ANALYSIS AND SUPPRESSION OF INSTABILITIES …ANALYSIS AND SUPPRESSION OF INSTABILITIES IN...
Transcript of ANALYSIS AND SUPPRESSION OF INSTABILITIES …ANALYSIS AND SUPPRESSION OF INSTABILITIES IN...
ANALYSIS AND SUPPRESSION OF
INSTABILITIES IN VISCOELASTIC FLOWS
By
Karkala Arun Kumar
A DISSERTATION SUBMITTED IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
Doctor of Philosophy
(Chemical Engineering)
at the
UNIVERSITY OF WISCONSIN – MADISON
2001
c© Copyright 2001 by Karkala Arun Kumar
All Rights Reserved
i
Abstract
The viscoelastic character of polymer solutions and melts gives rise to instabilities that are
not seen in the flows of Newtonian liquids. In industrial applications such as coating and
extrusion, these so-called “elastic” instabilities can impose a limitation on the throughput.
Hence, it is important to understand, and if possible, to suppress them.
The first instability we study is the phenomenon of melt fracture, which occurs in the
extrusion of polymer melts and takes the form of gross distortions of the surface of the
extrudate. This instability is linked to the phenomenon of wall-slip, i.e., the velocity of the
polymer at the wall relative to the velocity of the wall itself (also called the slip velocity)
is non-zero. Several slip relations based on microscopic theories for polymers predict
regions in which the slip velocity is multivalued. The expectation is that a multivalued
slip relation will result in a multivalued flow curve, which in turn causes melt fracture.
Using a simple slip relation, we show that when the dependence of the slip velocity on
the pressure is taken into account, this is not necessarily true: a multivalued slip law does
not necessarily imply a multivalued flow curve.
The second instability we study is the “filament stretching instability,” which occurs
in the extension of a polymeric liquid bridge between two parallel plates. This insta-
bility takes the form of a bifurcation to a non-axisymmetric shape near the endplates at
ii
high extensions. Motivated by the idea of stress localization near the free surface, we
model the portion of the filament near the endplates as an elastic membrane enclosing an
incompressible fluid and show this is unstable to non-axisymmetric disturbances.
The third instability that we present results for occurs when polymeric liquids flow
along curved streamlines. Such flows are common in industrial coating operations, and
concentric cylinder geometries where the fluid flows along curved streamlines, such as
circular Couette flow (where the flow is driven by the rotation of one of the cylinders)
and Dean flow (where an azimuthal pressure gradient drives the flow), serve as model
geometries for the more complicated coating flows. In the context of Dean flow, we show
how the addition of a steady or time-periodic axial flow of small magnitude compared
to the primary azimuthal flow, and applied either in Couette or Poiseuille form, can sig-
nificantly delay the onset of the instability. The stabilization mechanism is related to the
generation of axial normal stresses induced by the secondary flow which suppress radial
velocity perturbations.
Recent experimental observations by Groisman and Steinberg (1997) and Baumert
and Muller (1999) in the nonlinear regime of viscoelastic circular Couette flow have
shown the formation of stationary, spatially isolated, axisymmetric patterns, termed “di-
whirls” or “flame patterns.” These patterns are very long wavelength vortex pairs, with a
core region of strong radial inflow, surrounded by a larger region of much weaker radial
outflow. These structures may be connected to localized defects seen in coating flows,
and may also form the building blocks for more complicated patterns seen in viscoelastic
flows. Modeling of these patterns is complicated by the absence of a stationary bifurca-
tion in isothermal circular Couette flow. We show how these solutions can be accessed
by numerical continuation from stationary bifurcations in flows that are geometrically
iii
similar to circular Couette flow using a simple constitutive equation for the polymer. Al-
though the stationary solutions we compute are unstable, they are very similar, both qual-
itatively and quantitatively, to the experimentally observed diwhirls and flame patterns.
We also use the results from our computations to propose a fully nonlinear self-sustaining
mechanism for these patterns.
iv
Acknowledgments
I wish to express my deep gratitude to my advisor, Professor Michael Graham, without
whose guidance this work would not have been possible. The breadth and depth of his
knowledge, his scientific curiosity, and his attention to detail have been, and will continue
to be, a source of inspiration for me. I would also like to thank Bill Black and Venkat
Ramanan for their patience in answering my questions, and for helping me when I first
started my research.
It was a privilege to have worked with such talented colleagues as John Kasab, Richard
Jendrejack, Gretchen Baier, Philip Stone, Guiyu Bai, and Jun Sato. Along with Bill and
Venkat, they helped create a stimulating atmosphere for research and contributed im-
mensely to my learning experience. Outside of work, Prasad, Shreyas, Rahul, Sirjana,
Ramesh, Kamal, Sanjay, Mukund, Mala, and others too numerous to name made my stay
in Madison enjoyable.
None of what I have accomplished would have been possible without the support of
my parents and my brother. For their love, and for my parents' belief in the value of a
good education, I shall be forever grateful.
v
Contents
Abstract i
Acknowledgments iv
1 Overview 1
1.1 Constitutive equations for polymeric liquids . . . . . . . . . . . . . 3
1.2 Instabilities in polymeric liquids . . . . . . . . . . . . . . . . . . . . 10
2 Pressure dependent slip and flow curve multiplicity 25
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.2 Mathematical Model . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.3 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . 38
2.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
3 The filament stretching instability 54
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
3.2 Planar elongation: a model problem . . . . . . . . . . . . . . . . . 61
3.3 Elongation of a truncated cone . . . . . . . . . . . . . . . . . . . . 64
3.4 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . 70
vi
3.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
4 Stabilization of Dean flow instability 76
4.1 Instabilities in coating flows . . . . . . . . . . . . . . . . . . . . . . 76
4.2 Elastic instability in Dean flow . . . . . . . . . . . . . . . . . . . . . 81
4.3 Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
4.4 Stability and Numerical Analysis . . . . . . . . . . . . . . . . . . . 91
4.5 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . 97
4.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
5 Localized solutions in viscoelastic shear flows 116
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
5.2 Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
5.3 Discretization and solution methods . . . . . . . . . . . . . . . . . 124
5.4 Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . 138
5.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
6 Conclusions and future work 168
A Introduction to finite elasticity 175
A.1 The basic equations of finite elasticity . . . . . . . . . . . . . . . . 175
B Base state solutions and matrix components in Dean flow 182
B.1 Base state solutions . . . . . . . . . . . . . . . . . . . . . . . . . . 182
B.2 Disturbance equations . . . . . . . . . . . . . . . . . . . . . . . . . 184
C Velocity and Stress scalings for Couette-Dean flow 189
vii
D Time Integration of viscoelastic flows 193
D.1 Governing equations . . . . . . . . . . . . . . . . . . . . . . . . . . 195
D.2 Numerical Method . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
D.3 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . 201
E Branch tracing in three dimensional plane Couette flow 205
E.1 Introduction and Formulation . . . . . . . . . . . . . . . . . . . . . 205
E.2 Discretization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
E.3 Preliminary Results . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
E.4 Proposals for future work . . . . . . . . . . . . . . . . . . . . . . . 212
Nomenclature 216
Bibliography 226
viii
List of Figures
1 An illustration of the rod climbing effect. The beaker in (a) contains a
Newtonian liquid, glycerin, which shows a vortex. The beaker in (b)
contains a solution of polyacrylamide in glycerin, which climbs the rod
(figure scanned from Bird et al. (1987a)). . . . . . . . . . . . . . . . . . 2
2 Diagram of simple steady shear flow. As shown here, the two plates are
moved with equal speed in opposite directions. . . . . . . . . . . . . . . 3
3 Diagram of elongational flow with l = exp(ε t). . . . . . . . . . . . . . . 5
4 (a) A random walk calculation showing one of a very large number of
conformations of a polymer molecule (b) A dumbbell model which only
captures the longest relaxation time. . . . . . . . . . . . . . . . . . . . . 5
5 Pictorial representation of the two types of pitchfork bifurcations: (a)
supercritical and (b) subcritical. [y] is some measure of y that captures
the features of the transition. A solid line indicates a stable branch while
a dashed line indicates an unstable branch. . . . . . . . . . . . . . . . . . 13
6 Extrusion related instabilities (a) sharkskin (b) melt fracture. (Agassant
et al., 1991). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
ix
7 A typical flow curve for polymers exhibiting spurt and melt fracture. The
y coordinate 8V/D is proportional to the exit flow rate (Kalika and Denn,
1987). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
8 The filament stretching instability (a) The non-axisymmetric bifurcation
seen from below the bottom plate (b) side view of the instability at a later
stage (Spiegelberg and McKinley, 1996). . . . . . . . . . . . . . . . . . . 18
9 Flow visualization of the purely elastic Taylor-Couette instability (Larson
et al., 1990). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
10 Flow visualization of a blade coating geometry. The figure clearly shows
the presence of an upstream recirculation region with curved stream-
lines (Davard and Dupuis, 2000). . . . . . . . . . . . . . . . . . . . . . . 20
11 Sequence of snapshots showing the transition from non-axisymmetric
disordered flow to solitary vortex structures. On the left in (a), the en-
tire flow geometry is shown, with the box showing the cross section
being visualized. On the right, in (b), the actual transition sequence is
shown (Groisman and Steinberg, 1998). . . . . . . . . . . . . . . . . . . 21
12 Sequence of snapshots showing the transition from non-axisymmetric
flow to the predominantly axisymmetric and localized flame patterns (Baumert
and Muller, 1999). The flow geometry is the same as in figure 11. . . . . 22
13 Plot of equation 30 for different values of pressure (A2 = 4, A3 = 122.24,
β = 0.0102). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
14 Schematic diagram of the constant piston speed experiment. . . . . . . . 36
x
15 Flow curve for a Newtonian Fluid (Λ = 10, A2 = 3, A3 = 122.24,
β = 0.0102, H = 1). The profiles at the points marked `X' are shown in
figures 18 and 19. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
16 Behavior of critical stresses τc2 and τc3 with increasing Λ (A2 = 3, A3 =
122.24): (a) Newtonian model (β = 0.0102, H = 1) (b) shear thinning
model (β = 0.0102, H = 1, n = 0.56) (c) UCM model (β = 0.0102,
H = 1) (d) PTT model (ε = 10−1, β = 0.0102, H = 1). The upper curve
corresponds to τc = τc2 and the lower one to τc = τc3. The curves are
not extended to Λ = 0 because the approximations used are not valid for
small Λ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
17 Flow curve for a Newtonian fluid showing no multiplicity (Λ = 238.62,
A2 = 3, A3 = 122.24, β = 0.0102, H = 1). . . . . . . . . . . . . . . . . 42
18 Profiles of pressure, slip velocity and shear stress on the low flow rate
branch for the Newtonian fluid of figure 15. (∆P =5.22, Λ = 10, A2 =
3, A3 = 122.24, β = 0.0102, H = 1). . . . . . . . . . . . . . . . . . . . 43
19 Profiles of pressure, slip velocity and shear stress on the high flow rate
branch for the Newtonian fluid of figure 15 (∆P =6.39, Λ = 10, A2 = 3,
A3 = 122.24, β = 0.0102, H = 1). . . . . . . . . . . . . . . . . . . . . . 44
20 Flow profiles on the three dimensional slip surface for a shear thinning
fluid: (a) upper and lower limit points (∆P = 6.39, Λ = 10) (b) cusp
point (∆P = 132.24,Λ = 230.46) (c) point in the central portion without
multiplicity (∆P = 140.0, Λ = 240.82). Other parameters are A2 = 3,
A3 = 122.24, β = 0.0102, H = 1, and n = 0.56 for all three cases. . . . . 46
xi
21 Oscillatory flow for a Newtonian fluid with Qp = 0.21:(a) Pressure drop
and exit flow rate vs. time (b) Plot of Qe vs. Pb superimposed on the
steady state flow curve. . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
22 Non-oscillatory flow for a Newtonian fluid with Qp = 0.15: Pressure
drop and exit flow rate vs. time. . . . . . . . . . . . . . . . . . . . . . . 49
23 Non-oscillatory flow of a Newtonian fluid with Λ = 250, Qp = 0.1862
and other parameters as in figure 21: Pressure drop and exit flow rate vs.
time. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
24 Oscillatory flow of a UCM fluid (C = 106, Λ = 10, Re = 10−4, κ =
10−5, De = De∗ = 10, λ1 = λ2 = 0.01, Qp = 0.081): (a) Barrel
pressure and exit flow rate vs. time (b) Plot of Qe vs. Pb superimposed
on the steady state flow curve. . . . . . . . . . . . . . . . . . . . . . . . 52
25 Schematic of a filament stretching rheometer. The setup on the left shows
the undeformed state of the liquid bridge. . . . . . . . . . . . . . . . . . 58
26 Schematic of a planar elongation setup and flow field. . . . . . . . . . . . 62
27 Coordinate system for the truncated cone. . . . . . . . . . . . . . . . . . 66
28 Evolution of the shape of a truncated cone under elongation: (a) un-
deformed configuration (b) axisymmetric configuration at l = 0.2 (c)
axisymmetric configuration at l = lc = 0.5 (d) post bifurcation non-
axisymmetric shape at l = lc = 0.5. The figures on the right track the
change in a cross section originally at a distance of 0.75 units from the
left edge. In (d), the perturbation has been exaggerated for clarity. . . . . 72
29 Spatial profile of the hoop stress, τθθ, and the amplitude of the bifurcating
solution ‖ λ1 ‖. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
xii
30 Some commonly used coating industrial coating processes: (a) Dip coat-
ing and rod coating. (b) Blade coating and air knife coating. (c) Gravure
coating. (d) Reverse roll coating. (e) Extrusion coating. (f) Slide coating
and curtain coating (Cohen, 1992) . . . . . . . . . . . . . . . . . . . . . 78
31 Photograph of ribbing instability in forward roll coating (Coyle et al.,
1990b). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
32 Mechanism of the elastic instability in Dean flow. . . . . . . . . . . . . . 84
33 Illustration of the mechanism by which additional axial stresses gener-
ated by a superimposed axial flow stabilize the viscoelastic Dean flow
instability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
34 Dean flow geometry, shown with superimposed Poiseuille flow. . . . . . . 87
35 Neutral stability curves for DAC flow (S=0). In each case, the position of
Wpc,min is denoted by a •. . . . . . . . . . . . . . . . . . . . . . . . . . 102
36 Neutral stability curves for DAP flow (S=0). In each case, the position of
Wpc,min is denoted by a •. . . . . . . . . . . . . . . . . . . . . . . . . . 103
37 Plot of Wpc,min vs. Wez for DAC and DAP flows (S=0). Note the linear
scaling at high Wez. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
38 Plots of Wpc,min vs. Wez for S = 0 and S = 10, displaying the stabilizing
influence of solvent viscosity. . . . . . . . . . . . . . . . . . . . . . . . . 104
39 Plot of Wpc,min vs. n for different values of Wez for DAC flow (S = 0). . 104
40 Plot of Wpc,min vs. n for different values of Wez for DAP flow (S = 0). . 105
41 Plot of Wpc,min vs. Wez for DAC flow with n = 1.0 (S = 0). Note the
linear scaling for Wez > 1. . . . . . . . . . . . . . . . . . . . . . . . . . 105
xiii
42 Neutral stability curve of pure Dean flow at high α. The Takens-Bogdanov
bifurcation point is indicated by a •. . . . . . . . . . . . . . . . . . . . . 107
43 Decay of the perturbation hoop stress τθθ when axial flow is imposed.
Parameters are: Wez = 1.0, ω1 = 0.5, Wp = 4.06, S = 0. . . . . . . . . . 110
44 Time sequence of density plots of the perturbation hoop stress τθθ when
axial flow is imposed. The parameters are identical to figure 43, so that
without axial flow, the flow is neutrally stable. Each frame shows a z− r
cross-section of the geometry. . . . . . . . . . . . . . . . . . . . . . . . 111
45 Plot of the magnitude of the Floquet multiplier |β| vs. ω1 for different
values of α. (Wez = 0.5, Wp = 4.06, S = 0). . . . . . . . . . . . . . . . 112
46 Plot of the magnitude of the Floquet multiplier |β| vs. ω1 for different
values of Wez. (Wp = 4.06, α = 6.6, S = 0). |β| asymptotes to 1 at
large ω1 in agreement with the asymptotic prediction. . . . . . . . . . . . 113
47 Plot of the magnitude of the periodic component of the hoop stress τθθ
over a cycle of the forcing for ω1 1 (Wez = 1.0, Wp = 4.06, α = 6.6,
ω1 = 0.01). Note the large increases in magnitude. . . . . . . . . . . . . 114
48 Plot of the magnitude of the periodic component of the hoop stress τθθ
over a cycle of the forcing for ω1 = O(1) (Wez = 1.0, Wp = 4.06,
α = 6.6, ω1 = 1.0). The magnitude remains O(1) over the entire cycle. . 115
49 Geometry of Couette-Dean flow in an annulus . . . . . . . . . . . . . . . 122
50 A spectral element mesh with 16 axial and 16 radial elements with fifth
order polynomials in each direction in each domain. Note the dense con-
centration of points near r = 1 and z = L/2. The high resolution is
necessary to capture the intense stress localization in these regions. . . . . 128
xiv
51 Comparison of ILUT∗ and ILU(0) preconditioners. The test problem was
the calculation of the unit tangent for a point on the nontrivial branch in
Dean flow. The matrix A had a dimension of 21987. . . . . . . . . . . . 137
52 Linear stability curves at δ = 1 (Dean flow) computed using the FENE-P
model. The points marked “TB” are Takens-Bogdanov points. The lines
correspond to points where the base state flow loses stability to stationary
axisymmetric perturbations. . . . . . . . . . . . . . . . . . . . . . . . . 140
53 Linear stability curves at δ = 1 (Dean flow) computed using the FENE-
CR model. As in figure 52, only stationary axisymmetric perturbations
are considered. Note the complete absence of non-stationary bifurcations. 140
54 Continuation in Weθ of a stationary solution in Dean flow. The parameter
values are L = 1.05, b = 700, ε = 0.20, and S = 1.2. At the Hopf point,
a pair of complex conjugate eigenvalues become unstable. These collide
and form two real eigenvalues, one of which re-crosses the imaginary axis
at Weθ = 29.57, where the stationary branch originates. The solution
amplitude used here differs from that used in subsequent figures and is
defined in equation 150. . . . . . . . . . . . . . . . . . . . . . . . . . . 142
55 A path to stationary solutions in circular Couette flow. The parameters
are Weθ = 25.15, L = 3.07, b = 1830, ε = 0.2, and S = 1.2. . . . . . . . 143
56 Results from continuing the stationary circular Couette flow solutions in
L. The parameters are Weθ = 25.15, b = 1830, ε = 0.2, and S = 1.2.
The gaps in the lower branch correspond to places where we changed
the mesh. Note the flatness of the branches as L increases. We have
computed extensions of the upper branch at lower values of Weθ. . . . . 145
xv
57 Density plot of 〈QQ〉θθ (white is large stretch, black small) and contour
plot of the streamfunction at L = 116.52 (Weθ = 24.29, b = 1830,
S = 1.2, and ε = 0.2). For clarity, most of the flow domain is not
shown. Note the very strong localization of 〈QQ〉θθ near the center. The
maximum value of 〈QQ〉θθ at the core is 1589 which gives τθθ = 12722.
Compared to this, the maximum value of 〈QQ〉θθ in the circular Couette
base state is 706, which gives τθθ = 1150. Away from the core of the
diwhirl, the structure is pure circular Couette flow. The streamlines show
striking similarity to those in figure 10 of Groisman and Steinberg (1998).
This point was generated by stretching the point at the corresponding Weθ
on the upper branch of the curve for L = 9.11 in figure 58. . . . . . . . . 146
58 Diwhirl solution amplitudes as functions of Weθ and L. Note that the
curves at L = 9.11 and L = 4.74 are very close together, while both
curves are well separated from the curve at L = 3.07 (b = 1830, S = 1.2,
and ε = 0.2). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
59 Plot of the location of the turning point, Weθ,c, versus L at S = 1.2 and
ε = 0.2. Note the flatness of the curve at large L. . . . . . . . . . . . . . 149
60 Plot of the position of the linear stability limit in circular Couette flow
with respect to axisymmetric disturbances and the turning point in Weθ
for the diwhirls as a function of b. The parameters are S = 1.2 and
ε = 0.2. The computations for the diwhirls were performed at L = 4.74,
which is close to the minimum in figure 59. . . . . . . . . . . . . . . . . 150
xvi
61 (a) The axial variation of vr at r = 0.6 for L = 4.74 and Weθ = 23.50 on
the upper branch. (b) Figure 9 on page 2457 of Groisman and Steinberg
(1998), shown here for purposes of comparison. We have shifted the axial
coordinate so that the symmetry axis of the computed diwhirl in (a) is at
z = 0, to make comparison with (b) easier . . . . . . . . . . . . . . . . . 151
62 Variation of solution amplitudes with ε. Here, Weθ = 25.15, and the
other parameters as in figure 58. . . . . . . . . . . . . . . . . . . . . . . 152
63 Diwhirl solution amplitudes as a function of the parameter cr for two
different wavelengths. Weθ = 25.15 and the other parameters are as in
figure 58. The existence of turning points demonstrates that these solu-
tions cannot be extended to the FENE-CR model. . . . . . . . . . . . . . 153
64 Intensity plot of the dimensionless viscous dissipation for Weθ = 23.73
on the upper branch at L = 4.74. Light areas represent areas of large
viscous dissipation and dark areas represent regions of low viscous dissi-
pation. The horizontal axis is stretched by a factor of two relative to the
vertical axis for clarity. . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
65 (a) Vector plot of v near the outer cylinder at the center of the diwhirl
structure (oblique arrows) and the base state (straight arrows). The length
of the arrows is proportional to the magnitude of the velocity. The axial
velocity is identically zero in the base state, and is zero by symmetry
at the center of the diwhirl. (b) Principal stress directions at the same
location as for (a). The Couette flow stress is not shown because it is
very small in comparison. This figure shows how fluid elements at larger
radii are pulled down and forward sustaining the increase in vθ. . . . . . . 157
xvii
66 Nonlinear self-sustaining mechanism for the diwhirl patterns. . . . . . . . 158
67 Density plot of vr showing the (a) real and (b) imaginary parts of the
destabilizing axisymmetric disturbance, and (c) the streamlines of the
base diwhirl. Note that the core of the diwhirl is entirely unaffected by
the disturbance. The parameters are Weθ = 23.87, S = 1.2, ε = 0.20,
b = 1830, and L = 4.74. . . . . . . . . . . . . . . . . . . . . . . . . . . 161
68 Density plot of the perturbation radial velocity for the three non-axisymmetric
unstable eigenmodes with n = 1. The parameters are identical to those
in figure 67. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
69 Density plot of the perturbation radial velocity for the three non-axisymmetric
unstable eigenmodes with n = 2. The parameters are identical to those
in figure 67. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
70 Bifurcation diagram for Dean flow at b = 1830, S = 1.2, ε = 0.2, and
L = 1.963. We could not continue the branch beyond Weθ = 13.94, for
reasons discussed in section 5.4.6. Radial velocity profiles corresponding
to the points marked by the filled and open circles are shown in figure 71. 164
71 Density plots of the radial velocity at the points circled in figure 70. (a)
Weθ = 15.486 (b) Weθ = 13.930. . . . . . . . . . . . . . . . . . . . . . 164
72 Linear stability curves and existence boundaries for nonlinear solutions
in Dean flow at S = 1.2, ε = 0.2, and b = 1830. The filled triangle shows
the data at L = 1.795 where the solution terminates via collision with the
L/2 branch in a pitchfork bifurcation before the turning point. . . . . . . 165
73 Streamlines for equation 214 with A = 1 and k = π. . . . . . . . . . . . 194
xviii
74 Poincare map of the trajectory of a point in the flow field given by equa-
tion 215. The parameters chosen were k = π, ω = 2π, and ε = 0.25. . . . 195
75 Illustration of the effect of adding a diffusion term to the constitutive
equation on a mesh of 144×72. The parameter values used were We = 1,
Re = 1, ε = 0.2, k = π, ω = 2π, b = 10, and S = 6.67. . . . . . . . . . . 202
76 Effect of mesh refinement on numerical stability. Runs with both meshes
were performed with a = 10−4 and the same parameter values as in
figure 75. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
77 Effect of artificial diffusion when equation 223 in integrated in a known
velocity field. The parameters and mesh are identical to those in figure 75.
Note the earlier blow up when compared to the coupled case. . . . . . . . 203
78 Contours and density plot of the streamwise velocity u in the y− z plane
in the base state. The value of Re = 150, Λ = 0.002, and Fr = 5.08. . . . 212
79 Contours and density plot of the u component of the two unstable eigen-
vectors for the flow shown in figure 78 in the x − z plane at y = 0. The
mode shown in (a) has the higher growth rate of the two unstable modes. . 213
80 Plot of the profile of the u component of the streamwise velocity of the
two unstable eigenvectors for the flow shown in figure 78 at y = 0, and
z = π/γ. The eigenvalue corresponding to eigenvector 1 has the higher
growth rate. Eigenvector 2 is simply a vertically shifted and scaled ver-
sion of eigenvector 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214
81 Contours and density plot in the x − z plane of the u component of the
real and imaginary part of the unstable eigenvector at Re = 150, Λ = 0.1,
and Fr = 5.01. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
1
Chapter 1
Overview
Unlike Newtonian fluids, polymeric liquids have some memory of the deformation they
have experienced. Newtonian fluids respond virtually instantaneously to an imposed de-
formation rate, whereas polymeric fluids respond on a macroscopically large time scale,
known as the relaxation time. When subjected to deformation rates much larger than
the inverse relaxation time, their behavior resembles that of elastic solids, whereas their
response to deformation rates much smaller in magnitude than the inverse relaxation
time resembles that of viscous liquids. For this reason, they are known as “viscoelastic”
liquids. The viscoelastic nature of polymer melts and solutions causes them to exhibit
behavior not seen in Newtonian liquids. An interesting example of this behavior is the so
called rod-climbing effect. An illustration of this effect is shown below in figure 1. The
figure shows two liquids being stirred in two different beakers. The liquid in figure 1(a)
is Newtonian, while the liquid in figure 1(b) is a polymer solution. The Newtonian liquid
shows a vortex, with free surface being depressed in the region near the stirrer. We would
expect this intuitively, based on the fact that the centrifugal force pushes liquid near the
2
stirrer outward. In contrast, the polymer solution shows the opposite effect, with the liq-
uid climbing up the stirring rod. This is because the polymer molecules are stretched
along the circular streamlines of the flow. The extra tension in the streamlines exerts an
inward force on the fluid, which acts against the centrifugal force and gravity and pushes
the liquid up the rod. Other examples of viscoelastic behavior include die swell and
elastic recoil (Bird et al., 1987a).
(a) (b)
Figure 1: An illustration of the rod climbing effect. The beaker in (a) contains a New-tonian liquid, glycerin, which shows a vortex. The beaker in (b) contains a solutionof polyacrylamide in glycerin, which climbs the rod (figure scanned from Bird et al.(1987a)).
In industrial applications involving polymeric liquids, instabilities that arise from
their viscoelastic nature can have important consequences, and this document details a
computational investigation of a few of these instabilities. In this chapter, we present an
overview of the topics covered in the rest of the thesis. Before discussing instabilities
in polymerflows however, it is instructive to look at some of the elementary approaches
taken to modeling polymeric liquids, which lead to the constitutive equations that we use
in our analyses.
3
1.1 Constitutive equations for polymeric liquids
Unlike Newtonian liquids, which are very well described by Newton's law of viscosity,
there is no single constitutive equation that describes the entire range of polymeric liq-
uids. Consequently, the approach taken is to use simpleflows to check the predictions of
constitutive equations with the actual behavior of the polymeric liquid being studied. We
will, for the most part, be interested in the behavior of dilute solutions of linear polymers.
In this section, we discuss some of theflows used to characterize these liquids, and then
proceed to describe a few of the simple constitutive equations used to model them.
x
y
z
Figure 2: Diagram of simple steady shearflow. As shown here, the two plates are movedwith equal speed in opposite directions.
The simplestflow used for rheological characterization is the steady shearflow shown
in figure 2. Denoting the velocity components in the three coordinate directions asvx,
vy, andvz, the only component that is not zero isvx. For a Newtonianfluid, the force
required to maintain theflow would only have a component in theflow direction. The
only nonzero components of the stress tensor are the shear stresses,τyx andτxy. Since
the stress tensor is symmetric, these are equal to each other and are given by Newton's
4
law of viscosity as
τxy = τyx = ηγyx, (1)
whereγyx = dvx/dy is the shear rate,γ = |γyx|, andη is the viscosity, which is indepen-
dent of the shear rate. For a polymer solution, the viscosityη, would, in general, depend
on the shear rate, i.e.,η = η(γ). In addition, the stress componentsτxx, τyy, andτzz are
not all zero, unlike in the shearflow of Newtonian liquids. A direct physical consequence
of this is that there are components of thefluid stresses that act to push the plates apart.
Therefore, in addition to the shearing force, a force acting normal to the plates has to be
applied to maintain theflow. The additional nonzero stress components allow us to define
two newfluid properties using the relations
τxx − τyy = Ψ1(γ)γ2yx, (2)
τyy − τzz = Ψ2(γ)γ2yx. (3)
The quantityτxx − τyy is known as thefirst normal stress difference, whileτyy − τzz
is known as the second normal stress difference.Ψ1 andΨ2 are known as thefirst and
second normal stress coefficients respectively, and are functions ofγ.
Another flow used for rheological characterization is elongationalflow (figure 3).
Here, the velocityfield is given by
vx = −1
2εx
vy = −1
2εy
vz = εz, (4)
so thatfluid elements are being stretched exponentially in one direction and compressed
exponentially in the two directions perpendicular to it. The material function of relevance
5
l
1/
z
x
y
l
1/ l
Figure 3: Diagram of elongationalflow with l = exp(ε t).
here is the elongational viscosityη, which is defined by the relation
η =τzz − τxx
ε. (5)
For Newtonian liquids,η = 3η, and the ratioη/η is called the Trouton ratio.
coarsegraining
Q
(a) (b)
Figure 4: (a) A random walk calculation showing one of a very large number of confor-mations of a polymer molecule (b) A dumbbell model which only captures the longestrelaxation time.
6
The steady shear and elongationalflows are used as modelflow fields for testing con-
stitutive equations for polymers. Experimental measurements of the behavior of specific
polymer solutions and melts can be compared with the predictions of the constitutive
equations. The expectation is that constitutive equations that can capture the behavior of
a polymer in these modelflows might reasonably be expected to work in more compli-
catedflow fields as well. Also, the modelflows are used to obtain rheological parameters
needed in constitutive equations. For example, simple shearflow can be used to obtain
the zero shear viscosity and the relaxation time.
We now proceed to discuss simple constitutive equations that are used for modeling
dilute solutions of linear polymers. The constitutive equations that we describe below
are based on molecular models. Figure 4(a) shows a random walk model of a polymer
molecule, which represents one of a large number of conformations that the polymer
molecule can take. These conformations change continually due to thermal motion, with
different time scales associated with motion on different length scales. For example, it is
easy to see that small internal reorientations would take place on a shorter time scale than
a change involving a large number of segments. Thus, the polymer molecule responds on
a wholespectrum of time scales. For the simple models that we will use in this work, we
only consider the upper limit of this spectrum, i.e., the longest relaxation time.
At the level of coarse graining described above, the polymer molecules are modeled
as elastic dumbbells (a pair of beads connected by a spring), as shown infigure 4(b), with
Q representing the end to end vector of the dumbbell. The elastic force here represents
the restoring force due to entropy. A solution of such dumbbells is described by the
probability distribution function forQ, denoted byψ(Q, t). For a dilute solution of these
dumbbells, neglecting inertia and hydrodynamic interaction between the beads allows us
7
to use kinetic theory (Bird et al., 1987b) to derive a diffusion equation forψ(Q, t), which
is given by
∂ψ
∂t+ v ·∇ψ = −
(∂
∂Q·
[κ ·Q]ψ −
2kT
ζ
∂ψ
∂Q−2
ζF cψ
). (6)
Here,v is the velocity vector,k is Boltzmann's constantt represents time,T represents
temperature,F c the spring force,ζ the friction coefficient due to hydrodynamic drag, and
κ is a second order tensor which specifies the local velocityfield.
Equation 6 serves as the starting point for the calculation of the averages of quantities
of physical interest. For the purpose of this work, we are mainly interested in the stress
tensorτ . For elastic dumbbells, this is given by
τ = N 〈QF c〉 − NkT I, (7)
whereN is the number density of dumbbells, and we have once again assumed that the
external forces acting on the two beads of the dumbbell cancel each other out. For simple
models,τ can be approximated from the ensemble average of the second order tensor
QQ. This quantity,〈QQ〉, is given by the equation
〈QQ〉(1) =4kT
ζI −
4
ζ〈QF c〉 . (8)
Here,〈QQ〉(1) is theupper convected derivative of 〈QQ〉 given by
〈QQ〉(1) =∂〈QQ〉
∂t+ v ·∇〈QQ〉−
(∇v)t · 〈QQ〉+ 〈QQ〉 · (∇v)
, (9)
wherev is the velocity vector, and the superscriptt is used to denote the transpose.
In our discussion above, we have not specified a form for the spring force. The sim-
plest case occurs when we assume that the spring is Hookean, i.e.,F c = HQ, where
H is the spring constant. This assumption yields the upper convected Maxwell (UCM)
8
equation. This equation assumes that the beads can stretch to an infinite extent, which
is clearly not a reasonable assumption, especially for so called strongflows, wherefluid
elements undergo exponentially large degrees of stretching. However, the UCM equation
has the virtue of being simple, and for this reason is often used to get qualitative informa-
tion on the effect of viscoelasticity. This simplicity of form is in some respects deceptive,
for it masks the fact that the UCM equation is extremely difficult to use in the numerical
simulation of strongflows. This in turn gives rise to another important use of the UCM
equation: as a test case for numerical methods for solving viscoelasticflow problems.
Since polymers havefinite maximum lengths, a nonlinear spring law makes for a
more realistic approximation. The Finitely Extensible Nonlinear Elastic (FENE) spring
law takes the form
F c = HQ
1−Q2/Q20, Q ≤ Q0, (10)
whereQ2 = tr(QQ) is the square of the dumbbell extension withtr used to denote the
trace of its argument. For small extensions, the spring is nearly Hookean, but increases in
stiffness for larger extensions, up to a maximum extension ofQ0. In the limitQ0 → ∞,
equation 10 becomes Hooke's law. Substituting equation 10 into equation 8 gives
〈QQ〉(1) =4kT
ζI −
4H
ζ
1−Q2/Q20
⟩. (11)
The second term on the right hand side prevents equation 11 from being an explicit equa-
tion for 〈QQ〉. To obtain a closed form equation for〈QQ〉, it is necessary to make
a closure approximation. One such approximation is the Peterlin closure (Bird et al.,
1987b) which takes the form⟨QQ
1−Q2/Q20
⟩
〈QQ〉1− tr(〈QQ〉)/Q20
. (12)
9
Substituting this relation into equation 11 gives the FENE-P constitutive equation. A
related FENE type equation is the FENE-CR equation (Chilcott and Rallison, 1988),
which is a good model for dilute solutions of some polymers. This equation is
〈QQ〉(1) =1
1− tr(〈QQ〉)/Q20
(4kT
ζI −
4
ζ〈QQ〉
). (13)
In all three equations, knowing〈QQ〉 enables us to get expressions for the stress tensor.
For the UCM equation, this is given by the Kramers expression
τ = NH〈QQ〉−NkT I. (14)
The quantityNkT is equal toG, the relaxation modulus of the polymer. Substituting
equation 14 and the Hookean spring law into equation 8 yields an explicit equation forτ ,
given by
τ + λ τ (1) = ηγ, (15)
which is linear inτ . Here,λ is the relaxation time, given by
λ =ζ
4H, (16)
η is the viscosity, given by
η = λG, (17)
and
γ = I(1) =∇v + (∇v)t (18)
is the shear rate tensor. For the FENE-P equation, the Kramers expression for the stress
tensor is
τ = NH〈QQ〉
1− tr(〈QQ〉)/Q20−NkT I, (19)
10
and for the FENE-CR equation, the stress tensor can be calculated using the relation
τ =1
1− tr(〈QQ〉)/Q20(NH〈QQ〉−NkT I) . (20)
In the latter two cases, it is possible to substitute equations 19 and 20 into the constitutive
equations for〈QQ〉 and get explicit constitutive equations forτ . However, it is much
simpler to work with〈QQ〉, and use equation 19 or 20 to getτ when needed.
Table 1 summarizes the qualitative behavior of the UCM, FENE-P and FENE-CR
equations in shear and elongation. Also presented in table 1 is the behavior of the exact
FENE model in theseflows, obtained from Brownian dynamics simulations. In this work,
we will use the FENE-P and UCM equations for most of our calculations. Occasionally,
we will use the FENE-CR equation when testing for model dependence.
PropertyModel η Ψ1 η
UCM Constant Constant Blows up atfiniteelongation rate
FENE Shear thins Shear thins Saturates at highelongation rate
FENE-P Shear thins Shear thins Saturates at highelongation rate
FENE-CR Constant Shear thins Saturates at highelongation rate
Table 1: Summary of behavior of the UCM, FENE, FENE-P, and FENE-CR constitutiveequations in shear and elongation. All four models predictΨ2 = 0.
1.2 Instabilities in polymeric liquids
Before embarking on a discussion offlow instabilities in polymeric liquids, it is instruc-
tive to clarify what we mean by the term“flow instability.” This is best done with the aid
11
of a simple example. Consider theflow of a Newtonian liquid (such as water) in the gap
between two very long concentric cylinders, driven by the motion of the inner cylinder.
At low rotation rates, we would see aflow where the velocities and pressure only varied
with radial position and are constant along the azimuthal and axial directions. We will
refer to thisflow as the baseflow. As the rotation speed is increased, a transition occurs
to an axisymmetric (i.e. no variation in the azimuthal direction), axially periodic vortex
flow. Thisflow is qualitatively different from the baseflow: in particular, unlike in the
baseflow, the velocities and stresses vary in the axial direction. What is interesting is
that the baseflow is an admissible solution to the governing equations at all values of the
rotation rate. Beyond a critical value of the rotation rate however, perturbations (even in-
finitesimal ones) applied to the baseflow grow in magnitude until a qualitatively different
steady state is reached. We say that the baseflow is unstable beyond the critical rotation
speed, hence the phrase“flow instability.” The qualitative change caused by the variation
of a parameter (in this case, the rotation rate of the inner cylinder) is called abifurcation.
In this work, we will mainly be interested in two types of bifurcations: pitchfork and
Hopf. Pitchfork bifurcations generally occur in systems with symmetry. As a simple
one-dimensional example, consider the differential equation
y = µy − y3. (21)
This equation is invariant with respect to the transformationy → −y, i.e., replacingy by
−y results in the same equation. Steady states are obtained by settingy to zero. They are
solutions to
f(y) = µys − ys3 = 0, (22)
which areys = 0 andys = ±√
µ, with the subscripts being used to denote a steady state.
12
While ys = 0 is a solution for all values ofµ, the solutionsys = ±√µ are valid only for
µ > 0. In this problem,ys =√µ andys = −
õ bifurcate from the solutionys = 0
at µ = 0. In a small neighborhood of the bifurcation point, the branchesys =√µ and
ys = −√
µ are one sided, i.e., they only exist forµ > 0. This type of bifurcation is called
a pitchfork. Note that the bifurcating branchesys =√µ andys = −
õ are related by
symmetry.
For the one-dimensional example presented here, the stability of the bifurcating branches
is determined by the sign of the linearization of the steady state equations, evaluated at
the point whose stability is to be determined. If the linearization is positive, the solu-
tion is unstable, because small disturbances grow. If the linearization is negative, small
disturbances decay, and the solution is stable. For equation 21, the linearization is given
by
∂f
∂y= fy = µ− 3y2s . (23)
Substitutingys = 0 givesfy = µ. Thus, the solutionys = 0 is stable for negative values
of µ and unstable for positiveµ. Note that the change in stability occurs at the bifurcation
point,µ = 0. Substitutingys = ±√µ in equation 23 givesfy = −2µ, which is negative
for µ > 0, the only regime where these solutions exist. Therefore, these solutions are
stable. When the bifurcating branch is stable, the bifurcation is said to besupercritical.
The opposite case, asubcritical bifurcation, occurs when the bifurcating branches are
unstable. An example of this is the system
y = y3 + µ y, (24)
which has the steady statesys = 0 andys = ±√−µ. An analysis similar to the one pre-
sented above shows that the steady statesys = ±√−µ bifurcate fromys = 0 at µ = 0,
13
only exist for negative values ofµ and are unstable. Figure 5 shows a pictorial repre-
sentation of supercritical and subcritical bifurcations. For higher dimensional systems,
stability is determined by the eigenvalues of the matrix that results from linearizing the
steady state governing equations, with the linearization being performed about the point
whose stability is being determined. This matrix is called the Jacobian. If all the eigenval-
ues of the Jacobian are negative, the solution is stable. If one or more of the eigenvalues
are positive, the solution is unstable. A bifurcation occurs when the real part of one or
more eigenvalues changes sign.
[y]
µ
[y]
µ
[y]
µ
[y]
µ
(a)
(b)
Figure 5: Pictorial representation of the two types of pitchfork bifurcations: (a) super-critical and (b) subcritical.[y] is some measure ofy that captures the features of thetransition. A solid line indicates a stable branch while a dashed line indicates an unstablebranch.
The other bifurcation that we will present here is the Hopf bifurcation. A Hopf bi-
furcation occurs when a pair of complex conjugate eigenvalues of the Jacobian crosses
14
the imaginary axis. In this case, the bifurcating branches are not steady states, but time-
periodic oscillations (also called limit cycles). As with the pitchfork bifurcation, a super-
critical Hopf bifurcation occurs if the bifurcating branch of periodic solutions is stable,
and a subcritical Hopf bifurcation occurs when they are unstable. Since a Hopf bifurca-
tion requires that a pair of eigenvalues cross the imaginary axis, it follows that a system
must be at least two-dimensional to show a Hopf bifurcation. An example of a system
that undergoes a Hopf bifurcation is
y1 = −y2 + y1(µ− y12 − y2
2),
y2 = y1 + y2(µ− y12 − y2
2), (25)
which can be written in vector form as
y = f(y). (26)
It is trivial to show that(y1s, y2s) = (0, 0) is a steady state. Linearizing about this steady
state, we get the Jacobian matrix
J =∂f
∂y=
µ −1
1 µ
, (27)
which has the complex conjuagte pair of eigenvaluesµ±i. This pair crosses the imaginary
axis atµ = 0, giving rise to a Hopf bifurcation. It can be shown (see Seydel (1994)) that
the bifurcating branch of limit cycles is stable, so this is a supercritical Hopf bifurcation.
With this background, we can now move on to a discussion offlow instabilities in
polymer solutions. Theflow instability that we described at the beginning of this section
as an example is called the Taylor-Couette instability. It is driven by inertial effects,
specifically, the unstable stratification of angular momentum. Another example of an
15
inertia driven instability is the phenomenon of turbulence in pipe and plane Couetteflow
at high Reynolds numbers. In most practical applications, the high viscosity of polymer
melts and solutions results in a low Reynolds number, which means that inertial effects
are of secondary importance. However, polymericfluids display an entirely separate class
of instabilities which have their origin in the elastic nature of thesefluids. In an industrial
setting, these so called“elastic” instabilities, seen in many commercially importantflows,
are detrimental to the quality of thefinal product, and avoiding them involves imposing
limitations on throughput, or modifying theflow apparatus. It is therefore of practical
importance to understand these instabilities, and if possible, come up with methods to
delay their onset. The work presented in this document focuses on four of these: melt
fracture, thefilament stretching instability, the viscoelastic Couette-Dean instability, and
nonlinear pattern formation in Couette-Deanflow. These instabilities are seen either in
industrial processing operations or in simpleflows used to characterize the rheological
properties of viscoelastic liquids.
Thefirst part of our work concentrates on melt fracture, which is an instability seen
in extrusion processes. In an extrusion process, a polymer melt is forced out of a die
by applying a pressure gradient or a constant volumetricflow rate. At lowflow rates
(or pressure drops), the shape of the extrudate is smooth. As theflow rate is increased,
the extrudate, for linear polymers, begins to show surface distortions. These arefirst
seen in the form of small amplitude, small wavelength distortions parallel to the surface
(figure 6(a)). The effect is called sharkskin, and a detailed review of this phenomenon
may be found in Graham (1999). At somewhat higher pressure drops, if theflow is
being driven by a constant pressure gradient, theflow rate shows a sudden large jump
to a higher value. This phenomena is known as spurtflow. If the flow is driven by an
16
imposed constantflow rate, the pressure drop and exit massflow rate show oscillations.
These are accompanied by gross distortions in the shape of the extrudate (figure 6(b)).
This phenomena is known as melt fracture. The dependence of the onsetflow rate (or
pressure drop) for both sharkskin as well as melt fracture (Ramamurthy, 1986; Wang
and Drda, 1997; Piau et al., 1995) suggests that both phenomena have their origin in
the fact thatflows of polymericfluids exhibit slip at the wall, i.e., the velocity of the
polymer at the wall is non-zero relative to the velocity of the wall itself. This is different
from Newtonian liquids which obey at no-slip condition at a solid boundary. Currently,
there exist several slip expressions relating the slip velocity to the wall shear stress which
are based on microscopic theories for polymers (Leonov, 1990; Brochard and de Gennes,
1992; Adjari et al., 1994; Yarin and Graham, 1998; Mhetar and Archer, 1997; Hill, 1998).
These theories predict certain regions in which the slip velocity is multivalued. They do
not, however, consider the effect of pressure on slip velocity.
(a)
(b)
Figure 6: Extrusion related instabilities (a) sharkskin (b) melt fracture. (Agassant et al.,1991).
While sharkskin is a phenomenon related toflow at the exit of the die, melt fracture
is related to theflow profile over the entire length of the die. Aflow curve is generated by
17
plotting theflow rate (or equivalently, the shear rate) versus the pressure drop (figure 7),
and if the slip relation is multivalued (i.e., it has a region where more than one slip
velocity is possible for a given shear rate), theflow curve itself can have a multivalued
region. Spurtflow results from a jump from the lower to the upper branch of theflow
curve when the operation is at constant pressure. Melt fracture arises due to relaxation
oscillations arising from the interaction betweenfluid compressibility and slip, when an
attempt is made to operate on the decreasing part of theflow curve at constant imposed
volumetricflow rate. Since many theoretical models for slip have a multivalued region,
the expectation is that theflow curve will also be multivalued. We show in chapter 2
that this is not always the case. If the slip velocity decreases with increasing pressure
(for which there is experimental evidence in the literature), we get the surprising result
that multivaluedness in the slip relation does not imply multivaluedness in theflow curve.
Thus, there could be certain cases where melt fracture would not be seen even if the slip
relation is multivalued.
Shear Stress (MPa)
8V/D
Figure 7: A typicalflow curve for polymers exhibiting spurt and melt fracture. The ycoordinate8V/D is proportional to the exitflow rate (Kalika and Denn, 1987).
18
The second instability that we present here is related to theflow of polymeric solu-
tions in extension. This so-called“filament stretching instability” wasfirst observed by
Spiegelberg and McKinley (1996) in the extension of a liquid bridge between two parallel
plates. Thisflow has been used to measure the extensional viscosity of polymericfluids.
A visualization of the instability is shown infigure 8. The instability takes the form of
a bifurcation to a non-axisymmetric shape near the endplates, followed by a break-up of
thefilament, and ultimately by complete detachment from one of the endplates. There is
evidence to show that elasticity plays a significant role in this instability. The instability
is never seen in Newtonian liquids, and always occurs when elastic effects are large (ex-
tensional rates on the order of the inverse polymer relaxation time). The mechanism of
this instability is not well understood, and it has been hypothesized that it is related to
the classical Saffman-Taylor instability in Newtonian liquids (Saffman and Taylor, 1958),
which occurs when a less viscous liquid displaces a more viscous one. In chapter 3, we
propose a different mechanism, much more closely related to elastic effects. We model
the region near the endplates (where the instability is seen), as a membrane in the shape
of a thin truncated cone enclosing an incompressiblefluid, and show that this is unstable
to non-axisymmetric disturbances.
(a) (b )
Figure 8: Thefilament stretching instability (a) The non-axisymmetric bifurcation seenfrom below the bottom plate (b) side view of the instability at a later stage (Spiegelbergand McKinley, 1996).
19
The third viscoelastic instability for which we present results is the purely elastic
instability in flows of viscoelasticfluids with curved streamlines. This instability was
first observed in a circular Couette geometry (two concentric cylinders with liquidfilling
the annulus, and theflow driven by one of the cylinders) by Larson et al. (1990), and a
visualization of the instability is shown infigure 9. At a certain criticalflow rate, the
smooth homogeneousflow breaks up into axially and azimuthally periodic cells. The
Reynolds number at the onset of instability in their experiment was close to zero, so
clearly the mechanism is different from the classical inertial Taylor-Couette instability
described at the beginning of this section. The mechanism of this instability is a purely
elastic one, and is related to the coupling of perturbations in the hoop stress with the base
state velocity gradients. Joo and Shaqfeh (1991) showed that a similar instability arises
when theflow is driven by imposing an azimuthal pressure gradient (Deanflow). In this
case, the mechanism is related to the coupling between a radial velocity perturbation and
the base state hoop stress.
Figure 9: Flow visualization of the purely elastic Taylor-Couette instability (Larson et al.,1990).
20
Flows with curved streamlines are common in coating operations. In coating tech-
niques such as forward and reverse roll coating (seefigure 30, and the discussion in sec-
tion 4.1), they arise as a result of the curvature of the geometry. Even in techniques such
as blade coating where the geometry is not curved, curved streamlines can be present in
recirculation regions as shown infigure 10. Instabilities that occur in theseflows impose
a limitation on the operating speed of the processing apparatus and so limit throughput.
Flows in a circular Couette geometry are also used to characterize the properties of poly-
mer solutions in shear. Given the importance of theseflows, any scheme to suppress the
instability may have considerable practical utility. Graham (1998) found that the addition
of a steady axialflow of small magnitude (compared to the azimuthal forcing) either in
Couette or Poiseuille form significantly delays the onset of the elastic instability. The
mechanism is related to the development of an additional axial normal stress induced by
the secondaryflow, which suppresses radial velocity perturbations. In chapter 4 we build
on the work of Graham (1998) by showing that the elastic instability in Deanflow can
be suppressed by the same technique, thus demonstrating its utility for a broader class of
flows. Further, we demonstrate that an oscillatory axial forcing also yields stabilization.
Figure 10: Flow visualization of a blade coating geometry. Thefigure clearly showsthe presence of an upstream recirculation region with curved streamlines (Davard andDupuis, 2000).
21
In chapter 6, we present work on nonlinear pattern formation in viscoelastic circu-
lar Couetteflow. Recent experimental observations by Groisman and Steinberg (1997)
and Baumert and Muller (1999) have shown the formation of stationary, spatially iso-
lated, axisymmetric patterns in circular Couetteflow. These patterns have been termed
“diwhirls” by Groisman and Steinberg, and as“flame patterns” by Baumert and Muller.
These patterns are very long wavelength axisymmetric vortex pairs, with a core region
of strong radial inflow, surrounded by a much larger region of weaker radial outflow. In
the absence of non-isothermal effects, the primary bifurcation in circular Couetteflow is
to a non-axisymmetric, time dependent mode (i.e., a Hopf bifurcartion). Upon further
increasing the strength of theflow, there is a secondary transition to the diwhirl structures
or flame patterns. This transition is shown infigure 11 for the diwhirls and infigure 12
for theflame patterns. If, at this point, theflow strength is reduced, these patterns persist,
until eventually the baseflow is recovered at a shear rate much lower than where thefirst
transition to the non-axisymmetric mode occurred.
t
z
(a) (b)
Figure 11: Sequence of snapshots showing the transition from non-axisymmetric disor-deredflow to solitary vortex structures. On the left in (a), the entireflow geometry isshown, with the box showing the cross section being visualized. On the right, in (b), theactual transition sequence is shown (Groisman and Steinberg, 1998).
Localized structures such as those described above are important for several reasons.
22
Figure 12: Sequence of snapshots showing the transition from non-axisymmetricflowto the predominantly axisymmetric and localizedflame patterns (Baumert and Muller,1999). Theflow geometry is the same as infigure 11.
23
Firstly, they may be connected to localized defects seen in coatingflows. Secondly, they
may form the building blocks of more complex patterns seen in viscoelasticflows, such as
the recently observed phenomenon of elastic turbulence (Groisman and Steinberg, 2000).
Finally, they can serve as a test for the ability of the approximate constitutive equations
described in the previous section to model complexflows. The modeling of these struc-
tures is complicated by the fact that the transition in the space of circular Couetteflow
occurs from a non-axisymmetric, time dependent state. Rather than undertake the ap-
proach of modeling a three-dimensional time dependentflow, we adopt an approach that
is computationally simpler. We start from stationary bifurcations inflows that are geomet-
rically similar to circular Couetteflow, and continue the bifurcating stationary solutions
into the regime of circular Couetteflow. These stationary solutions are very similar to
the experimentally observed diwhirls andflame patterns. We also compute the stability
of these solutions with respect to axisymmetric and non-axisymmetric perturbations.
So far, we have discussed purely elastic instabilities, i.e., instabilities that are caused
by the viscoelastic character of polymers. Polymers can also have significant effects on
instabilities driven by inertia. Of particular interest to us is the transition to turbulence
in plane Couetteflow. It is well known that adding a small quantity of polymer shifts
this transition to higher Reynolds numbers (Giles and Pettit, 1967; White and McEligot,
1970). Since plane Couetteflow is stable to small perturbations at all Reynolds numbers,
non-trivial solutions that may be related to coherent structures seen in turbulence have
to be obtained indirectly. Waleffe (1998) obtained such non-trivial solutions by adding
a forcing term to the Navier-Stokes equations, so that the modifiedflow had a stationary
bifurcation. What is interesting is that the non-trivial solutions emanating from the sta-
tionary bifurcation in the system with forcing persist even when the forcing is removed:
24
they exist as isolated steady states in plane Couetteflow, and show similarities to coher-
ent structures seen in turbulence. Determining the behavior of these solutions when a
polymer is added can help us understand what effect polymers have on turbulence, and
in Appendix E, we present a brief description of some preliminary work done in set-
ting up branch tracing of three dimensional Newtonian plane Couetteflow with Waleffe's
forcing. Future goals would involve coupling this with a polymer constitutive equation.
The main body of this document is divided into four chapters. Chapter 2 describes the
work on melt fracture. Chapter 3 concentrates on the instability infilament stretching.
Chapter 4 describes the stabilization of the elastic instability in Deanflow. Chapter 5
describes the work on the modeling of the diwhirl andflame patterns in circular Couette
flow. In chapter 6, we present a discussion on future directions in the area of elastic insta-
bilities. In Appendix D, we briefly discuss methods to integrate viscoelastic constitutive
equations, andfinally, in Appendix E, we present an application of continuation methods
to track solutions in three dimensional plane Couetteflow.
25
Chapter 2
Pressure dependent slip and flow
curve multiplicity†
The melt fracture instability in the extrusion of polymer melts is often linked to the fact
that the polymer does not obey a no slip condition at the wall of the die. It is thought that
if the slip law, which is a relationship between the velocity of the polymer with respect to
the wall and the shear stress, has a multivalued region, theflow curve will be multivalued
as well. In this chapter, we demonstrate that this is not always the case. In particular, we
show that adding a pressure dependence which preserves the multivaluedness of the slip
law can give rise toflow curve that is not multivalued.
2.1 Introduction
Severalflow instabilities occur during the process of extrusion of melts of linear poly-
mers, and are manifested in the form of distortions in the extrudate. A discussion of
† Most of the material in this chapter has been published in Kumar and Graham (1998a)
26
these may be found in several reviews (Denn, 1990; Larson, 1992; Leonov and Prokunin,
1994). In this work, we consider the phenomena of hysteresis and spurtflow. We can
get a clearer picture of these phenomena by examining theflow curve, which is a plot of
apparent wall shear rate (γA) versus wall shear stress (τw) during capillary extrusion. For
steadyflows in a circular capillary,
γA =4Q
πR3+4usR
, (28)
and
τw =∆pR
2L, (29)
with R andL being the radius and length of the capillary respectively,Q theflow rate,∆p
the pressure drop, andus the slip velocity of thefluid at the wall. Thus,τw is a measure
of the pressure drop, andγA is a measure of theflow rate. The unusual behavior seen
in these curves is usually attributed to the effect of wall slip (i.e., a nonzero value ofus)
because of the dependence of the phenomena on the materials of construction of the die
(Ramamurthy, 1986; Wang and Drda, 1997; Piau et al., 1995). A plot oflog(γA) versus
log(τw) starts as a straight line, indicating a power law regime. At a certain critical stress
τc1, there is a distinct increase in slope. This is followed by a sharp jump in the value
of γA at a second critical shear stressτc2. A rheometer operated in the constant pressure
mode would show a sudden jump to a higherflow rate at this value ofτw. If the pressure
is increased further,γA continues along the highflow rate branch. At this point, if the
pressure is decreased, theflow curve decreases along the highflow rate branch even below
τc2, until there is a jump to the lowflow rate branch at a third critical stressτc3, less than
τc2. Thus, theflow curve exhibitshysteresis (Bagley et al., 1958; Tordella, 1956, 1963;
El Kissi and Piau, 1990). This hysteresis implies that theflow curve displays a region of
27
multiplicity: two stableflow rates are possible at the same pressure drop. If a rheometer is
operated in the constant piston speed mode in the region betweenτc2 andτc3, the pressure
and extrudateflow rate show oscillations (Lupton and Regester, 1965; Kalika and Denn,
1987; Ramamurthy, 1986; Hatzikiriakos and Dealy, 1992a). This phenomenon is known
as spurtflow.
The dependence of the critical shear stressτc2 on the length to diameter ratio (L/D)
of the die can be inferred from several experimental works in the literature. The data
of Vinogradov et al. (1984) for polybutadienes and Wang and Drda (1996) for entangled
linear polyethylenes, which are taken at lowL/D (≤ 25), suggest thatτc2 is virtually
independent ofL/D. The data of El Kissi and Piau (1990) taken at somewhat higher
L/D ratios (L/D = 20 and 40) for polydimethylsiloxanes, corrected for entrance losses
show that there is a small decrease in this value (from 0.61 bar to 0.59 bar) for the constant
pressure experiments where theL/D ratio was varied byfixing D and varyingL. Kalika
and Denn (1987) report data for the constant piston speed experiments using LLDPE
which show thatτc2 decreases asL/D is varied from 33.2 to 66.2, and atL/D = 100.1,
their reported value ofτc3 is greater thanτc2 (in this case, theflow curve shows a large
jump in flow rate in the relatively small region betweenτc2 andτc3) which indicates that
there may not be multiplicity if the rheometer is operated in the constant pressure mode at
highL/D. Finally, Hatzikiriakos and Dealy (1992a) report that they observe an increase
in τc2 with increasingL/D (from L/D = 10 to L/D = 100) for linear high density
polyethylene, and take this as evidence of a pressure dependence of wall slip. Thus, there
seems to be evidence in the literature to support theL/D dependence, but there does not
seem to be an agreement on the trend, and this may differ depending on the polymer used
and other experimental conditions.
28
In addition to experimental evidence for multiplicity in slip behavior, the motivation
for using a multivalued slip model arises from the slip relations obtained by several re-
searchers using microscopic theories for polymers. Several of these theories explicitly
predict multivaluedness in the relation between slip velocity and shear stress, i.e., there
is a region in which three distinct values of slip velocity are possible for the same value
of shear stress. Specific examples of such relations include those proposed by Leonov
(1990); Yarin and Graham (1998); Mhetar and Archer (1997); Hill (1998). In particular,
in the model of Yarin and Graham (1998), the limit point at which the jump from a small
slip to a large slip regime occurs arises from an imbalance between the increasing force
per adsorbed chain and the decrease in the concentration of adsorbed chains, as shear
stress increases. Multivalued slip relations such as these have been used, for example, to
model spurt and oscillations in capillary and Couetteflows of molten polymers (Adewale
and Leonov, 1997).
Georgiou and Crochet (1994) have proposed a computationally convenient phenomeno-
logical slip equation which also shows a maximum and minimum in wall shear stress. Al-
though not derived fromfirst principles, this model has a closed form expression, and can
be viewed as an approximate version of models based on molecular parameters which do
not have closed form expressions. They showed that this slip relation, when taken along
with finite compressibility of the polymer melt, can lead to self sustained oscillations of
the pressure drop and massflow rate at the exit of the die for Newtonianflow in a slit.
However, they do not include a pressure dependence in their slip relation.
Several slip models relating the slip velocity to the wall shear stress and pressure
(or total compressive normal stress on the wall) have been proposed in the literature. In
particular, Hill et al. (1990); Denn (1992); Person and Denn (1997) propose slip relations
29
which show a power law dependence on the wall shear stress and an exponential one
on pressure. Hatzikiriakos and Dealy (1992b) have proposed a similar slip model which
shows saturation at high pressures. They used this model for the lowflow rate branch and
a simple power law dependence on the shear stress for the upper branch of theflow curve
to modelflow oscillations in a capillary rheometer (Hatzikiriakos and Dealy, 1992a).
Thus, their model is in effect a discontinuous multivalued slip relation. The slip model
of Stewart (1993) brings in the pressure dependence through changes in the density. All
these models predict a decrease in slip velocity with increase in pressure. There is also
experimental evidence from White et al. (1991) that this is indeed the case.
In this work, we study the combined effects of adding pressure dependent and mul-
tivalued slip onflow curve multiplicity in capillary extrusion. We do this by modifying
the approximate slip relation proposed by Georgiou and Crochet and adding a pressure
dependence. Consistent with observations, the pressure dependence is such that slip ve-
locity decreases with increasing pressure. As with the model proposed by Georgiou and
Crochet, our relation is not derived fromfirst principles, but contains the same qualita-
tive features found in more complex models derived from molecular considerations. We
apply this equation to the steadyflow of incompressible Newtonian and shear thinning
fluids through a cylindrical die at a constant pressure drop. As also done by Person and
Denn (1997), we simulate the entire axial profile of pressure, stress and slip velocity. The
flow curve obtained shows multiplicity, with the critical shear stressτc2 showing a small
decrease at lowL/D ratios and a more pronounced one at higherL/D ratios. Most im-
portantly, we see that at sufficiently highL/D, theflow curve is no longer multivalued,
despite the fact that the slip relation is multivalued, i.e., it predicts a maximum and mini-
mum in shear stress at all pressures between the entry and exit pressures in the die. Thus
30
we obtain the result that multiplicity in the slip relation does not guarantee multiplicity
in theflow curve.
2.2 Mathematical Model
2.2.1 Slip Model
In terms of dimensionless quantities, our slip model may be written as
τw = usp
(usp +
A21 +A3u2sp
), (30)
where
usp = us(1 + exp(β P )). (31)
Here,τw is the shear stress exerted by thefluid on the wall andus is the slip velocity. The
scaling factors used areu∗ = G/a1, P ∗ = G andτ∗w = G for the velocity, pressure and
shear stress respectively,a1, A2, A3 andβ are parameters of the slip model andG is the
shear modulus of thefluid. For smallusp, this model reduces to the Navier slip condition,
τw = A2usp.
If we arbitrarily chooseA3 = 122.24, we can calculate that the slip model loses
multiplicity for A2 < 2.29. Figure 13 shows a plot of the slip relation for various values
of P for A2 = 3, A3 = 122.24, andβ = 0.0102. Unless otherwise mentioned, we
will work with A2 = 3 andA3 = 122.24 in subsequent calculations. Note that the
pressure dependence is such that the slip velocity decreases with increase in pressure, as
shown infigure 13, but the multiplicity remains. Since the slip model is multivalued by
construction, the natural expectation is that theflow curve will be as well. We shall see
that this expectation is not always fulfiled.
31
P=10
P=50
τw
us
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.00 0.04 0.08 0.12 0.16 0.20
Figure 13: Plot of equation 30 for different values of pressure (A2 = 4, A3 = 122.24,β = 0.0102).
2.2.2 Constant Pressure Formulation
We now develop the governing equations for the constant pressure case using Newto-
nian, shear thinning and UCM constitutive equations. We have already described the
UCM equation in the introduction as a constitutive equation for obtaining the polymer
component of the stress tensor in dilute solutions. It can also be derived using network
models for polymer melts (Bird et al., 1987b). In this case,Q can be thought of as the
length of a polymer segment between two junctions in the network, with the segment
being modeled as a Hookean spring. As shown by Lupton and Regester (1965),fluid
compressibility does not play an important part in the constant pressure case, and we can
simplify our analysis slightly if we assume that thefluid is incompressible. However, for
the constant piston speed case,fluid compressibility plays a crucial role in providing a
mode for storing energy, which is essential for the generation of relaxation oscillations.
32
We first consider the pressure drivenflow of an incompressible Newtonianfluid
through a cylindrical die of radiusR (diameterD) and lengthL. The pressure is de-
fined so that it is zero at the exit of the die. Using the lubrication approximation (Pearson
and Petrie, 1968), the dimensionless volumetricflow rate is given by
Q = us +τwH
, (32)
Here, we assume that the diameter of the die is kept constant and choose to scale the
volumetricflow rate usingQ∗ = πR2u∗. In addition, we scale the lengths byL and
retain the scaling factors for pressure, slip velocity and shear stress discussed above. The
dimensionless numberH = 8ηu∗/GR, whereη is thefluid viscosity. Note that since
γA = 4Q/πR3, QH measures the dimensionless apparent shear rate. The use of the
incompressibility conditiondQ/dz = 0 gives us the following equations for the slip
velocity and pressure:
dusdz
= −dP
dz
∂τw∂P
H + ∂τw∂us
, (33)
dP
dz= −4Λτw, (34)
whereΛ = L/D. Equation 33 is obtained by using equation 32 and differentiating
equation 30 with respect toz. This set of equations is to be solved using the boundary
conditions
P (0) = ∆P, (35)
P (1) = 0. (36)
Since it will be convenient to calculate theflowrate simultaneously with the other
33
quantities, we add the redundant equation and boundary condition
dQ
dz= 0 (37)
Q(0) = us(0) +τw(0)
H. (38)
For a power lawfluid, the constitutive relation between the shear stress exerted by the
fluid on the wall and the radial velocity gradient is given by Bird et al. (1987a) as
τrz = −K
∣∣∣∣∂vz∂r
∣∣∣∣n−1
∂vz∂r
. (39)
Using the lubrication approximation once again, the dimensionless volumetricflow rate
is given by
Q = us +nτ1/nw
H, (40)
whereτw is obtained from equation 30. We define the dimensionless numberH in a
manner similar to the Newtonian case as
H =n(3 + 1/n)K1/n
a1G1/n−1R, (41)
which results in the following set of equations:
dusdz
= −∂τw∂P
∂τw∂us+ H
τ(1/n−1)w
dP
dz, (42)
dP
dz= −4Λτw, (43)
dQ
dz= 0. (44)
These equations have to be solved together with the boundary conditions:
P (0) = ∆P, (45)
P (1) = 0, (46)
Q(0) = us(0) +nτ1/nw (0)
H. (47)
34
The governing equations for the UCM constitutive equation are similar to those for the
Newtonian model. The inclusion of a normal stress only changes the boundary conditions
at the entrance and exit. Neglecting the elastic normal stresses in the barrel, and entrance
and exit effects, these are given by:
P (0) + τzw(0) = Pb, (48)
P (1) + τzw(1) = 0. (49)
Here,τzw is the normal stressτzz evaluated at the wall. At steady state,τzw = −2 τ2w.
We can examine the combined effect of viscoelasticity and shear thinning by using the
Phan-Thien-Tanner (PTT) equation, which can also be derived from network theory (Bird
et al., 1987b). The general form of the PTT equation is given by Bird et al. (1987a) as
Z(tr τ )τ + λ τ (1) +ξ
2λγ · τ + τ · γ = −η γ. (50)
We consider the case of affine motion (ξ = 0) and use the linear form of the functionZ,
i.e,
Z = 1− ε λ tr τ/η.
In this case, the governing equations can be written as
dus
dz= −
dP
dz
∂τw∂P(1 + 4ε/H τ2w)
H + ∂τw∂us(1 + 4ε/H τ2w)
, (51)
dP
dz= 0, (52)
dQ
dz= 0. (53)
35
The boundary conditions are given by
Q(0)− us(0) =τw(0)
H+4ε
3Hτ3w(0), (54)
P (0) + τzw(0) = Pb, (55)
P (1) + τzw(1) = 0. (56)
Our simulations show that the addition of viscoelasticity in the form of either the UCM
or the PTT equation does not change the qualitative behavior of the model.
2.2.3 Constant Piston Speed Formulation
We now consider theflow of a Newtonianfluid through a capillary die as shown infig-
ure 14. Thefluid is driven by a piston moving at constant speed. We assume that the melt
has a constant compressibilityχ, and use a linear relation for the density, i.e.,
ρ = ρ0(1 + χP ), (57)
whereρ0 is the density atP = 0.
Wefirst write the governing equations for the Newtonian case. A mass balance on the
barrel gives
CdPb
dt=1
κ(Qp −Q0) , (58)
wherePb is the dimensionless pressure in the barrel (scaled byG), t is the dimensionless
time (scaled by the residence timet∗ = Qp/πR2L), C = Vb/πR
2u∗t∗ andκ = Gχ
(cf. Lupton and Regester (1965) and Molenaar and Koopmans (1994)). The volumetric
rate of displacement of the piston,Qp is constant during the simulation andQ0 is the
volumetricflow rate at the exit of the barrel. We assume thatC remains constant in the
36
L
diameter = D
die
barrel
plunger
Vb
dVb / dt =Qp
Figure 14: Schematic diagram of the constant piston speed experiment.
time scale of the experiment. A mass balance on the capillary die gives
∂P
∂t= −λ1
(Q
∂P
∂z+1
κ
∂Q
∂z
), (59)
with λ1 = t∗u∗/L. The boundary conditions areQ = Q0, andP = Pb at the entrance of
the die. Finally, integrating the momentum balance overr gives
Re∂Q
∂t= −λ2
(1
4Λ
∂P
∂z+ τw
), (60)
whereλ2 = t∗G/η, Re = ρ0u∗D/4η is the Reynolds number andη is thefluid viscosity.
The momentum equation in this form is also valid for viscoelasticflow if ∂τzz∂z ∂P
∂z.
Numerical results for viscoelasticflow show that this condition is satisfied.
We now need an equation relating the volumetricflow rate to the slip velocity. If the
compressibility is small, we can assume that equation 32 still holds. Hence, the governing
equations for the Newtonian case are equations 58 to 60 together with equation 32 and
the slip relation, equation 30. The same set of equations hold for the shear thinning case,
with the exception that equation 32 is replaced by equation 40.
37
As mentioned before, viscoelasticity does not significantly affect the steady state
equations. However, it is possible that using an evolution equation for the normal and
shear stresses could yield time-dependent results that are qualitatively different from
those for the Newtonian and shear thinning cases. Thus, we felt that it was necessary
to examine the time-dependent case using a simple viscoelastic model. This section dis-
cusses the formulation of the governing equations using the UCM model for viscoelastic-
ity (Bird et al. (1987a)). Assuming that the Reynolds number is small, the UCM model
gives the following evolution equation forτw:
De
2ΛDe∗∂τw∂t=
τwDe∗
− 8(Q− us). (61)
Evaluating the evolution equation forτzz at the wall gives
De
2Λ
∂τzw∂t
+De τwΛ
∂τw∂t= −2τ2w − τzw. (62)
Here,τzw represents the value ofτzz at the wall of the capillary andDe∗ = λu∗/R. The
Deborah number is given byDe = λ〈v〉/R, where〈v〉 = Qp/πR2.
The slip velocity and shear and normal stresses are related by the slip model. The
boundary conditions are:
P (0) + τzw(0) = Pb, (63)
P (1) + τzw(1) = 0. (64)
In the first boundary condition, we assume that the elastic normal stresses in the barrel
are negligible. These boundary conditions neglect entrance and exit effects.
38
2.3 Results and Discussion
2.3.1 Constant Pressure Case
We wish to determine the parameter regimes in whichflow curve multiplicity occurs. A
natural way to do this is through bifurcation analysis. AtfixedΛ, we find the turning
points of theflow curve (i.e. the points where the curve turns back on itself). These
determine the boundariesτc2 andτc3 of the multiplicity region at thatΛ (in the model
we consider, slip occurs at all shear stress levels and henceτc1 does not exist). Then
we track the motion of these points asΛ varies, thus outlining the region in(∆P,Λ)
space whereflow curve multiplicity occurs. The AUTO software package (Doedel, 1981;
Taylor and Kevrekidis, 1990) automatically performs this type of analysis for boundary
value problems like the one we consider here. AUTO uses a domain decomposition
collocation method for spatial discretization and a pseudo-arclength continuation scheme
to trace out steady state solution curves in one parameter or curves of bifurcation points
in two parameters.
Figure 15 shows theflow curve computed by AUTO atΛ = 10,A2 = 3, A3 = 122.24,
β = 0.0102 andH = 1. We see that this curve is multivalued, as expected because the
slip model is multivalued. The turning points occur at∆P = 5.8090 and∆P = 5.6506,
corresponding to critical shear stresses ofτc2 = 0.1452 andτc3 = 0.1413.
The results of the turning point continuation are shown infigure 16a, where the upper
curve corresponds toτc2, and the lower one toτc3. Note that our slip law depends on
pressure, so the slip velocity, and henceτw changes through the length of the die. The
values ofτc2 andτc3 that we report here are computed by dividing the pressure drop∆P
by 4Λ. This is what the (constant) wall shear stress would be in the absence of slip,
39
X
X
Q
∆ P
0.16
0.20
0.24
0.28
0.32
5.0 5.5 6.0 6.5
Figure 15: Flow curve for a Newtonian Fluid (Λ = 10, A2 = 3, A3 = 122.24, β =0.0102, H = 1). The profiles at the points marked `X' are shown infigures 18 and 19.
40
under the same pressure drop. As thefigure shows, the value ofτc2 shows a decrease and
the two critical shear stresses approach one another as the length is increased, and this
effect is more pronounced at higherΛ values. Finally, atΛ = 230.39, we see that the the
two turning points come together in a cusp, indicating that theflow curve beyond thisΛ
value has no multiplicity. The pressure drop at this value is132.19 and the slip relation
still predicts a multiple valued curve for all pressures in this range. Thus, the differential
effect of pressure in the capillary has resulted in the absence of multivaluedness in the
flow curve, although the slip relation itself is multivalued for the entire pressure range
experienced in the die. We also note that at the cusp, the critical stresses have a value of
0.1434, which corresponds to a change of only1.24% from the value ofτc2 atΛ = 10.
Finally, for completeness, we show a plot of theflow curve forΛ = 238.62 in figure 17,
where there is no multiplicity, just as shown onfigure 16a.
Results for the shear thinning case withn = 0.56 and the other parameter values as
for the Newtonian model are similar. The results of the turning point continuation are
shown infigure 16b. The behavior is similar to the Newtonian case where there is a more
pronounced drop inτc2 at higherΛ values. Also, theflow curve loses multiplicity at
Λ = 230.46 and a pressure drop of132.19, which corresponds toτc2 = τc3 = 0.1434.
As with the Newtonian case discussed in the preceding paragraph, the slip relation itself
remains multivalued for all pressures experienced in the die, and the loss of multiplicity
in theflow curve is a result of the differential effect of pressure in the capillary.
As mentioned earlier, wefind that including the effect of viscoelasticity does not
affect the qualitative behavior of the model, in particular, the loss of multiplicity at high
Λ. The result of a turning point continuation using the UCM constitutive equation is
41
τ c
τ c
Three steady states
Three steady states
Three steady states
Three steady states
One steady state
One steady state
One steady state
One steady state
One steady state
One steady state
One steady state
One steady state
Λ
τ c(d)
(c)
(b)
τ c(a)
0.141
0.143
0.145
0.141
0.143
0.145
0.141
0.143
0.145
0.141
0.143
0.145
50 100 150 200
Figure 16: Behavior of critical stressesτc2 andτc3 with increasingΛ (A2 = 3, A3 =122.24): (a) Newtonian model (β = 0.0102, H = 1) (b) shear thinning model (β =0.0102, H = 1, n = 0.56) (c) UCM model (β = 0.0102, H = 1) (d) PTT model(ε = 10−1, β = 0.0102, H = 1). The upper curve corresponds toτc = τc2 and the lowerone toτc = τc3. The curves are not extended toΛ = 0 because the approximations usedare not valid for smallΛ.
42
Q
∆ P
0.16
0.18
0.20
0.22
0.24
125 130 135 140 145 150
Figure 17: Flow curve for a Newtonianfluid showing no multiplicity (Λ = 238.62,A2 = 3, A3 = 122.24, β = 0.0102, H = 1).
shown infigure 16(c). A similar result for the PTT model, withε = 0.1 is shown in
figure 16(d). In both cases, the same parameters were used as for the Newtonian model.
Figures 18 and 19 show the spatial profiles of the pressure, shear stress and slip veloc-
ity at points on the low and highflow rate branches of theflow curve for the Newtonian
fluid shown infigure 15. The profiles at a point on the lowflow rate branch are shown in
figure 18 and those at a point on the highflow rate branch are shown infigure 19. In both
cases, the pressure profiles are nearly linear and the shear stress decreases as we move
towards the exit of the die. The degree of variation in the magnitude of the shear stress is
much smaller than that of the pressure and correspondingly, the slip velocity increases to-
wards the exit where the pressure is lowest. This behavior is expected intuitively and also
seen by Hatzikiriakos and Dealy (1992b) in their simulations at constant piston speed.
43
However, it differs from the results of Person and Denn (1997) where slip velocity has
the smallest magnitude at the exit, which is the region of lowest shear stress. They inter-
pret this result as arising from the strong power law dependence of slip velocity on shear
stress. Finally, the small variation inτw with z provides ana posteriori validation of our
neglect of axial gradients.
τ w
P
us
z
0
2
4
6
0.1302
0.1304
0.1306
0.0 0.2 0.4 0.6 0.8 1.0
0.0304
0.0306
0.0308
Figure 18: Profiles of pressure, slip velocity and shear stress on the lowflow rate branchfor the Newtonianfluid of figure 15. (∆P =5.22,Λ = 10, A2 = 3, A3 = 122.24,β = 0.0102, H = 1).
44
z
τ w
us
P
0
2
4
6
0.139
0.140
0.141
0.142
0.158
0.159
0.160
0.161
0.0 0.2 0.4 0.6 0.8 1.0
Figure 19: Profiles of pressure, slip velocity and shear stress on the highflow rate branchfor the Newtonianfluid of figure 15 (∆P =6.39,Λ = 10, A2 = 3, A3 = 122.24,β = 0.0102, H = 1).
45
It is of interest to observe the location of the profiles of the two turning pointsτc2
andτc3 on the three-dimensional surface defined by the slip model (equation 30). The
pressure profile is almost linear for all cases, and hence, the plots may also be viewed as
the profiles superimposed on the slip curve in(z, us, τw) space, with the higher pressures
corresponding to lowerz values, i.e., points close to the entrance of the die. Such a plot
is shown for the shear thinning case withΛ = 10 in figure 20(a). At this value ofΛ, the
pressure drop is relatively small, and the profiles at the two turning points lie very close
to the location of the maxima and minima of equation 30 at all pressures in the die.
The location of the profile at the cusp is shown infigure 20(b). Here, the high pressure
points (which lie near the entrance of the die) are located to the right of the minimum at
the corresponding pressure. As we move down the die to regions of lower pressures, the
points tend to move toward the right, i.e., closer to the maxima, till at the exit, the points
are located to the left of the maximum. Finally,figure 20(c) shows the profile of a point
in the central region (i.e., the region where there is a sharp increase in the slope of the
flow curve, although there is no multiplicity).
2.3.2 Constant Piston Speed case
For the constant piston speed case, we performed a spatial discretization of the governing
equations using a Chebyshev collocation scheme (Canuto et al., 1988) withN = 30 collo-
cation points for the pressure, volumetricflow rate and slip velocity. The resulting system
of differential-algebraic equations was solved using the integrator DDASAC (Caracotsios
and Stewart, 1985), for imposed values ofQp and zero initial conditions. For the base
case, we chose values ofχ = 10−9m2/N andG = 105N/m2 andRe = 10−10. In all
46
02
4P 0.050.1
0.15
us
0.05
0.15Τw
02
4
050
100P0.05
0.1
us
0.10.20.3
Τw
050
100
050
100P0.05
0.1
us
0.10.20.3
Τw
050
100
(a)
(b)
(c)
Figure 20: Flow profiles on the three dimensional slip surface for a shear thinningfluid:(a) upper and lower limit points (∆P = 6.39, Λ = 10) (b) cusp point (∆P = 132.24,Λ = 230.46) (c) point in the central portion without multiplicity (∆P = 140.0, Λ =240.82). Other parameters areA2 = 3, A3 = 122.24, β = 0.0102, H = 1, andn = 0.56for all three cases.
47
runs,C wasfixed at106 and the parametersλ1 andλ2 werefixed at0.01 and0.1 re-
spectively. All other parameters were chosen to have the same values as in the constant
pressure case.
Results for the Newtonian and shear thinning cases are similar and we only show them
for the Newtonian case. As expected because of the multiplicity in the constant pressure
flow curve, attempting to operate at aflow rate corresponding to the branch betweenτc2
andτc3 of the flow curve (which we call the unstable region) results in oscillations in
the pressure dropPb and the exitflow rateQe. This is shown infigure 21, for a value
of Qp = 0.21 where the steady state solution lies in the unstable region of theflow
curve. Operating on the low or highflow rate branches results in stableflow as shown in
figure 22, forQp = 0.15, where the steady state solution lies on the lowflow rate branch
of theflow curve. These results are similar to those obtained by Georgiou and Crochet
(1994) and Adewale and Leonov (1997), and are to be expected when multivalued slip
models are used.
As long asRe 1, the Reynolds number has virtually no effect on the frequency
of the oscillations. This can be understood by noting that, forRe 1, these are basi-
cally classical relaxation oscillations, as occur in problems with widely separated time
scales (Nayfeh and Mook, 1979). Here, the time scales correspond to the compressibility
(κC) and inertia (Re). Hence,Re must be nonzero for oscillations to be observed, but
is otherwise not important as long as it is small. Decreasing the compressibility has the
effect of increasing the frequency of the oscillations. This can be understood on the same
lines as theRe dependence by noting that the time spent on the high and lowflow rate
branches of theflow curve is directly related to the compressibility. Fluids with lower
48
(b)
(a)
Qe
Pb
Qe
Pb
time
5.6
5.7
5.8
5.9
0.16
0.18
0.20
0.22
0.24
0.26
0.28
0.30
6000 6500 7000 7500 8000 8500 9000 9500
0.16
0.18
0.20
0.22
0.24
0.26
0.28
0.30
5.6 5.7 5.8 5.9
Figure 21: Oscillatoryflow for a Newtonianfluid with Qp = 0.21:(a) Pressure drop andexit flow rate vs. time (b) Plot ofQe vs.Pb superimposed on the steady stateflow curve.
49
time
Qe
Pb
0
1
2
3
4
5
0.00
0.05
0.10
0.15
0 5000 10000
Figure 22: Non-oscillatoryflow for a Newtonianfluid with Qp = 0.15: Pressure dropand exitflow rate vs. time.
50
values ofχ would spend less time on these branches and hence exhibit oscillations of a
higher frequency.
Figure 23 shows the results for a simulation atΛ = 250, where the steady state
flow curve shows no multiplicity. Recall, however, that the slip model itself remains
multivalued, and as in the constant pressure case, the loss of multiplicity is a result of
the differential effect of pressure along the capillary. The value ofQp was chosen to be
0.1862, which lies in the steepest region of theflow curve. As expected, no oscillations
are found.
We conclude with a brief discussion of the results obtained using the UCM consti-
tutive equation, with the same slip model parameters as for the Newtonian case. In this
case, high frequency shock waves are seen at very small values of the Reynolds number
and large values ofκ. To avoid these shock waves, we present results for a parameter set
different from that used for simulations with the Newtonian model. Figure 24 shows the
result of a simulation withRe = 10−4. As before, wefind that Reynolds number does not
affect the frequency of the oscillations provided that it is sufficiently small. We alsofind
that, as for the Newtonian case, decreasing the compressibility increases the frequency of
the oscillations.
2.4 Conclusions
In this chapter, we have studied the behavior of extrusion when a multivalued, pressure
dependent model for wall slip is used. The main features of spurtflow and multiplicity
can be modeled using this relation. The steady stateflow curves using Newtonian, shear
thinning and viscoelastic models predict a critical shear stressτc2 for the onset of spurt
51
time
Qe
Pb
0
50
100
150
0.00
0.05
0.10
0.15
0.20
0 50000 100000 150000 200000 250000 300000
Figure 23: Non-oscillatoryflow of a Newtonianfluid with Λ = 250, Qp = 0.1862 andother parameters as infigure 21: Pressure drop and exitflow rate vs. time.
52
Qe
Pb
Qe
Pb
time
5.6
5.7
5.8
5.9
0.04
0.06
0.08
0.10
0.12
500 600 700 800 900 1000 1100
0.04
0.06
0.08
0.10
0.12
5.6 5.7 5.8 5.9
Figure 24: Oscillatoryflow of a UCMfluid (C = 106, Λ = 10, Re = 10−4, κ = 10−5,De = De∗ = 10, λ1 = λ2 = 0.01, Qp = 0.081): (a) Barrel pressure and exitflow ratevs. time (b) Plot ofQe vs.Pb superimposed on the steady stateflow curve.
53
flow that decreases with increase inΛ. In all cases, we see a loss in multiplicity at highΛ,
although the slip relation itself gives a multivalued slip velocity for a given shear stress
at all pressures in the range involved.
Time-dependent simulations takingfluid compressibility into account show that at-
tempting to operate on the decreasing branch of theflow curve gives rise to oscillations
in the exitflow rate and pressure drop, which is in accord with the results of previous
researchers using multivalued slip models. The frequency of these oscillations is virtu-
ally independent of the Reynolds number, provided that it is small enough. Decreasing
compressibility increases the frequency of the oscillations. No oscillations are observed
if the flow curve does not show multiplicity.
Although there is little experimental data in the literature for experiments conducted
at highΛ values at constant pressure, the behavior we observe is similar to that reported
in Kalika and Denn (1987) for a constant piston speed experiment. The main conclusion
of this work is that multiplicity in the slip model does not guarantee multiplicity in the
flow curve.
54
Chapter 3
The filament stretching instability †
In this chapter, we discuss the instability that occurs during the elongation of a polymer
filament. Here, an initially axisymmetricfilament undergoes a buckling instability pro-
ducing a non-axisymmetric shape. We use the idea of stress localization to model the
process as the stretching of an elastic membrane enclosing a passive, but incompressible
fluid. We show that this simple model exhibits an instability that is similar to that seen in
the stretching of a polymerfilament.
3.1 Introduction
Flows with significant elongational components are common in industrial applications
such asfiber spinning, as well as in rheometry, during the measurement of elongational
viscosity. The most common types of elongationalflow instabilities are draw resonance,
necking, capillary breakup, and elasticfilament breakup, Detailed reviews of which may
be found in Petrie and Denn (1976) and more recently by Larson (1992), so we content
† Most of the material in this chapter has been published in Kumar and Graham (2000)
55
ourselves with a brief overview of these instabilities. We then summarize the observa-
tions in the literature that are related to elasticfilament breakup, and postulate a new
mechanism by which this instability could occur.
Fiber spinning is defined (Petrie and Denn, 1976) as a process whereby afilament or
sheet is extruded from a die and drawn down in cross sectional area by being taken up at
a velocity greater than the extrusion velocity. The draw ratioDr, is defined as the ratio
of the take-up velocity to the extrusion velocity. This also equals the ratio of the initial
to drawnfiber area. Draw resonance is described by Christensen (1962) as a periodic
variation infiber diameter, which occurs at constant take-up speed, when small variations
in the diameter offiber at the take-up spool produce oscillations in thefiber tension. The
analysis of Pearson and Matovich (1969) for a Newtonianfluid with negligible inertia,
gravity, and surface tension showed that the steady state solution is unstable to time-
periodic disturbances whenDr exceeds20.21. A later analysis by Pearson and Shah
(1972) showed that inertia acts to stabilize theflow, while surface tension is destabilizing.
The corresponding stability analysis for viscoelasticfluids has been performed by Ze-
ichner (1973), Fisher and Denn (1975, 1976), and Chang and Denn (1980). An important
parameter in this case is the Deborah number, defined asDe = λ v(0)/L, whereλ is
the relaxation time of thefluid, v(0) is the extrusion velocity, andL is the length of the
fiber. The analysis of Fisher and Denn (1976) for a UCMfluid shows that elasticity has a
stabilizing influence. The analysis also shows that there is no steady solution for certain
values ofDe, an effect related to the unbounded growth of normal stress predicted by the
UCM equation for elongationalflow, at certain Deborah numbers. Nonlinear analyses
carried out for Newtonianfluids (Kase, 1974; Fisher and Denn, 1975) and for power law
56
and UCMfluids (Fisher and Denn, 1976) show thatfinite amplitude effects are stabiliz-
ing. Finally, fluids that thin in elongation are found to be less stable to draw resonance
than Newtonianfluids, while elongation thickeningfluids are more stable (Pearson and
Shah, 1974; Fisher and Denn, 1976).
When thefiber is extruded at a constant tension rather than a constant take-up speed,
Fisher and Denn (1976) found that theflow is stable to draw resonance. However, small
local indentations can grow as they are convected down the length of thefiber. This
phenomenon is known as necking. Necking occurs because the constant force applied
along thefiber length translates to a greater elongational stress in the neck as compared to
adjoining regions in thefiber. Thus, the neck thins faster than the neighboring regions. It
follows that elongation thinning, which speeds up the reduction in area, is destabilizing,
while elongation thickening leads to increased stability. They also showed that if the
residence time of the neck in thefiber is long enough, necking can lead tofiber breakage.
The phenomenon of capillary breakup wasfirst analyzed by Rayleigh (1879). He
showed that a jet of Newtonian liquid with nonzero surface tension is unstable to ax-
isymmetric disturbances whose wavelength is greater than the circumference of the jet.
For fiber spinning, the capillary breakup instability augments necking, but the additional
effect can be shown to be negligible (Larson, 1992). For UCMfluids, the analysis of
Goldin et al. (1969) showed that in the linear limit, disturbances have higher growth rates
than for Newtonianfluids. However, nonlinear effects quickly become significant, and
the lubrication andfinite element analysis of Bousfield et al. (1986) which takes these
into account shows that the growth rates slow down significantly, so that the net effect is
a stabilization compared to Newtonianfluids. Once again, it is the elongation thickening
property of the UCMfluid that leads to increased stability, by resisting the thinning and
57
breakup of threads. While the capillary breakup instability is not of significance infiber
spinning, it has been suggested (Larson, 1992) that the measurement of growth rates of
the disturbances may be used to estimate the surface tension of viscoelasticfluids.
The next instability that we discuss is the breakup of a viscoelastic liquid bridge
in elongation. Unlike the instabilities we have discussed so far, this instability is purely
elastic, in the sense that it is not seen in corresponding experiments with Newtonianfluids
with similar viscosity. The elongation of a liquid bridge confined between two endplates
is used in rheometry to measure the elongational viscosity of afluid (Tirtaatmadja and
Sridhar, 1993; McKinley et al., 1996). A schematic of the setup is shown infigure 25.
In ideal uniaxial elongationalflow, a liquid column of initial lengthL0 is extended at
an exponential rate so that its length at timet is given byLt = L0 exp(ε0 t), whereε0
is the strain rate. The radius of thefilament remains uniform throughout its length. The
elongational viscosity is then defined as
η ≡ (τzz − τrr) /ε0, (65)
whereτzz andτrr are respectively the axial and radial normal stresses. This stress differ-
ence, and henceη, may be calculated from the force required to produce the elongation.
A measure of the displacement is given by the Hencky strainε, given byloge(Lt/L0).
The Deborah number is given byDe = λ ε0, whereλ is the relaxation time of thefluid.
A filament stretching rheometer, such as the one shown infigure 25 attempts to repro-
duce uniaxial elongation by separating the endplates at an exponential rate. The actual
kinematics differ from this due to the presence of the two endplates. While conducting
experiments with polyisobutylene based Bogerfluids in afilament stretching rheometer,
McKinley et al. (1996) found that the liquid bridge partially decohered from one of the
endplates at high strain rates (corresponding toDe > 1). The instability began as a
58
polymer solutionL0
Lt
Figure 25: Schematic of afilament stretching rheometer. The setup on the left shows theundeformed state of the liquid bridge.
non-axisymmetric deformation near the endplates at a certain critical strain, with a well
defined azimuthal wavenumber of 4. Upon further elongation, the column broke into
fibrils, which themselves bifurcated into smallerfibrils at larger elongations. The critical
strain for onset of the instability was found to decrease with increasing Deborah num-
ber. The dependence on the initial aspect ratio (defined asL0/R0, whereR0 is the initial
filament radius) was weaker, with the critical strain increasing slightly at larger aspect
ratios. They report that Newtonian liquid columns of similar viscosity do not display the
instability, and in fact break up due to the capillary instability at comparable strains.
McKinley et al. (1996) propose that the mechanism is similar to the one seen in the
Saffman-Taylor instability (Saffman and Taylor, 1958), which occurs when a less viscous
fluid attempts to displace a more viscous one. In this mechanism, a small perturbation
at the interface between the twofluids causes the less viscousfluid to find itself in a
region of lower pressure, hence amplifying the perturbation. The Bogerfluids used in
59
the experiment of McKinley et al. (1996) are strongly strain hardening. This means that,
at large strains, the central portions of the liquid bridge which are highly stretched resist
further decrease in diameter. This in turn means that a large portion of thefilament is
being extended as a cylinder of almost constant radius, with the additional volume of
fluid being supplied by the rapidly depleting pool of liquid near the two endplates. At
large strains, the thickness of this pool is small compared to its radius, so that theflow of
liquid in this region is radially inward. This indicates that the pressure at the center of the
pool must be lower than at the edge. If the free surface near this region is perturbed, the
surrounding medium (in this case, air at atmospheric pressure)finds itself in a region of
lower pressure, which causes the interface to deform further.
This instability is relatively new, so not much work has been done in terms of model-
ing it. Spiegelberg and McKinley (1998) performed a numerical simulation of the stretch-
ing of afilament of Oldroyd-Bfluid using a commercial software program based on the
finite element method. They reported that the initialflow inhomogeneity caused the for-
mation of stress boundary layers near the free surface of thefilament at large strains.
For Hencky strainsε > 4, they found that it became increasingly difficult to model the
rapidly draining region near the endplates. Significantly, the elastic instability sets in
close toε = 4 for large Deborah numbers.
Rasmussen and Hassager (1999) performed an analysis of the stretching of a UCM
filament. They considered two cases: thefirst was was the elongation of a purely cylin-
drical filament, and the second was the elongation a cylindricalfilament with a small
uniform non-axisymmetric perturbation superimposed on it. For the non-axisymmetric
initial state, they found that while the central region and the region near the endplates
remained almost circular, the deviation from axisymmetry in the region just above the
60
endplates grew as thefilament was extended. They also found that stress boundary lay-
ers were formed near the free surface for the axisymmetric initial condition. For non-
axisymmetric initial conditions, the stress boundary layer was relieved in regions which
bulged out as compared to the axisymmetric state, while it was more pronounced in the
regions that had retracted. Care must be taken when interpreting their results, however,
since the initial condition was non-axisymmetric, in contrast to the axisymmetric base
state in the experiments. Also, numerical difficulties restricted their analysis to Hencky
strains less than2.5 for Deborah numbers larger than1, so their simulations could not
capture the behavior of thefilament at large elongations and high Deborah numbers seen
in the experiments by Spiegelberg and McKinley (1998).
While the mechanism suggested by McKinley et al. (1996) is plausible, the impor-
tance offluid elasticity, as evidenced by the fact that the instability is not observed at small
Deborah numbers or in Newtonianfluids, indicates that a purely elastic mechanism may
be at work. The formation of stress boundary layers near the free surface as seen in the
simulations of Spiegelberg and McKinley (1998) and Rasmussen and Hassager (1999)
suggests that the physics of the instability may be captured by examining a small region
near the interface. Since theflow in the region near the endplates is radially inward, com-
pressive hoop stresses develop in thefluid (Kumar and Graham, 2000). Further, at high
Deborah numbers, the Oldroyd-Bfluid behaves like an elastic solid. Hence, we expect
the region near the free surface to behave like a membrane subjected to a compressive
hoop stress.
We conduct a preliminary investigation of our proposed mechanism by modeling the
region of the liquid bridge near the endplates as an elastic membrane enclosing a passive
incompressiblefluid. The membrane is then stretched with the radii at the two ends kept
61
fixed. The incompressiblefluid then exerts a force of constraint on the membrane, which
acts to keep the enclosed volumefixed. The constraint forces cause compressive hoop
stresses to develop in the membrane, and thus we expect the initial axisymmetric shape
to bifurcate into non-axisymmetric modes at sufficiently large strains.
In the subsequent section, we demonstrate the build up of stress near the surface
by modeling an ideal planar elongation experiment. This serves to justify the use of
our membrane analogy. We next present the formulation of the membrane model for a
reference configuration obtained by chopping away the region near the apex of a cone.
This shape approximates the geometry near near the endplates. We conclude the chapter
with a discussion of the important results and conclusions of our analysis.
3.2 Planar elongation: a model problem
In this section, we examine a model problem that captures the main features of theflow
field near the central region in afilament stretching experiment. Figure 26 shows a por-
tion of the central region in the planar elongation of a UCMfluid. Theflow field is
Hamiltonian, with the stream functionψ = ε x y. Thus, the velocity components are
given by
vx =∂ψ
∂y= ε x
vy = −∂ψ
∂x= −ε y. (66)
The origin is a stagnation point, which means that polymer molecules which are carried
by theflow near the origin will experience large elongations. Thus, on physical grounds,
we would expect high stresses to develop in the neighborhood of the origin.
62
y=- l
y=0
x=0
y
x
Figure 26: Schematic of a planar elongation setup andflow field.
We can get a quantitative picture of the stress buildup by following the motion of a
fluid element initially located at(x0,−l) (which we assume is a point of zero stress) as it
moves along a streamline. The evolution of the stress tensorτ is described by the UCM
equation as
τ + λ
(D τ
D t− τ · ∇v − τ · ∇vt
)= η0 γ, (67)
wherev is the velocity vector,γ = (∇v + (∇v)t), λ is the relaxation time,η is the
viscosity andD/Dt denotes the material derivative∂/∂t+v·∇. We make these equations
dimensionless by scaling the length withl, velocity with l ε , time with ε, and stress with
the shear modulusη/λ. With these scalings, equation 67 is written as
1
Weτ +
(D τ
D t− τ · ∇ v − τ · ∇ vt
)= ˆγ, (68)
with the Weissenberg numberWe is given byλε. Here, the hats above the variables are
used to denote the scaled versions of the respective quantities, and will be dropped in the
rest of this section for convenience.
63
For theflow described above, the UCM equation gives
dτxxdt−
(2−
1
We
)− 2 = 0, (69)
for the evolution ofτxx on a particular streamline. Thefluid element is chosen so that it
is initially located at(x0,−1), wherex0 ∈ (0, 1), with the stressτxx(t = 0) = 0. The
fluid element is carried along with theflow according to
dx
dt= x, (70)
dy
dt= −y. (71)
Equations (69) to (71) may be readily solved to yield
x(t) = x0 exp(t), (72)
y(t) = − exp(−t), (73)
τxx(t) =2We
2We− 1
(−1 + exp
(2−
1
We
)t
). (74)
Equation 73 may be used to eliminatet from equation 74 to yield
τxx(y) =2We
2We− 1
((−1
y
)2−1/We
− 1
). (75)
Equations 74 and 75 yield the well known result that the axial stress in planar elonga-
tion of a UCMfluid blows up atWe = 0.5. It is also evident from equation 75 thatτxx is
independent ofx for the initial condition of zero stress aty = −1. In the limitWe→∞,
equation 75 givesτxx = 1/y2 − 1, which indicates that at high Weissenberg numbers, a
stress boundary layer exists near the free surfacey = 0. This serves as justification for
using the membrane approximation for largeWe, where we assume that the physics of
the instability can be captured by modeling a small region near the interface.
64
At first sight, it seems somewhat surprising that the limiting value ofτxx, or equiv-
alently, the thickness of the boundary layer, is independent ofWe. This is resolved by
observing that, in general, the Weissenberg number governs the lengthl, at which the
normal stresses are close to zero. This in turn means thatWe would affect the length
scale fory, and hence the limiting value ofτxx.
3.3 Elongation of a truncated cone
In this section, we apply the theory of membrane elasticity to determine the evolution of
the shape of a truncated cone under axial elongation. This geometry is chosen because
it approximates the“reservoir” region near the endplates in afilament stretching exper-
iment. The cone encloses an incompressiblefluid, so that its volume remains constant
under elongation. Thus, the liquid exerts a force on the membrane to keep the volume
fixed, but otherwise plays no role in the formulation. This force exerted by the liquid is a
force of constraint which does not contribute to the total energy of the system (Goldstein,
1980). Our goal is to demonstrate that compressive hoop stresses develop in this system,
resulting in a symmetry breaking bifurcation to a non-axisymmetric shape.
3.3.1 Problem Formulation
The mathematical formulation uses aspects of tensor analysis, differential geometry, and
finite elasticity. An overview offinite elasticity, to the extent needed to formulate this
problem, is presented in Appendix A. In keeping with the Einstein notation, we will use
65
superscripts to denote the components of vectors or tensors that transform contravari-
antly, and subscripts for those that transform covariantly. Scalars and constants are in-
variant under transformation, and subscripts used in conjunction with them merely serve
to distinguish between different quantities.
Consider the portion of the cone shown infigure 27. In its undeformed (rest) config-
uration, that the radius varies linearly fromr1 to r2 over a lengthL. Thus, at any axial
locationx3 (with x3 < L), the radius of the cone is given by
r(x3) = r1 − s x3, (76)
wheres = (r1 − r2)/L. In the limit of s = 0, this reduces to a cylindrical membrane.
The volume enclosed by the membrane is given byV0 = π(r21 L− r1 sL2 + s2
3L3)
.
The membrane is subjected to a deformation at constant enclosed volume. This can be
achieved by extending either end or both ends, but it is convenient from our point of view
to think of the left end (the end near the origin) as being heldfixed, while the right end is
extended by an amountl.
We begin by defining the Gaussian surface coordinates for the reference configura-
tion. Letv1 be the distance along a circular cross section of the cone measured from the
intersection of the cone with thex2 = 0 plane. Thus, at any axial location, the azimuthal
angle (measured anti-clockwise) is given byθ = v1/r(x3). The second surface coordi-
nate,v2 is defined to be the distance of a point from the intersection of the cone with
thex3 = 0 plane, measured along its intersection with thex1 x3 plane (seefigure 27).
Thus, thex3 coordinate of a point whose second surface coordinate isv2 is given by
x3 = v2 cos(φ), wheretan(φ) = (r1 − r2)/L. Hence we have0 ≤ v1 < 2 π r(x3),
and0 ≤ v2 ≤ L sec(φ). The position vector in Cartesian coordinates of a point whose
66
x
r1
r2
L
v2
v1
1
x 3
x 2
Figure 27: Coordinate system for the truncated cone.
surface coordinates are(v1, v2) is given by
a = r(v2) cos(θ) e1 + r(v2) sin(θ) e2 + v2 cos(φ) e3. (77)
We scale all lengths byL and write
a∗ = r∗(v2∗) cos(θ) e1 + r(v2∗) sin(θ) e2 + v2∗ cos(φ) e3. (78)
In future, we will drop the asterisks for convenience. We now consider an axisymmetric
deformation such that the position vector of a point whose surface coordinates in the
undeformed configuration are(v1, v2) is given by
A = λ1(v2) r(v2) cos(θ) e1 + λ1(v
2) r(v2) sin(θ) e2 + λ2(v2) e3. (79)
67
The boundary conditions on the stretches are:
λ1(0) = 1,
λ1(sec(φ)) = 1,
λ2(0) = 0,
λ2(sec(φ)) = 1 + l. (80)
Given the position vectorsa andA, we can compute the covariant and contravariant
components of the surface metric tensors in the reference and deformed states, as de-
scribed in section A.1.1. Let us denote byaαβ the covariant components, and byaαβ the
contravariant components of the surface metric tensor in the reference state. The corre-
sponding components in the deformed state are given byAαβ andAαβ respectively. The
strain invariantI1 is then given by
I1 = aαβAαβ + a/A. (81)
Thus, the dimensionless strain energy stored in a neo-Hookean form is given by the equa-
tion
E =
∫ 10
∫ 2π (r1−s z)0
(I1 − 3) dv1 dz, (82)
wherez = v2 cos(φ), and the neo-Hookean coefficientC1, and a factor ofcosφ have been
absorbed intoE to make it dimensionless. Note that the integral is over theundeformed
state of the membrane.
The next step in the formulation is to compute the volume enclosed by the membrane.
One way to do this is to assume that the cone is a solid body, so that the Cartesian
coordinates(x1, x2, x3) of a point enclosed by the membrane are given in terms ofv1 and
68
v2 by
x1 = λ1(v2) r cos(v1/r(v2)),
x2 = λ1(v2) r sin(v1/r(v2)),
x3 = λ2(v2), (83)
where0 < r < r1 − s v2 cos(φ). The Cartesian volume elementdV = dx1dx2dx3 is
related to the differential surface coordinates by
dx1dx2dx3 =| J | dr dv1 dv2, (84)
whereJ is the determinant of the Jacobian matrix of the transformation. Since| J | is
not a function ofr, we can perform the integration overr. The volume enclosed by the
deformed membrane is given by
Vl =
∫ 10
∫ 2π(r1−s z)0
1
2(r1 − s z)λ21 λ2,z dv
1 dz. (85)
Hence, the dimensionless strain energy function is given by
F = E + Λ (Vl − V0), (86)
whereE is given by equation 82 andV by equation 85, andΛ is a Lagrange multiplier.
For a given value ofl, λ1 andλ2 take values so that thefirst variation ofF is zero, i.e.,
λ1 andλ2 are solutions to the Euler-Lagrange equations corresponding to the variational
principleδ F = 0 (Greenberg, 1978).
3.3.2 Method of solution and stability analysis
The standard method of solving variational problems is by means of the Ritz method.
Here, we expandλ1 andλ2 as polynomial series which satisfy the boundary conditions
69
and solve for the coefficients such thatF is extremized. We choose the following func-
tional form:
λ1 = 1 + (z2 − 1)N∑i=0
ai Ti(η),
λ2 = (1 + l) z +N∑i=0
bi Ti(η), (87)
whereη = 2 z − 1 andTi(η) is the Chebyshev polynomial ofith order with argument
η. Substitution of these expressions into equation 86 givesF in terms ofai, bi andΛ.
The variational statementδ F = 0 then corresponds to the requirement that the partial
derivatives ofF with respect to each of theai, bi andΛ vanish. This gives us a set of
2N + 3 nonlinear equations forai, bi andΛ which we can solve by Newton iteration.
Having obtained the axisymmetric solution, we now check for bifurcations to non-
axisymmetric states. To do this, we assume that the position vector in the deformed state
can be written as
A = (λ1(v2) + λ1(v
1, v2)) r(v2) cos(θ + λ3(v1, v2)/r(v2)) e1 +
(λ1(v2) + λ1(v
1, v2)) r(v2) sin(θ + λ3(v1, v2)/r(v2)) e2 +
(λ2(v2) + λ2(v
1, v2)) e3. (88)
The perturbationsλ1, λ2 and λ3 are set to be0 at the two ends of the membrane. In
addition, we impose a periodicity condition in the azimuthal direction.
We can now perform an analysis similar to the one described in the previous section
to determine the constrained strain energy function for non-axisymmetric deformations
Fa = Ea + Λa (Vl,a − V0), (89)
on the same lines as equation 86. Note thatFa reduces toF if we takeλ1, λ2 andλ3 to be
zero (the axisymmetric shape). In the linear limit, the azimuthal modes decouple, so we
70
can examine each mode separately. We therefore choose the following functional form
for the perturbations:
λ1 = (z2 − 1)N∑i=0
ai Ti(η) cos(n θ),
λ2 = (z2 − 1)
(N∑i=0
bi Ti(η) cos(n θ) +N∑i=0
di Ti(η) sin(n θ)
),
λ3 = (z2 − 1)
(N∑i=0
ei Ti(η) cos(n θ) +N∑i=0
fi Ti(η) sin(n θ)
), (90)
whereθ = v2/(r1 − s z), andn is the azimuthal mode number. The form ofλ1 is chosen
so as tofix the phase of the perturbation.
Substituting equations 87 and 90 into equation 89 gives usFa in terms ofai, bi, ai, bi,
di, ei, fi, andΛ. The requirement that the partial derivatives ofFa with respect to these
variables be zero gives us a set of7N + 8 equations in these variables. One solution to
this set of equations corresponds to the axisymmetric solution calculated above, i.e.,ai,
bi andΛ corresponding to the axisymmetric solution, andai, bi, di, ei, fi all equal to zero.
A bifurcation to a different solution occurs when one or more eigenvalues of the Hessian
matrix of Fa, evaluated at the axisymmetric solution, crosses the origin with non-zero
slope.
3.4 Results and Discussion
The axisymmetric shape is obtained using a Newton iteration. In general, 9 Chebyshev
polynomials provide enough accuracy to capture the shape. Since the equations are highly
nonlinear, we need a good initial guess, so we start atl = 0 where the exact solution is
71
given byai = 0, bi = 0, andΛ = 0, and then do a continuation inl, with the initial guess
at each point computed from the tangent at the previous point. We use a Broyden update
method (Broyden, 1965) to avoid having to recompute the Hessian between successive
iterations of the Newton method at a given value ofl. We use two different methods to
check our results. First, fors = 0, the truncated cone is a cylinder, and the evolution
of the shape in the absence of the volume constraint was computed by Yang and Feng
(1970). Our results for the corresponding case are in agreement with theirs. We also
check that equation 179, which is in the nature of a momentum conservation equation
for the membrane, is satisfied both in the base state, and also in linearized form by the
bifurcating solution.
We first consider the specific case ofr1 = 0.5 ands = 0.45, and examine its stability
with respect to a mode3 disturbance. Figures 28(a)–(c) show the development of the
three dimensional structure of the cone as we increasel from 0 to the bifurcation point,
l = 0.5. It is clear from thesefigures that much of the deformation is concentrated in the
central portion of thefilament, and the right end of thefilament stretches into a cylindrical
portion of nearly constant radius. This is what we desire if we wish to match this region
in some way to the constant thickness central region in afilament stretching experiment.
Denoting the elongationl at which the bifurcation occurs bylc, wefind that the cone
undergoes a bifurcation to then = 3 mode atlc = 0.50. The structure of the bifurcat-
ing solution is shown infigure 28(d). This is obtained by adding a small multiple of the
eigenvector corresponding to the bifurcating eigenvalue and adding it to the base eigen-
vector. As mentioned in the introduction, the instability results because portions of the
membrane are in compression. To verify this, we plot the amplitude of the perturbation
radial stretchλ1 and the hoop stressτθθ (calculated from equation 163) infigure 29. We
72
(c)
(d)
(a)
(b)
Figure 28: Evolution of the shape of a truncated cone under elongation: (a) undeformedconfiguration (b) axisymmetric configuration atl = 0.2 (c) axisymmetric configurationat l = lc = 0.5 (d) post bifurcation non-axisymmetric shape atl = lc = 0.5. Thefigureson the right track the change in a cross section originally at a distance of 0.75 units fromthe left edge. In (d), the perturbation has been exaggerated for clarity.
73
0.0 0.5 1.0 1.5z
−0.2
0.0
0.2
0.4
0.6
0.8
τθθ ||
^λ1||
Figure 29: Spatial profile of the hoop stress,τθθ, and the amplitude of the bifurcatingsolution‖ λ1 ‖.
observe that the perturbation has its maximum amplitude close to the point where the
hoop stress has its largest negative value, i.e., near the point of largest compressive stress.
Table 2 shows the values oflc at which then = 3 mode bifurcates for various values of
r1 ands. While the dependence ons is non-monotonic, we see that for afixed value of
s, increasingr1 pusheslc to higher values.
r1 s lc0.5 0.45 0.500.5 0.3 0.310.5 0.2 0.360.7 0.2 0.691.0 0.2 1.08
Table 2: Variation of the bifurcation point for then = 3mode of the truncated cone withr1 ands.
So far, we have confined our discussion to the behavior of then = 3 mode. While
computing values oflc for different modes is a computationally intensive task, we can
74
get an idea of the dependence ofn by looking at the eigenvalue spectrum of different
modes for thelc of then = 3 mode. This computation indicates that modes with higher
wavenumbers bifurcate sooner, an observation consistent with the fact that even very
small increases in length result in compressive hoop stresses in parts of the membrane.
Thus, the elongation of this geometric shape atfixed volume is always unstable to non-
axisymmetric perturbations, with a critical wavenumbern→∞. At first sight, this seems
unphysical, but a similar situation exists in the Saffman-Taylor problem (Saffman and
Taylor, 1958) in the absence of surface tension. What is absent in the present formulation
is a mechanism to damp out large wavenumber disturbances. In a viscoelasticfilament,
two such mechanisms would be surface tension and bending moments arising from non-
zero boundary layer thickness. In fact, an approximate analysis of the radial inflow of a
thin layer of an Oldroyd-Bfluid in a washer shaped domain with free surface boundary
conditions at the top and bottom (Kumar and Graham, 2000), which models theflow the
polymer solution just above the end plates, shows that disturbances of large wavelength
are the most unstable, thus providing evidence for the stabilizing role of surface tension.
Another example of this is in the Saffman-Taylor problem, in which when surface tension
is taken into account, the fastest growing disturbance has afinite wavenumber.
3.5 Conclusions
In this chapter, we have proposed a new mechanism to explain the purely elastic insta-
bility seen infilament stretching at large Weissenberg numbers. In thefirst part of this
chapter, we used the simple model of a planar elongation experiment to show how large
stresses build up in free surfaceflows. Based on this, and upon evidence in the literature
75
(Rasmussen and Hassager, 1999; Spiegelberg and McKinley, 1998) on the formation of
stress boundary layers near the free surface at large Weissenberg numbers, we postulated
that the physics of the instability can be captured by modeling the region near the inter-
face as an elastic membrane enclosing an incompressiblefluid. We showed that such a
membrane is unstable in elongation to non-axisymmetric disturbances. The instability
is related to the formation of compressive hoop stresses in the membrane, and these are
largest close to the point where the perturbation has the greatest magnitude. Our calcu-
lations show that disturbances withn → ∞ are most unstable, but we expect that the
inclusion of selection mechanisms like surface tension and a non-zero bending moment
which impose an energy penalty on large wavenumber disturbances will yield afinite crit-
ical wavenumber. Validation of the mechanism proposed here would require numerical
simulations using a viscoelastic constitutive equation at large Deborah number and up to
large Hencky strains. For the mechanism proposed here to be effective, the simulations
would need to show stress boundary layers with regions of compressive hoop stress close
to the portions of thefilament that have buckled.
76
Chapter 4
Stabilization of Dean flow
instability †
In this chapter, we will discuss the stabilization of the purely elastic instability in the
Deanflow geometry. Instabilities in Deanflow, and the closely related circular Couette
flow geometry are interesting because they serve as models for more complexflows in
polymer coating operations. Therefore, we will motivate the discussion by providing an
overview of instabilities in coatingflows. We will then describe the main features of the
elastic instability in Deanflow and present a means of suppressing this instability.
4.1 Instabilities in coating flows
Coating may be defined as the process of replacing air with a new material on the sub-
strate (Cohen and Gutoff, 1992). Coatings play an important part in everyday life, being
† Most of the material in this chapter has been published in Ramanan et al. (1999)
77
used in the production of common products such as photographicfilm, paper, and mag-
netic media used for data storage. The most visible part of the industry is the manufac-
ture of paints and protectivefilms for automobiles, houses, and other structures. In the
United States, this segment of the industry alone has an annual revenue in of about $17
billion (The Freedonia Group, 2000). When the less visible part of the industry, which
includes such areas as print and publishing and printed circuit boards for electronics for
is taken into account, the value exceeds $300 billion.
There are several coating methods currently in use. The most common single layer
coating methods are rod coating, forward and reverse roll coating, blade coating, air knife
coating, gravure coating, slot coating, and extrusion coating. Multiple layer coatings pro-
cesses include slide and curtain coating. Schematics of some of these processes are shown
in figure 30. The choice of coating method to use for a given application is determined by
several factors: these include the type of substrate (porous or non-porous), coating speed,
the thickness and accuracy desired, the viscosity and viscoelasticity of the coatingfluid,
and the number of layers to be deposited. Booth (Booth, 1958a,b) was thefirst to publish
a guide to choose the right coating method based on the criteria mentioned above. More
recently, a simpler guide has been proposed by Cohen (1992).
Throughput in coating operations is generally limited by interfacial instabilities which
give rise to surface distortions on the coating (Strenger et al., 1997). Perhaps the most
well known of these interfacial instabilities is the so called“ribbing” instability in forward
and reverse roll coating (Saffman and Taylor, 1958; Pearson, 1960; Pitts and Greiller,
1961). A photograph of this instability in forward roll coating is shown infigure 31.
The photograph clearly shows the spatially periodic patterns on the surface of the rollers,
78
(a)
(b)
(e)
(f)
(c)
(d)
Figure 30: Some commonly used coating industrial coating processes: (a) Dip coatingand rod coating. (b) Blade coating and air knife coating. (c) Gravure coating. (d) Reverseroll coating. (e) Extrusion coating. (f) Slide coating and curtain coating (Cohen, 1992)
79
which translate to surface distortions on the coating. Since theflow in the gap in be-
tween the rollers is complicated, most of the earlier analyses of the ribbing instability
used simplified models of theflow (Pitts and Greiller, 1961; Savage, 1977a,b; Gokhale,
1981, 1983b,a; Savage, 1984; Benkreira et al., 1982). With the advent of more pow-
erful computers, it became possible to perform proper asympototic and fully numerical
calculations usingfinite element methods, and to analyze its stability for Newtonian liq-
uids (Ruschak, 1985; Coyle et al., 1986, 1990b). As a result, the critical conditions for
the onset of ribbing for Newtonian liquids are now very well predicted by theory (Coyle,
1992). These results are typically reported in terms of the capillary number, which is a
measure of the ratio of viscous to surface tension effects.
Figure 31: Photograph of ribbing instability in forward roll coating (Coyle et al., 1990b).
Many industrial coating processes use polymericfluids, and this can have a significant
effect on the instability. Bauman et al. (1982) showed that adding as small an amount as
10 ppm of high molecular weight polyacrylamide can reduce the critical capillary number
for instability by a factor of 2 to 5 in forward roll coating. Slot coating (Ning et al., 1996)
and reverse roll coating (Coyle et al., 1990a) are also significantly destabilized when
polymericfluids are used. Strong destabilization has also been observed recently in free
80
surfaceflows in eccentric cylinder geometries (Grillet et al., 1999) with dilute solutions
of polyisobutylene in polybutene/kerosene mixtures. Soules et al. (1988) and Fernando
and Glass (1988) performed a series of experiments which demonstrated that the earlier
onset of ribbing correlates well with the extensional viscosity of the polymer.
It is clear from the observations listed above that the addition of polymer has a marked
effect on the stability of coatingflows in industrial use. Therefore, it is important to
understand the mechanisms by which polymers can causeflow instabilities. Rather than
attempt the complex and computationally demanding task of analyzing coatingflows, we
look instead for simpler geometries where some of these mechanisms are manifested.
Since we are interested in instabilities caused by elastic effects alone, we restrict our
attention toflows where inertial effects are negligible. Examples of these include theflow
of polymer solutions in cone-and-plate and plate-and-plate geometries (Larson, 1988;
Byars et al., 1994), the circular Couette geometry (Larson et al., 1990), and the Dean and
Couette-Dean geometries (Joo and Shaqfeh, 1991, 1994). In the latter three geometries,
thefluid flows in the gap between two concentric cylinders. In circular Couetteflow, the
flow is driven by the motion of one of the cylinders, while in Deanflow, a pressure drop
is applied in the azimuthal direction. In Couette-Deanflow, a combination of the two
methods is used to drive theflow. Apart from these, recirculationflows, such asflow
in a lid-driven cavity (McKinley et al., 1996) also exhibit elastic instabilities. In each
of theseflows, polymer molecules are stretched along the curved streamlines in much
the manner as in the rod climbing effect discussed in chapter 1. This extra tension in
the streamlines, or the so called hoop stress, is what causes the instability. It is obvious
from an examination offigure 30 that curved streamlines exist in many coatingflows. In
addition, recentflow visualizations (figure 10) have shown the presence of recirculation
81
regions in blade coating. Thus, it is plausible that the instabilities seen in theseflows are
driven by mechanisms similar to those that drive the instability in simpler geometries,
and that any stabilization mechanism that suppresses such instabilities in the simpler
geometries such as Couette and Deanflows carries over to the more complex coating
flows as well.
4.2 Elastic instability in Dean flow
Before discussing the elastic instability in Deanflow, we first present an overview of
the instability in the related geometry of circular Couetteflow. Larson et al. (1990)
performed thefirst theoretical and experimental analyses of elaticity-driven instabilities
arising in circular Couetteflow. Their theoretical analysis, using an Oldroyd-Bfluid,
showed that this“viscoelastic Taylor-Couette” (VETC) instability occured even at negli-
gible Reynolds number. The criterion for the instability to occur is thatε1/2Weθ = O(1),
whereWeθ is the azimuthal Weissenberg number (ratio between material andflow time
scales) andε is the gap width, non-dimensionalized with respect to the radius of the cylin-
der. The scaling reflects how largeWeθ must be for the azimuthal normal (hoop) stress
to contribute to the leading order radial momentum balance. For comparison, the well
known inertial Taylor-Couette instability occurs whenε1/2Re is O(1), whereRe is the
Reynolds number. The mechanism of destabilization proposed by Larson et al. (1990)
and later refined by Joo and Shaqfeh (1994) to include non-axisymmetric modes, was
based on the coupling of stress perturbation to the base state velocity gradient to pro-
duce an azimuthal normal stress that drives radial and transverse motions leading to the
formation of cells. The equations governing perturbations to the basic Couetteflow are
82
O(2) symmetric, i.e., they are invariant under both reflections and translations. Theory
(Golubitsky et al., 1985) indicates that bifurcations in such systems are either pitchfork
or degenerate Hopf (i.e., the Jacobian matrix has two pairs of complex conjugate eigen-
values crossing the imaginary axis at the onset of the instability). In agreement with this
prediction, Larson et al. (1990) found that the instability took the form of a degenerate
Hopf bifurcation. While Larson et al. (1990) only considered axisymmetric modes, it
was later found (Joo and Shaqfeh, 1994; Sureshkumar et al., 1994) that the most unsta-
ble disturbance is in fact a non-axisymmetric mode, with a wavenumber between 1 and
3. Further, Renardy et al. (1996) conducted a nonlinear analysis to study mode inter-
actions arising from the introduction of inertia– and thus the classical Taylor-Couette
instability – into the system. Computations (Northey et al., 1992; Avgousti and Beris,
1993; Avgousti et al., 1993) and experiments (Shaqfeh et al., 1992) have shown thatfi-
nite gap effects tend to stabilize theflow. Recently, Al-Mubaiyedh et al. (2000) have
shown that non-isothermal effects can give rise to an entirely new mode of instability
at very low values of the Weissenberg number. In contrast to the isothermal case, this
instability takes the form of a stationary, axisymmetric bifurcation. Al-Mubaiyedh et al.
(2000) also investigated the effect offluid rheology, in particular, accounting for multi-
ple relaxation times, on the viscoelastic circular Couette instability. They found that the
critical Weissenberg number asymptotes to a value that is half that predicted by a single
mode constitutive equation, but that the mechanism of the instability and structure of the
bifurcated modes remain the same. While the analyses discussed above have focused on
the behavior close to the bifurcation point, i.e., the linear and weakly nonlinear regimes,
experiments (Groisman and Steinberg, 1997; Baumert and Muller, 1999) have shown that
a rich variety of dynamical phenomena can occur far away from the bifurcation point. In
83
chapter 5, we describe how a fully nonlinear analysis can capture some aspects of these
observations.
In the context of Deanflow, Joo and Shaqfeh (1991) performed a theoretical investi-
gation for an Oldroyd-Bfluid to show that such an elastic instability did indeed occur. Joo
and Shaqfeh (1992b) generalized this to Couette-Deanflow. In line with the fact that this
instability is purely elastic, they found that the most unstablefluid was one where there
was no contribution to the extra stress tensor from the Newtonian solvent, and that the
onset of the instability was delayed when solvent viscosity was present. Later, Joo and
Shaqfeh (1992a) investigated the effect of inertia on Dean and circular Couetteflows.
They found that Deanflow was destabilized by inertial effects, while circular Couette
flow was stabilized if theflow was driven by the rotation of the outer cylinder, and desta-
bilized if it was driven by the inner cylinder rotation, consistent with the mechanism of
the inertial instability. Joo and Shaqfeh (1994) presented experimental confirmation of
the elastic instability in Deanflow, and also performed an experimental and theoretical in-
vestigation of the effects offinite gap width on the instability. As in circular Couetteflow,
they found both theoretically and experimentally thatfinite gap effects were stabilizing.
The perturbation equations in Deanflow are also O(2) symmetric, and in this case,
Joo and Shaqfeh (1991) found that the axisymmetric mode bifurcates as a pitchfork.
They performed an energy analysis and determined that the mechanism of the instability
involved the coupling of base state hoop stresses with radial velocity perturbations. In
a later publication (Joo and Shaqfeh, 1994), they found experimental confirmation that
the primary bifurcation in Deanflow was a stationary wave, and performed a theoretical
analysis to show that non-axisymmetric modes are always more stable than axisymmetric
ones.
84
The mechanism of the instability may be understood fromfigure 32. Polymer molecules
are modeled as pairs of beads connected by springs. Molecules near the outer cylinder are
more highly stretched than those closer to the center. An inwardly directed radial velocity
perturbation convects highly stretched molecules towards the center, and thus increases
the local hoop stress. This has the effect of reinforcing the radial velocity perturbation,
hence driving the secondaryflow. A simpler way of understanding this mechanism is
to imagine that the azimuthalflow stretches polymer molecules so that they act like a
stretched cylindrical membrane. This“membrane” in turn exerts an inward compressive
force on thefluid, which causesfluid columns to buckle in the axial direction, causing
the instability.
v
ab
dc
Figure 32: Mechanism of the elastic instability in Deanflow.
Knowledge of the mechanism of theflow can be used to develop methods to stabilize
it. A recent study by Graham (1998) has shown that the addition of a relatively weak ax-
ial flow results in significant stabilization of the isothermal VETC instability. The axial
flow increases the critical Weissenberg number for onset of the instability and moves the
most unstable mode to longer wavelengths. The stabilization is due to the additional axial
normal stress generated by the axialflow, which suppresses radial displacements. Scal-
ing analyses and numerical simulation showed that non-axisymmetric disturbances are
85
strongly suppressed. A weakly nonlinear analysis was performed to determine whether
the bifurcation was subcritical or supercritical. In the narrow gap limit, it was found that
the primary bifurcation in circular Couetteflow is subcritical. The bifurcation remains
subcritical if an axial Couetteflow is added, while an axial Poiseuilleflow changes the
bifurcation to a supercritical one.
Given that the effect of an axialflow is to suppress radial velocity perturbations, it
seems reasonable to suppose that the same mechanism would stabilize the Deanflow
instability as well. In the context of the simple instability mechanism stated above, axial
flow has the effect of stretching the polymer“membrane” in the axial direction. This
creates a restoring force that acts against radial buckling, as shown infigure 33. . In
thefirst part of this work we extend the analysis of Graham (1998) to demonstrate that,
as expected, the addition of axialflow stabilizes the viscoelastic Deanflow instability.
In particular, we show that at large axial shear rates, the critical value ofε1/2Weθ scales
linearly with the axial shear rate, thus confirming that a relatively weak axialflow results
in dramatic stabilization. This is similar to the behavior in circular Couetteflow. We also
perform a weakly nonlinear analysis for Deanflow to examine the effect of adding axial
flow on the criticality of the bifurcation.
Having established that the addition of a steady axialflow does stabilize viscoelastic
Deanflow, we next examine the effect of adding a time-periodic axialflow. There are
two reasons to study this. Firstly, it is more practical to superimpose an oscillatory axial
motion in an industrial polymer processing operation than it is to superimpose steady
axial motion. Secondly, viscoelasticfluids have characteristic relaxation time scales,
and these might be expected to interact with the time scale of the parametric forcing to
produce interesting dynamical phenomena. Note that the axial normal stress depends
86
τrz , τzzτrz , τzz
τθθ/r
Figure 33: Illustration of the mechanism by which additional axial stresses generated bya superimposed axialflow stabilize the viscoelastic Deanflow instability
on the square of the axial shear rate in the baseflow. Thus, the average of this stress
over one period of the oscillation is non-zero. Superimposing an oscillatoryflow on the
circular Couetteflow has been shown to stabilize the inertial Taylor-Couette instability
both experimentally (Weisberg et al., 1997) as well as theoretically (Hu and Kelly, 1995;
Marques and Lopez, 1997). No corresponding analysis is available for Newtonian Dean
flow.
The organization of this chapter is as follows. Wefirst begin by formulating the prob-
lem for an Oldroyd-Bfluid in the narrow gap approximation in section 4.2. Section 4.3
describes the details of the linear stability analysis used in the study. To examine the
effect of adding a steady axialflow, we use a standard linear stability analysis. How-
ever, adding an oscillatory axialflow results in time-periodic coefficients in the stability
equations. In this case, the stability analysis is performed using Floquet theory. Rather
than construct the entire monodromy matrix, we use the Arnoldi technique to determine
its dominant eigenvalues. Finally, in section 4.4, we present results from our numerical
analysis and compare them to predictions from the asymptotics.
87
4.3 Formulation
We consider the isothermalflow of an Oldroyd-Bfluid in the annular region between two
infinitely long circular cylinders. Consideration of isothermalflow ensures that only the
purely elastic mechanism of destabilization operates. The annulus is assumed to be open,
i.e., no constraint is imposed on the net axialflow rate. Thefluid has a relaxation time
λ; the polymer and solvent contributions to the viscosity areηp andηs respectively, and
the ratioηs/ηp is denoted byS. The primaryflow is produced by imposing a constant
pressure drop in the azimuthal direction,Kθ = ∂P/∂θ. A schematic of this geometry is
shown infigure 34.
Κθ R 1
R 2
Figure 34: Deanflow geometry, shown with superimposed Poiseuilleflow.
The equations governing theflow are the momentum, constitutive and continuity
equations. In dimensionless form, they are given by
∇ · τ −∇p+WeθS∇2v = 0, (91)
88
τ +Weθ
(∂τ
∂t+ v · ∇τ −
(τ · ∇v + (τ · ∇v)t
))= Weθ
(∇v +∇vt
), (92)
∇ · v = 0, (93)
wherev is the velocity,p is the pressure andτ is the polymer stress tensor. No-slip
boundary conditions are imposed at the two cylinders. The Weissenberg number is de-
fined in terms of the pressure drop as
Weθ = −Kθλε
2 ηp, (94)
whereε is the dimensionless gap width,
ε =R2 −R1
R2. (95)
The Newtonianfluid is recovered in the limitS → ∞, with an appropriate rescaling of
pressure. We measure length in terms of distance from the inner cylinder, scaled by the
gap width, i.e.,
r =r∗ −R1R2 −R1
, (96)
wherer∗ is the dimensional radial position, andr is the dimensionless radial coordinate,
scaled and shifted so thatr = 0 is the inner cylinder andr = 1 is the outer cylinder. Stress
is scaled byηp/λ, and velocity is scaled by−Kθε2R2/2 ηp. Time is made dimensionless
by the ratio of the gap width,εR2, to the velocity scale. In terms of the dimensionless
variables, theflow domain is given by(r, θ, z) : 0 ≤ r ≤ 1, 0 ≤ θ < 2π, −∞ < z <
∞.
Axial Couetteflow is imposed by moving the inner cylinder with the velocityV cos(ωt),
whereV is the amplitude of the modulation andω is the frequency. The scaling used for
89
ω is the inverse of that used for time, so thatωt is O(1). Settingω = 0 is tantamount to
imposing steady axial Couetteflow. Later, we will see thatω → 0 is a singular limit. The
imposition of axialflow introduces an axial Weissenberg number,Wez, defined as
Wez =λV (1− ε)
R2ε. (97)
When the axialflow is imposed by means of a constant axial pressure drop,Pz, we can
similarly define another Weissenberg number
Wez = −PzλεR22 ηp
. (98)
For convenience, we will use the following acronyms:
• DAC: Deanflow with steady Axial Couetteflow superimposed (ω = 0 ).
• DMAC: Deanflow with Modulated Axial Couetteflow (ω = 0).
• DAP: Deanflow with steady Axial Poiseuilleflow.
We only consider the case where the axial pressure drop is steady.
For an Oldroyd-Bfluid, it is possible to obtain exact analytical expressions for the
base state velocitiesv and polymer stressesτ in Deanflow subjected to superimposed
time-periodic or steady axialflow. The narrow gap limit of these expressions (i.e. the
leading term in a Taylor series aboutε = 0) are shown in Appendix B. We are inter-
ested in the stability of these steady and periodic solutions to infinitesimal perturbations.
We define a vector of perturbationsu = (τrr, τrθ, τrz, τθθ, τθz, τzz, vr, vθ, vz, p), where,
for example,τθθ = τθθ − τθθ. It was shown by Joo and Shaqfeh (1991) that the elastic
instability occurs whenWeθ is O(ε−1/2). We enforce this by defining a scaled Weis-
senberg numberWp = ε−1/2Weθ, which isO(1) in the limit ε → 0. We now scale the
90
perturbation stresses and velocities by requiring that all terms in the Oldroyd-B equa-
tions appear in the perturbation equations for steady axialflow. This may be obtained
by imposing the following scalings on the perturbation velocities and stresses:τrr =
O(1), τrθ = O(ε−1/2), τrz = O(1), τθθ = O(ε−1), τθz = O(ε−1/2), τzz = O(1), vr =
O(ε1/2), vθ = O(1), vz = O(ε1/2), p = O(1) andWez = O(1). Terms involving the
azimuthal wavenumbern do not appear in the governing equations at leading order. To
get around this, we must either consider terms ofO(ε1/2) relative to the leading order
terms in the equations (as done by Joo and Shaqfeh (1994)) or equivalently, define a new
azimuthal wavenumbern by n = nε−1/2, wheren is O(1) (as done by Graham (1998)).
We choose the latter approach. Thus, in our formulation,n is not necessarily an integer,
althoughn must be. The scaling on the stresses and velocities is identical to that on the
base state stresses and velocities, and also to the scaling obtained by Graham (1998) for
circular Couetteflow with imposed axialflow. For time-periodic axialflow, a balance of
terms is obtained by rescaling time and frequency so thatω = ω1ε1/2 andt = t1ε
−1/2
whereω1 and t1 areO(1). With this scaling, the dimensionless relaxation time of the
polymer isWp. This is the“low frequency” regime that includes all possible terms in
the narrow gap approximation. Settingω1 = 0 in this regime, we recover the equations
of DAC flow. We also investigate the effect of imposing a“high frequency” oscillation
by scalingω to be anO(1) quantity i.e. we specifyω1 = O(ε−1/2). Dimensionally, this
corresponds to a forcing frequency on the order of the azimuthal shear rate. In this regime
the stability characteristics depend on the order of magnitude of the axial shear rate. In
section 4.5 we show that the relative magnitude ofWez with respect toω1 determines
the appropriate balance of terms that affects the stability of the system. Asymptotic and
numerical results are presented for both regimes.
91
4.4 Stability and Numerical Analysis
4.4.1 Linear Analysis and Floquet Theory
To analyze the stability of the baseflow, we expand the perturbation vector as
u = δφ(r, t)eiαz +O(δ2) + c.c., (99)
wherec.c. stands for“complex conjugate” andα is the axial wavenumber which is a
parameter in the problem. Wefirst consider the imposition of steady axialflow. In this
case, the vectorφ(r, t) is expanded out as
φ = ξ(r)ei(nθ−αct), (100)
whereξ(r) = (τrr(r), τrθ(r), τrz(r), τθθ(r), τθz(r), τzz(r), vr(r), vθ(r), vz(r)), andc =
cr + i ci is the growth rate. This expression is added to the base state stresses and ve-
locities, and the result substituted into equations 91 to 93. The equations may then be
expanded as a series inδ, and atO(δ0) we recover the equations governing the base
state stresses and velocities. AtO(δ), we get the equations that govern the evolution of
perturbations to the base state. We may write these succinctly as
i α cEu = L(α,Weθ,Wez, n)u, (101)
whereL is the linearization of the governing equations about the base state, andE is a
diagonal matrix with1 on the diagonal entries corresponding to the Oldroyd-B equations
and zeros elsewhere. The nonzero components ofE are shown in Appendix B forn = 0.
Equation 101 is in the form of a generalized eigenvalue problem for the growth ratec.
If the other parameters are given, we can numerically solve for the eigenvaluec. For
stability, ci must be less than zero.
92
The numerical technique we use is a Chebyshev collocation method (Canuto et al.,
1988). We use a primitive variable formulation of the conservation, constitutive and con-
tinuity equations, i.e., our unknowns are the six stress components, the three components
of the velocity, and the pressure. Since the conservation equations only contain deriva-
tives of the pressure, it is necessary to define pressure on a staggered grid (Canuto et al.,
1988) to avoid spurious modes. In the formulation we use, the velocity components and
stresses are defined at theN + 1 Gauss-Lobatto points (for anNth order collocation),
which are
rk = cos
(πk
N
), k = 0, . . . , N, (102)
while the pressure is defined at theN Chebyshev-Gauss points, which are given by
r′
k = cos
(π(k + 1/2)
N
), k = 0, . . . , N − 1, (103)
so that pressure is defined at one less point than the other variables. In evaluating the
conservation equations, pressure is interpolated onto the Gauss-Lobatto points from the
staggered grid, while the continuity equation is evaluated on the staggered grid by in-
terpolating the velocity components from the Gauss-Lobatto points. Wefind that it is
sufficient to useN = 32 for accurate calculation ofc and spatial resolution of the eigen-
vectors.
We now move on to the stability analysis for the time-periodic case. In this case, the
vectorφ(r, t) is expanded as
φ(r, t) = (τrr(r, t), τrθ(r, t), τrz(r, t), τθθ(r, t), τθz(r, t), τzz(r, t),
vr(r, t), vθ(r, t), vz(r, t), p(r, t)). (104)
93
We will only consider axisymmetric perturbations because, as will be shown in sec-
tion 4.5, it is the axisymmetric mode which is most unstable whether or not axialflow is
present. The equations governing evolution of the perturbations are given by
−Eu = A(t)u. (105)
Here,A(t) is a matrix with time-periodic coefficients such thatA(t) = A(t+T ), where
T = 2π/ω is the period of the forcing function. The nonzero components ofA are shown
in Appendix B. The problem lends itself to analysis by Floquet theory (Iooss and Joseph,
1989). The solutionu(T ) at t = T , given the initial vectoru(0), is
u(T ) = Φ(T )u(0), (106)
whereΦ(T ) is the monodromy matrix, whose eigenvaluesβ, known as Floquet multipli-
ers, determine the stability of the system. Stable and unstable behavior is indicated by
|β| < 1 and|β| > 1 respectively. The Floquet exponentσ is defined by the relation
β = exp(σ T ). (107)
SupposeΨ is an eigenvector ofΦ(T ) corresponding to the Floquet exponentσ. Then,
it can be shown (Iooss and Joseph, 1989) that the solutionw(t) to E = A(t)w with
w(0) = Ψ has aT periodic componentζ(t) given by
ζ(t) = e−σtw(t), (108)
such that
ζ(t) = ζ(t+ T ). (109)
Thus, theζ(t) corresponding to the dominant Floquet multiplier gives us information on
the spatial structure and time evolution of the disturbance.
94
The standard way to calculateΦ(T ) is to do it column by column, integrating equa-
tion 105 withu(0) = ek, whereek is thekth column of the unit matrix of same dimension
asE or A(t). Given the size of our problem, this is a computationally expensive task.
However, we only need the dominant eigenvalues (those with the largest modulus) of
Φ(T ), and we can conveniently obtain these using the Arnoldi method (Arnoldi, 1951).
We use a primitive variable formulation as in the steady case, with pressure defined on a
staggered grid. We use the public domain code ARPACK (Lehoucq et al., 1997) to carry
out the Arnoldi calculations. The advantage of the Arnoldi method lies in the fact that we
do not need to explicitly constructΦ(T ) in its entirety; we need only determine the action
of the matrix on a vector,q ← Φ(T )p. In our problemq is the solution vector obtained
by integrating the time evolution equations for a given initial vectorp. The choice of the
initial vectorp is arbitrary, with the only requirement being that it satisfy the algebraic
components of equation 105. In our case, this is ensured by choosing an arbitrary stress
profile and solving the linear momentum and continuity equations for the correspond-
ing velocities and pressure. To compute the periodic componentζ(t), we calculate the
disturbance vectorw(t) by integrating the time evolution equation 105 with initial con-
ditions given by the eigenvector of the monodromy matrix obtained from ARPACK. This
choice of starting value ensures that the periodicity condition given by equation 109 is
not violated.
The integration of the time-dependent viscoelastic equations is performed using the
EVSS decomposition (Rajagopalan et al., 1990) of the total stress into elastic and viscous
components, coupled with a fully implicitfirst-order time stepping procedure. The time
integration and eigenvalue routines were benchmarked against known eigenvalues for the
95
steady case (ω = 0) from the DAC results presented later in this chapter. Most calcu-
lations are performed with a time step∆t = T/100, while at very low frequencies we
employ increased temporal resolution of∆t = T/1000. This high resolution is necessary
to ensure the accuracy of the eigenvalue computation, which is checked by monitoring
the periodicity ofζ(t).
4.4.2 Weakly nonlinear analysis
Following the methodology of Graham (1998) which in turn follows Iooss and Joseph
(1989), we conduct a weakly nonlinear analysis to determine the criticality of the bifur-
cation in DAC and DAPflows. In the presence of either type of axialflow, the loss of
stability takes the form of a Hopf bifurcation, while in the absence of axialflow the bifur-
cation is a pitchfork. In either case, the vectors take the form of complex conjugates, and
since we are only interested in real valued solutions, the analysis described below works
whether or not an axialflow is present.
The nonlinear solutions are constructed by expanding all variables as a power series
in the amplitudeδ, substituting in the governing equations, and applying the narrow gap
approximation at each order inδ. We letµ = Wp−Wpc, ω0 = −αcr, ands = ωt. Thus,
the solution takes the form
u(z, s, δ)
µ(δ)
ω(δ)− ω0
=
∞∑k=1
δk
k!
uk(z, s)
µk
ωk
, (110)
whereu is taken to be the solution at the Chebyshev points. The linear operatorL can
96
be written as
L(µ) = L0 + δµ1L′+
δ2
2µ2L
′′+ ... (111)
Further, we define the following inner products:
〈a(z, s), b(z, s)〉 ≡α
4π2
∫ 2π/α0
∫ 2π0
a(z, s) b(z, s) ds dz, (112)
and
[a(z, s), b(z, s)] ≡
9(N+1)+N−1∑l=0
〈al(z, s), bl(z, s)〉. (113)
At O(δ), we recover the linear stability problem, which has the real valued solutionu1 =
z + z∗. Note thatu is time-dependent in the absence of axialflow, and a traveling wave
otherwise. The eigenvectors are normalized so that[z, z] = 1, wherez is the solution to
the adjoint problem. A solvability condition atO(δ2) gives
−2 i ω1 [Eu1, z∗] + 2µ1[L
′u1, z
∗] = −2 [N2(u), z∗]. (114)
The structure of the solutions atfirst order implies that the right hand side of this equation
vanishes. Upon separating the real and imaginary parts, the coefficient matrix of the
resulting2×2 real system has, in general, a nonvanishing determinant, and hence admits
only the trivial solution forµ1 andω1.
Substituting these values into theO(δ2) problem, we see that the particular solution
at that order can be written as
u2 = u20 + u22e2i(α z+s) + u22e
−2i(α z+s). (115)
Here,u20 is given by
L0,0u20 = −2〈N2(u),1〉, (116)
97
whereL0,0 is obtained by replacingα by 0 in L0. The vectoru22 is given by
(−2iω0E +L0,2α)u22 = −2〈N2(u),1e2i(αz+s)〉, (117)
where1 is a column vector with 1 in all its entries, andL0,2α is obtained by replacing
α with 2α in L0. Unlike in Graham (1998), the present formulation retains pressure
in the momentum equations. Hence, neither of the above equations is singular and the
solution is obtained by LU decomposition. A solvability condition atO(δ3) gives the
single complex equation (or the2× 2 real system)
−iω2[Eu1, z∗] + µ2[L
′u1, z
∗] = −2[N3(u), z∗], (118)
from whichµ2 andω2 can be obtained.
4.5 Results and Discussion
4.5.1 Scaling Analysis
As discussed in the introduction, the elastic instability is caused by an unstable stratifica-
tion of the hoop stress,τθθ. Therefore, if this term is absent from the momentum balance
in some asymptotic limit, theflow is stable. It is instructive to perform an asymptotic
analysis to determine how largeWp must be for the hoop stress to enter into the mo-
mentum balance at leading order. This can also be used to check the results from the
numerical analysis.
Wefirst consider the imposition of steady axialflow, withWez 1. The scalings for
both DAC and DAPflows are identical, and the effect ofO(1) solvent viscosity does not
affect them. Wefirst consider the results for axisymmetric perturbations. For these, the
98
scaling regimes and asymptotics are identical to the VETCflow scaling regimes discussed
in Graham (1998), so we merely restate them here. The method of determining these
scalings is to assume thatWp scales asWemz , and then use the constitutive equations to
determine the stress scalings. These are then substituted into the momentum balance, and
the requirement thatτθθ appear at leading order in the radial momentum balance is used
to solve form. Depending on the value ofα, three scaling regimes exist for axisymmetric
disturbances:
αWez 1 : Wp = O(α1/2We3/2z ), (119)
Wez 1, αWez = O(1) : Wp = O(Wez), (120)
Wez 1, αWez 1 : Wp = O(1/α). (121)
Clearly, it is the second scaling which has the lowestWp of the three regimes. This
means that forWez 1, the most unstable wavenumber will beO(1/Wez), and the
critical Wp will scale linearly withWez. For non-axisymmetric modes, wefirst consider
n = O(1), andα = O(1). For this case, the dominant balance reveals thatm = 1, or
Wp = O(Wez). This is the only regime forn = O(1), and indicates that for a givenn,
Wp scales linearly withWez, as in the axisymmetric case. Note that the scaling for non-
axisymmetric modes differs from the corresponding scaling in VETCflow determined
by Graham (1998), whereWp must beO(We2z) for non-axisymmetric disturbances to
become unstable. Disturbances withn 1 are very strongly suppressed. For example,
whenn = O(ε−1), Wp has to beO(ε−1) to keepτθθ in the momentum balance.
The linear scaling ofWp with Wez has an important practical consequence. Recalling
thatWeθ = Wp ε−1/2, we see that the ratioWez/Weθ, which is a measure of the mag-
nitude of axialflow relative to the azimuthalflow, isO(ε1/2). Sinceε 1, this means
99
that a small amount of axialflow provides significant stabilization. Although we have as-
sumed thatWez 1 in the asymptotic analysis above, numerical computations indicate
that these results are valid down to aboutWez = 1, yet another instance of the power of
asymptotics in supplying information about qualitative behavior.
We now present scaling results for the imposition of time-periodic axialflow. Our
aim here is to determine how the imposition of a time-periodic axialflow changes the
stability characteristics of the system. We have already observed that the stabilization
is due to the stressτzz induced by the axialflow. From Appendix B, we see that the
expression forτzz has aWe2z term multiplied by a factor that is the sum of a constant
term and a term that depends on the frequency and time. From our analysis above, we
know that the constant term has a stabilizing influence. We seek to determine the effect
of the frequency and time dependent terms on the stability characteristics. Wefind that,
depending on the frequency, this term can either stabilize or destabilize theflow.
We first consider the high frequency regime,ω1 = O(ε−1/2), with time rescaled ap-
propriately to reflect the shorter time scales. Further, we also retain the scalings on the
axial and azimuthal Weissenberg numbersWez = Wp = O(1) and restrictα to beO(1).
In this regime, the axial strain rate is large, but the strain amplitude is small (O(ε1/2)),
and wefind that the leading order evolution equation forτzz does not contain theWe2z
contribution that has been established to be the source of high normal axial stresses that
contribute to the stabilization in theflow. TheWe2z term comes from the time-averaged
non-zero base state axial stress, which tends to zero with increasingω andfixedWez.
Thus, we expect decreased stabilization in this regime. Further, theω1 = O(ε−1/2) regime
also lets us apply the method of averaging, a rigorous asymptotic technique (Sanders and
100
Verhulst, 1985), to the time dependent equations. Briefly, the averaging method trans-
forms the system
x = εf(x, t, ε) with f(x, t, ε) = f(x, t+ T, ε), (122)
to an autonomous system
y = ε1
T
∫ T
0
f(y, t, 0) dt = ε f(y). (123)
In other words, the averaging procedure replaces the time-dependent coefficients with
their averages over the period of oscillation. This procedure is valid for arbitrary ampli-
tude of oscillation as long asε 1. It can be shown that in our problem the equations
governing the stress perturbations in this regime are of the form shown in equation 122
with ε = ε1/2, ε being the dimensionless gap width defined by equation 95 in section 4.3.
Importantly, we note the absence of the contribution due to the zero frequency terms in
τzz in the leading order equations forτrz andτzz. The remaining coefficients ofWez are
periodic with zero mean and drop out in the averaging procedure, leaving a system that
is independent ofWez andω1. Hence in theω1 = O(ε−1/2) andWez = O(1) regime,
the system is reduced to the steady Deanflow limit with no axial motion. Our numerical
computations also indicate that in the limit of large frequencies, we recover the steady
Deanflow results.
In the second high frequency regime of interest we investigate the effect ofO(1) axial
deformation on the dynamics of the system. In this regime we haveω1 = O(ε−1/2) and
Wez = O(ε−1/2), which corresponds to high frequency and large deformation. By con-
sidering time scales of the order of the relaxation time of thefluid, we discover the pres-
ence of stress boundary layers near the cylinders of sizeO(1/Wez) for O(1) wavenum-
bers and of sizeO(1/αWez) for O(ε−1/2) wavenumbers. The existence of boundary
101
layers at high shear rates is consistent with the analysis of Renardy (1997) and Graham
(1998). The asymptotic balances reveal that the destabilizing azimuthal stress,τθθ, drops
out of the radial momentum balance, thus resulting in stabilization. Again, computational
results confirm this prediction.
4.5.2 Numerical Results
We begin by presenting results for the steady case. As mentioned in the introduction,
the basic Deanflow profile is invariant under translations and reflections, i.e., it has O(2)
symmetry. Symmetry breaking bifurcations in such systems take the form of pitchforks
or degenerate Hopf bifurcations (Golubitsky et al., 1985). In agreement with this pre-
diction, Joo and Shaqfeh (1991) found that the primary bifurcation in Deanflow was a
pitchfork. When axialflow is added, the reflection symmetry is lost, and theflow is SO(2)
symmetric, and theory (Iooss and Joseph, 1989) predicts that symmetry breaking takes
the form of Hopf bifurcations.
In line with the predictions above, wefind that the primary bifurcation in pure Dean
flow is a pitchfork, and when axialflow is added, symmetry breaking takes the form of a
Hopf bifurcation. Figure 35 shows the results of a linear stability analysis for DACflow
whenS = 0. The curves are plots of the critical value ofWp (denoted byWpc) versus
the wavenumberα for givenWez. For each curve, there is a global minimum value of
Wpc, which we denote byWpc,min, with the corresponding wavenumber denoted byαmin.
ThusWpc,min denotes the smallest value ofWp at which the baseflow becomes unstable,
and the destabilizing disturbance has a wavenumberαmin. Thefigure clearly shows that
Wpc,min increases withWez. A similar plot is shown for DAPflow in figure 36. Here too,
the effect is increased stability at highWez. Note also that in both DAC and DAPflows,
102
αmin decreases with increasingWez. Figure 37 shows a plot ofWpc,min versusWez for
both DAC and DAPflows. In accordance with the asymptotic analysis,Wpc,min increases
linearly with Wez. Note also that in both theseflows,Wpc,min increases monotonically
asWez is increased, in contrast to the initial destabilization observed in circular Couette
flow (Graham, 1998). Solvent viscosity has a stabilizing effect in both DAC and DAP
flows, as indicated infigure 38 which plotsWpc,min versusWez for S = 0 andS = 10.
2.0 4.0 6.0 8.0 10.0α
4.0
6.0
8.0
10.0
12.0
14.0
16.0
18.0
20.0
Wp
Wez=0Wez=1Wez=2Wez=3
Figure 35: Neutral stability curves for DACflow (S=0). In each case, the position ofWpc,min is denoted by a•.
As predicted by the asymptotic analysis, non-axisymmetric modes are similarly sta-
bilized. Solvent viscosity has a stabilizing effect on non-axisymmetric modes as well,
and we only show the results forS = 0 . Plots ofWpc,min versusn for differentWez are
shown infigure 39 for DACflow andfigure 40 for DAPflow. We see increased stabi-
lization asn increases and also asWez increases, in accord with the asymptotic analysis.
Figure 41 shows a plot ofWpc,min versusWez for n = 1. The linear scaling predicted by
the asymptotics is evident.
103
2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0α
4.0
5.0
6.0
7.0
8.0
9.0
10.0
11.0
12.0
13.0
Wp c
Wez = 0Wez = 1 Wez = 2Wez = 3
Figure 36: Neutral stability curves for DAPflow (S=0). In each case, the position ofWpc,min is denoted by a•.
0.0 1.0 2.0 3.0 4.0 5.0 Wez
0.0
5.0
10.0
15.0
Wp c,
min
DAP flow DAC flow
Figure 37: Plot ofWpc,min vs. Wez for DAC and DAPflows (S=0). Note the linearscaling at highWez.
104
0.0 1.0 2.0 3.0 4.0 5.0Wez
0.0
10.0
20.0
30.0
Wp c,
min
Axial Couette (S=0)Axial Couette (S=10)Axial Poiseuille (S=0)Axial Poiseuille (S=10)
Figure 38: Plots ofWpc,min vs. Wez for S = 0 andS = 10, displaying the stabilizinginfluence of solvent viscosity.
0.0 0.2 0.4 0.6 0.8 1.0n
2.5
7.5
12.5
Wp c,
min
Wez=0Wez=0.5Wez=1.0Wez=1.5
~
Figure 39: Plot ofWpc,min vs. n for different values ofWez for DAC flow (S = 0).
105
0.0 0.1 0.3 0.4 0.6 0.8 0.9n
3.0
5.0
7.0
9.0
11.0
Wp c,
min
Wez=0Wez=0.5Wez=1.0Wez=2.0
~
Figure 40: Plot ofWpc,min vs. n for different values ofWez for DAP flow (S = 0).
0.0 0.5 1.0 1.5 2.0Wez
0.0
5.0
10.0
15.0
20.0
25.0
30.0
Wp c,
min
Figure 41: Plot ofWpc,min vs. Wez for DAC flow with n = 1.0 (S = 0). Note the linearscaling forWez > 1.
106
Figure 42 shows a portion of the neutral stability curve for pure Deanflow. We point
out that there is a change in the slope of the curve atα = 21.9. At this point, four eigen-
values with zero real parts are simultaneously neutrally stable. This is a codimension-2
Takens-Bogdanov bifurcation point (Takens, 1974; Bogdanov, 1975; Knobloch and Proc-
tor, 1981; Guckenheimer and Knobloch, 1983). Hereafter, we refer to this value ofα as
αtb. For wavenumbers greater thanαtb, the bifurcation is a Hopf, with a four dimensional
center manifold, similar to the one in circular Couetteflow. As we decreaseα towardsαtb,
the period of the Hopf bifurcation increases, reaching infinity atαtb, where the form of the
bifurcation changes to a pitchfork. From a physical point of view, this bifurcation point
is expected, since large wavenumber disturbances are localized in a small region near the
outer cylinder, and hence experience a base velocity profile that is locally linear, similar
to the profile in circular Couetteflow. Thus, we expect that for large wavenumbers, the
destabilization mechanism, and hence the nature of the bifurcation, will be similar to the
one in circular Couetteflow.
Experimental results by Genieser (1997) on theflow of viscoelasticfluids through a
planar contraction display instabilities that occur as transitions from steadyflow to ei-
ther steady or oscillatoryflow, depending on contraction ratio. The oscillatoryflows
have very low frequency, so it may be that these observations are another manifestation
of Takens-Bogdanov bifurcation and concomitant change in destabilization mechanism
found here. The unfolding of a Takens-Bogdanov bifurcation in an O(2) symmetric sys-
tem was performed by Dangelmayr and Knobloch (1987). They showed that, depending
on the value of the bifurcation parameters, observable patterns include a nontrivial steady
state, traveling waves, standing waves and modulated waves.
The results of the weakly nonlinear analysis atWpc,min andαmin are summarized in
107
20.0 20.5 21.0 21.5 22.0 22.5 23.0 23.5 24.0α
6.5
7.0
7.5
8.0
Wp c pitchfork
(steady)
Hopf
(oscillatory)
Takens−Bogdanov point
Figure 42: Neutral stability curve of pure Dean flow at highα. The Takens-Bogdanovbifurcation point is indicated by a•.
Tables 3 and 4 for DAC flow and Tables 5 and 6 for DAP flow. Negative values ofµ2
indicate subcritical behavior while positive values imply supercriticality. We note that
in the absence of solvent viscosity (S = 0), the bifurcation changes from subcritical
to supercritical atWez ≈ 0.05 while at finite solvent viscosity the change occurs at
Wez ≈ 1. Results for higherWez show that forS = 0 the bifurcations revert back to
being subcritical. Clearly, by varying the extent of axial motion we may not only change
the position of the bifurcation point but also ensure that we stay in a regime that admits
stable solutions. Similar behavior arises for the DAP flow, where there is a change in the
sign ofµ2 at atWez ≈ 0.25 for S = 0, andWez ≈ 1.0 whenS = 1. Here, however,
nonzeroS does not change the criticality of the dominant unstable mode at higher values
of Wez. We note that a change in the criticality of bifurcation has also been observed in
108
Wez αmin Wpc,min µ2 ω20 6.6 4.06 −11.21 0
0.05 6.6 4.07 −6.91 −0.740.1 6.5 4.07 0.27 −0.581.0 4.4 5.25 7.05 −0.482.0 2.9 7.39 26.92 −0.423.0 2.2 9.73 46.21 0.244.0 1.8 12.18 52.65 0.575.0 1.5 14.68 54.15 0.66
Table 3:µ2 andω2 for DAC flow (S = 0).
Wez αmin Wpc,min µ2 ω20 6.3 5.88 −484.67 01.0 4.8 7.03 −38.17 −2.582.0 3.6 9.36 144.25 0.403.0 3.0 11.81 123.18 4.154.0 2.6 14.28 −145.25 4.775.0 2.3 16.76 −528.80 3.69
Table 4:µ2 andω2 for DAC flow (S = 1).
the weakly nonlinear analysis of cone and plate flow of an Oldroyd-B fluid (Olagunju,
1997).
In the final portion of this section, we present results for the imposition of time-
periodic axial flow. As in the steady case, solvent viscosity has no qualitative effect on
the nature of the results, and we confine our discussion to the fluid with zero solvent
viscosity. Addition of solvent viscosity simply results in increased stabilization relative
to S = 0, but does not otherwise change the nature of the results. Also, our analysis
for the imposition of steady axial flow indicates that non-axisymmetric modes are always
more stable than axisymmetric ones, so we will only consider axisymmetric perturbations
here.
The addition of time-periodic axial flow introduces a new parameterω1, which in-
creases the dimension of the parameter space. Scanning the entire regime ofWez, α and
109
Wez αmin Wpc,min µ2 ω20 6.6 4.06 −11.21 00.2 6.5 4.08 −3.95 −0.420.3 6.5 4.10 2.20 −0.141.0 5.6 4.54 7.73 −0.382.0 4.1 5.91 22.55 −2.053.0 3.2 7.60 40.10 −3.654.0 2.6 9.63 47.64 −2.835.0 2.1 11.70 20.18 0.46
Table 5:µ2 andω2 for DAP flow (S = 0).
Wez αmin Wpc,min µ2 ω20 6.3 5.88 −484.67 01.0 5.7 6.42 −95.63 −1.352.0 4.5 7.92 151.33 −1.203.0 3.7 9.89 459.44 7.994.0 3.2 11.99 717.29 19.515.0 2.8 14.24 813.98 25.62
Table 6:µ2 andω2 for DAP flow (S = 1).
ω1 is a computationally demanding task. Since our primary goal is to determine the ef-
fect of the periodic axial flow on the stability of pure Dean flow, we limit ourselves to a
smaller parameter space. For the most part, we chooseWp = 4.06, which corresponds to
Wpc,min for pure Dean flow.
For all the values ofα, Wez andω1 that we considered, we always found increased
stabilization relative to pure Dean flow. Thus, the effect of time-periodic axial flow is
always stabilizing. This can be seen quite dramatically in figure 43, which shows the
decay of the Euclidean norm of the destabilizing perturbation hoop stress when time-
periodic axial flow is imposed. The forcing chosen wasWez = 1 andω1 = 0.5. The
initial condition was chosen to be the eigenvector for pure Dean flow withWp = 4.06
andα = 6.6, which corresponds to the most unstable disturbance in pure Dean flow. A
time sequence of density plots of the perturbation hoop stress,τθθ, corresponding to these
110
parameters is shows in figure 44. Byt = 12T , the perturbation has essentially decayed
to the point where the base flow is recovered. The complete stabilization found in Dean
flow is in contrast to the results reported by Ramanan et al. (1999) for the imposition
of time-periodic axial flow in the circular Couette case. While they find that axial flow
results in increased stability for most cases, they also find that the system shows instability
(synchronous resonance) for certain values ofω1 andWez.
0.0 2.0 4.0 6.0 8.0 10.0t/T
0.0
2.0
t/T
0.0 2.0 4.0 6.0 8.0 10.0t/T
0.0
2.0
t/T
Figure 43: Decay of the perturbation hoop stress τθθ when axial flow is imposed. Param-eters are: Wez = 1.0, ω1 = 0.5, Wp = 4.06, S = 0.
As mentioned in section 4.4, stability requires that the magnitude of the Floquet mul-
tiplier be less than one. We show a plot of the magnitude of the Floquet multiplier, |β|
versus the frequency ω1 for different wavenumbers in figure 45. The other initial condi-
tions were Wez = 0.5, and Wp = 4.06. In each case, the magnitude of the Floquet mul-
tiplier is less than 1, indicating stabilization relative to pure Dean flow. Figure 46 shows
the effect of increasing Wez. The parameters chosen were Wp = 4.06 and α = 6.6. As
in the steady case, increasing Wez increases stabilization, reflected here in the smaller
111
t=0
t=T/4
t=3T/4
t=T t=3T
t=6T
t=9T
t=12T
Figure 44: Time sequence of density plots of the perturbation hoop stress τθθ when axialflow is imposed. The parameters are identical to figure 43, so that without axial flow, theflow is neutrally stable. Each frame shows a z − r cross-section of the geometry.
112
magnitude of the Floquet multipliers for larger axial forcing. The figure also shows that
for large ω1, the Floquet multiplier tends asymptotically to unity. This is in agreement
with the scaling analysis for high frequency discussed in the preceding section, which
indicates that for ω1 1, the O(We2z) contribution to τzz in the base flow has a zero
average over a single period of the forcing, thus reducing the problem to zero axial forc-
ing. For the values of Wp and α we chose in this simulation, the base flow is marginally
stable, and hence |β| → 1.
0.0 0.5 1.0 1.5ω1
0.0
0.2
0.4
0.6
0.8
1.0
|β|
α=2.0α=4.0α=6.6α=8.0
Figure 45: Plot of the magnitude of the Floquet multiplier |β| vs. ω1 for different valuesof α. (Wez = 0.5, Wp = 4.06, S = 0).
In the limit ω1 → 0, we might expect the results from our simulations with time-
periodic forcing to reduce to those with steady axial forcing. It turns out however, that
this is not the case, i.e., the limit ω1 → 0 is singular. This has also been observed in the
stabilization of Newtonian circular Couette flow (Marques and Lopez, 1997), as well as in
the linear stability analysis of other time-dependent flows (Rosenblat, 1968; von Kerczek
and Davis, 1974). Davis and Rosenblat (1977) analyzed a damped Mathieu-Hill equation
113
0.0 0.5 1.0 1.5 2.0ω1
0.0
0.2
0.4
0.6
0.8
1.0
|β|
Wez=0.5Wez=1.0Wez=2.0
Figure 46: Plot of the magnitude of the Floquet multiplier |β| vs. ω1 for different valuesof Wez. (Wp = 4.06, α = 6.6, S = 0). |β| asymptotes to 1 at large ω1 in agreement withthe asymptotic prediction.
with externally imposed modulation and showed that at low frequencies the eigenfunc-
tions have large temporal increases in magnitude within the period of oscillation; these
become unbounded as ω1 → 0. Thus, while the Floquet analysis may predict overall
stable behavior for a suitable choice of parameters at low frequencies, the long periods
of oscillation allow for large transient increases in the response. Since the perturbations
in the system are now no longer infinitesimally small as is required for linear stability
analysis, one would have to resort to a nonlinear theory to be able to adequately represent
the solution of the system. This type of behavior is seen in DMAC flow at small ω1, and
is shown in figure 47, where the Euclidean norm of the periodic component of the hoop
stress is plotted against time (normalized by the period of the oscillation) for ω1 = 0.01.
The hoop stress is normalized with respect to its minimum value. We see in the figure
that there are periods where the stress increases close to 500 times its minimum value in
114
the cycle. In contrast, a similar plot for ω1 = 1.0 (figure 48) does not show this behavior,
with the stress remaining O(1) relative to its minimum value in the cycle. Similar be-
havior has been observed in the context of periodic axial flow applied to circular Couette
flow (Ramanan et al., 1999).
0.0 0.2 0.4 0.6 0.8 1.0t/T
0.0
100.0
200.0
300.0
400.0
500.0
||~ τ θθ||
Figure 47: Plot of the magnitude of the periodic component of the hoop stress τθθ over acycle of the forcing for ω1 1 (Wez = 1.0, Wp = 4.06, α = 6.6, ω1 = 0.01). Note thelarge increases in magnitude.
The results presented above for periodic axial forcing reflect the stability properties of
the forced system relative to the critical Wp of pure Dean flow. We conclude this section
with a sample calculation showing how the minimum critical value of Wp is shifted by the
presence of axial oscillations. For DMAC flow, simulations for S = 0, ω1 = 0.8,Wez =
2.0 yield Wpc,min = 4.37 and αmin = 6.0. This is greater than Wpc,min for pure Dean
flow (4.06), thus indicating increased stabilization. For comparison, the value of Wpc,min
for Wez = 2 in steady axial flow would be 7.39 for DAC flow.
115
0.0 0.2 0.4 0.6 0.8 1.0ω1
1.0
1.0
1.0
1.1
||~ τ θθ||
Figure 48: Plot of the magnitude of the periodic component of the hoop stress τθθ over acycle of the forcing for ω1 = O(1) (Wez = 1.0, Wp = 4.06, α = 6.6, ω1 = 1.0). Themagnitude remains O(1) over the entire cycle.
4.6 Conclusions
In this chapter, we showed that the elastic instability in isothermal Dean flow could be
delayed by the addition of steady axial flow either in Couette or Poiseuille form. Both
axisymmetric and non-axisymmetric disturbances are suppressed. The stabilization is a
result of the additional axial normal stress resulting from the axial flow. We conducted
a weakly nonlinear analysis to determine the criticality of the bifurcation in Dean flow
with and without axial flow. Our results indicate that the bifurcation is subcritical for pure
Dean flow, and the subsequent nature of the bifurcation depends on Wez and S. We also
showed that time-periodic axial Couette flow can be used to stabilize Dean flow. Finally,
we report a codimension-2 Takens-Bogdanov bifurcation point at an axial wavenumber
of 21.9 for Dean flow without axial forcing. This bifurcation point represents a change in
the mechanism of the instability.
116
Chapter 5
Localized solutions in viscoelastic
shear flows†
In the previous chapter, we discussed how the primary instability in Dean flow can be
suppressed by the addition of an axial flow that is small in magnitude when compared to
the primary flow. The tools that we used were asymptotic analysis, linear stability analy-
sis, and weakly nonlinear analysis. These tools are useful in giving us information about
the dynamics close to the point where the flow first loses stability. As recent experimental
work (Groisman and Steinberg, 1997, 1998; Baumert and Muller, 1999; Groisman and
Steinberg, 2000) shows, the dynamics far from the bifurcation point can be very complex.
Of particular interest to us are the observations of stationary, long wavelength structures
in circular Couette flow by Groisman and Steinberg (1997), Groisman and Steinberg
(1998), and Baumert and Muller (1999). This chapter describes how a fully nonlinear
† Most of the material in this chapter has been published in Kumar and Graham (2000a) and submittedfor publication in J. Fluid Mech.
117
analysis can capture the main features of these localized solutions in circular Couette
flow.
5.1 Introduction
Spatially localized structures are common in pattern forming physical systems (Cross
and Hohenberg, 1993). Such patterns are interesting and important because they are an
indication of significant nonlinear effects, and their interaction with other patterns may
give information on spatiotemporal behavior. Examples of oscillatory localized struc-
tures can be found in binary liquid mixtures (Moses et al., 1987; Heinrichs et al., 1987;
Kolodner et al., 1988), parametrically excited surface waves (Wu et al., 1984), elas-
tic media (Wu et al., 1987), granular media (Umbanhowar et al., 1996; Lioubashevski
et al., 1996; Fineberg and Lioubashevski, 1998), and colloidal suspensions (Liouba-
shevski et al., 1999). Recently, stationary, two dimensional finite amplitude localized
states have been computed in Newtonian plane Couette flow (Cherhabili and Ehrenstein,
1995, 1997). These solutions are isolated from the base Couette flow branch and were
computed by numerical continuation of traveling wave solutions in plane Poiseuille flow.
Although unstable, these may be related to coherent structures observed in turbulent plane
Couette flow.
In flows of viscoelastic liquids, long wavelength structures were first observed by
Beavers and Joseph (1974) in a circular Couette device. These structures, termed “ tall
Taylor cells” , are primarily inertia driven patterns (Taylor vortices) modified by elastic-
ity. Similar patterns have been computed by Lange and Eckhardt (2000). In contrast,
the structures seen by Groisman and Steinberg (1997), Groisman and Steinberg (1998),
118
and Baumert and Muller (1999) are driven purely by elasticity, since the Reynolds num-
ber is negligibly small in their experiments. There are three interesting aspects to their
observations: (1) isothermal linear stability analysis in this geometry never predicts sta-
tionary bifurcations, (2) these vortex pairs, dubbed “diwhirls” by Groisman and Steinberg
(1997), and “fl ame patterns” by Baumert and Muller (1999) and are very localized, i.e.,
there does not seem to be a selected axial wavelength for these patterns, and (3) the tran-
sition back to the base Couette flow is hysteretic, i.e., the shear rate at which the Couette
flow base state is recovered is much lower than the onset point at which it loses stability.
Here, we seek answers to the following questions, motivated by these observations: (1)
Do isolated branches of stationary solutions exist in a simple model for a viscoelastic
fluid? (2) Are such solutions, if they exist, localized in space? (3) Can the results from
the computations be used to postulate a self sustaining mechanism for these structures?
We address these questions by fully nonlinear computations of the branching behav-
ior of an isothermal inertialess Finitely Extensible Nonlinear Elastic (FENE) dumbbell
fluid in the circular Couette geometry. Our computations show that an isolated branch of
stationary solutions does indeed exist in the circular Couette geometry. In common with
the experimentally observed patterns (which, adopting the nomenclature of Groisman
and Steinberg (1997), we term diwhirls), they are long wavelength solutions, exhibit sig-
nificant asymmetry between radial inflow and outflow, and show hysteresis. In addition,
these solutions persist at arbitrarily large wavelengths: some of the solutions we have
computed have an axial wavelength that is more than a hundred times larger than the gap
width. We also use the results from our computations to propose a self-sustaining mech-
anism for these patterns. Along with the circular Couette flow base state, these structures
119
may form the building blocks for complex spatiotemporal dynamics in the flow of elas-
tic liquids, such as the recently observed phenomenon of elastic turbulence (Groisman
and Steinberg, 2000). In addition, they may be linked to localized defects seen in poly-
mer processing operations and possibly to the strongly nonlinear and long-wave features
observed in free surface flows (Grillet et al., 1999).
In the previous chapter, we discussed the mechanisms by which elastic instabilities
arise in Dean and circular Couette flows due to the inward radial force associated with ten-
sile stresses along curved streamlines. As mentioned earlier, non-isothermal effects can
give rise to stationary, axisymmetric bifurcations. However, the shear rates at which these
bifurcations occur are an order of magnitude lower that those at which the diwhirls are ob-
served. Therefore, the non-isothermal mechanism appears to have limited relevance for
the diwhirls, but seems to explain the very weak stationary vortices seen experimentally
by Baumert and Muller (1995, 1997). Hereafter we consider isothermal flow.
The absence of a stationary bifurcation from the circular Couette flow base state
means that any branch of stationary solutions that exists in this flow must be isolated
from the base state flow, i.e., there can be no direct path from the base state flow to this
branch of solutions. One way of accessing such an isolated branch is to use a technique
known as “homotopy.” The idea in this technique is to start with a problem different from
the original one, but whose solution has the desired properties. For example, the modified
problem may be easier to solve than the original one. After the solution to the modified
problem is computed, it is tracked as the problem is morphed to the original one. Re-
cently, homotopy has been used to find isolated solutions in plane Couette flow (Waleffe,
1998). For our problem, we seek a flow whose base state has a stationary bifurcation, and
which can be easily morphed to circular Couette flow. Clearly, Dean flow is a very good
120
candidate to satisfy this criterion, since the flow geometry is identical to that of circular
Couette flow, and there is a stationary bifurcation from the base state for a wide range of
parameters. Starting from Dean flow, we can approach circular Couette flow in a smooth
way by progressively decreasing the pressure drop, while simultaneously increasing the
rotation speed of one of the cylinders. The linear stability characteristics of viscoelastic
Couette-Dean flow were studied by Joo and Shaqfeh (1992b). As we might expect, this
flow is unstable to a stationary axisymmetric mode when the pressure gradient is the dom-
inant driving force, whereas a non-axisymmetric oscillatory mode is the most dangerous
when cylinder rotation dominates.
Work on nonlinear analysis in viscoelastic circular Couette and Dean flows has con-
centrated on regimes close to the bifurcation point - there have been no extensive com-
putational studies of fully nonlinear behavior in these flows. Renardy et al. (1996) con-
ducted a nonlinear analysis to study mode interactions arising from the introduction of
inertia into the system. Graham (1998) performed a weakly nonlinear analysis to de-
termine the criticality of the bifurcation in circular Couette flow in the narrow gap limit
upon addition of axial flow. Later, Ramanan et al. (1999) extended this analysis to Dean
flow. Khayat (1999) used a low dimensional model in an attempt to determine the dy-
namical behavior in purely elastic and inertio-elastic circular Couette flow. It should be
noted, however, that stress localization (a striking example of which will be seen below)
is common in viscoelastic flows, and it is questionable whether a simple low dimensional
model, which is essentially a low-resolution Galerkin projection, can adequately capture
such behavior.
The strategy we adopt to search for isolated branches of stationary solutions in cir-
cular Couette flow is a fully nonlinear analysis of the governing equations. We use a
121
numerical continuation procedure (Seydel, 1994) to trace out stationary nontrivial solu-
tions bifurcating from the trivial branch in Dean or Couette-Dean flow and see if these
solutions persist as a parameter is varied smoothly to change the flow from Dean or
Couette-Dean to pure circular Couette flow. Any such stationary solutions that persist in
the limit of circular Couette flow have to be part of an isolated branch since there is no
stationary bifurcation from the base state isothermal circular Couette flow. In the remain-
der of the chapter, we report our procedure and results as follows. Section 5.2 contains a
discussion of the geometry, governing equations, and scalings that are used in the com-
putations. In section 5.3, we present a discussion of the discretization scheme and the
numerical method that we use to solve the sparse linear systems arising in the Newton
iterations during the continuation process. This section includes discussion on a precon-
ditioner that we have found to be especially useful. In section 5.4, we discuss the results
of continuation in the various parameters, mechanism of the diwhirl solutions we com-
pute, their stability with respect to time dependent axisymmetric and non-axisymmetric
disturbances, and present a quantitative comparison of our computed diwhirls with exper-
imental data (Groisman and Steinberg, 1998). Finally, we conclude in section 5.5 with a
summary of our main findings.
5.2 Formulation
We consider the flow of an inertialess polymer solution between two concentric cylinders
(figure 49). The inner cylinder has radius R1 and the outer cylinder has radius R2.The
fluid has a relaxation time λ; the polymer and solvent contributions to the viscosity are
122
ΩR 1
R 2Kθ
Figure 49: Geometry of Couette-Dean flow in an annulus
denoted respectively by ηp and ηs, with the ratio ηs/ηp denoted by S. The solution vis-
cosity ηt, is given by the sum of the solvent and polymer viscosities, ηs + ηp. The flow
is created by a combination of the motion of the inner cylinder at a velocity ΩR1 and by
the application of an azimuthal pressure gradient Kθ = ∂P/∂θ.
The equations governing the flow are the dimensionless momentum and continuity
equations
∇ · τ −∇p+WeθS∇2v = 0, (124)
∇ · v = 0, (125)
where v is the velocity, p is the pressure and τ is the polymer stress tensor. The polymer
molecules are modeled as dumbbells connected by finitely extensible springs. Approx-
imate constitutive equations for this model include the FENE-P equation (Bird et al.,
1987b) and the FENE-CR equation (Chilcott and Rallison, 1988), which were described
123
in chapter 1. In dimensionless form, they are
Weθ
(∂〈QQ〉
∂t+ v · ∇〈QQ〉− 〈QQ〉 · ∇vt − 〈QQ〉 · ∇v
)
+
(〈QQ〉
(1− tr(〈QQ〉)/b)−
I
(1− crtr(〈QQ〉)/b)
)= 0, (126)
where 〈QQ〉 is the ensemble average of the polymer conformation tensor,√b is a dimen-
sionless measure of the maximum extensibility of the dumbbells (the dimensionless form
of Q0 in chapter 1), Weθ is the Weissenberg number, which is the product of the polymer
relaxation time and a characteristic shear rate, and cr is a parameter which takes the value
1 for the FENE-CR model and 0 for the FENE-P model. Values of cr between 0 and 1 do
not correspond to any standard constitutive equation; this parameter merely serves as a
convenient way of performing numerical continuation between the FENE-P and FENE-
CR equations. We have already discussed the behavior of the FENE-P and FENE-CR
models in shear and extension in chapter 1. Here, we simply mention that the FENE-
P model has been found to better approximate the behavior of the kinetic theory based
FENE model in steady shear and elongational flows than the FENE-CR model (Herrchen
and Ottinger, 1997). Given the differences between the FENE-P and FENE-CR models
even in simple flows, we would expect them to exhibit different behavior in complex
flows as well, and our computations confirm this.
For both models, Q20 and the components of 〈QQ〉 are scaled by kT/H , where k
is Boltzmann's constant, T is the temperature, and H is the spring constant. Distance
is scaled by the gap width, and time by the inverse of a characteristic shear rate. Since
the FENE-P model does not yield an analytical solution for the base state in Couette-
Dean flow, we take the characteristic shear rate to be the shear rate at the outer cylin-
der for an Oldroyd-B fluid i.e., 1/b = 0) flowing through the same geometry. The
124
velocity scale is chosen to be the product of the time scale and the gap width. The
polymer stress, scaled by the shear modulus, is obtained from 〈QQ〉 using the relation
τ = 〈QQ〉/(1−tr(〈QQ〉)/b)−I/(1−crtr(〈QQ〉)/b). Other parameters of importance
are the dimensionless gap width ε = (R2 − R1)/R2, and δ, which measures the relative
importance of the pressure gradient as the driving force for the flow, given by
δ =−Kθε
2R2/(2 ηt)
(1− ε)R2Ω−Kθε2R2/(2 ηt), (127)
so that δ = 0 is circular Couette flow and δ = 1 is Dean flow. Explicit forms for
the scalings used are presented in Appendix C. The velocity satisfies no slip boundary
conditions on the walls of the cylinder.
5.3 Discretization and solution methods
Equations 124, 125 and 126 form a set of partial differential equations for the three com-
ponents of the velocity, the pressure, and the six components of 〈QQ〉. We look for
steady, axisymmetric solutions that are periodic in the axial direction with a dimension-
less period (scaled by the gap width) of L, so each variable only depends on two spatial
directions, the radial direction r (shifted and scaled so that r = 0 is the inner cylinder
and r = 1 the outer cylinder), and the axial direction z. In performing the numerical
discretization, we can take advantage of certain symmetry properties of the solutions
we seek. In particular, we take the radial and azimuthal velocities to be reflection sym-
metric about the plane z = L/2, and the axial velocity to be reflection anti-symmetric.
This implies that 〈QQ〉rr, 〈QQ〉rθ, 〈QQ〉θθ, 〈QQ〉zz, and p are reflection symmetric, while
〈QQ〉rz and 〈QQ〉θz are reflection anti-symmetric. Thus, the computational domain is
Γ = 0 ≤ r ≤ 1, L/2 ≤ z ≤ L, which is half the size of the physical domain.
125
Considerable care needs to be exercised when choosing a discretization scheme.
Since we are looking for localized solutions, our primary consideration is to choose a
discretization scheme that can place a high concentration of points in regions of strong
velocity and stress localization. In our work, we experimented with two discretization
schemes. The first scheme that we used was a global spectral method. Spectral meth-
ods enjoy the very desirable property of exponential convergence as long as all the fea-
tures of the solution are captured. For our problem, we used Chebyshev discretization
in both directions. In the axial direction, we implemented the symmetry conditions by
choosing odd Chebyshev interpolants for the reflection anti-symmetric components, and
even Chebyshev interpolants for the reflection symmetric components. Chebyshev in-
terpolants work better than Fourier interpolants in the periodic direction because they
have an uneven distribution of collocation points. Regions at the two ends of the domain
have a higher concentration of collocation points, and we can get an improvement in con-
vergence compared to Fourier method by making these regions coincide with the areas
of stress localization. The early computations were performed with this global method.
However, we found that as the degree of localization of the solutions increased, very high
order polynomials were needed, which resulted in unacceptable increases in memory re-
quirements, and presented difficulties in solving the linear systems associated with the
continuation scheme outlined below due to the poor condition number of the matrices.
Our experience with the spectral method indicated that the appropriate discretization
scheme would permit efficient local concentration of points. To this end, we attempted a
domain decomposition Chebyshev collocation scheme (Canuto et al., 1988) in which we
split the computational domain into four conforming rectangular sub-domains and used
Chebyshev collocation in both directions in each of the domains. The interface conditions
126
for each variable were continuity of the value of the variable and its normal derivative be-
tween the domains. Since the governing equation for 〈QQ〉 is hyperbolic, continituity of
the normal derivative is not technically required for its components. However, we found
that the solutions on the non-trivial branch would not converge unless this condition was
imposed. Even when these conditions were imposed, the performance of this scheme was
worse than the global spectral method. One reason could be that the interface conditions
on the components of 〈QQ〉 impose stronger continuity requirements than needed, thus
inhibiting convergence. Also, there is a degree of arbitrariness in the direction of the
normal derivative at the corner point common to all four domains.
The next scheme that we tried was the spectral element method (Patera, 1984). This
method, which we describe below, allows efficient local refinement by subdividing the
domain into several sub-domains or elements. Within each element, the solution com-
ponents are approximated by tensor products of high order orthogonal polynomials. As
long as the solution components within each element are well resolved, this scheme pre-
serves the exponential convergence properties of the global spectral method. As will
be evident from the description below, interface and corner points are treated naturally
within this formulation: the governing equations at these points are the sum of contri-
butions from all elements that border them. We found that this method was superior to
both the global spectral method as well as the domain decomposition collocation scheme
in terms of resolving the stress and velocity localization. In addition, we found that the
linear systems associated with the continuation scheme presented later in this section are
also better conditioned and more sparse for a spectral element scheme than for a global
spectral method. Finally, the computational domain is rectangular in shape, which makes
this method relatively easy to implement.
127
In the spectral element formulation, we apply Galerkin weighting on the conser-
vation and continuity equations (i.e., the test and weight functions are the same), and
streamline-upwind/Petrov-Galerkin weighting (Brooks and Hughes, 1982; Marchal and
Crochet, 1987) on the constitutive equations. The formulation, which is based on the
weak form of the governing equations, is given by∫Γ
(−p I + τ +WeθS∇v) : ∇u dΓ =
∫∂Γ
u · (−pI + τ +WeθS∇v) · nds, ∀u ∈ U,
(128)
∫Γ
q∇ · vdΓ = 0, ∀ q ∈ Q, (129)
∫Γ
(Weθ
(∂〈QQ〉
∂t+ v · ∇〈QQ〉− 〈QQ〉 · ∇vt − 〈QQ〉 · ∇v
)+
〈QQ〉(1− tr(〈QQ〉)/b)
−I
(1− crtr(〈QQ〉)/b)
)(w + c
′v · ∇w
)dΓ = 0, ∀w ∈W,
(130)
where U ∈ H1(Γ), the space of functions whose first derivatives are square integrable
over Γ, W,Q ∈ L2(Γ), the space of functions which are square integrable overΓ, and c′=
h′/V
′where h
′is a characteristic length scale of an element, and V
′is a characteristic
velocity. We take h′
to be the square root of the area of the element, and V′
to be the
average of the magnitude of the velocity at the four corners of each element. We take
u, q, and w to be the same as the interpolating functions for the velocity, pressure and
polymer stress components. These functions must be chosen to satisfy the symmetry
conditions discussed above.
In each element, the variables are approximated by tensor products of Lagrange poly-
nomials defined on the Gauss-Lobatto-Legendre (GLL) grid. We take the symmetry prop-
erties into account by treating the elements bordering the axial edges in a different way
128
z
r
2.5 3 3.5 4 4.50
0.2
0.4
0.6
0.8
1
Figure 50: A spectral element mesh with 16 axial and 16 radial elements with fifth orderpolynomials in each direction in each domain. Note the dense concentration of pointsnear r = 1 and z = L/2. The high resolution is necessary to capture the intense stresslocalization in these regions.
than the interior elements. Since the GLL grid is defined in the domain −1 ≤ ξ ≤ 1,
we map each interior element to −1 ≤ ξr ≤ 1,−1 ≤ ξz ≤ 1. We map the elements
bordering the left edge (z = L/2) to the range −1 ≤ ξr ≤ 1, 0 ≤ ξz ≤ 1, and use
even axial interpolants for the reflection symmetric components and odd interpolants for
the reflection anti-symmetric components. Similarly, we map the elements bordering the
right edge (z = L) to −1 ≤ ξr ≤ 1,−1 ≤ ξz ≤ 0 and use even or odd interpolants as
appropriate. A sample spectral element mesh is shown in figure 50.
In order to avoid spurious pressure modes, the relative approximation orders of ve-
locity and pressure need to satisfy the Ladyzheskaya, Babuska and Brezzi condition. We
ensure this by using interpolants for the pressure that are based on a grid that has two
129
fewer points in each direction than the velocity grid in each element. We choose the ap-
proximation orders for the components of 〈QQ〉 to be the same as for the velocity. We
perform the integrations in equations 128 to 130 using Gauss-Legendre quadrature on the
GLL grid in each element and construct the final system by direct stiffness summation.
This procedure reduces the system of nonlinear partial differential equations to a system
of nonlinear algebraic equations for the nodal values of the variables on the appropriate
GLL grid in each element. These equations can be written in compact form as
E∂y
∂t= f(y,Weθ, S, ε, b, L, δ). (131)
The matrixE has zeros in the rows corresponding to the momentum and continuity equa-
tions. Steady states correspond to solving
f(y,Weθ, S, ε, b, L, δ) = 0. (132)
Solutions to equation 132 are tracked using a numerical continuation procedure, the
starting point for which is the base state Oldroyd-B solution. This solution is used as an
initial guess for the FENE-P or FENE-CR base state solution and refined using a Newton
iteration. We calculate subsequent points along the branch using a pseudo-arclength con-
tinuation algorithm (Seydel, 1994), which we briefly describe here. Let us denote the set
of the values of the variables at the collocation points by the vector y, and the continua-
tion parameter by µ. Here, µ could be Weθ, b, or L. In pseudo-arclength continuation, we
consider both y and µ to be functions of a step length parameter s. Thus, we can write
the set of discretized equations in the compact form
f(y(s), µ(s)) = 0. (133)
130
Given a point (y0, µ0) on the solution branch, the idea is to find the next point (y1, µ1)
such that, apart from satisfying the governing equations, it obeys an additional constraint
N(y, µ) = y0 · (y − y0) + µ0(µ− µ0)−∆s = 0, (134)
where (y0, µ0) is the unit tangent at the point (y0, µ0) , and∆s is a specified step length.
This equation requires that the next computed solution point lie a distance ∆s from the
current point, in the direction of the tangent to the solution curve. At each step, we use a
Newton iteration to solve the augmented set of equations
f(y(s), µ(s)) = 0, (135)
N(y(s), µ(s)) = 0.
The Jacobian matrix of this system is given by
J =
fy fµ
y0 µ0
(136)
and is not singular at a turning point.
While tracing a solution branch, we check for stationary bifurcations using a test
function method (Seydel, 1994). The test function is a scalar function that changes sign
at a stationary bifurcation point, and is relatively inexpensive to compute. Suppose that
(y0, µ0) is a stationary bifurcation point. Then, it follows that
fy(y0, µ0)h = 0, (137)
where h is the eigenvector corresponding to the zero eigenvalue of fy(y0, µ0). Suppose
now that we are at a point (y, µ) different from the bifurcation point. Then equation 137
with (y0, µ0) replaced by (y, µ) has no nontrivial solution for h. However, we can get
131
a solution to an equation that resembles this closely. We arbitrarily choose two indices l
and k, and require that hk = 1. We do this by replacing the lth row in equation 137 by the
equation etkh = 1, where ek is the column vector with a 1 in the kth position and zeros
elsewhere. After this substitution, equation 137 becomes
J lkh = el, (138)
where J lk is the matrix obtained after performing the indicated substitutions in fy. If we
are exactly at the bifurcation point, then h is simply the eigenvector corresponding to the
zero eigenvalue, normalized so that its kth component is 1. If equation 138 is solved close
to a bifurcation point, then h is a good approximation to the eigenvector corresponding
to the zero eigenvalue. In particular, the scalar function
tlk = etlfy(y, µ)h (139)
is zero at a bifurcation point, and changes sign as a bifurcation point is crossed. We use
tlk with l = k to check for stationary bifurcations.
If a stationary bifurcation is detected, as for example when the trivial branch in Dean
flow becomes unstable, we need to begin tracking the new branch. To compute a first
approximation to a point on the new branch, we use the fact that h closely approximates
the eigenvector corresponding to the zero eigenvalue, and write
z = y + δ0h, (140)
for some small value of δ0, as an approximation to a point on the new branch. However, if
we perform a Newton iteration starting with z as the initial guess, we will likely converge
back to the old branch. Instead, we perform a Newton iteration on the augmented system
132
of equations f(z, µ)
zk − z
= 0 (141)
to solve for a point (z∗, µ∗) on the new branch. In equation 141, we are simply specifying
the value of a solution component on the new branch and solving for the point (z∗, µ∗)
where this holds. The method is especially effective when switching from the base solu-
tion to the nontrivial branch because we can choose k to be a component of vr or vz, both
of which are zero in the base state.
The question of determining stability with respect to oscillatory disturbances can be
divided into two parts. For axisymmetric disturbances, we need to find the eigenvalues ω
of the generalized eigenvalue problem
fy q = ωE q. (142)
If any of the eigenvalues have positive real parts, the solution is unstable with respect to
disturbances that have the same symmetry properties that y does and have wavelengths L∗
such that L/L∗ is an integer, otherwise it is stable with respect to such disturbances. We
do not attempt here the much more demanding task of determining stability with respect
to disturbances of arbitrary wavelength. Since we are only interested in determining
stability, we do not need to find the entire spectrum of eigenvalues. It is only necessary
to check if any of the ω have positive real parts. To do this, we use an Arnoldi scheme,
as implemented in the public domain software package ARPACK (Lehoucq et al., 1997),
to calculate a few such eigenvalues. Since ARPACK does not have a built in option
to calculate a specified number of eigenvalues with positive real parts, we use a spectral
transformation suggested by Christodolou and Scriven (1988). Using this transformation,
133
we find the eigenvalues κ of the matrix P = (E − fy)−1(E + fy). These eigenvalues
are related to the the eigenvalues ω in equation 142 by means of the transformation
κi =1 + ωi
1− ωi
. (143)
The eigenvectors of P are identical to those of equation 142. This transformation maps
the eigenvalues in the left half of the complex plane to the interior of the unit circle.
Thus, the eigenvalues of equation 142 with positive real parts map on to the eigenvalues
of P with the largest magnitude, and are easily found by ARPACK. We should mention
here that the Arnoldi scheme constructs a Krylov subspace by the successive action of P
on a vector. As is evident from the definition of P , the construction of each such vector
requires the solution of a linear system. Thus, the eigenvalue computation is an expensive
process, and we only perform it for a few points.
The process for determining stability with respect to non-axisymmetric modes is
somewhat more complicated. We write the solution vector φ as
φ(r, z, θ) = φ(r, z) + ε∗φ(r, z) exp(c t+ i n θ), (144)
where ε∗φ is a small perturbation, c is the growth rate, and n is the azimuthal wavenumber
of the perturbation, assumed to be an integer. Substituting this in the governing equations,
and retaining terms at O(ε∗) gives a complex generalized problem for the growth rate c:
J u = cE u. (145)
As in the axisymmetric case, we can reduce this to a regular eigenvalue problem using
the spectral transformationK = (E − J)−1(E + J).
At this point, it is clear that every step of the procedure involves the solution of a
134
system of linear equations. Let us denote the generic equation we solve as
Ax = b. (146)
It is well worth expending effort to make the solution process efficient. Firstly, A is a
sparse matrix, so considerable savings in memory result from just storing the nonzero
entries together with integer pointer arrays that store information about the coordinates
of each stored entry. The sparse nature of A also means that a properly implemented
iterative method could be more efficient at solving equation 146 than a direct scheme.
The iterative scheme we use is GMRES (Saad and Schultz, 1986). However, GMRES
will converge only if A is well conditioned, and this is generally far from being true for
the systems we solve. Therefore, we solve a preconditioned system
MAx =Mb, (147)
where M is an approximation to A−1. In the remainder of this section, we present a
discussion on some of the preconditioners that we have used.
The first preconditioner that we describe is the incomplete LU decomposition precon-
ditioner with zero fill level, or ILU(0) for short. This preconditioner is a sparse version of
the full LU decomposition algorithm, with sparsity being preserved by only keeping those
entries in theL andU matrices where the corresponding positions inA are nonzero. This
technique gives no consideration at all to the size of the entries that are dropped, so we
would not expect this to be a good preconditioner. However, our experience has been
that this preconditioner is surprisingly robust, and works reasonably well up to b = 1500,
where b is the finite extensibility parameter. At this value of b, we need about twenty
five thousand degrees of freedom to resolve the components of 〈QQ〉, and solving these
linear systems takes about 450 GMRES iterations. At larger values of b, however, ILU(0)
135
fails due to numerical instability: the dropping of entries causes the buildup of very large
or very small numbers in the LU decomposition, which makes it impossible to complete
it. Pivoting actually makes the situation worse by giving rise to zero pivots. Although
we can complete the decomposition by replacing these with small numbers, the resulting
matrix is useless as a preconditioner.
The discussion above suggests that a preconditioner that drops entries based on nu-
merical size would be more effective than ILU(0). A variant of ILU which takes this into
account is threshold ILU, or ILUT (Saad, 1996). In this method, entries in the incomplete
LU decomposition are dropped based on tolerances applied at two different stages. The
algorithm is shown below, where we denote the dimension of the matrix by N and use
the symbols l and u to represent the entries in the L and U matrices respectively. In each
entry, the first subscript represents the row, and the second represents the column. The
algorithm used a one-dimensional work array of length N , which we denote by w below.
1. Do i=1, N
2. w:=ai,∗ (Copy the nonzero entries of row i into w)
3. Do k=1, i− 1 and when wk = 0
4. wk := wk/ukk
5. Apply a dropping rule to wk
6. If wk = 0 then
7. wj = wj − wk ∗ ukj
8. EndIf
136
9. EndDo
10. Apply a dropping rule to row w
11. lij = wj for j = 1, . . . , i− 1
12. uij = wj for j = i, . . . ,N
13. w := 0
14. EndDo
The decomposition is sparse because of the dropping rules applied in steps 5 and
10. The usual choice in step 5 is to keep entries which are larger than some fraction
(tol) of the norm of the row in the original matrix. In step 10, it is usual to keep the
p entries which have the largest magnitude in order to control the amount of fill in. In
using ILUT, we found that the algorithm would generate zero pivots unless p was made
unacceptably large. Again, we can complete the decomposition by replacing these pivots
with small numbers, but the resulting preconditioner is completely ineffective. We found
that this problem could be overcome by performing ILUT on a modified version of A.
We construct this matrix, denoted by A, by setting to zero inA the entries corresponding
to the velocities and pressure in the rows corresponding to the constitutive equation, and
the entries corresponding to the components of 〈QQ〉 in the rows corresponding to the
momentum and continuity equations. We then construct an ILUT decomposition of A
and apply this as a preconditioner for A. We find that this is a very effective technique,
and the largest linear systems that we have solved (O(60000) unknowns) converge to a
relative accuracy of 10−6 in about 350 GMRES iterations, with the preconditioner having
about three times the number of nonzero entries that A does. Smaller systems converge
137
faster and fewer nonzero entries can be kept in the preconditioner. Henceforth, we shall
refer to this preconditioner as ILUT∗. Figure 51 shows a comparison between ILU(0) and
ILUT∗ for a sample problem, which clearly demonstrates the superiority of the ILUT∗
preconditioner over ILU(0).
0.0 50.0 100.0 150.0 200.0 iterations
10−8
10−6
10−4
10−2
100
Log
10(r
esid
ual)
ILUT*
ILU(0)
Figure 51: Comparison of ILUT∗ and ILU(0) preconditioners. The test problem was thecalculation of the unit tangent for a point on the nontrivial branch in Dean flow. Thematrix A had a dimension of 21987.
In using ILUT∗ to solve the complex generalized eigenvalue problem indicated by
equation 145, we need to modify the problem so that it only involves real numbers. Recall
that the spectral transformation requires the solution of the linear system
(E − J)x = b (148)
for each iteration of ARPACK. Instead, we rewrite equation 145 as a real valued problem
c
E 0
0 E
ur
ui
=
Jr J i
J i Jr
ur
ui
, (149)
138
where the subscripts r and i denote respectively the real and imaginary parts of the vec-
tor or matrix. The spectral transformation now requires the solution of linear systems
involving the matrix
S =
(E − J)r (E − J)i
−(E − J)i (E − J)r
,
which is real. To precondition this matrix, we note that the matrix (E−J)i has far fewer
entries than (E−J)r. Therefore, as a first approximation, we can neglect it in computing
the preconditioner. We construct the preconditioner by performing ILUT∗ on (E − J)r.
If M∗ represents this decomposition, we precondition S using the matrix M ∗−1 0
0 M ∗−1
.
5.4 Results and discussion
5.4.1 Stationary bifurcations from the Dean and Couette-Dean
flows
Rather than explore large volumes of parameter space, we restrict our attention to values
close those used in the experiments by Groisman and Steinberg (1997). Specifically, we
fix the value of S at 1.2, and except when we examine the effect of varying the gap width,
set ε = 0.2. For most of our work, we use the FENE-P model which has been shown
to be a better approximation to the exact FENE model (Herrchen and Ottinger, 1997), at
least in simple flows.
In order to track stationary nontrivial branches, it is first necessary to find out from
where they bifurcate. Therefore, the logical starting point of our investigation is the
139
linear stability diagram for Dean flow. Figure 52 shows such a diagram computed using
the FENE-P model. Takens-Bogdanov points, where the bifurcation switches from a
stationary mode to an oscillatory mode, are marked as “TB” . Unlike in the Oldroyd-B
model, where one such point is seen only at very small wavelengths (Ramanan et al.,
1999), we see that, as the polymer becomes stiffer (i.e., b decreases), these points are
shifted to larger wavelengths. We also see that for a sufficiently small value of b, there is
no stationary bifurcation from the base state at all. From the point of view of numerical
simulation, we would prefer to work with as small a value of b as possible, since we would
expect stress boundary layers to be less sharp for smaller values of b, which in turn makes
the computations easier. However, decreasing b tends to lower the elastic character of the
fluid and suppresses elastic instabilities. Before moving on to the nontrivial solutions, we
present some linear stability results for Dean flow of the FENE-CR model in figure 53.
Note that, unlike in the FENE-P model, stationary bifurcations are seen even at low values
of b. Thus, the FENE-CR model predicts linear stability behavior that is qualitatively
different from that predicted by the FENE-P model. We will have more to say on the
differences between the two models in section 5.4.3. For now, we simply concentrate on
the FENE-P model.
5.4.2 The branch structure of viscoelastic Couette-Dean flows
We begin the nonlinear analysis by tracking the bifurcating branch of stationary solu-
tions in Dean flow at b = 700 and L = 1.05. Since this value of b is small, numerical
continuation is not difficult and a crude numerical scheme suffices. We performed these
calculations using a global Chebyshev collocation scheme in both the radial and axial
directions. At L = 1.05, a pair of complex conjugate eigenvalues crosses the imaginary
140
3.5 4.5 5.5α
15.0
20.0
25.0
30.0
35.0
40.0
We θ
b=600b=700b=800b=900
TBTB
TB
Unstable
Stable
Figure 52: Linear stability curves at δ = 1 (Dean flow) computed using the FENE-Pmodel. The points marked “TB” are Takens-Bogdanov points. The lines correspond topoints where the base state flow loses stability to stationary axisymmetric perturbations.
4.0 4.5 5.0 5.5 6.02π/L
15.0
19.0
23.0
27.0
31.0
We θ
b=500b=600b=700
Figure 53: Linear stability curves at δ = 1 (Dean flow) computed using the FENE-CRmodel. As in figure 52, only stationary axisymmetric perturbations are considered. Notethe complete absence of non-stationary bifurcations.
141
axis at Weθ = 28.73. Upon further increasing Weθ, the two unstable eigenvalues coalesce
and form a pair of unstable real eigenvalues which then move in opposite directions. The
smaller one of these re-crosses the imaginary axis at Weθ = 29.57. We track in Weθ the
stationary branch bifurcating as result of this crossing. The result is shown in figure 54,
where we plot the solution amplitude, measured by the quantity
‖ vr ‖=
(Nr∑i=0
Nz∑j=0
|vr,ij|2
)1/2, (150)
as a function of Weθ, where Nr +1 and Nz +1 are the number of Chebyshev collocation
points used in the radial and axial directions respectively. We see that the bifurcation
is mildly supercritical, but quickly turns back and shows a marked hysteretic character.
The turning point at Weθ = 22.34 is much lower than the value of 28.73 where the base
solution loses stability. When we pick a point on this branch and continue it down in
δ, we find however, that it does not extend all the way to δ = 0. For instance, picking
Weθ = 23.35 on the upper branch in figure 54 and tracking it in δ, we find that the branch
turns back at δ = 0.69. We have tried this for other points as well, but in all cases, they
turn back well before reaching δ = 0. Therefore, at least for this value of b and L, there is
no direct path from nontrivial solutions in Dean flow to those (if any) in circular Couette
flow.
Given the apparent absence of a direct route from δ = 1 to δ = 0, we focussed
on smaller values of δ and larger values of b and L. As b increases, the solutions be-
come more localized, and the global scheme we originally used is inefficient. Hence, we
switched to the spectral element/SUPG method described in section 5.3.
At δ = 0.576, b = 1830, and L = 2.71, a stationary bifurcation occurs at Weθ =
25.15. We tracked the bifurcating branch at this value of Weθ up to L = 3.07 and then
142
22.0 24.0 26.0 28.0 30.0Weθ
0.00
0.02
0.04
0.06
0.08
||vr||
Hopf point
Figure 54: Continuation in Weθ of a stationary solution in Dean flow. The parametervalues are L = 1.05, b = 700, ε = 0.20, and S = 1.2. At the Hopf point, a pairof complex conjugate eigenvalues become unstable. These collide and form two realeigenvalues, one of which re-crosses the imaginary axis at Weθ = 29.57, where thestationary branch originates. The solution amplitude used here differs from that used insubsequent figures and is defined in equation 150.
143
−0.10 0.10 0.30 0.50δ
0.000
0.002
0.004
0.006
0.008
0.010
||vr||
Multiple steady states incircular Couette flow
Continuationto L=2.71.
δ=0
Figure 55: A path to stationary solutions in circular Couette flow. The parameters areWeθ = 25.15, L = 3.07, b = 1830, ε = 0.2, and S = 1.2.
down in δ. This path is shown in figure 55. The velocity norm used in this and all
subsequent figures is defined as
‖ vr ‖=
(∫Γ
v2r dΓ
)1/2. (151)
As figure 55 shows, we found that this branch persists all the way to δ = 0. This compu-
tation demonstrates that an isolated branch of nontrivial solutions does indeed exist in the
circular Couette geometry. This is the first time that stationary nontrivial solutions have
been computed in zero Reynolds number circular Couette flow.
5.4.3 Nontrivial stationary solutions in Couette-Dean flow -
Diwhirls
We now proceed to discuss the effect of changing parameters on the stationary solution
at δ = 0. The first parameter we focus on is the wavelength. The results of continuing
144
our solution at Weθ = 25.15 in L are shown in figure 56. The upper branch, being a
stronger flow, is much harder to track than the lower branch, and the end point of this
branch represents the largest value of L at which we could obtain converged solutions
on the upper branch for this value of Weθ. The lower branch presents fewer problems
and we were able to track it with relative ease. The key observation from figure 56 is
that as L increases, both the lower and upper branches become flat, suggesting that the
spatial patterns are becoming independent of the size of the computational domain, i.e.,
they are becoming localized. Examination of the solution components confirmed that this
was indeed the case, with the localization occurring in the region near z = L/2. Since
the components of the solution show little or no axial variation far away from z = L/2,
we can simply use their values at the collocation points for a lower value of L as an
initial guess for the solution at a larger value of L, while increasing the axial extent of
the domains bordering the edges (i.e., z = L). This method of remeshing captures the
localization effectively and avoids the necessity of computing solutions at intermediate
values of L. Using this technique, we were able to get converged solutions on the lower
branch for wavelengths that are in excess of 100 times the width of the gap between the
cylinders. We have also used this technique to compute such long wavelength solutions
on the upper branch at lower values of Weθ. We show the results of one such computation
in figure 57. This figure shows the streamfunction contours and a density plot of 〈QQ〉θθ
at L = 116.52 and Weθ = 24.29. For clarity, we only show the center and edges of
the domain. The streamfunction contours are strongly localized near the center of the
flow cell, which is a region of very strong inflow. Away from the core is a region of
weak outflow, and even further away, the solution is pure circular Couette flow. The
〈QQ〉θθ field shows an even stronger localization. It is the necessity of capturing this
145
2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0L
0.000
0.002
0.004
0.006
0.008
0.010
0.012
||vr||
Figure 56: Results from continuing the stationary circular Couette flow solutions in L.The parameters are Weθ = 25.15, b = 1830, ε = 0.2, and S = 1.2. The gaps in thelower branch correspond to places where we changed the mesh. Note the flatness of thebranches as L increases. We have computed extensions of the upper branch at lowervalues of Weθ.
strong localization that requires the use of a numerical method that permits efficient local
refinement. The streamlines at the core are remarkably similar to those in figure 10 on
page 2457 of Groisman and Steinberg (1998). Henceforth, we will call our solutions
diwhirls as well.
Figure 58 shows the results of continuation in Weθ for solutions at three different
values of L. We see that all three curves show turning points in Weθ, i.e., there is a lower
limit in Weθ below which the diwhirls are not seen. Note that the curves at L = 9.11
and 4.74 are close together, and are both well separated from the curve at L = 3.07.
This further highlights the independence of the solutions on L for large enough L. At
L = 3.07, the critical Weθ at which the base circular Couette flow solution loses stability
to an axisymmetric time-dependent mode is 29.65, which is significantly higher than the
146
L=116.52
2.8
1
1
Outer cylinder
Inner cylinder
Outer cylinder
Inner cylinder
Figure 57: Density plot of 〈QQ〉θθ (white is large stretch, black small) and contour plotof the streamfunction at L = 116.52 (Weθ = 24.29, b = 1830, S = 1.2, and ε = 0.2).For clarity, most of the flow domain is not shown. Note the very strong localization of〈QQ〉θθ near the center. The maximum value of 〈QQ〉θθ at the core is 1589 which givesτθθ = 12722. Compared to this, the maximum value of 〈QQ〉θθ in the circular Couettebase state is 706, which gives τθθ = 1150. Away from the core of the diwhirl, the structureis pure circular Couette flow. The streamlines show striking similarity to those in figure10 of Groisman and Steinberg (1998). This point was generated by stretching the pointat the corresponding Weθ on the upper branch of the curve for L = 9.11 in figure 58.
147
23.0 23.5 24.0 24.5 25.0Weθ
0.000
0.002
0.004
0.006
0.008
0.010
||vr||
L=9.11L=4.74L=3.07
Figure 58: Diwhirl solution amplitudes as functions of Weθ and L. Note that the curvesat L = 9.11 and L = 4.74 are very close together, while both curves are well separatedfrom the curve at L = 3.07 (b = 1830, S = 1.2, and ε = 0.2).
location of the turning point, located at 24.98. The critical Weθ for the other values of L
are even higher. Therefore, the overall bifurcation structure shows hysteresis.
In figure 59, we show a plot of the location of the turning point in Weissenberg num-
ber (Weθ,c) as a function of the wavelength. The most interesting feature in figure 59 is
the flatness of the curve at large L, indicating yet again that for large L, the character-
istics of the solution are independent of the wavelength. Another interesting feature in
figure 59 is that the curve shows a minimum, i.e., the diwhirl patterns exhibit wavelength
selection. This minimum, which occurs at a Weissenberg number of approximately 23.3,
is therefore the lowest Weissenberg number at which the FENE-P model with the cho-
sen parameters predicts diwhirls to occur. More important than the absolute value is the
relative position of the turning point and linear stability limits. The base state circular
Couette flow is unstable with respect to axisymmetric disturbances above Weθ = 20.37.
148
This means that all the solutions that we compute lie above the linear stability limit of
circular Couette flow. In contrast, Groisman and Steinberg (1998) observe diwhirls at
Weissenberg numbers as low as 10, well below the linear stability limit of the base flow.
One reason for this discrepancy could be the approximate nature of the FENE-P model,
which does not take into account the internal degrees of freedom of a real long chain
polymer. Yet another could be that our numerical simulations have not been able to ac-
cess a sufficiently high value of b. In figure 60, we plot the position of the turning point
in Weθ as a function for b for L = 4.74, which is close to the minimum in figure 59. Also
plotted in the figure is the minimum critical value of Weθ at which the base state circular
Couette flow loses stability with respect to axisymmetric perturbations. This figure shows
that as b increases, the position of the turning point shifts to lower Weissenberg numbers
at a faster rate than the shift in the minimum of the linear stability curve. This result is
not unexpected, because the polymer molecules are much more highly stretched in the
core of the diwhirl than in the base state. Therefore, we would expect the nonlinearity
of the FENE-P spring law to have a greater effect on the diwhirls than on the base state
Couette flow. It is conceivable to suppose, based on the results shown in figure 59, that
the two curves would cross at larger values of b (which we are not able to access due to
limitations in the numerical scheme) and that the diwhirls would come into existence be-
low the linear stability limit of the base flow. We point out here that Baumert and Muller
(1999) and Groisman and Steinberg (1997) performed their experiments with very high
molecular weight polymers, for which the values of b are likely to be much higher than
we have been able to access in our simulations.
We now present a rough quantitative comparison of our patterns with those from
the experimental observations of Groisman and Steinberg (1998). In figure 9 of their
149
2.0 4.0 6.0 8.0 10.0 12.0 14.0L
23.0
23.5
24.0
24.5
25.0
We θ,
c
diwhirls
no diwhirls
Figure 59: Plot of the location of the turning point, Weθ,c, versus L at S = 1.2 andε = 0.2. Note the flatness of the curve at large L.
paper, Groisman and Steinberg present the radial velocity profile as a function of z at a
constant radial position near the middle of the gap, where vr has maximum amplitude.
To compare our results with this figure, we chose a point on the upper branch of the
curve for L = 4.74 with Weθ = 23.50 in figure 58 of this work. We then converted
our radial velocity into dimensional units by using values for the physical parameters
from Groisman and Steinberg (1998). Specifically, the values we used were λ = 1 s
and R2 = 41 mm. For ε = 0.2 used in our computations, this gives a gap width of
8.2 mm, slightly higher than the 7 mm gap used in the experiments. In figure 61(a), we
present a profile of the radial velocity as a function of z for r = 0.6, where the radial
velocity is maximum. The peak inflow velocity we find is 5.9 mm/sec, which should
be compared to the value of 3.8 mm/sec that Groisman and Steinberg (1998) show in
their figure 9, which we reproduce here as figure 61(b). The qualitative and quantitative
150
1750 1770 1790 1810 1830 b
19.0
20.0
21.0
22.0
23.0
24.0
25.0
26.0
27.0
We θ,
c
diwhirl onset
linear onset
Figure 60: Plot of the position of the linear stability limit in circular Couette flow withrespect to axisymmetric disturbances and the turning point in Weθ for the diwhirls as afunction of b. The parameters are S = 1.2 and ε = 0.2. The computations for the diwhirlswere performed at L = 4.74, which is close to the minimum in figure 59.
151
similarity between these two values is remarkable, more so when we consider how simple
the FENE-P model is, and that the Weissenberg number used in figure 61(a) is roughly
twice that at which Groisman and Steinberg (1998) report their results, which means a
larger radial velocity should be expected. Furthermore, the wiggles at the shoulders of
the peak in figure 61(a) are not numerical artifacts; similar features in figure 61(b).
0 1 2-1-2
1
0
-1
-2
-3
-4
-5
-6
Vr
, mm
/s
z , d
(a) (b)
Figure 61: (a) The axial variation of vr at r = 0.6 for L = 4.74 and Weθ = 23.50 on theupper branch. (b) Figure 9 on page 2457 of Groisman and Steinberg (1998), shown herefor purposes of comparison. We have shifted the axial coordinate so that the symmetryaxis of the computed diwhirl in (a) is at z = 0, to make comparison with (b) easier
The dimensionless gap width or curvature ε plays a critical role in generating elastic
instabilities. Based on the generic mechanism of elastic instabilities, we expect a decrease
in curvature to have a stabilizing effect, i.e., keeping other parameters fixed, we would
expect the diwhirl pattern to vanish at small enough values of ε. In figure 62, we show
the dependence of the diwhirl solution amplitudes on ε. In agreement with expectations,
we observe turning points as ε decreases. The role of streamline curvature will become
clear in section 5.4.4 where we propose a mechanism for the diwhirls.
As mentioned in section 5.1, an important reason for attempting to numerically simu-
late experimentally observed flow patterns is to determine whether a constitutive equation
152
0.192 0.194 0.196 0.198 0.200ε
0.000
0.002
0.004
0.006
0.008
0.010
||vr||
L=4.74L=9.11
Figure 62: Variation of solution amplitudes with ε. Here, Weθ = 25.15, and the otherparameters as in figure 58.
can model complex flows of viscoelastic liquids. Both the FENE-P and the FENE-CR
equations are derived by applying closures to the evolution equation for 〈QQ〉 for a dilute
solution of noninteracting dumbbells connected by nonlinear springs. We have already
seen that the linear stability curves predicted by the FENE-P and FENE-CR models show
significant differences, and we have seen that the FENE-P model equation has station-
ary solutions in circular Couette flow, which indicates that it is able to capture, at least
qualitatively, the mechanism behind the diwhirls. A natural question to ask is whether
the FENE-CR model can do so as well. To this end, we perform a continuation of the
diwhirl solutions in the parameter cr, starting from the FENE-P solutions (cr = 0). If the
solutions persist at cr = 1, then we will have obtained solitary solutions for the FENE-
CR model in circular Couette flow. Figure 63 shows the results of these computations.
Both values of L that we chose exhibit turning points at small values of cr, indicating that
these solutions do not exist for the FENE-CR model, at least for the parameter values
153
0.0 0.2 0.4 0.6 0.8 1.0cr
0.000
0.002
0.004
0.006
0.008
0.010
||vr||
L=3.57L=4.74
Figure 63: Diwhirl solution amplitudes as a function of the parameter cr for two differentwavelengths. Weθ = 25.15 and the other parameters are as in figure 58. The existenceof turning points demonstrates that these solutions cannot be extended to the FENE-CRmodel.
that we have chosen. Note, though, that in the limit of b → ∞, both the FENE-P and
FENE-CR models are equivalent to the Oldroyd-B model. Hence, at sufficiently large
values of b, the FENE-CR model should yield these localized solutions as well. What
these computations show is that the parameter values at which the solutions come into
existence depends very strongly on the details of the model being used.
The work of Al-Mubaiyedh et al. (2000) has shown that non-isothermal effects intro-
duce a new mode of instability in circular Couette flow. While this mode of instability
does not appear to be relevant for the diwhirls, it is still of interest to quantify the effect
of viscous dissipation at the core of the diwhirl, where we would expect dissipation to be
largest. In dimensionless form the viscous dissipation, scaled by the product of the shear
154
modulus G and the characteristic shear rate γs, is given by the relation
Φ =(τ +WeθS(∇v +∇v
t)):∇v. (152)
Here Φ represents the work done by viscous dissipation per unit time, per unit volume
of fluid. In figure 64, we show a density plot of Φ for a point on the upper branch at
L = 4.74. For the sake of clarity, we have only shown a twentieth of the wavelength,
centered around the symmetry plane. As expected, Φ has its largest amplitude in the core.
2.25 2.35 2.450
0.2
0.4
0.6
0.8
1
z
r
Figure 64: Intensity plot of the dimensionless viscous dissipation for Weθ = 23.73 on theupper branch at L = 4.74. Light areas represent areas of large viscous dissipation anddark areas represent regions of low viscous dissipation. The horizontal axis is stretchedby a factor of two relative to the vertical axis for clarity.
To get an estimate of the magnitude of the viscous dissipation, we can compute the
work done on a fluid element as it traverses the core of the diwhirl, and use that to calcu-
late its temperature rise in the absence of heat loss to the neighboring fluid. This can be
155
calculated using
Ev =
∫ t1
t0
Φdt. (153)
Since the diwhirl is axisymmetric and vz = 0 at the symmetry plane, we can replace dt
by dr/vr and write
Ev =
∫ r1
r0
Φ
vrdr. (154)
Once Ev is known, we can calculate the temperature rise using
∆T =Gγs(R2 −R1)Ev
ρCp
, (155)
where ρ is the mass density of the polymer solution, Cp its specific heat capacity at
constant pressure on a mass basis, and the product Gγs(R2 − R1) is required to convert
Ev to dimensional form. We performed this integral using Simpson's rule for the data
in figure 155 choosing (arbitrarily) r0 = 0.98 and r1 = 0.2. From the data in Groisman
and Steinberg (1998), we use a solvent viscosity of 0.1 Pa s, and a relaxation time of
1 s. As a first (and rough) approximation, we assume the heat capacity and density of the
sugar syrup to be the same as that of water (≈ 4.186× 103J kg−1C−1 and 1000 kg m−3
respectively). This yields a temperature rise of about 10−4C in one cycle. For a PIB/PB
Boger fluid such as the one used in the experiments by Baumert and Muller (1999),
however, the temperature rise would be much higher. For instance, using the data for
their medium viscosity Boger fluid (ηs = 6.5 Pa s, λ = 0.23 s, ρ = 880 kg m3, and Cp ≈
2.0 × 103J kg−1C−1) yields a temperature rise of 0.27C if the geometrical parameters
are taken to be the same as those used by Groisman and Steinberg (1998). Given the
low magnitude of the temperature rise in the PAA Boger fluids, it would appear that
156
the experiments of Baumert and Muller (1999) are more likely to be subject to non-
isothermal effects than those of Groisman and Steinberg (1998). However, since similar
structures are seen in both fluids, this would seem to indicate that non-isothermal effects
do not significantly influence either the diwhirls or the flame patterns.
5.4.4 Self-sustaining mechanism
Since there is no stationary bifurcation in circular Couette flow, the diwhirl solutions that
we have computed are part of an isolated branch that does not connect directly to the base
state. Therefore, the sustaining mechanism for these patterns must be inherently nonlin-
ear. Following their experiments, Groisman and Steinberg (1998) proposed one such
mechanism. They argued that the difference in symmetry between inflow and outflow
results in the elastic forces performing net positive work on the fluid. While this argu-
ment shows that finite amplitude stationary structures that exhibit significant asymmetry
between inflow and outflow are physically plausible, it does not explain the mechanism
by such structures sustain themselves. Having the detailed velocity and stress fields avail-
able to us from our computations, we propose a more complete mechanism. Figure 65
shows a vector plot of v at the axial centerline of the vortex. The azimuthal velocity field
has a parabolic structure near the outer cylinder, similar to the velocity field in the outer
half of the channel in Dean flow. This velocity field results in an unstable stratification of
azimuthal normal stress (Joo and Shaqfeh, 1992b). We therefore propose the following
fully nonlinear mechanism for the diwhirls: a finite amplitude perturbation near the outer
cylinder results in a locally parabolic velocity profile. This in turn creates an unstable
stratification of hoop stress (visible in figure 65(b) for r <∼ 0.99), as in Dean flow, which
drives inward radial motion. As the fluid moves inward, it accelerates azimuthally due to
157
0.97
1.0 1.0
r r
0.97θ θ
(a) (b)
Figure 65: (a) Vector plot of v near the outer cylinder at the center of the diwhirl structure(oblique arrows) and the base state (straight arrows). The length of the arrows is propor-tional to the magnitude of the velocity. The axial velocity is identically zero in the basestate, and is zero by symmetry at the center of the diwhirl. (b) Principal stress directionsat the same location as for (a). The Couette flow stress is not shown because it is verysmall in comparison. This figure shows how fluid elements at larger radii are pulled downand forward sustaining the increase in vθ.
the base state velocity gradient. The azimuthal tension in polymer chains drags the fluid
at larger r forward and down (figure 65(b)). This maintains the increase in vθ and results
in a self-sustaining mechanism (figure 66). Thus we see that, in common with other elas-
tic instabilities, the mechanism is based on the inward radial force associated with tensile
stresses along curved streamlines. This mechanism shows that we have come a full cir-
cle. We began our search for stationary, finite amplitude solutions in circular Couette
flow by starting from Dean flow. We now find that these solutions sustain themselves by
a mechanism similar to the one resulting in the primary instability in Dean flow.
158
local increase in vθ near outer cylinder
unstable stratificationof τθθ, generates inward radial motion
fluid moves azimuthally due to base state shear
tension in the streamlines pulls fluid near the outerouter cylinder forward and downFigure 66: Nonlinear self-sustaining mechanism for the diwhirl patterns.
5.4.5 Stability
We now address the stability of the diwhirl patterns with respect to axisymmetric and
non-axisymmetric perturbations. We determine this by finding the eigenvalues ω in equa-
tion 142 for axisymmetric perturbations or equation 145 for non-axisymmetric perturba-
tions. These computations are expensive, so we only perform them for a few points along
the upper branch for L = 4.74 in figure 58. Here, we report the results from one such
computation, performed at Weθ = 23.87. To resolve this point, we needed 16 radial
elements and 14 axial elements using fifth order polynomials in both directions in each
element, which is a slightly coarser mesh than the one shown in figure 50, and results in
a system with 53687 unknowns for axisymmetric disturbances, and double that number
for non-axisymmetric perturbations. The computation requires the storage of two sparse
matrices, the preconditioner for one linear system, and the Krylov basis. This was well
above the storage capacity available to us on a single computer, so we stored the Krylov
basis and one of the matrices on one machine and the second matrix and preconditioner
159
on another, and used Message Passing Interface (MPI) subroutines to communicate be-
tween the two machines. We used a 600 vector basis (ARPACK would not converge if
significantly fewer vectors were used) and asked for the most unstable eigenvalues.
For axisymmetric disturbances, we found two pairs of complex conjugate eigenvalues
that had positive real parts. The eigenvectors corresponding to one pair had an irregular
grid scale structure, suggesting that they are part of the continuous spectrum of eigen-
values (Graham, 1998; Wilson et al., 1999; Renardy, 2000). These modes are expected
to be stable, but since the eigenvectors are nonintegrable (Graham, 1998), they will not
converge exponentially in a spectral element scheme and can display spurious instability.
The structure of the other two eigenvectors is shown in figure 67. They show that although
the branch is unstable, the destabilizing disturbance has significant amplitude only near
the ends of the domain, where the flow is essentially circular Couette flow. The base
circular Couette flow has a minimum critical Weθ of 20.37 with respect to axisymmet-
ric disturbances, and so is linearly unstable at Weθ = 23.87. Hence, it is not surprising
that the portion of flow pattern where the flow is essentially circular Couette would be
susceptible to destabilizing disturbances. What is interesting, however, is that the core of
the pattern, where the diwhirl lies, is entirely unaffected. This shows that the diwhirl pat-
tern is dynamically distinct from the oscillatory finite wavelength axisymmetric pattern
arising from the linear instability of circular Couette flow.
For non-axisymmetric disturbances, the stability picture is slightly more complicated.
We performed computations with n = 1 and n = 2 and found three pairs of unstable
complex conjugate eigenvalues. Their growth rates are summarized in table 7, while
their structures are shown in figure 68 and 69. For n = 1, we see that there are two
modes that have their largest amplitude close to the core of the diwhirl (figures 68 (a) and
160
(b)), while the third (figure 68 (c)) has a large amplitude away from the core. The third
mode is directly related to the linear instability of circular Couette flow with respect to
non-axisymmetric disturbances with n = 1. For n = 2, the picture is slightly different.
There is still one mode (figure 68(a)) that is largely concentrated outside the core, and
which therefore seems related to the linear instability of circular Couette flow. The mode
in figure 68 (b) is largely concentrated at the core, and shows similarities to figures 68 (a)
and (b). In addition, there is non-localized mode (figure 69(c)) that is absent for n = 1.
n = 1 n = 21. 3.878× 10−2 ± 9.276× 10−3 i 3.674× 10−2 ± 1.742× 10−3 i2. 3.590× 10−2 ± 1.843× 10−2 i 2.968× 10−2 ± 8.189× 10−2 i3. 3.279× 10−2 ± 4.878× 10−2 i 2.788× 10−2 ± 5.722× 10−5 i
Table 7: Growth rates for the unstable nonaxisymmetric modes at Weθ = 23.87, L =4.74, S = 1.2, and b = 1830.
5.4.6 Dean flow revisited
In the previous sections, we have demonstrated that stationary, localized solutions of
very large wavelength exist in circular Couette flow. Here, we return to Dean flow to
investigate whether such solitary solutions are possible there. We do this by tracking
long wavelength steady states that bifurcate from the base state in Dean flow. To make
comparison with the circular Couette flow results easier, we choose b = 1830, S = 1.2,
and ε = 0.2, which are the same parameter values that we used for circular Couette
flow. In figure 70, we show the bifurcation diagram for L = 1.963. This diagram is
similar to figure 54, in that there is a large subcritical region, and the nontrivial branch
exhibits a turning point, i.e., there is a critical value of the Weissenberg number below
which the nontrivial solution does not exist. In figure 71, we show an density plot of
161
z
r
z
r
z
r
(a)
(b)
(c)
Figure 67: Density plot of vr showing the (a) real and (b) imaginary parts of the desta-bilizing axisymmetric disturbance, and (c) the streamlines of the base diwhirl. Note thatthe core of the diwhirl is entirely unaffected by the disturbance. The parameters areWeθ = 23.87, S = 1.2, ε = 0.20, b = 1830, and L = 4.74.
162
t=0 t=T/4
t=0 t=T/4
t=0 t=T/4
(a)
(b)
(c)
r r
r r
r r
z z
z z
z z
Figure 68: Density plot of the perturbation radial velocity for the three non-axisymmetricunstable eigenmodes with n = 1. The parameters are identical to those in figure 67.
t=0 t=T/4
t=0 t=T/4
t=0 t=T/4
(a)
(b)
(c)
r r
r r
r r
z z
z z
z z
Figure 69: Density plot of the perturbation radial velocity for the three non-axisymmetricunstable eigenmodes with n = 2. The parameters are identical to those in figure 67.
163
the radial velocity at Weθ = 15.486 (marked by an open circle in figure 70) and at
Weθ = 13.930 (marked by a filled circle in figure 70). The figure shows that rather than
becoming localized, the nonlinear patterns have a tendency to split into two vortices, each
one having half the wavelength of the original vortex. In fact, for the particular case of
L = 1.963, we were not able to advance the continuation beyond Weθ = 13.94 on the
upper branch. Examination of figure 71(b) reveals why: it is clear from the symmetry of
the solution that the L = 1.963 path crosses a branch of solutions with L = 1.963/2 in
a pitchfork bifurcation. This intersection is the reason the continuation fails beyond this
point. We wish to stress here that the absence of localization in Dean flow is not limited
to the particular case of L = 1.963: we have seen this for larger values of L as well.
One possible reason for this could be that localized structures would have areas in the far
field that would be very similar to base state Dean flow, much like the case with solitary
steady states in circular Couette flow where the far field is essentially base state circular
Couette flow (see figure 57). However, since Dean flow is linearly unstable to stationary
disturbances of smaller wavelengths, there would be a tendency for vortex structures of
smaller wavelengths to form in the far field and prevent true localization.
As with circular Couette flow in figure 59 , we can plot an existence boundary for the
non-trivial Dean flow solutions. This curve is shown in figure 72. Note that there is a
discontinuity at L = 1.795. At this wavelength, we get a collision with the L = 1.795/2
branch in a pitchfork bifurcation as was the case with L = 1.963, but this time, the col-
lision occurs before the turning point is reached. Therefore, no turning point exists for
this value of L, resulting in the discontinuity. For reference, we also show two linear
stability curves of Dean flow. The curve on the left (the dashed line) is the usual linear
stability curve of Weθ,c versus L. The curve on the right (the dotted line) is the linear
164
13.0 14.0 15.0 16.0 17.0 18.0 19.0Weθ
0.0000
0.0020
0.0040
0.0060
||vr||
Figure 70: Bifurcation diagram for Dean flow at b = 1830, S = 1.2, ε = 0.2, andL = 1.963. We could not continue the branch beyond Weθ = 13.94, for reasons discussedin section 5.4.6. Radial velocity profiles corresponding to the points marked by the filledand open circles are shown in figure 71.
0 0.5 1 1.5 20
0.2
0.4
0.6
0.8
1
z
r
0 0.5 1 1.5 20
0.2
0.4
0.6
0.8
1
z
r
(a)
(b)
Figure 71: Density plots of the radial velocity at the points circled in figure 70. (a)Weθ = 15.486 (b) Weθ = 13.930.
165
stability curve with respect to disturbances which have two vortices, each of wavelength
L/2. This curve is trivially obtained from the regular linear stability diagram by multi-
plying the horizontal coordinate by a factor of two. At the intersection of the two curves,
denoted by Lt, is a codimension 2 point, where disturbances of wavelength Lt and Lt/2
bifurcate simultaneously. We would expect interactions between modes to play a signif-
icant role in the nonlinear behavior close to this wavelength. Note that the discontinuity
in the nonlinear existence curve occurs close to L = Lt.
0.5 1.5 2.5L
12.0
13.0
14.0
15.0
16.0
17.0
18.0
19.0
20.0
21.0
22.0
We θ,
c
Linear Stability (single vortex) Linear Stability (two vortices) Nonlinear Existence boundary
L t
Figure 72: Linear stability curves and existence boundaries for nonlinear solutions inDean flow at S = 1.2, ε = 0.2, and b = 1830. The filled triangle shows the data atL = 1.795 where the solution terminates via collision with the L/2 branch in a pitchforkbifurcation before the turning point.
5.5 Conclusions
In this chapter, we have computed stationary, nontrivial solutions in viscoelastic circu-
lar Couette flow by numerical continuation from stationary bifurcations in Couette-Dean
166
flow by using a FENE-P model. These stationary solutions show very strong spatial lo-
calization with a marked asymmetry between inflow and outflow, exist only for large
values of the polymer extensibility parameter and large wavelengths, show a hysteretic
character in the Weissenberg number, and are isolated from the base Couette flow branch
of solutions, all of which are characteristics shared by similar patterns (“diwhirls” or
“fl ame patterns” ) observed experimentally by Groisman and Steinberg (1997, 1998) and
Baumert and Muller (1999). Direct comparison shows that the streamlines and magni-
tudes of the radial velocity for the patterns we have computed are very similar to the
corresponding experimentally measured quantities. The only major difference between
the experimentally observed patterns and our solutions is in the degree of hysteresis they
exhibit: the experimentally observed diwhirls exist at Weissenberg numbers as low as 10,
whereas the computed diwhirls do not exist at Weissenberg numbers lower than about
23.3. We attribute this discrepancy to the approximate nature of the FENE-P model.
We have also performed computations with the FENE-CR model. These show that the
FENE-P and FENE-CR model produce different results in both the linear and nonlinear
regimes. In particular, the diwhirls are not predicted with FENE-CR model, at least for
the parameter values that we have examined. Note that the “exact” FENE molecular
model is better approximated by the FENE-P model than the FENE-CR model (Herrchen
and Ottinger, 1997). Our result demonstrates that the choice of constitutive equation is
critical in modeling nonlinear pattern formation in viscoelastic flows.
We have used the velocity and stress fields generated from our computations to pro-
pose a self sustaining mechanism for the diwhirls. The mechanism is related to the mech-
anism of instability in viscoelastic Dean flow and arises from a finite amplitude pertur-
bation giving rise to a locally parabolic profile of the azimuthal velocity near the outer
167
cylinder at the symmetry axis of the vortices. Strictly speaking, we find that the computed
diwhirls are unstable with respect to axisymmetric and non-axisymmetric perturbations.
The axisymmetric perturbations only have significant amplitude away from the core of
the diwhirl, whereas the non-axisymmetric perturbations can introduce modulations in
the core itself. Although unstable, the qualitative and quantitative similarities between
the computed structures and the experimentally observed diwhirls and flame patterns in-
dicates a direct link between the two.
Finally, we investigated whether structures similar to diwhirls or flame patterns could
form in Dean flow by tracking steady states bifurcating from the base flow in Weissenberg
number. Our computations demonstrate that long wavelength solutions in Dean flow tend
to show transitions to shorter wavelength structures rather than becoming localized.
Along with the solutions arising from the linear instability of the circular Couette flow
base state, we propose that the solitary solutions we have computed form building blocks
for spatiotemporal dynamics in the flow of viscoelastic liquids. We believe that com-
putations presented here are a first step towards understanding these complex nonlinear
dynamics.
168
Chapter 6
Conclusions and future work
In this thesis, we have performed analyses of some of the instabilities observed in the
flows of viscoelastic fluids. In chapter 2, we showed that multiplicity in the slip law does
not necessarily imply multiplicity in the flow curve. We performed simulations with the
UCM, PTT, Newtonian, and shear thinning models, all of which yielded similar results,
thus showing that the result did not depend on the particular slip model being used.
In chapter 3, we performed an analysis of the buckling instability in elongational flow
by modeling the polymer filament as a thin elastic membrane enclosing a passive fluid
that only serves to enforce the incompressibility constraint. We demonstrated that such
a set up will show instabilities that are similar to those seen in the elongational flow of
polymeric liquids. This simple model makes several assumptions that need to be checked
by direct numerical simulation of elongational flow. In particular, the numerical results
would need to show stress localization at the free surface, and regions of compressive
hoop stress near areas where the filament has buckled. It is only recently that such nu-
merical simulations have been attempted (Rasmussen and Hassager, 1999) for a UCM
169
liquid with no solvent viscosity. A direct comparison between our results and these sim-
ulations is not possible for reasons which we outline below.
Firstly, the computations in Rasmussen and Hassager (1999) were performed for a
non-axisymmetric base state. In particular, they added a small (0.5% of the original
radius) non-axisymmetric perturbation to the unstretched filament and then performed
the stretching. Thus, their work cannot be classified as a true stability analysis of the
stretching of an axisymmetric filament. Secondly, the stress contours for their simulations
at high Deborah numbers exhibit boundary layers in the component along the direction
of elongation, not the hoop stress. However, the high Deborah number results are only
shown at low values of the Hencky strain (2.5 for the case when the Deborah number
equals 1, and smaller strains at higher Deborah numbers), since numerical difficulties
prevented exploration of the regimes of larger Hencky strain.
Given the problems outlined in the preceding paragraph, we propose that future sim-
ulations take the following suggestions into account. In performing the stability analysis
of filament stretching, the primary difficulty is that the base state is time dependent. One
way of handling this problem is to perform what is called a “momentary stability” anal-
ysis, first described by Shen (1961) in the context of time dependent parallel shear flows,
and recently used for stability analysis of the spreading of surfactant coated films (Matar
and Troian, 1999). In this scheme, the governing equations are linearized about a base
state, just as in traditional linear stability analysis. A small disturbance is added to the
base flow at some reference time t0, and the linearized equations are then integrated si-
multaneously with the base state equations. The energy in the disturbance relative to its
initial energy at time t0 is then compared with the energy in the base state, relative to its
value at t0. If the disturbance grows faster than the base state, the flow is considered to
170
be unstable. In the context of filament stretching, this technique will ensure that the base
state being considered is purely axisymmetric.
The second problem that needs to be addressed is the numerical difficulty in achieving
large values of the Hencky strain at high Deborah numbers. One improvement that can
be made with relative ease is to use the FENE-P or FENE-CR model for the simulations.
Both are more realistic models for polymer solutions than the UCM equation, and are also
easier to simulate. Yet another improvement would be to take solvent viscosity into ac-
count. Besides improving the stability of the calculations, solvent viscosity would make
the simulations more representative of the experiments, since Spiegelberg and McKin-
ley (1998) used polymer solutions where the solvent and polymer contributions to the
viscosity were similar in magnitude.
In chapter 4, we demonstrated how an axial flow of small magnitude can signifi-
cantly stabilize the purely elastic instability in Dean flow. The next logical step would
be to perform experiments to verify that the stabilization mechanism works in practice.
As demonstrated by Joo and Shaqfeh (1994), a single concentric cylinder geometry can
be used to generate both Dean and circular Couette flow, simply by inserting a small
impenetrable block that spans the gap between the cylinders and extends across the en-
tire length. Moving both cylinders at equal angular velocities approximates Dean flow,
whereas moving them in opposite directions at equal angular speeds approximates circu-
lar Couette flow. By adjusting the relative angular velocities of the cylinders, it is possible
to get the entire range of flows between Dean and circular Couette. Axial flow can be
added either by imposing a pressure drop or by oscillating the inner cylinder. Thus, using
this apparatus, it would be possible to investigate the effect of axial flow on both Dean
and circular Couette flows. In a recent paper, Grillet et al. (1999) have described how
171
forward and reverse roll coating flows can be approximated in an eccentric cylinder ge-
ometry. It follows that if we make this modification in the set up described above, we can
determine the effect of axial forcing on free surface coating flows as well.
From the purely computational point of view, there are still some issues that remain
to be addressed. Firstly, as shown by Al-Mubaiyedh et al. (2000), non-isothermal effects
create a new mechanism of instability in viscoelastic circular Couette flow. The instabil-
ity associated with this mechanism sets in at a Weissenberg number that is an order of
magnitude lower than the one at which the purely elastic instability does. Their results
also showed that incorporation of additional relaxation times can have a large effect on
the Weissenberg number at which the instability sets in, although the mechanism of in-
stability is not affected. Given this fact, it seems logical to suppose that the mechanism of
suppression would work even when more than one relaxation time is taken into account.
This supposition must be checked with computations, as also what effect the axial flow
has on the non-isothermal mode of instability.
Finally, in chapter 5, we used numerical continuation from Couette-Dean flows to ac-
cess isolated, long wavelength, stationary, axisymmetric steady states in circular Couette
flow. We showed that these steady states are remarkably similar to the diwhirls (Grois-
man and Steinberg, 1997, 1998) and flame patterns (Baumert and Muller, 1999) observed
in experiments. Our computations demonstrated that such structures are absent in Dean
flow. There are several future directions that are worth exploring, within the context of the
FENE-P constitutive equation. Firstly, these solitary structures are unstable with respect
to axisymmetric and non-axisymmetric time dependent disturbances. While simulation
of fully three dimensional viscoelastic flow is far too computationally intensive, it would
be interesting to perform two dimensional axisymmetric time integrations of the unstable
172
steady states in order to gain insight into the interactions between the diwhirls and the
axisymmetric instability in circular Couette flow. Secondly, we have eased the computa-
tional burden by assuming that the structures that we are trying to compute have certain
symmetry properties and only modeled half the computational domain. This excludes
solutions that do not satisfy these properties. For instance, Baumert and Muller (1999)
have observed that the vortices in the flame patterns have a tendency to merge. Removing
the symmetry constraint may help us detect this phenomena in our simulations.
The main discrepancy between our calculations and the experimental results lies in
the fact that the solutions we compute lose existence at Weθ = 23.3, whereas diwhirls
and flame patterns have been observed at Weissenberg numbers as low as 10. There could
be several reasons for this lack of agreement. Firstly, we have not been able to access the
regime of large b due to limitations in our numerical scheme, and as discussed in chap-
ter 5, the point at which the diwhirls lose existence moves to lower values of Weθ as b
increases. Secondly, the FENE-P model that we have used for our simulations only incor-
porates a single relaxation time. Using the multimode Oldroyd-B, Giesekus, and Phan-
Thein and Tanner models, Al-Mubaiyedh et al. (2000) demonstrated that incorporating
multiple relaxation times can significantly change the point at which the linear instability
sets in. It is quite possible that incorporation of multiple relaxation times could have an
effect on the characteristics of the nonlinear solutions as well. Performing computations
with multimode constitutive equations such as the ones mentioned above can help us bet-
ter understand the effect of additional relaxation times. Another approach to including
multiple relaxation times is to use a FENE or Hookean chain, and perform microstructural
simulations using a technique like Operator Splitting Coefficient Matching (OSCM) (Jen-
drejack et al., 2000) or the Adaptive Lagrangian Particle Method (ALPM) (Gallez et al.,
173
2000) to get the stress. These techniques, however, are still in their infancy. ALPM
has been successfully used for a strong flow (a 1:4:1 contraction) for a FENE chain at
low Weissenberg numbers (Gallez et al., 2000), but requires enormous computational
power. OSCM appears to be less computationally intensive, but is still in the test phase,
with simulations being restricted to obtaining the stress field for a given velocity profile.
Non-isothermal effects could be another possible source of the quantitative discrepancy
between the experiments and our results. In chapter 5, we demonstrated that there is a
large difference between polyacrylamide based Boger fluid and polyisobutylene based
ones in terms of thermal effects. Since localized patterns are observed in both fluids, it is
unlikely that thermal effects play a role in the mechanism. However, they may still play
a qualitative role, for instance by changing the location of the turning point. Computa-
tions using non-isothermal models, such as the model developed by Crochet and Naghdi
(1969), or the more recent models based on phase space kinetic theory for dilute solutions
of dumbbells under non-isothermal conditions (Ottinger and Petrillo, 1996; Curtiss and
Bird, 1996), can elucidate the role of thermal effects.
In the computations described in chapter 5, the chief bottleneck was the development
of an effective preconditioning technique. The preconditioner that we developed took us
close the limits of what is achievable using a single computer. Further progress, such as
adding additional relaxation times, will require the use of more advanced numerical tech-
niques. A promising area in this regard is parallel computing. In performing the Newton
iterations, the computation of the Jacobian is an easily parallelizable step, since we can
simply split the domains between processors. However, the ILUT∗ preconditioner that
we have developed is not easy to parallelize, either in its computation or its application.
Therefore, it is critical to develop parallelizable preconditioners that are effective. In this
174
context, approximate inverse preconditioners (Benzi et al., 2000; Duff and van der Vorst,
1999; Benzi and Tuma, 1998; Chow and Saad, 1998) are worth investigating. The ILU
class of preconditioners are implicit preconditioners, in the sense that their application
requires the solution of a linear system of equations. On the other hand, approximate
inverse preconditioners are explicit, i.e., they involve the construction of a matrix that
approximates the inverse of the original matrix. Both the construction and application
involve the action of a matrix on a vector, and these are steps that are easily paralleliz-
able. Recently, Leriche and Phillips (2000) have discussed preconditioners for spectral
element solutions to viscoelastic problems that are based on linear finite element precon-
ditioning. The linear finite element systems are solved using Schur complement methods
or overlapped Schwartz methods, both of which are parallelizable. This last precondi-
tioner, however, is used in the context of time integration, where operator splitting is
used to update the convection terms in the constitutive and momentum equations explic-
itly. How effective it will be for a steady state continuation scheme where the momentum,
continuity, and constitutive equations are coupled remains to be determined.
175
Appendix A
Introduction to finite elasticity
This section of the appendix gives an overview of a few of the basic equations of finite
elasticity. This material is condensed from Corneliussen and Shield (1961).
A.1 The basic equations of finite elasticity
Consider a homogeneous elastic body which is isotropic in its undeformed state. The
coordinates of a typical point in the undeformed state in a fixed Cartesian coordinate
system x are given by (x1, x2, x3). Let (v1, v2, v3) denote the coordinates of this point in
an arbitrary curvilinear coordinate system. We assume that a transformation rule of the
form
xi = xi(v1, v2, v3) (156)
exists. Let r denote the position vector from the origin of the system x to a point of the
body. The covariant base vectors of the system v of the undeformed body are given by
g(k) =∂r
∂vk= r,k, (157)
176
where in this notation, a comma followed by k denotes differentiation with respect to vk.
In future, we will drop the parentheses on the subscripts of base vectors for convenience.
An element of length ds is given by
ds2 = dxkdxk = gijdvidvj, (158)
where gij are the covariant components of the metric tensor of the system v. The con-
travariant components of the metric tensor are denoted by gij and are the components of
the inverse of the matrix gij. As before, g denotes the determinant of the matrix gij .
At some future time t,the body has been deformed, and the coordinates of a typical
point in the system x are (X1, X2, X3). The curvilinear system moves with the body (i.e.
we can think of the vi as particle labels), and we have
Xk = Xk(v1, v2, v3, t). (159)
Thus, at time t, the covariant base vectors of the system v are
Gk = R,k. (160)
The element of length dS is given by
dS2 = dXkdXk = Gijdvidvj, (161)
where the Gij are the covariant components of the metric tensor of of the system v at time
t. Gij denote the contravariant components of the metric tensor and G the determinant of
Gij .
For some materials, the stress work acts to change a state function called the strain
energy. If stress work is done, the strain energy changes by the same amount, and this
strain energy is recoverable as stress work. The strain energy function W gives the change
177
in strain energy from its value at the reference state, per unit volume of the reference state.
It is a function of the three strain invariants I1, I2 and I3,
W = W (I1, I2, I3),
with the invariants being given by
I1 = grsGrs
I2 = grsGrsI3
I3 = G/g. (162)
The contravariant components of the stress tensor are now given by
τ ij = Φ gij +ΨDij + P Gij, (163)
where
Φ = 2√
I3∂W
∂I1,
Ψ = 2√
I3∂W
∂I2,
P = 2√
I3∂W
∂I3,
Dij = I1 gij − gir gjsGrs. (164)
For an incompressible material I3 = 1, and
W = W (I1, I2).
For certain rubber-like materials, the strain energy function takes the Mooney form
W = C1(I1 − 2) + C2(I2 − 3), (165)
178
where C1 and C2 are positive constants. A special case of the Mooney material is the
neo-Hookean material, whose strain energy function takes the form
W = C1(I1 − 3). (166)
In the absence of body forces, the equations of motion are given by
ρ f j = ∇iτij , (167)
where the f j are the components of the acceleration vector in the system v.
A.1.1 The equations of membrane elasticity
In this section, we look at the simplifications that result when the equations of finite
elasticity are applied to a thin membrane. Let (v1, v2) be the coordinates of a general
curvilinear coordinate system defined on a surface in space. Suppose that a is the position
vector from the origin of a Cartesian coordinate system to a point on the surface, the base
vectors of this system are
aα =∂a
∂vα= a,α . (168)
We now define a unit vector a3(v1, v2) perpendicular to the surface at each point, in the
direction a1 × a2. The distance from the point (v1, v2) on the surface to a point on the
normal to the surface at that point is denoted by v3, and is considered positive in the
direction of a3. If we now consider a point on the middle surface of a shell of thickness
h with v3 = 0, the position vector can be written as
r = a(v1, v2) + v3a3(v1, v2),
179
where
−1
2h(v1, v2) ≤ v3 ≤
1
2h(v1, v2). (169)
Thus, in the undeformed (or reference) state, we get
gαβ = aαβ, gα3 = 0, g33 = 1,
g = a,
gαβ = aαβ, gα3 = 0 , g33 = 1,
where aαβ are the covariant components of the surface metric tensor
aαβ = aα · aβ, (170)
a = |aαβ|, and aαβ is the inverse of the matrix whose elements are aαβ .
Now consider the surface at the later time t, when it has been deformed. Let the
position vectors of the deformed surface be given by A(v1, v2). The base vectors of the
deformed surface are given by
Aα = A,α . (171)
As before, we define at each point on the surface, a unit vector A3 perpendicular to
the surface at that point and in the direction A1 × A2, with the same convention for
positivity as before. The unit normal vector N is thus given by (A1 ×A2)/|A1 ×A2|.
The covariant components bαβ of the curvature tensor B are then given by
bαβ =N ·Aα,β =N ·A,αβ. (172)
180
For a shell composed of an isotropic material, whose principal radii of curvature of
the deformed middle surface are large compared to the thickness, and whose thickness
varies slowly with v1 and v2, it can be shown that the position vector to a point on the
deformed membrane can be written as
R = A(v1, v2) + λ3(v1, v2) v3A3(v
1, v2), (173)
where λ3 is positive and is the extension ratio in the direction of the normal to the middle
surface. The components of the metric tensor are now given as
Gαβ = Aαβ, Gα3 = 0, G33 = λ23,
G = λ23A,
Gαβ = Aαβ, Gα3 = 0, G33 =1
λ23,
where Aαβ and Aαβ are respectively the covariant and contravariant components of the
surface metric tensor and A = |Aαβ|. The strain invariants are given as
I1 = aαβAαβ + λ23,
I3 = λ23A/a, (174)
I2 = I3
(aαβA
αβ +1
λ23
). (175)
For an incompressible material, I3 = 1, which implies that W =W (I1, I2) and that
λ3 =√(a/A). (176)
This in turn implies that the remaining two invariants are given by
I1 = aαβAαβ + a/A,
I2 = aαβAαβ +A/a. (177)
181
The contravariant components of the stress tensor in the coordinate system v is then given
by
nαβ =√(a/A3)h
((Aaαβ − aAαβ
)Φ +
(ADαβ − aAαβ D33
)Ψ), (178)
where h is the thickness of the membrane, and
Dαβ =a
Aaαβ +
(aαβ aµν − aαµ aβν
)Aµν
D33 = aµν Aµν .
Let P = pjaj be the external force acting on the membrane. Then the equations of
motion for the membrane are given by
∇α nαβ + pβ = h√
a/Aρ0 fβ,
bαβ nαβ + p3 = h√
a/Aρ0f3, (179)
where ρ0 is the density of the undeformed membrane.
182
Appendix B
Base state solutions and matrix
components in Dean flow
We present the expressions for the base state velocities and stresses, as well as the nonzero
entries of the matrices L and A for the axisymmetric case.
B.1 Base state solutions
The following are the steady state velocities and polymer stresses of an Oldroyd-B fluid.
DAP flow
vr = 0 (180)
vθ = r(1− r) (181)
vz =Wez
Wpr(1− r)ε1/2 (182)
τrr = 0 (183)
183
τrθ =Wp (1− 2r)
ε1/2(184)
τrz = Wez(1− 2r) (185)
τθθ =2(1− 2r)2Wp2
ε(186)
τθz =2(1− 2r)2WezWp
ε1/2(187)
τzz = 2(1− 2r)2We2z (188)
DMAC flow: The equations correpond to the regime where ω = O(ε1/2) and t =
O(ε−1/2)with all other variables scaled as in section 4.3. We define ω1 = ε−1/2ω = O(1)
and t1 = ε1/2t = O(1). The steady axial flow (DAC) equations are obtained by setting
ω1 = 0.
vr = 0 (189)
vθ = (r − r2) (190)
vz =Wez(1− r)ε1/2
Wpcos(ω1t) (191)
τrr = 0 (192)
τrθ =Wp (1− 2 r)
ε1/2(193)
τrz = −Wezcos(ω1t1) +Wpω1 sin(ω1t1)
(1 + ω21Wp2)(194)
τθθ =2Wp2(1− 2r)2
ε(195)
τθz =WezWp(−1 + 2r)(2 cos(ω1t1) +Wpω1(3 +Wp2ω21) sin(ω1t1))
ε1/2(1 +Wp2ω12)2(196)
τzz =−We2z
(1 + 4Wp2ω21)(1 +Wp2ω21)(−1− 4ω21Wp2 − cos(2ω1t1) (197)
+2ω21Wp2 cos(2ω1t1)− 3ω1Wp sin(2ω1t1))
184
B.2 Disturbance equations
The non-zero entries of WpA(t) in equation (105) for the DMAC cases are given below.
The operator D represents the derivative w.r.t. the gap coordinate r. The matrixE(t) has
−1 on the diagonal corresponding to the constitutive equations and zeros elsewhere. The
EVSS formulation (not shown here) is derived by substituting
τrr = Σrr + 2WpD vr,
τrθ = Σrθ +WpD vθ,
τrz = Σrz + i αWp vr +WpD vz,
τθθ = Σθθ,
τθz = Σθz +Wp vθ,
τzz = Σzz + 2 i αWp vz,
where the components of Σ represent the elastic part of the extra stress tensor τ . The
equations for DAC flow are recovered by setting ω1 = 0.
WpA1,1 = 1 + iWez α (1− r) cos(ω1t1)
WpA1,7 =2 iWez Wpα (cos(ω1t1) +Wpω1 sin(ω1t1))
(1 +Wp2 ω21)− 2WpD
185
WpA2,1 = Wp (−1 + 2r)
WpA2,2 = 1 + i αWez (1− r) cos(ω1t1)
WpA2,7 = −Wp22 (1 +Wp2 ω21)
2 + 2 iWez α (−1 + 2 r) cos(ω1t1)
(1 +Wp2 ω21)2
−Wp2iWez Wpα(−1 + 2 r)ω1 (3 +Wp2 ω21) sin(ω1t1)
(1 +Wp2 ω21)2
+
Wp2 (−1 + 2r)D
WpA2,8 = iαWez Wpcos(ω1t1) +Wpω1 sin(ω1t1)
(1 +Wp2 ω21)−WpD
WpA3,1 = Wez cos(ω1t1)
WpA3,3 = 1 + iWez α (1− r) cos(ω1t1)
WpA3,7 =−i αWp
(1 + 4Wp2 ω21)(1 +Wp2 ω21)
((1 +We2z +Wp2 ω21)(1 + 4Wp2 ω21)
−We2z (−1 + 2Wp2 ω21) cos(2ω1t1))
−i αWp3We2z Wpω1 sin(2ω1t1)
(1 + 4Wp2 ω21)(1 + ω21Wp2)
+Wez Wp (cos(ω1t1) +Wpω1 sin(ω1t1))D
(1 +Wp2 ω21)
WpA3,9 = −WpD
WpA4,2 = 2Wp (−1 + 2 r)
WpA4,4 = 1 + i αWez (1− r) cos(ω1 t1)
WpA4,7 = 8Wp3 (−1 + 2 r)
WpA4,8 =−2 iWez Wp2 α
(1 +Wp2ω21)((−1 + 2 r)(2 cos(ω1t1)
+Wpω1 (3 +Wp2 ω21) sin(ω1t1)))+ 2Wp2 (−1 + 2 r)D
186
WpA5,2 = Wez cos(ω1t1)
WpA5,3 = Wp (−1 + 2 r)
WpA5,5 = 1 + i αWez (1− r) cos(ω1t1)
WpA5,7 = 2Wez Wp22 cos(ω1t1) +Wpω1(3 +Wp2 ω21) sin(ω1t1)
(1 +Wp2 ω21)
WpA5,8 =i αWp
(1 + 4Wp2 ω21)(1 +Wp2 ω21)((1 +We2z +Wp2 ω21)(1 + 4Wp2 ω21)
−We2z (−1 + 2Wp2 ω21) cos(2ω1t1) + 3We2z Wpω1 sin(2ω1 t1)) +
Wez Wp (cos(ω1t1) +Wpω1 sin(ω1t1))
(1 +Wp2 ω21)D
WpA5,9 = iWez Wp2 α (−1 + 2 r)2 cos(ω1t1) +Wpω1 (3 +Wp2 ω21) sin(t1 ω1)
(1 +Wp2 ω21)2
+Wp2 (−1 + 2 r)D
WpA6,3 = 2Wez cos(ω1t1)
WpA6,6 = 1 + iWez α(1− r) cos(ω1t1)
WpA6,9 =−2 iWpα
(1 + 4Wp2 ω21)(1 + 4Wp2 ω21)
((1 +We2z +Wp2 ω21)(1 + 4Wp2 ω21)
−We2z (−1 + 2Wp2 ω21) cos(2 ω1t1))
−2 iWpα3We2z ω1Wp sin(2ω1t1)
(1 + 4Wp2 ω21)(1 + 4Wp2 ω21)
+2Wez Wpcos(ω1t1) +Wpω1 sin(2ω1t1)
(1 +Wp2 ω21)D
WpA7,1 = D WpA7,3 = i α WpA7,4 = −1 WpA7,7 = S Wp(−α2 +D2
)WpA7,10 = −D
WpA8,2 = D WpA8,5 = i α WpA8,8 = S Wp(−α2 +D2
)
187
WpA9,3 = D WpA9,6 = i α WpA9,9 = S Wp(−α2 +D2
)WpA9,10 = −α
WpA10,7 = D WpA10,9 = i α
DAP flow: Below are given the nonzero components of the linear operator WpL for DAP
flow.The components of the momentum and continuity equations are the same as for the
DMAC case, as are the terms in the matrix E.
WpL1,1 = 1 +Wez i α r (1− r)
WpL1,7 = 2 i, Wez Wpα(−1 + 2 r)− 2WpD
WpL2,1 = Wp (−1 + 2 r)
WpL2,2 = WpL1,1
WpL2,7 = −2Wp2 − 2 iWez Wp2 α (−1 + 2 r)2 +Wp2 (−1 + 2 r)D
WpL2,8 = iWez Wpα (−1 + 2 r)− WpD
WpL3,1 = Wez (−1 + 2 r)
WpL3,3 = WpL1,1
WpL3,7 = −2WezWp+ 8 i rWe2z Wpα(1− r)− i αWp (1 + 2We2z)
+Wez Wp (−1 + 2 r)D
WpL3,9 = iWez α (−1 + 2 r)D
WpL4,2 = 2Wp (−1 + 2 r)
WpL4,4 = WpL1,1
188
WpL4,7 = 8Wp3(−1 + 2 r)
WpL4,8 = −4 iWez Wp2 α(1− 2 r)2 + 2Wp2(−1 + 2 r)D
WpL5,2 = Wez (−1 + 2 r)
WpL5,3 = Wp (−1 + 2 r)
WpL5,5 = WpL1,1
WpL5,7 = 8Wez Wp2(−1 + 2 r)
WpL5,8 = −i αWp (1 + 2We2z) + 8 i rWe2z Wp (1− r) +Wez(−1 + 2 r)D
WpL5,9 = −2 iWez Wp2 α(1− 2 r)2 +Wp2(−1 + 2 r)D
WpL6,3 = 2Wez (−1 + 2 r)
WpL6,6 = WpL1,1
WpL6,7 = 8We2z Wp (−1 + 2 r)
WpL6,9 = −2 i αWp (1 + 2We2z (1− 2 r)2) + 2Wez Wp (−1 + 2 r)D
189
Appendix C
Velocity and Stress scalings for
Couette-Dean flow
We found base state solutions to the FENE-P equation starting from the analytical so-
lutions for the Oldroyd-B model in Couette-Dean flow. We also used these solutions to
obtain scalings for the velocity and time. For the Oldroyd-B model, the azimuthal veloc-
ity in dimensional form can be written as the sum of a Couette contribution, v∗θ,c, and a
Dean contribution v∗θ,d, where
v∗θ,c = ΩR2
((1− ε)
ε(1− ε+ rε)(2− ε)−(1− ε)(1− ε+ rε)
ε(2− ε)
), (198)
and
v∗θ,d =−Kθε
2R22ηt
((1− ε)2 log(1− ε)
ε3(1− ε+ r ε) (2− ε)−(1− ε+ rε)
ε3 (2− ε)
((1− ε)2 log(1− ε)
+ε(2− ε) log(1− ε+ rε))) , (199)
190
with r being the radial coordinate scaled by the gap width, so that r = 0 is the inner
cylinder and r = 0 is the outer cylinder. Adding these two equations gives
v∗θ =−Kθε
2R22ηt
fd +ΩR2(1− ε)fc, (200)
where
fc =(1− ε)
ε(1− ε+ rε)(2− ε)−(1− ε)(1− ε+ rε)
ε(2− ε), (201)
and
fd =(1− ε)2 log(1− ε)
ε3(1− ε+ r ε) (2− ε)−(1− ε+ rε)
ε3 (2− ε)
((1− ε)2 log(1− ε)
+ ε(2− ε) log(1− ε+ rε)) . (202)
If we now choose
Vd∗ =−Kθε
2R22ηt
(203)
to be a characteristic velocity for Dean flow and
Vc∗ = ΩR2 (1− ε), (204)
which is the velocity at the inner cylinder, to be a characteristic velocity for circular
Couette flow, can define parameter δ, which measures the relative contribution of the
azimuthal pressure gradient to the total flow as
δ =Vd∗
Vd∗ + Vc
∗ . (205)
We will only consider the case where Vd∗ and Vc
∗ have the same sign, i.e., where the
pressure gradient and inner cylinder rotation act to drive the flow in the same direction.
191
This means that δ = 0 is circular Couette flow and δ = 1 is Dean flow. With this definition
of δ, we can write the azimuthal velocity as
v∗θ = (V∗c + V ∗d )(δ fd + (1− δ) fc). (206)
We choose the velocity scale to be the magnitude of the shear rate at the outer cylinder
times the length scale. Since Vd∗ and Vc
∗ are both taken to be positive, the shear rate at
the outer cylinder is negative, so that the magnitude of the shear rate is −γ2, where γ2 is
the shear rate at the outer cylinder. This gives the velocity scale as
v∗ = (V ∗c + V ∗d )v, (207)
where
v =
(δ2 (1− ε)2 log(1− ε) + ε (2− ε)
ε2(2− ε)+ (1− δ)
2 (1− ε)
2− ε
). (208)
Therefore, the azimuthal velocity can be written in dimensionless form as
vθ =δ fd + (1− δ) fc
v. (209)
The Weissenberg number is the product of the characteristic shear rate, εR2 v∗, and
the relaxation time. The non-zero components of the polymer stress tensor and the radial
pressure gradient can be computed from the relations
τrθ = Weθ
(∂vθ∂r−
vθr − 1 + 1/ε
), (210)
τθθ = −2Weθτrθ
(vθ
r − 1 + 1/ε−
∂vθ
∂r
), (211)
∂p
∂r= −
τθθr − 1 + 1/ε
. (212)
Having calculated the stresses, the components of the 〈QQ〉 for the Oldroyd-B model
can be obtained using the Kramers equation as
〈QQ〉 = τ + I. (213)
192
We use these components as the initial approximations for the FENE-P model.
193
Appendix D
Time Integration of viscoelastic
flows
In chapter 5, we have described steady state solution tracking for viscoelastic flows. Here,
we will describe time integration of viscoelastic flows where inertia is important. This has
applications in the simulation of the effect of polymers on turbulence, and also in studying
mixing flows. Our goal here is to gain expertise in the use of spectral methods in time
integration of viscoelastic flows, and for that reason, we will work with a model problem.
We consider here the classical problem of Rayleigh-Benard convection in which a layer of
fluid of infinite extent in both horizontal directions is heated from below. For simplicity,
we assume free slip conditions at both vertical boundaries. There are several convection
patterns possible with these boundary conditions, but we will restrict ourselves to the
two dimensional rolls whose axes are aligned with one of the horizontal directions. We
choose this direction to be the y axis, the horizontal direction perpendicular to it to be the
x axis, and the vertical direction to be the z axis. Thus, the flow is in the x−z plane. The
194
streamfunction for this flow is given by Chandrasekhar (1961) as
ψ0(x, z) =A
ksin(πz) sin(kx), (214)
where A is the maximum velocity of the flow, k = 2πL
, where L is the wavelength of
the cell. In equation 214, length has been nondimensionalized by the depth of the layer
so that the upper boundary is at z = 1 and the lower boundary is at z = 0. Figure 73
shows a contour plot of equation 214. The roll cells are themselves susceptible to a two
dimensional, time dependent instability as the temperature drop is increased. Solomon
and Gollub (1988) modeled the stream function of the flow that develops after the onset
of the instability as
ψ(x, z, t) =A
ksin(πz) sin(kx) + εA cos(ωt) cos(kx) sin(πz), (215)
where ω is the frequency of the instability, and ε = (Ra − Rac)1/2, where Ra is the
Rayleigh number, and Rac is the value of Ra at the onset of the instability. This time-
periodic flow pattern creates regions within the flow cell where passive scalars have
chaotic trajectories (see figure 74).
0 1 20
1
x
z
Figure 73: Streamlines for equation 214 with A = 1 and k = π.
195
0.0 0.5 1.0 1.5 2.0x
0.0
0.2
0.4
0.6
0.8
1.0
z
Figure 74: Poincare map of the trajectory of a point in the flow field given by equa-tion 215. The parameters chosen were k = π, ω = 2π, and ε = 0.25.
D.1 Governing equations
In the Rayleigh-Benard problem, the convection patterns are driven by heating the fluid
from below. Our interest lies only in obtaining the velocity profiles corresponding to
equation 215, and we do this by the computationally simpler approach of adding forcing
terms to the momentum equations. For Newtonian fluids, the governing equations are the
Navier-Stokes equations and the continuity equation. In dimensionless form, they are
Re
(∂u
∂t+ u
∂u
∂x+ w
∂u
∂z
)= −
∂p
∂x+
∂2u
∂x2+
∂2u
∂z2+ fx, (216)
Re
(∂w
∂t+ u
∂w
∂x+ w
∂w
∂z
)= −
∂p
∂z+
∂2w
∂x2+
∂2w
∂z2+ fz, (217)
∂u
∂x+
∂w
∂z= 0. (218)
Here, u and w are respectively the x and z components of the velocity, and p is the
pressure. The velocity has been scaled by A, distance by the height of the layer, and
196
time by the ratio of the distance scale to the velocity scale. The forcings fx and fz can
be obtained by requiring that the governing equations be satisfied by the desired velocity
field and zero pressure. We calculate them to be
fx =1
4 k(−(π (−4 ε k π Re cos(ω t) cos2(k x) +
4 ε k cos(k x) cos(π z) ((k2 + π2) cos(ω t)
−ω Re sin(ω t))+sin(k x) (2 (−2+ ε2 k2) πRe cos(k x)+ ε2 k2 π Re cos(2ω t−k x)+
ε2 k2 π Re cos(2ω t+ k x) + 4 k2 cos(π z) + 4 π2 cos(π z)−
2 ε k π Re sin(ω t− k x) + 2 ε k π Re sin(ω t+ k x))))) (219)
fz = sin(π z) ((k2 + π2) cos(k x) + πRe cos2(k x) cos(π z) (1 + k2 ε2 cos2(t ω))
+ sin(k x) (−(k (k2 + π2) ε cos(t ω)) + k2 πRe ε2 cos(π z) cos2(t ω) sin(k x) +
Re (π cos(π z) sin(k x) + k ε ω sin(t ω)))) (220)
When polymer is added to the system, the momentum conservation equations are
modified to incorporate the contribution of the polymer stresses:
Re
(∂u
∂t+ u
∂u
∂x+ w
∂u
∂z
)= −
∂p
∂x+
∂2u
∂x2+
∂2u
∂z2+ fx (221)
+1
WeS
(∂τp,xx∂x
+∂τp,xz∂z
),
Re
(∂w
∂t+ u
∂w
∂x+ w
∂w
∂z
)= −
∂p
∂z+
∂2w
∂x2+
∂2w
∂z2+ fz (222)
+1
WeS
(∂τp,xz
∂z+
∂τp,zz
∂z
),
where τp,xx, τp,xz, and τp,zz are the three components of the polymer stress τp, We =
λA/d is the Weissenberg number, λ is the relaxation time of the polymer, and S is the
ratio of the solvent to polymer viscosity. We choose the FENE-P constitutive equation
197
for the polymer with the same scalings for the conformation tensor 〈QQ〉 as in chapter
5. Once again, the stress tensor is calculated from the Kramers relation. We use periodic
boundary conditions in the x direction. The z direction is periodic as well, except that
here, the period length is 2. This means that the solutions exhibit symmetries in this
direction. In particular, u, p, 〈QQ〉xx, and 〈QQ〉yz can be represented in terms of even
functions, while w and 〈QQ〉xz can be represented in terms of odd functions.
D.2 Numerical Method
Periodic boundary conditions naturally suggest the use of a Fourier series representation.
Here, we use Fourier series in both the x and z directions. We take advantage of the
symmetry in the z direction by representing the even functions as a Fourier cosine series
in the vertical direction, and the odd functions as a Fourier sine series. The primary
representation for all variables is in Fourier space. We compute Fourier expansions of
nonlinear terms by transforming the variables involved into physical space, computing
the nonlinear terms, and transforming back into Fourier space. This can be done very
efficiently using a the Fast Fourier Transform (FFT). A one dimensional FFT on an N
point grid requires O(N log2(N)) operations, where N is the number of grid points.
Transforming data on a two dimensional grid with Nx points in the x direction and Nz
points in the z direction requires Nz transformations of one dimensional data that is Nx
points long, followed by Nx transformations involving data that is Nz points long. Thus,
the work required for transforming two dimensional data is a semi-linear function of
the number of grid points (NxNz). We use the highly efficient public domain package,
FFTW (Frigo and Johnson, 1999), to perform these transformations. Equation 215 gives
198
a velocity field that has stagnation points at x = L/2, z = 0 and x = L/2, z = 1 (i.e., the
midpoints of the upper and lower walls). Polymer molecules are highly stretched in the
neighborhood of these points, which gives rise to sharp stress gradients and requires the
use of high resolution.
We use a time splitting scheme to perform the integration. For purposes of numerical
stability, we found it necessary to add an isotropic artificial diffusion term to the con-
stitutive equations. Similar observations have been made in the context of simulating
high Reynolds number Poiseuille flow of a polymer solution in a channel (Sureshkumar
and Beris, 1995) and turbulent flow in a channel (Sureshkumar et al., 1997). Thus, the
modified constitutive equation is written
〈QQ〉(1) =1
We
(〈QQ〉
1− tr(〈QQ〉)/b− I
)+D∇2〈QQ〉, (223)
where D is the diffusion coefficient. The diffusion coefficient is chosen to have the form
D = a(∆h)2
t∗, (224)
where ∆h is the mesh size, t∗ is a time scale, and a is a constant. Thus, in the limit
∆h → 0 (i.e., an infinitesimally small mesh size) with a fixed, D = 0, so that the
magnitude of the diffusion becomes smaller as the mesh is refined. Since the flow is
periodic in both x and z, we do not need to specify additional boundary conditions for
equation 223.
We use a splitting scheme for the time integration. The first stage involves updating
〈QQ〉 in two steps. The first of these involves an explicit update of the convection and
nonlinear terms using an Adams-Bashforth scheme. Denoting the vector (u, w) by v, this
199
can be written as
〈QQ〉∗ = 〈QQ〉(n) +∆ t
1∑j=0
βj[−v · ∇〈QQ〉+ 〈QQ〉 · ∇v + (∇v)t · 〈QQ〉
+1
We
(〈QQ〉
1− tr(〈QQ〉)/b− I
)](n−j), (225)
where β0 = 3/2 and β1 = −1/2, and the superscripts denote values at the corresponding
timestep. The second step involves a semi-implicit update to include the diffusion term.
We do this using a Crank-Nicholson method:
〈QQ〉(n+1) − 〈QQ〉∗
∆t=
D
2(∇2〈QQ〉(n+1) +∇2〈QQ〉(n)). (226)
Note that although equation 226 is an implicit equation for 〈QQ〉(n+1), it is trivially
solvable in Fourier space, since the individual Fourier modes decouple.
The next stage involves the calculation of v(n+1). This is done in three steps. The first
step involves the update of the convection and stress terms. We use an Adams-Bashforth
scheme for the convection terms and Crank-Nicholson for the stress gradients:
v − v(n)
∆t= −
1∑j=0
βj(v · ∇v)(n−j) +
1
2WeSRe
(∇ · τ (n+1)p +∇ · τ (n)p
). (227)
The second step is a pressure correction, given by the equation
ˆv − v
∆t= −
1
Re∇p(n+1). (228)
This equation can be used to obtain the pressure by requiring that ˆv be divergence free,
which yields the Poisson equation
∇2p =∆t
Re∇ · v. (229)
In flows which have non-periodic directions, the boundary conditions for the pressure
equation must be chosen with care in order to preserve the overall temporal accuracy of
200
the integration scheme, and Karniadakis et al. (1991) present a discussion on high-order
pressure boundary conditions for such flows. In our case, however, the only indetermi-
nacy is in choosing the coefficient for the pressure mode with zero wavenumbers in both
directions. What we choose this coefficient to be is completely irrelevant, since this has
no effect on the pressure gradient, which is the quantity that appears in the momentum
equation.
The final step in the integration scheme is to account for the viscous terms. We use a
Crank-Nicholson scheme for these terms. The update is given by
v(n+1) − ˆv
∆t=
1
2Re(∇2v(n+1) +∇2v(n)). (230)
As in the case of the diffusion correction in equation 226, equations 229 and 230 are
trivially solvable in Fourier space.
The scheme as outlined above is second order in time. The time step is chosen based
on the requirements of accuracy and stability. The stability of the scheme is governed
by the Courant-Freidrichs-Levy (CFL) condition for the two explicit steps (equation 225
and 227). We choose the value of t∗ in equation 224 to be the largest timestep permitted
by the CFL condition. This essentially makes the diffusivity independent of the actual
choice of timestep for a given spatial resolution.
We can get a rough estimate of the total work required per time step by keeping track
of the number of FFTs that need to be performed. Given the initial vectors v and 〈QQ〉
in Fourier space, we need to perform a total of 15 two dimensional transformations to
get their values and those of their derivatives in two directions in physical space, in order
to compute the nonlinear terms. Once the nonlinear terms are computed, a total of 5
transformations need to be performed to get their values in Fourier space, yielding a total
of 20 transforms per timestep. In the absence of artificial diffusion in the constitutive
201
equation, this number is reduced to 17, since in this case, the update in equation 225
gives the physical space components directly.
D.3 Results and Discussion
We restrict our discussion here to the necessity of using artificial diffusion in integrating
the constitutive equation. For our test case, we used a resolution of 144 points in the
horizontal direction and 72 points in the vertical direction and started our integration with
initial conditions being the Newtonian velocity field and zero values for the components
of the conformation tensor. We set ∆t to 10−4, which is well below the upper limit set
by the CFL condition. The Weissenberg and Reynolds numbers were both chosen to be
1, and the results are shown in figure 75, for various values of the diffusivity multiplier,
a. The quantity plotted is the value of 〈QQ〉xx at x = L/4 and z = 0.5. We see from the
figure that the integration blows up if artificial diffusion is not used, or if the diffusion
constant a is much smaller than 10−3. At this value of a, the diffusion coefficient, D
was calculated to 1.640 × 10−5. The integration for a = 10−3 was stable indefinitely.
Figure 75 also shows that the quantitative effect of artificial diffusion is quite small: the
curves for the integrations with and without artificial diffusion virtually coincide before
the numerical instability sets in.
To determine the effect of increasing mesh resolution, we refined the mesh to a grid
of 256 points in the horizontal direction and 128 points in the vertical direction. We then
integrated the governing equations for the same parameter values and time step as the
coarser mesh, with a = 10−4. At this value of a, the diffusion coefficient D has a value
of 9.223× 10−7 for the fine mesh, and 1.640× 10−6 for the coarse mesh. The results are
202
shown in figure 76. Although both integrations blow up due to numerical instability, it is
clear that the fine mesh with the smaller diffusion coefficient is stable for a longer period
of time than the coarse mesh. This shows that refining the mesh improves numerical
stability even if the diffusion coefficient is lowered.
0.0 5.0 10.0 15.0 20.0t
−1.0
0.0
1.0
2.0
3.0
4.0
5.0
<Q
Q>
xx
*
a=10−3
a=10−4
a=10−5
a=10−6
a=0
Figure 75: Illustration of the effect of adding a diffusion term to the constitutive equationon a mesh of 144×72. The parameter values used were We = 1, Re = 1, ε = 0.2, k = π,ω = 2π, b = 10, and S = 6.67.
A possible question to ask is if the numerical instability would still appear if the con-
stitutive and momentum equations are decoupled. To answer this question, we integrated
the constitutive equation using the steps outlined in equations 225 and 226, with the ve-
locity field supplied by equation 215, with the same parameters as for the coupled case.
The results, shown in figure 77 for a = 0 and a = 10−3, show that the instability is
indeed associated with the integration of the constitutive equation, since the case with
a = 0 blows up. In fact, comparing figures 75 and 77 shows that the instability is delayed
203
0.0 5.0 10.0 15.0t
−1.0
0.0
1.0
2.0
3.0
4.0
5.0
<Q
Q>
xx
*
144 x 72 256 x 128
Figure 76: Effect of mesh refinement on numerical stability. Runs with both meshes wereperformed with a = 10−4 and the same parameter values as in figure 75.
0.0 5.0 10.0 15.0 20.0t
0.0
1.0
2.0
3.0
<Q
Q>
xx
*
a=10−3
a=0
Figure 77: Effect of artificial diffusion when equation 223 in integrated in a known ve-locity field. The parameters and mesh are identical to those in figure 75. Note the earlierblow up when compared to the coupled case.
204
in the coupled case. This may possibly be due to the stabilizing influence of the viscous
terms in the momentum equations.
205
Appendix E
Branch tracing in three dimensional
plane Couette flow
E.1 Introduction and Formulation
Analysis of turbulence in plane Couette flow is complicated by the absence of a linear
instability, i.e., plane Couette flow is stable to infinitesimal perturbations at all Reynolds
numbers. This means that the transition to turbulence must occur as a finite amplitude
effect. Homotopy (see the description in section 5.1) offers one approach to access non-
trivial steady states in plane Couette flow that may have relevance to coherent structures
seen in turbulence. This approach was taken by Nagata (Nagata, 1988, 1990, 1997) who
tracked finite amplitude solutions from Taylor-Couette flow to plane Couette flow. A
slightly different approach was taken by Waleffe (Waleffe, 1995, 1997), who started by
proposing a self sustaining mechanism to maintain turbulence, which we describe below
in some detail.
206
Consider the schematic diagram of plane Couette flow shown in figure 2. We shall
call the direction of flow (the x direction) as the streamwise direction, the direction of ve-
locity change (the y direction) as the wall-normal direction, and the neutral direction (the
z direction) as the spanwise direction. The Waleffe mechanism consists of three steps. In
the first step, rolls with their axes in the streamwise direction redistribute the streamwise
momentum to create spanwise fluctuations in the streamwise velocity. The presence of in-
flections due to this spanwise fluctuation causes a wake-like instability (Drazin and Reid,
1981), which takes the form of vortices that have their axes oriented in the wall-normal
direction. The mean shear then advects these vortices into the streamwise direction, rein-
forcing the original streamwise vortices and completing the self-sustaining mechanism.
In a later work (Waleffe, 1998), Waleffe extended this mechanism into a continuation
strategy for obtaining non-trivial solutions in plane Couette flow by adding an explicit
forcing term, f , in the Navier-Stokes equations of the form
f = κFr
Re2[0, γ cos(βy) cos(γz), β sin(βy) sin(γz)], (231)
where Re is the Reynolds number, β = π/2, κ = (β2 + γ2)/γ, and Fr is a continuation
parameter.
Waleffe considered plane Couette flow, with free-slip boundary conditions. In dimen-
sionless form, the governing equations are the Navier-Stokes equations
Dv
Dt= −∇p+
1
Re∇2v + f , (232)
∇ · v = 0, (233)
where p is the pressure and v = (u, v, w) is the velocity vector. The boundary conditions
207
are the so-called free-slip boundary conditions
∂u
∂y= 1, v =
∂w
∂y= 0 (234)
at the lower and upper boundaries in the wall-normal direction, taken to be at y = −1
and y = 1 respectively. Waleffe looked for solutions that satisfied certain symmetry
constraints (listed in Waleffe (1997)), and was able to show the existence of a subcritical
bifurcation at Fr = 5, Re = 150, α = 0.49, and γ = 1.5. By tracking this branch as Fr
was reduced, he found that the solution existed at Fr = 0. In the absence of forcing, the
self-sustaining mechanism discussed above drives the rolls.
Our ultimate interest here is in investigating what effect polymers have on the Waleffe
solutions. Since these structures are three-dimensional, we reduce the computational
burden by splitting the problem into two parts. First, we rewrite equation 233 to take into
account the polymer stresses as
Dv
Dt= −∇p+
1
Re∇2v + f +
1
ReWeS∇ · τ , (235)
∇ · v = 0, (236)
where τ is the polymer component of the stress tensor, We is the Weissenberg number,
and S is the ratio of the solvent to polymer viscosity. We would like to examine the effect
of both free slip and no slip boundary conditions, so we write the boundary conditions as
Λ(u− 1)− (1− Λ)
(∂u
∂y− 1
)= 0 at y = 1,
Λ(u+ 1) + (1− Λ)
(∂u
∂y− 1
)= 0 at y = −1,
v = 0 at y = ±1,
Λw + (1− Λ)∂w
∂y= 0 at y = ±1, (237)
208
so that Λ = 1 is no slip and Λ = 0 is free slip. The difference in the boundary conditions
on u at y = 1 and y = −1 is to ensure that they are identical when u is replaced by −u
and y by −y. As before, we have scaled the polymer component of the stress tensor by
the shear modulus G. Our strategy will be to treat the polymer contribution as a forcing
term in the momentum equations. Thus, for a given value of the polymer stress tensor, we
plan to use equation 236 to solve for v and p. We will then use v in an as yet unspecified
polymer constitutive equation to update τ , and iterate back and forth till the velocity and
stress fields are consistent. Here we describe the first step in this process, which is the
solution of equations 237 with no polymer component, i.e., with τ = 0.
E.2 Discretization
We use a three-dimensional spectral method, with Chebyshev collocation in the y di-
rection, and Fourier collocation in the x and z directions. It is common to use Fourier
collocation based on an even number of grid points. However, the first derivative matrix
for this discretization has two zero eigenvalues, one corresponding to the constant mode,
and one corresponding to the cos(Nx/2) where N is the number of collocation points.
The second eigenmode gives rise to spurious solutions in a collocation scheme. For this
reason, we use Fourier collocation based on an odd number of collocation points. This
scheme includes the cos(Nx/2) mode, and thus only has one zero eigenvalue, which
corresponds to the constant mode. As in chapter 4, we use a primitive variable formula-
tion, and define pressure on a staggered grid in the non-periodic (y) direction. Thus, our
unknowns are the three components of the velocity and the pressure.
If we denote the collocation order in the x direction by Nx, in the y direction by
209
Ny, and in the z direction by Nz, equations 236 yield a system of 3(Nx + 1)(Ny +
1)(Nz + 1) + (Nx + 1)(Nz + 1)Ny nonlinear equations for the three components of the
velocity and the pressure. As in chapter 5, we construct solutions to these equations by
using pseudo-arclength continuation. Since the Navier-Stokes equations do not contain
mixed derivatives, the Jacobian matrix, J , is sparse: for instance, the x derivative of a
variable at a collocation point only depends on the value of the variable at the collocation
points which have the same values of y and z as the point where the derivative is being
computed. Thus, we can store the matrix in sparse format, i.e., we only store the non-zero
entries and information about the row and column numbers of these entries. We solve
the linear systems arising from the Newton iteration using GMRES (Saad and Schultz,
1986). Once again, it is necessary to precondition the system in order to make GMRES
converge, and we use the dual threshold incomplete LU decomposition preconditioner,
or ILUT (Saad, 1996) described in section 5.3. We use a drop tolerance of 10−3 and in
each row of the preconditioner, keep twice the number of entries as in the corresponding
row of J .
The continuation scheme needs a known solution as a starting point, and the Couette
flow solution, u(x, y, z) = y, v = 0, w = 0, valid for Fr = 0, provides a convenient
one. Although this solution is valid for all values of Λ, the continuation method fails for
Λ = 0. This is because, for free slip boundary conditions, u(y, z) + U is a solution to
the governing equations for any constant U . Thus, the absolute value of u is not fixed,
which results in failure to converge. Adding a small component of no-slip, i.e., setting Λ
to some non-zero number, removes this degeneracy, and all the results presented here will
be for Λ different from zero. Since we do not impose any symmetry constraints, we do
not know beforehand whether the loss of stability is through a stationary bifurcation, or
210
through a Hopf bifurcation. For this reason, we determine the unstable eigenvalues using
ARPACK (Lehoucq et al., 1997), using the technique described earlier in section 5.3, for
determining the stability of the diwhirls.
E.3 Preliminary Results
We describe the results from our simulations using Re = 150, β = π/2, γ = 1.5, and
α = 0.49. These are identical to the values used by Waleffe (1997) in his quest for co-
herent structures in Couette flow. The base flow is two-dimensional, homogeneous in
the x direction, and has a y − z dependence. We use 15 Chebyshev collocation points
in the wall normal direction, and 15 Fourier collocation points in the z direction. Us-
ing fewer points gives rise to spurious unstable eigenvalues, and using more collocation
points, particularly in the wall-normal direction, results in ill-conditioned matrices. In
the streamwise direction, we use three Fourier collocation points, since the base flow has
no variation in this direction, and we are, for now, only interested in capturing the struc-
ture of the destabilizing disturbance close to the bifurcation point. Keeping three Fourier
collocation points is equivalent to keeping the cos(αx), sin(αx), and the constant modes
in the x direction.
We first show results for the run with Λ = 0.002, which is very close to the free slip
case. Figure 78 shows the streamwise component of the velocity in the y − z plane. The
redistribution of the streamwise momentum due to the forcing is evident. Consistent with
the findings of Waleffe (1997), we find that the base flow becomes unstable to a three-
dimensional disturbance at Fr = 5. However, since we do not assume any symmetries,
we find that two modes bifurcate close to Fr = 5. The modes have different growth rates,
211
and are thus not degenerate. The structure of these two modes is shown in figure 79,
where we have plotted the streamwise components of the two unstable eigenvectors in
the x−z plane at y = 0. Both modes are consistent with the mechanism of the instability,
i.e., the generation of vortices that have their axes in the y direction due to the existence
of inflection points in the spanwise direction. The difference in the symmetries between
the two modes becomes evident if we plot the streamwise velocity component of the two
unstable eigenvectors as a function of x at y = 0 and z = π/γ. This plot is shown in
figure 80. Clearly, this component in the first eigenvalue (which has the higher growth
rate of the two) has a non-zero mean, while in the case of the second eigenvalue, the
mean is zero. The fact that two eigenvalues bifurcate very close to each other has the
unfortunate effect of making the test function method of detecting stationary bifurcations
(discussed in section 5.3) very unreliable. Although it is possible to construct a test
function that changes sign close to the bifurcation point, the sign change occurs abruptly,
and the test function quickly reverts back to the original sign, so that it is easy to miss the
bifurcation unless the step size is very small.
As we increase Λ, the bifurcation changes from stationary to Hopf, suggesting that
a Takens-Bogdanov point exists close to Λ = 0, where the two real eigenvalues collide
and are transformed to a complex conjugate pair. This phenomenon was discussed in
section 4.5.2 in the context of Dean flow. In figure 81, we show the real and imaginary
parts of the streamwise velocity component of the unstable mode in the x − z plane at
y = 0. Note the similarity between this and figure 79. The similarity in the structure of
the unstable modes between Λ = 0.002 and Λ = 0.1 lends support to the existence of a
Takens-Bogdanov point.
212
z
y
0 1 2 3-1
-0.5
0
0.5
1
Figure 78: Contours and density plot of the streamwise velocity u in the y − z plane inthe base state. The value of Re = 150, Λ = 0.002, and Fr = 5.08.
E.4 Proposals for future work
In the preceding paragraphs, we have shown how base solutions can be tracked, and
bifurcations detected in plane Couette flow with a two-dimensional forcing. The next
step is to actually track the fully three-dimensional (non-trivial) branches. Since the test
function approach fails in this case, it will be necessary to approximate a point on the
non-trivial branch by adding a small component of the unstable eigenvector to the base
solution. We can then get an exact solution on the branch by using this approximation
as a first guess and performing a Newton iteration keeping the coefficient of the cos(αx)
or sin(α x) mode at a single point fixed to prevent convergence to the trivial branch.
Once a solution on the non-trivial branch is found, the solution, the branch can then be
tracked as desired. Since the bifurcation at larger values of Λ is of the Hopf type, non-
trivial solutions for these values of Λ will have to be accessed by continuation from lower
213
x
z
0 5 100
1
2
3
x
z
0 5 100
1
2
3
(a)
(b)
Figure 79: Contours and density plot of the u component of the two unstable eigenvectorsfor the flow shown in figure 78 in the x − z plane at y = 0. The mode shown in (a) hasthe higher growth rate of the two unstable modes.
214
0.0 5.0 10.0x
−0.06
−0.04
−0.02
0.00
0.02
u
Eigenvalue − 1Eigenvalue − 2
Figure 80: Plot of the profile of the u component of the streamwise velocity of the twounstable eigenvectors for the flow shown in figure 78 at y = 0, and z = π/γ. Theeigenvalue corresponding to eigenvector 1 has the higher growth rate. Eigenvector 2 issimply a vertically shifted and scaled version of eigenvector 1.
values of Λ, just as the non-trivial stationary solutions in viscoelastic circular Couette
flow were accessed by continuation of non-trivial solutions in viscoelastic Couette-Dean
flow in chapter 5.
One problem we had, even in the limited analysis that we performed, was the poor
condition number of the Jacobian matrix when the number of Chebyshev modes was
increased. A possible solution to this is to use a spectral element discretization in the
wall-normal direction, while preserving the Fourier discretization in the two periodic
directions. As the number of degrees of freedom increases, so will computational re-
quirements, and the parallelizable preconditioners discussed at the end of chapter 6 may
provide improvements in performance by permitting efficient parallelization of the solu-
tion phase.
215
x
z
0 5 100
1
2
3
x
z
0 5 100
1
2
3
(a)
(b)
Figure 81: Contours and density plot in the x − z plane of the u component of the realand imaginary part of the unstable eigenvector at Re = 150, Λ = 0.1, and Fr = 5.01.
216
Nomenclature
The following three tables summarize the notation used in the thesis. Since chapters 2
and 3 deal with subjects that follow a different notation from the material in the other
chapters, we have provided the notation for these chapters separately from the rest of the
thesis. The variables are listed in case insensitive alphabetical order, so the descriptions
of some of these variables may refer to later entries in the table. Sometimes, for the sake
of convenience, the same symbols are used in the text for dimensionless and dimensional
quantities. Such variables are specified as being either dimensional or dimensionless,
depending on where they are used in the text.
Notation used in Chapter 2
Variable Description
a1 parameter for the slip model (dimensional)
A2, A3, β scaled parameters for the slip model (dimensionless)
C capacitance, equals Vb/πR2u∗t∗ (dimensionless)
D diameter of die (dimensional)
De Deborah number (dimensionless)
Continued on next page
217
Continued from previous page
Variable Description
H dimensionless number, equals 8ηu∗/GR for Newtonian
and (3 + 1/n)K1/n/a1G1/n−1R for power law fluid
G shear modulus (dimensional)
P pressure (dimensional or dimensionless)
Pb barrel pressure (dimensionless)
Q volumetric flow rate (dimensional)
Q0 volumetric flow rate of polymer at die exit
(dimensional or dimensionless)
Qp volumetric rate of displacement of the piston
(dimensional or dimensionless)
R radius of the die (dimensional)
t∗ residence time, Qp/πR2L (dimensional)
u∗ scaling factor for the velocity, equals G/a1 (dimensional)
us slip velocity (dimensional or dimensionless)
Vb barrel volume (dimensional)
γA apparent shear rate (dimensional)
η shear viscosity of the melt (dimensional)
κ equals Gχ (dimensionless)
λ relaxation time (dimensional)
λ1 t∗u∗/L (dimensionless)
λ2 t∗G/η (dimensionless)
Continued on next page
218
Continued from previous page
Variable Description
Λ L/D (dimensionless)
ρ density (dimensional)
ρ0 density at reference pressure (dimensional)
τ stress tensor (dimensional or dimensionless)
τc2, τc3 lower and upper critical shear stresses on the flow curve
(dimensional or dimensionless)
τw wall shear stress (dimensional or dimensionless)
χ compressibility (dimensional)
Notation used in Chapter 3
Variable Description
a position vector in the undeformed state
(dimensionless)
a∗ position vector in the undeformed state
(dimensional)
a determinant of matrix with components aαβ
(dimensionless)
aαβ , aαβ covariant and contravariant components
of the metric tensor in the
undeformed state (dimensionless)
Continued on next page
219
Continued from previous page
Variable Description
A position vector in the deformed state
(dimensionless)
A∗ position vector in the deformed state
(dimensional)
A determinant of matrix with components Aαβ
(dimensionless)
Aαβ , Aαβ covariant and contravariant components
of the metric tensor in the
deformed state (dimensionless)
De Deborah number (dimensionless)
e1,e1, e3 Cartesian unit vectors (dimensionless)
E dimensionless strain energy
F strain energy functional
I1 strain invariant (dimensionless)
l amount by which cone is stretched
(dimensionless)
L length of the truncated cone (dimensional)
n azimuthal wavenumber (dimensionless)
r1, r2 radius of cone at x3 = 0 and x3 = L
(dimensional)
s slope of the cone, (r1 − r2)/L
Continued on next page
220
Continued from previous page
Variable Description
(dimensionless)
v velocity vector (dimensionless)
v velocity vector (dimensional)
vi component i of v
v1, v2 surface coordinates (dimensionless)
v1∗, v2∗ surface coordinates (dimensional)
V0 volume of cone in undeformed state
(dimensionless)
Vl volume of cone in deformed state
(dimensionless)
We Weissenberg number (dimensionless)
x1, x2, x3 Cartesian coordinate directions
z equals v2 cos(φ) (dimensionless)
γ rate of deformation tensor, equals∇v + (∇v)t(dimensionless)
ˆγ rate of deformation tensor (dimensional)
ε rate of elongation
η0 viscosity of UCM fluid
θ azimuthal angle (dimensionless)
λ relaxation time for UCM fluid
(dimensional)
λ1, λ2 stretch factors for the cone
Continued on next page
221
Continued from previous page
Variable Description
in the base state (dimensionless)
λ1, λ2, λ3 perturbation stretch factors
for the cone (dimensionless)
Λ Lagrange multiplier (dimensionless)
τ stress tensor
(dimensional or dimensionless)
τ stress tensor (dimensional)
τij component ij of the stress tensor
φ cone half angle, equals tan−1(s)
(dimensionless)
Notation used in remainder of main text
Variable Description
b maximum polymer extension squared (dimensionless)
c complex growth rate, equals cr + i ci (dimensionless)
cr continuation parameter linking
FENE-P and FENE-CR models (dimensionless)
Cp specific heat capacity at constant
pressure on a mass basis (dimensional)
d gap width, equals R2 −R1 (dimensional)
Continued on next page
222
Continued from previous page
Variable Description
Ev viscous dissipation per unit volume (dimensional)
F c spring force (dimensional)
G shear modulus (dimensional)
H spring constant (dimensional)
J Jacobian matrix (dimensionless)
Kθ azimuthal pressure gradient (dimensional)
L axial wavelength (dimensionless)
n azimuthal wavenumber (dimensionless)
n equals ε1/2n (dimensionless)
N number density of dumbbells (dimensional)
p pressure (dimensional or dimensionless)
Pz axial pressure drop (dimensional)
Q dumbbell end to end vector (dimensional)
Q0 maximum spring extension (dimensional)
Q2 tr(〈QQ〉) (dimensional or dimensionless)
〈QQ〉 ensemble average of conformation tensor
(dimensional or dimensionless)
〈QQ〉ij component ij of 〈QQ〉 (dimensional or dimensionless)
r shifted and scaled radial coordinate (r∗ −R1)/(R2 −R1)
r∗ radial coordinate (dimensional)
R1 inner cylinder radius (dimensional)
Continued on next page
223
Continued from previous page
Variable Description
R2 outer cylinder radius (dimensional)
S ratio of solvent to polymer viscosity
t time (dimensionless)
T temperature (dimensional)
T time period (dimensionless)
v velocity vector (dimensional or dimensionless)
vi component i of vector v (dimensional or dimensionless)
V axial velocity amplitude (dimensional)
Wez axial Weissenberg number (dimensionless)
Weθ azimuthal Weissenberg number (dimensionless)
Wp equals ε1/2Weθ
Wpc critical value of Wp for onset of instability (dimensionless)
Wpc,min minimum critical value of Wp for onset of instability
(dimensionless)
z axial coordinate
α axial wavenumber (dimensionless)
β Floquet multiplier (dimensionless)
γ rate of deformation tensor, equals∇v + (∇v)t
(dimensional or dimensionless)
γij component ij of γ (dimensional or dimensionless)
γ magnitude of shear rate (dimensional or dimensionless)
Continued on next page
224
Continued from previous page
Variable Description
γs characteristic shear rate (dimensional)
δ relative contribution of azimuthal pressure gradient
to total forcing (dimensionless)
ε gap width (dimensionless)
ε elongation rate (dimensional)
ρ density (dimensional)
σ Floquet exponent (dimensionless)
τ stress tensor (dimensional or dimensionless)
τij component ij of stress tensor τ
(dimensional or dimensionless)
ζ friction coefficient due to hydroynamic drag (dimensional)
η general symbol for shear viscosity (dimensional)
ηp polymer viscosity (dimensional)
ηs solvent viscosity (dimensional)
ηt total viscosity, equals ηs + ηp (dimensional)
κ equals (∇v)t (dimensional)
λ relaxation time (dimensional)
ψ distribution function (dimensionless)
Ψ1 First normal stress coefficient (dimensional)
Ψ2 Second normal stress coefficient (dimensional)
Continued on next page
225
Continued from previous page
Variable Description
ω frequency of axial forcing (dimensionless)
Ω angular velocity of inner cylinder rotation (dimensional)
226
Bibliography
Adewale, K. P. and A. I. Leonov (1997). Modeling spurt and stress oscillations in flows
of molten polymers. Rheol. Acta 36, 110–127.
Adjari, A., F. Brochard-Wyart, P.-G. de Gennes, L. Leibler, J.-L. Viovy, and M. Rubin-
stein (1994). Slippage of an entangled polymer melt on a grafted surface. Physica
A 204, 17–39.
Agassant, J.-F., P. Avenas, J.-P. Sergent, and P. J. Carreau (1991). Polymer Processing,
principles and modeling. New York: Hanser Publishers.
Al-Mubaiyedh, U. A., R. Sureshkumar, and B. Khomami (2000). Linear stability
of viscoelastic Taylor-Couette flow: Influence of fluid rheology and energetics. J.
Rheol. 44(5), 1121–1138.
Arnoldi, W. E. (1951). The principle of minimized iteration in the solution of matrix
eigenvalue problems. Q. Appl. Math 9, 17–29.
Avgousti, M. and A. N. Beris (1993). Non-axisymmetric modes in the viscoelastic
Taylor-Couette flow. J. Non-Newtonian Fluid Mech. 50, 225–252.
227
Avgousti, M., B. Liu, and A. N. Beris (1993). Spectral methods for tjhe viscoelastic
time-dependent flow equations with applications to Taylor-Couette flow. Int. J. Num.
Meth. Fluids 17, 49–74.
Bagley, E. F., I. M. Cabott, and D. C. West (1958). Discontinuity in the flow curve of
polyethylene. J. Appl. Phys. 29(1), 109–110.
Bauman, T., T. Sullivan, and S. Middleman (1982). Ribbing instability in coating flows:
Effect of polymer additives. Cgem. Eng. Commun. 14, 35–46.
Baumert, B. M. and S. J. Muller (1995). Flow visualization of the elastic Taylor-Couette
instability in Boger fluids. Rheol. Acta 34(2), 147–159.
Baumert, B. M. and S. J. Muller (1997). Flow regimes in model viscoelastic fluids in a
circular Couette system with independently rotating cyliders. Phys. Fluids 9(3), 566–
586.
Baumert, B. M. and S. J. Muller (1999). Axisymmetric and non-axisymmetric elas-
tic and inertio-elastic instabilities in Taylor-Couette flow. J. Non-Newtonian Fluid
Mech. 83(1–2), 33–69.
Beavers, G. S. and D. D. Joseph (1974). Tall Taylor cells in polyacrylamide solutions.
Phys. Fluids 17, 650–651.
Benkreira, H., M. F. Edwards, and W. L. Wilkinson (1982). Ribbing instability in the roll
coating of Newtonian fluids. Plastics Rubber Proc. Appl. 2, 137–144.
Benzi, M., J. C. Haws, and M. Tuma (2000). Preconditioning highly indefinite and non-
symmetric matrices. SIAM J. Sci. Comput. 22(4), 1333–1353.
228
Benzi, M. and M. Tuma (1998). Numerical experiments with two approximate inverse
preconditioners. BIT 38(2), 234–241.
Bird, R. B., R. C. Armstrong, and O. Hassager (1987a). Dynamics of Polymeric Liquids
(2nd ed.), Volume 1. New York: Wiley.
Bird, R. B., C. F. Curtiss, R. C. Armstrong, and O. Hassager (1987b). Dynamics of
Polymeric Liquids (2nd ed.), Volume 2. New York: Wiley.
Bogdanov, R. I. (1975). Versal deformations of a singular point on the plane in the case
of zero eigenvalues. Funct. Anal. Appl. 9, 144–145.
Booth, G. L. (1958a). Take your choice of coating method, Part 1. Modern Plastics 36(1),
91–95.
Booth, G. L. (1958b). Take your choice of coating method, Part 2. Modern Plastics 36(2),
90–99.
Bousfield, D. W., R. Keunings, G. Marrucci, and M. M. Denn (1986). Nonlinear analysis
of the surface tension driven breakup of viscoelastic filaments. J. Non-Newtonian Fluid
Mech. 21, 79–97.
Brochard, F. and P. G. de Gennes (1992). Shear-dependent slippage at a polymer/solid
interface. Langmuir 8, 3033–3037.
Brooks, A. N. and T. J. R. Hughes (1982). Streamline upwind/Petrov-Galerkin formula-
tions for convection dominated flows with particular emphasis on the incompressible
Navier-Stokes equations. Comp. Methods Appl. Mech. Eng. 32, 199–259.
229
Broyden, C. G. (1965). A class of methods for solving nonlinear simultaneous equations.
Math. Comput. 19, 577–593.
Byars, J. A., A. Oztekin, R. A. Brown, and G. H. McKinley (1994). Spiral instabilities
in the flow of highly elastic fluids between rotating parallel disks. J. Fluid Mech. 271,
173–218.
Canuto, C., M. Y. Hussaini, A. Quarteroni, and T. Zang (1988). Spectral methods in Fluid
Dynamics. New York: Springer-Verlag.
Caracotsios, M. and W. E. Stewart (1985). Sensitivity analysis of initial value problems
with mixed ODEs and algebraic equations. Computers chem Engng 9, 359–365.
Chandrasekhar, S. (1961). Hydrodynamic and Hydromagnetic Stability. New York, NY:
Dover.
Chang, J. C. and M. M. Denn (1980). In G. Astarita, G. Marucci, and L. Nicolais (Eds.),
Rheology, Volume 3. New York: Plenum Press.
Cherhabili, A. and U. Ehrenstein (1995). Spatially localized two-dimensional finite am-
plitude states in plane Couette flow. Eur. J. Mech. B/Fluids 14, 677–696.
Cherhabili, A. and U. Ehrenstein (1997). Finite-amplitude equilibrium states in plane
Couette flow. J. Fluid Mech. 342, 159–177.
Chilcott, M. D. and J. M. Rallison (1988). Creeping flow of dilute polymer solutions past
cylinders and spheres. J. Non-Newtonian Fluid Mech. 29, 381–432.
Chow, E. and Y. Saad (1998). Approximate inverse preconditioners via sparse-sparse
iterations. SIAM J. Sci. Comput. 19(3), 995–1023.
230
Christensen, R. E. (1962). Extrusion coating of polypropylene. S. P. E. J. 18, 751.
Christodolou, K. N. and L. E. Scriven (1988). Finding leading modes of a viscous free
surface flow: An asymmetric generalized eigenproblem. J. Sci. Comput. 3(4), 355–
406.
Cohen, E. D. (1992). Modern Coating and Drying Technology, Chapter 1. New York,
NY: VCH Publishers, Inc.
Cohen, E. D. and E. B. Gutoff (1992). Modern Coating and Drying Technology, Chap-
ter 0. New York, NY: VCH Publishers, Inc.
Corneliussen, A. H. and R. T. Shield (1961). Finite deformation of elastic membranes
with application to the stability of an inflated and extended tube. Arch. Rat. Mech.
Anal. 7, 273–304.
Coyle, D. J. (1992). Modern Coating and Drying Technology, Chapter 3. New York, NY:
VCH Publishers, Inc.
Coyle, D. J., C. W. Macosko, and L. E. Scriven (1986). Film-splitting flows in forward
roll coating. J. Fluid Mech. 171, 183–207.
Coyle, D. J., C. W. Macosko, and L. E. Scriven (1990a). Reverse roll coating of non-
Newtonian liquids. J. Rheol. 36(2), 615–636.
Coyle, D. J., C. W. Macosko, and L. E. Scriven (1990b). Stability of symmetric film-
splitting between counter-rotating cylinders. J. Fluid Mech. 216, 437–458.
Crochet, M. J. and P. M. Naghdi (1969). A class of simple solids with fading memory.
Int. J. Eng. Sci. 7, 1173.
231
Cross, M. C. and P. Hohenberg (1993). Pattern-formation outside of equilibrium. Rev.
Mod. Phys. 65(3), 851–1112.
Curtiss, C. F. and R. B. Bird (1996). Statistical mechanics of transport phenomena:
Polymeric liquid mixtures. Adv. Polym. Sci. 125, 1–101.
Dangelmayr, G. and E. Knobloch (1987). The Takens-Bogdanov bifurcation with O(2)
symmetry. Philos. Trans. R. Soc. London Ser. A. 322, 243–279.
Davard, F. and D. Dupuis (2000). Flow visualization experiments in a blade coating
process. J. Non-Newtonian Fluid Mech. 93(1), 17–28.
Davis, S. H. and S. Rosenblat (1977). On bifurcating periodic solutions at low frequency.
Studies in Appl. Math. 57, 59–76.
Denn, M. M. (1990). Issues in viscoelastic fluid mechanics. Ann. Rev. Fluid Mech. 22,
13–34.
Denn, M. M. (1992). Surface-induced effects in polymer melt flow. In P. Moldenaers
and R. Keunings (Eds.), Theoretical and Applied Rheology, Proc. XIth Int. Congr. on
Rheology, Brussels, Belgium.
Doedel, E. J. (1981). AUTO: A program for the automatic bifurcation analysis of au-
tonomous systems. Congressus Numeratum 30, 265–284.
Drazin, P. G. and W. H. Reid (1981). Hydrodynamic stability. New York: Cambridge
University Press.
Duff, I. S. and H. A. van der Vorst (1999). Developments and trends in the paralle solution
of linear systems. Parallel Comput. 25(13–14), 1931–1970.
232
El Kissi, N. and J. M. Piau (1990). The different capillary flow regimes of entangled
polydimethylsiloxane polymers: macroscopic slip at the wall, hysteresis and cork flow.
J. Non-Newtonian Fluid Mech. 37, 55–94.
Fernando, R. H. and J. E. Glass (1988). Dynamic Uniaxial Extensional Viscosity (DUEV)
effects on roll application I. Polymer blend studies. J. Rheol. 32(2), 199–213.
Fineberg, J. and O. Lioubashevski (1998). Propagating solitary states in highly dissipative
driven fluids. Physica A 249, 10–17.
Fisher, R. J. and M. M. Denn (1975). Finite-amplitude stability and draw resonance in
isothermal melt spinning. Chem. Eng. Sci. 30, 1129–1134.
Fisher, R. J. and M. M. Denn (1976). A theory of isothermal melt spinning and draw
resonance. AIChE J. 22, 236.
Frigo, M. and S. G. Johnson (1999). FFTW user's manual. http://www.fftw.org.
Gallez, X., A. N. Beris, R. Kuenings, and V. Legat (2000). Dissipative effects in complex
flows of dilute polymer solutions far from equilibrium: Modeling and numerical sim-
ulation. In Proceedings of the XIIIth International Congress on Rheology, Volume 2,
pp. 164–166.
Genieser, L. (1997, September). Stress and velocity field evolution in viscoelastic planar
contraction flow. Ph. D. thesis, MIT, Cambridge, MA.
Georgiou, G. C. and M. J. Crochet (1994). Compressible viscous flow in slits with slip at
the wall. J Rheol 38(3), 639–654.
233
Giles, W. B. and W. T. Pettit (1967). Stability of dilute viscoelastic flows. Nature 216,
470–472.
Gokhale, V. V. (1981). Bounds for region of ribbing instability in some free surface
coating flows. J. Rheol. 25(4), 421–432.
Gokhale, V. V. (1983a). Exact solution to the ribbing instability problem in lubrication
flow by invariant imbedding. Chem. Eng. Commun. 21, 81–87.
Gokhale, V. V. (1983b). Improved stability criterion for lubrication flow between counter-
rotating rollers. AICHE J. 29(5), 865–866.
Goldin, M., J. Yerushalmi, R. Pfeffer, and R. Shinnar (1969). Breakup of a laminar
capillary jet of a viscoelastic fluid. J. Fluid Mech. 38, 689–711.
Goldstein, H. (1980). Classical Mechanics. Reading, MA: Addison-Wesley.
Golubitsky, M., D. G. Schaeffer, and I. Stewart (1985). Singularities and Groups in
Bifurcation Theory. New York: Springer-Verlag.
Graham, M. D. (1998). Effect of axial flow on viscoelastic Taylor-Couette instability. J.
Fluid Mech. 360, 341–374.
Graham, M. D. (1999). The sharkskin instability of polymer melt flows. Chaos 9(1),
154–163.
Greenberg, M. D. (1978). Foundations of Applied Mathematics. Englewood Cliffs, N.J:
Prentice-Hall.
Grillet, A. M., A. G. Lee, and E. S. G. Shaqfeh (1999). Observations of ribbing instabil-
ities in elastic fluid flows with gravity stabilization. J. Fluid Mech. 399, 49–83.
234
Groisman, A. and V. Steinberg (1997). Solitary vortex pairs in viscoelastic Couette flow.
Phys. Rev. Lett. 78(8), 1460–1463.
Groisman, A. and V. Steinberg (1998). Mechanism of elastic instability in Couette flow
of polymer solutions: Experiment. Phys. Fluids 10(10), 2451–2463.
Groisman, A. and V. Steinberg (2000). Elastic turbulence in a polymer solution flow.
Nature 405, 53–55.
Guckenheimer, J. and E. Knobloch (1983). Nonlinear convection in a rotating layer:
amplitude expansions and normal form. Geophys. Astrophys. Fluid Dyn. 23, 247–272.
Hatzikiriakos, S. G. and J. M. Dealy (1992a). Role of slip and fracture in the oscillating
flow of HDPE in a capillary. J. Rheol. 36(5), 845–884.
Hatzikiriakos, S. G. and J. M. Dealy (1992b). Wall slip of molten high density
polyethylenes. II. Capillary rheometer studies. J. Rheol. 36(4), 703–741.
Heinrichs, R., G. Ahlers, and D. S. Cannell (1987). Traveling waves and spatial variation
in the convection of a binary mixture. Phys. Rev. A 35, 2761–2764.
Herrchen, M. and H. C. Ottinger (1997). A detailed comparison of various FENE dumb-
bell models. J. Non-Newtonian Fluid Mech. 68, 17–42.
Hill, D. A. (1998). Wall slip in polymer melts: A pseudo-chemical model. J. Rheol. 42(3),
581–601.
Hill, D. A., T. Hasegawa, and M. M. Denn (1990). On the apparent relation between
adhesive failiure and melt fracture. J. Rheol. 34(6), 891–918.
235
Hu, H. C. and R. E. Kelly (1995). Stabilization of longitudinal vortex instabilities by
means of transverse flow oscillations. Phys. Fluids 9, 648–654.
Iooss, G. and D. D. Joseph (1989). Elementary stability and bifurcation theory. New
York: Springer-Verlag.
Jendrejack, R. M., J. J. de Pablo, and M. D. Graham (2000). A multiscale approach to
computational fluid dynamics. In Proceedings of the XIIIth International Congress on
Rheology, Volume 2, pp. 214–216.
Joo, Y. L. and E. S. G. Shaqfeh (1991). Viscoelastic Poiseuille flow through a curved
channel: a new elastic instability. Phys. Fluids A 3(7), 1691–1694.
Joo, Y. L. and E. S. G. Shaqfeh (1992a). The effects of inertia on the viscoelastic Dean
and Taylor-Couette flow instabilities with application to coating flows. Phys. Fluids
A 4(11), 2415–2431.
Joo, Y. L. and E. S. G. Shaqfeh (1992b). A purely elastic instability in Dean and Taylor-
Dean flow. Phys. Fluids A 4(3), 524–543.
Joo, Y. L. and E. S. G. Shaqfeh (1994). Observations of purely elastic instabilities in the
Taylor-Dean flow of a Boger fluid. J. Fluid Mech. 262, 27–73.
Kalika, D. S. and M. M. Denn (1987). Wall-slip and extrudate distortion in linear low-
density polyethylene. J. Rheol. 31(8), 815–834.
Karniadakis, G. E., M. Israeli, and S. A. Orszag (1991). High-order splitting methods for
the incompressible Navier-Stokes equations. J. Comput. Phys. 97, 414–443.
236
Kase, S. (1974). Studies on melt spinning. IV. On the stability of melt spinning. J. Appl.
Polymer Sci. 18, 3279.
Khayat, R. E. (1999). Finite-amplitude Taylor-vortex flow of viscoelastic liquids. J.
Fluid. Mech 400, 33–58.
Knobloch, E. and M. R. E. Proctor (1981). Nonlinear periodic convection in double
diffusive systems. J. Fluid Mech. 108, 291–316.
Kolodner, P., D. Benison, and C. M. Surko (1988). Traveling-wave convection in an
annulus. Phys. Rev. Lett. 60, 1723–1726.
Kumar, K. A. and M. D. Graham (1998a). Effect of pressure-dependent slip on flow
curve multiplicity. Rheol. Acta 37(3), 245–255.
Kumar, K. A. and M. D. Graham (2000). Buckling instabilities in models of viscoelastic
free surface flows. J. Non-Newtonian Fluid Mech. 89, 337–351.
Kumar, K. A. and M. D. Graham (2000a). Solitary coherent structures in viscoelastic
shear flows: Computation and mechanism. Phys. Rev. Lett. 85(19), 4056–4059.
Lange, M. and B. Eckhardt (2000). Diwhirls in viscoelastic Couette-Taylor flow.
Preprint. Although these authors denote the solutions that they find as “diwhirls” , their
investigation is confined to regimes where Re 1 and Weθ/Re 1, where Re is
the Reynolds number. In these parameter regimes, fluid inertia plays a significant role.
Thus, the patterns they simulate cannot be classified as being driven by elasticity alone.
Larson, R. G. (1988). Constitutive Equations for Polymer Melts and Solutions. Boston:
Butterworths.
237
Larson, R. G. (1992). Instabilities in viscoelastic flows. Rheol. Acta 31, 213–263.
Larson, R. G., E. S. G. Shaqfeh, and S. J. Muller (1990). A purely elastic instability in
Taylor-Couette flow. J. Fluid Mech. 218, 573–600.
Lehoucq, R. B., D. C. Sorensen, and C. Yang (1997). Arpack Users Guide: Solu-
tion of Large Scale Eigenvalue Problems by Implicitly Restarted Arnoldi Methods.
ftp://ftp.caam.rice.edu/pub/software/ARPACK.
Leonov, A. I. (1990). On the dependence of friction force on sliding velocity in the theory
of adhesive friction of elastomers. Wear 141, 137–145.
Leonov, A. I. and A. N. Prokunin (1994). Nonlinear Phenomena in Flows of Viscoelastic
Polymer Fluids (1st ed.). London: Chapman & Hall.
Leriche, E. and T. N. Phillips (2000). Spectral element discretization of transient vis-
coelastic flows. In Proceedings of the XIIIth International Congress on Rheology,
Volume 2, pp. 134–136.
Lioubashevski, O., H. Arbell, and J. Fineberg (1996). Dissipative solitary states in driven
surface waves. Phys. Rev. Lett. 76, 3959–3962.
Lioubashevski, O., Y. Hamiel, A. Agnon, Z. Reches, and J. Fineberg (1999). Oscillons
and propagating solitary waves in a vertically vibrated colloidal suspension. Phys. Rev.
Lett. 83(16), 3190–3193.
Lupton, J. M. and J. W. Regester (1965). Melt flow of polyethylene at high rates. Polym.
Eng. Sci. 5, 235–245.
238
Marchal, J. M. and M. J. Crochet (1987). A new mixed finite-element for calculating
viscoelastic flow. J. Non-Newtonian Fluid Mech. 26, 77–114.
Marques, F. and J. M. Lopez (1997). Taylor-Couette flow with axial oscillations of the
inner cylinder. Floquet analysis of the base flow. J. Fluid Mech. 348, 153–176.
Matar, O. K. and S. M. Troian (1999). The development of transient fingering patterns
during the spreading of surfactant covered films. Phys. Fluids 11(11), 3232–3246.
McKinley, G. H., P. Pakdel, and A. Oztekin (1996). Rheological and geometric scaling
of purely elastic instabilities. J. Non-Newtonian Fluid Mech. 67, 19–47.
Mhetar, V. and L. A. Archer (1997). Molecular origin of polymer melt flow instabilities.
ANTEC, 1876–1880.
Molenaar, J. and R. J. Koopmans (1994). Modeling polymer melt-flow instabilities. J.
Rheol. 38(1), 99–109.
Moses, E., J. Feinberg, and V. Steinberg (1987). Multistability and confined traveling-
wave patterns in a convecting binary mixture. Phys. Rev. A 35, 2757–2760.
Nagata, M. (1988). On wavy instabilities of the Taylor-vortex flow between corotating
cylinders. J. Fluid Mech. 188, 585–598.
Nagata, M. (1990). Three dimensional finite-amplitude solutions in plane Couette flow:
bifurcation from infinity. J. Fluid Mech. 217, 519–527.
Nagata, M. (1997). Three-dimensional traveling-wave solutions in plane Couette flow.
Phys. Rev. E. 55, 2023–2025.
239
Nayfeh, A. H. and D. T. Mook (1979). Nonlinear Oscillations. New York: John Wiley
& Sons, Inc.
Ning, C.-Y., C.-C. Tsai, and ta Jo Liu (1996). The effect of polymer additives on extrusion
slot coating. Chem. Eng. Sci. 51(12), 3289–3297.
Northey, P. J., R. C. Armstrong, and R. A. Brown (1992). Finite-amplitude time-periodic
states in viscoelastic Taylor-Couette flow described by the UCM model. J. Non-
Newtonian Fluid Mech. 42, 117–139.
Olagunju, D. O. (1997). Hopf bifurcation in creeping cone-and-plate flow of a viscoelas-
tic fluid. Z. angew. Math. Phys. 48(3), 361–369.
Ottinger, H. C. and F. Petrillo (1996). Kinetic theory and transport phenomena for a
dumbbell model under nonisothermal conditions. J. Rheol. 40(5), 857–874.
Patera, A. T. (1984). A spectral element method for fluid dynamics: Laminar flow in a
channel expansion. J. Comput. Phys. 54(3), 468–488.
Pearson, J. R. A. (1960). The instability of uniform viscous flow under rollers and spread-
ers. J. Fluid Mech. 7, 481–500.
Pearson, J. R. A. and M. A. Matovich (1969). Spinning a molten threadline. Ind. Eng.
Chem. Fund. 8, 605–610.
Pearson, J. R. A. and C. J. S. Petrie (1968). On melt flow instability of extruded polymers.
In R. E. Wetton and R. W. Whorlow (Eds.), Polymer Systems, Deformation and Flow.
Macmillan, London.
240
Pearson, J. R. A. and Y. T. Shah (1972). Stability analysis of the fiber spinning process.
Trans. Soc. Rheol. 16, 519–533.
Pearson, J. R. A. and Y. T. Shah (1974). On the stability of isothermal and nonisothermal
fiber spinning of power-law fluids. Ind. Eng. Chem. Fund. 13, 134–138.
Person, T. J. and M. M. Denn (1997). The effect of die materials and pressure-dependent
slip on the extrusion of linear low-density polyethylene. J. Rheol. 41(2), 249–265.
Petrie, C. J. S. and M. M. Denn (1976). Instabilities in polymer processing. AIChE
J. 22(2), 209–236.
Piau, J. M., N. El Kissi, F. Tossaint, and A. Mezghani (1995). Distortions of polymer
melt extrudates and their elimination using slippery surfaces. Rheol. Acta 34, 40–57.
Pitts, E. and J. Greiller (1961). The flow of thin films between rollers. J. Fluid Mech. 11,
33–50.
Rajagopalan, D., R. C. Armstrong, and R. A. Brown (1990). Finite element methods for
calculation of steady, viscoelastic flow using constitutive equations with a Newtonian
viscosity. J. Non-Newtonian Fluid Mech. 36, 159–192.
Ramamurthy, A. V. (1986). Wall slip in viscous fluids and influence of materials of
construction. J. Rheol. 30(2), 337–357.
Ramanan, V. V., K. A. Kumar, and M. D. Graham (1999). Stability of viscoelastic shear
flows subjected to steady or oscillatory transverse flow. J. Fluid Mech. 379, 255–277.
Rasmussen, H. K. and O. Hassager (1999). Three-dimensional simulations of viscoelastic
instability in polymeric filament. J. Non-Newtonian Fluid Mech. 82, 189–202.
241
Rayleigh, L. (1879). On the instability of jets. Proc. London Math. Soc. 10, 4–13.
Renardy, M. (1997). High Weissenberg number boundary layers for the upper convected
Maxwell fluid. J. Non-Newtonian Fluid Mech. 68, 125–132.
Renardy, M. (2000). Location of the continuous spectrum in complex flows of the UCM
fluid. J. Non-Newtonian Fluid Mech. 94(2), 75–85.
Renardy, Y., M. Renardy, R. Sureshkumar, and A. N. Beris (1996). Hopf-Hopf and
steady-Hopf mode interactions in Taylor Couette flow of an upper convected Maxwell
liquid. J. Non-Newtonian Fluid Mech. 63, 1–31.
Rosenblat, S. (1968). Centrifugal instability of time-dependent flows. Part 1. Inviscid,
periodic flows. J. Fluid Mech. 33, 321–336.
Ruschak, K. J. (1985). Coating flows. Annu. Rev. Fluid Mech. 17, 65–89.
Saad, Y. (1996). Iterative methods for sparse linear systems. Boston: PWS Publishing
Company.
Saad, Y. and M. H. Schultz (1986). GMRES: A generalized minimal residual algorithm
for solving nonsymmetric linear systems. SIAM J. Sci. Stat Comput. 7(3), 856–869.
Saffman, P. G. and S. G. Taylor (1958). The penetration of a fluid into a porous medium
or Hele-Shaw cell containing a more viscous liquid. Proc. Roy. Soc. A. 245, 312–329.
Sanders, J. A. and F. Verhulst (1985). Averaging Methods in Nonlinear Dynamical Sys-
tems. Springer-Verlag.
Savage, M. D. (1977a). Cavitation in lubrication. Part 1. On boundary conditions and
cavity-fluid interfaces. J. Fluid Mech. 80, 743–755.
242
Savage, M. D. (1977b). Cavitation in lubrication. Part 2. Analysis of wavy interfaces. J.
Fluid Mech. 80, 757–767.
Savage, M. D. (1984). Mathematical model for the onset of ribbing. AIChE J. 30(6),
999–1002.
Seydel, R. (1994). Practical Bifurcation and Stability Analysis. New York: Springer-
Verlag.
Shaqfeh, E. S. G., S. J. Muller, and R. G. Larson (1992). The effects of gap width
and dilute solution properties on the viscoelastic Taylor-Couette instability. J. Fluid
Mech. 235, 285–317.
Shen, S. F. (1961). Some considerations of the laminar stability of of incompressible
time-dependent basic flows. J. Aerosp. Sci. 28, 397.
Solomon, T. H. and J. P. Gollub (1988). Chaotic particle transport in time-dependent
Rayleigh-Benard convection. Phys. Rev. A 38(2), 6280–6286.
Soules, D. A., R. H. Fernando, and J. E. Glass (1988). Dynamic Uniaxial Extensional
Viscosity (DUEV) effects on roll application I. Rib and web growth in commercial
coatings. J. Rheol. 32(2), 181–198.
Spiegelberg, S. H. and G. H. McKinley (1996). Stress relaxation and elastic decohesion of
viscoelastic polymer solutions in extensional flow. J. Non-Newtonian Fluid Mech. 67,
49–76.
243
Spiegelberg, S. H. and G. H. McKinley (1998). Numerical simulation of extensional
deformations of viscoelastic liquid bridges in filament stretching devices. J.Non-
Newtonian Fluid Mech. 74, 47–88.
Stewart, C. W. (1993). Wall slip in the extrusion of linear polyolefins. J. Rheol. 37(3),
499–513.
Strenger, M. R., R. B. Secor, and B. Lenz (1997). Knife coating method using ascension
of the fluid by its tension. U.S. Patent 5,612,092.
Sureshkumar, R. and A. N. Beris (1995). Effect of artificial stress diffusivity on the sta-
bility of numerical calculations and the flow dynamics of time-dependent viscoelastic
flows. J. Non-Newtonian Fluid Mech. 60(1), 53–80.
Sureshkumar, R., A. N. Beris, and M. Avgousti (1994). Non-axisymmetric subcritical
bifurcations in viscoelastic Taylor-Couette flow. Proc. Roy. Soc. London Ser. A 447,
135–153.
Sureshkumar, R., A. N. Beris, and R. A. Handler (1997). Direct numerical simulation of
the turbulent channel flow of a polymer solution. Phys. Fluids 9(3), 743–755.
Takens, F. (1974). Forced oscillations and bifurcations. In Commun. No. 3. Math. Inst.,
Rijksuniv. Utrecht, Neth.
Taylor, M. A. and I. G. Kevrekidis (1990). Interactive AUTO:A graphics interface for
AUTO86. Technical report, Department of Chemical Engineering, Princeton Univer-
sity.
244
The Freedonia Group (2000, August). Paints and coatings to 2004 (The Freedonia group
report: 1303). Cleveland, OH: The Freedonia Group.
Tirtaatmadja, V. and T. Sridhar (1993). A filament stretching device for the measurement
of extensional viscosity of polymer solutions. J. Rheol. 37(6), 1081–1102.
Tordella, J. P. (1956). Fracture in the extrusion of amorphous polymers through capillar-
ies. J. Appl. Phys. 27(5), 454–458.
Tordella, J. P. (1963). Unstable flow of molten polymers: A second site of melt fracture.
J. Appl. Polym. Sci. 7, 215–229.
Umbanhowar, P., F. Melo, and H. L. Swinney (1996). Localized excitations in a vertically
vibrated granular layer. Nature 382, 793–796.
Vinogradov, G. V., V. D. Protasov, and V. E. Dreval (1984). The rheological behavior
of flexible-chain polymers in the region of high shear rates and stresses, the critical
process of spurting, and supercritical conditions of their movement at T > Tg. Rheol.
Acta 23, 46–61.
von Kerczek, C. and S. H. Davis (1974). Linear stability theory of oscillatory Stokes
layers. J. Fluid Mech. 62, 753–773.
Waleffe, F. (1995). Hydrodynamic stability and turbulence: beyond transients to a self-
sustaining process. Studies in Appl. Math. 95, 319–343.
Waleffe, F. (1997). On a self-sustaining process in shear flows. Phys. Fluids 9, 883–900.
Waleffe, F. (1998). Three-dimensional coherent states in plane shear flows. Phys. Rev.
Lett. 81(19), 4140–4143.
245
Wang, S.-Q. and P. A. Drda (1996). Superfluid-like stick-slip transition in capillary flow
of linear polyethylene melts. 1. General features. Macromolecules 29, 2627–2632.
Wang, S.-Q. and P. A. Drda (1997). Stick-slip transition in capillary flow in linear
polyethylene: 3. Surface conditions. Rheol. Acta 36, 128–134.
Weisberg, A., I. G. Kevrikidis, and A. J. Smits (1997). Delaying transition in Taylor-
Coette flow with axial motion of the inner cylinder. J. Fluid Mech. 348, 141–152.
White, J. L., M. H. Han, N. Nakajima, and R. Brzoskowski (1991). The influence of ma-
terials of construction on biconical rotor and capillary measurements of shear viscosity
of rubber and its compounds and considerations of slippage. J. Rheol. 35(1), 167–189.
White, W. D. and D. M. McEligot (1970). Transition of mixtures of polymers in a dilute
aqueous solution. ASME J. Basic Engineering 92, 411–418.
Wilson, H. J., M. Renardy, and Y. Renardy (1999). Structure of the spectrum in zero
Reynolds number shear flow of the UCM and Oldroyd-B liquids. J. Non-Newtonian
Fluid Mech. 80, 251–268.
Wu, J., R. Keolian, and I. Rudnick (1984). Observation of a nonpropagating hydrody-
namic soliton. Phys. Rev. Lett. 52, 1421–1424.
Wu, J., J. Wheatley, S. Putterman, and I. Rudnick (1987). Observation of envelope soli-
tons in solids. Phys. Rev. Lett. 59, 2744–2747.
Yang, W. H. and W. W. Feng (1970). On axisymmetrical deformations of nonlinear
membranes. J. Appl. Mech. 37, 1002–1011.
246
Yarin, A. and M. D. Graham (1998). A model for slip at polymer/solid interfaces. J.
Rheol. 42(6), 1491–1504.
Zeichner, G. (1973). Spinnability of viscoelastic fluids. Master's thesis, Univ. of
Delaware, Newark.