form a layer at the wall. However, for this idea
to work, we have to establish that the less viscous liquid (e.g., water) rather
than the more viscous liquid (e.g., petroleum) will go to the wall region.
This question can be answered by invoking the principle that continuous
steady-state processes seek a state of minimum entropy production. Use this
principle along with the results of your scaling analysis in part (e) to deter-
mine whether the viscous or the less viscous liquid will go to the wall region.
3.P.22 Laminar Cylindrical Jet Flow
In Example Problem 3.E.8 we considered a jet of an incompressible Newtonian
liquid with constant physical properties issuing from a circular orifice with an
initial velocity U
0
and falling vertically under the influence of gravity in an inviscid
gas as shown in Figure 3.E.8-1. We scaled the describing equations to explore the
conditions for which quasi-parallel flow can be assumed; that is, when the axial
velocity profile can be assumed to depend only on the axial coordinate. In scaling
the describing equations we introduced a scale for the radial derivative of the axial
velocity since we did not anticipate that this derivative would scale as the ratio
of the characteristic axial velocity scale divided by the characteristic radial length
scale. We anticipated the need to scale this derivative with its own scale since the
axial velocity does not change significantly across the jet. It was stated that if we
had scaled this radial derivative with the ratio of the characteristic axial velocity
scale to the characteristic radial length scale, the forgiving nature of scaling would
have indicated a contradiction. To better understand what is meant by this, let us
assume (incorrectly!) that the radial derivative of the axial velocity scales as the
axial velocity scale u
zs
divided by the radial length scale r
s
. Show that this leads to
an inconsistency in the resulting dimensionless equations in that the dimensionless
group multiplying one term in the describing equations becomes very large, whereas