Visco-plastic fluids: from Theory to

Application
C. Ancey a Neil J. Balmforth b Ian Frigaard c
a School

of Architecture, Civil and Environmental Engineering, Ecole

Polytechnique
Federale de Lausanne, Lausanne, Switzerland

b Departments
c Departments

of Mathematics & Earth and Ocean Science, University of British

Columbia, Vancouver, Canada
of Mathematics & Mechanical Engineering, University of British
Columbia, Vancouver, Canada

Fig. 1. The participants enjoying the sunny conditions of Tecino.

Preprint submitted to Elsevier

12 November 2007

Viscoplastic materials are fluids that exhibit a yield stress: below a certain
critical threshold in the imposed stresses, there is no deformation and the
material behaves like a rigid solid, but when that yield value is exceeded, the
material flows like a fluid. Practically, such flow behavior appears in many
situations, including slurries and suspensions, certain polymer solutions, crystallizing lavas, muds and clays, heavy oils, avalanches, cosmetic creams, hair
gel, liquid chocolate, and some pastes. Consequently, the theory of the fluid
mechanics of such materials has applications in wide array of different fields,
ranging from the oil, gas and chemical industries, to food processing and to
geophysical fluid dynamics. The most commonly studied such fluid is the Bingham fluid, which is often uppermost in the minds of the scientists when they
think of non-Newtonian fluids, although in fact this model is usually a significant simplification of the true rheological behavior.
From the 14th to 18th October 2007, a workshop gathering 49 participants
from 13 countries was held in Monte Verit`a, a marvelous place in the Italianspeaking part of Switzerland, with spectacular scenery on lake Major and the
city of Locarno. The meeting was the follow-on of an earlier workshop organized at Banff (Canada) in 2005. This was the opportunity to have a closer
look at recent achievements in this hectic field and try to collectively think
of future developments. We also paid homage to the Greek and Cypriot communities for their enduring efforts in viscoplastic fluid mechanics. During an
enjoyable banquet, Evan Mitsoulis was awarded the Bingham Fluid Medal
(not to be confused with a more auspicious prize) for his research on computational aspects of viscoplasticity while John Tsamopoulos was rewarded
with a prize for successfully computing the shape of the yield surfaces around
bubbles in Bingham fluids. The next workshop will hold in Cyprus in October
2009.
Part of the objective of this workshop was to promote exchanges between
scientists from different origins, which made it possible to shed light on a
number of topical issues in different ways. Here is a rapid summary of the
different presentations and discussions.
The nature of yield stress is a longstanding problem, which has received considerable attention over the last 20 years. The emblematic question may be
the suitability of the Herschel-Bulkley equation to represent the true rheological behavior over a sufficiently wide range of shear rates. The tentative
answer provided by Coussot, Bonn, and Molder is that while the notion of
yield stress makes sense experimentally, there is also a critical shear rate below which no steady state is observed. This critical rate not only contrasts
with the predictions of the Herschel-Bulkley model, but also gives rise to flow
instabilities (e.g., shear banding) due to viscosity bifurcation. Understanding
the physical origin of yielding on the particle scale is of paramount importance to appreciate empirical bulk-scale idealizations such as the Bingham
2

model. Van Damme showed how adding adsorbing or non-adsorbing polymers

in a colloidal suspension changes the yield-stress and bulk-viscosity values,
which emphasizes the role played by particle contacts (lubricated or direct) in
yielding; a similar behavior was reported by Bardou with debris-flow samples.
Nicot summarized the various definitions used in geomechanics to define the
yield surface on the bulk and particle scales; a key result is the existence of
different modes of failure for a dry granular material. Coussot and Alexandrou
supplemented the Herschel-Bulkley constitutive equation with a scalar kinetic
equation reflecting the changes in particle arrangement (breakdown/builtup
of links between particles). A more formal continuum-mechanics approach was
proposed by Goddard, who showed that a host of phenomena can be described
using a generalized viscosity coefficient (fourth-rank viscosity tensor).
From this confrontation between experimental observations, phenomenological laws and theoretical views of particulate media in geomechanics and fluid
mechanics arose a number of questions that deserve mention. Are there hidden
issues with the shape of the yield surface in principal stress space (i.e. non-vonMises yield laws) that we have not appreciated so far? Are plasticity theorys
flow laws better than the current rate-dependent terms of viscoplastic constitutive models? Should we pay more attention to issues of symmetry/invariance
and violation of physical principles (e.g., the second law of thermodynamics)?
Is there interest for theoreticians to explore and understand more of the underlying physics, e.g. continue the search for microstructural theories to build
macroscopic flow models?
The notions of yield surface and post-yielding behavior has evolved differently
in geomechanics and fluid mechanics. As illustrated in Chevaliers experiments
and Pastors talk, yielding of saturated granular materials may be the consequence of a pore-pressure increase and particle fluidization. This highlights the
two-phase behavior of apparently single-phase bulk materials. Another nontrivial question, notably addressed in Wilsons squeezing flow experiments,
is related to the influence of the mean stress (first invariant) on yield stress:
once the material has yielded, does the shear stress still vary linearly with the
normal stress? Goddards theoretical analysis and Pouliquens phenomenological interpretation provided evidence of both shear-rate and normal-stress
dependence of the shear stress in dry and saturated granular materials. The
different talks underlined the gap between classical single-phase viscoplastic
models and plastic multi-phase Coulomb-like theories. To date, there is little
connection. Should we try to investigate Bingham-type models with porepressure-dependent yield stresses, and so forth? Is liquifaction something we
should be considering?
Whereas earlier experiments focused on viscometric flows in rheometers, most
recent experimental investigations have been done using more complex geometries. In parallel, image processing has been extensively used to take measure3

ments and visualize what is going on inside the material. Hogg reviewed recent
theoretical results for the dam-break problem, which can be seen as the prototypical time-dependent problem, while Cochard showed his experimental
results for this geometry. As shown by Balmforth, another interesting timedependent problem is the Stokes problem, where a viscoplastic material is
confined between two parallel walls, one being in oscillating motion, the other
being fixed. Chambon used a conveyor belt to investigate the inner structure of
a finite volume of Carbopol in a steady regime. Cheddadi, de Bruyn and Mitsoulis showed how flows past an obstacle or a falling sphere can be effectively
investigated experimentally and numerically.
Classical geometries such as the Couette cell still attract much attention. Compared to other rheometer geometries, wide-gap Couette cells are well suited to
particle suspensions, but flow-curve inference may be more delicate. Wiederseiner used transparent buoyant-particle suspensions and particle-tracking velocimetry to estimate the local flow curve. Naccache compared different boundaries (grooved surface, vane) to minimize slip effects at the wall. There was
little discussion about experimental techniques for measuring the flow properties inside the material; while magnetic resonance imaging is now arrived at
a mature stage, it remains an expensive and heavy technique. We are on the
verge of important developments in image processing that should make flow
visualization easier and cheaper, but probing what occurs in opaque materials
still offers substantial difficulties.
Since Glowinskis pioneering work, a great deal of work has been accomplished
to provide robust and accurate algorithms able to cope with viscoplastic behavior. As reviewed by Huilgol and exemplified in a number of applications
shown by Picasso and Muravleva, variational methods are in common use. In
particular, regularization techniques such as augmented Lagrangian and related methods have demonstrated their efficiency. Saramito provided another
perspective by showing how an adaptive meshing strategy can be used to
increase the order of a numerical scheme.
Since the Banff meeting, where we suggested using Carbopol 940 as a suitable
viscoplastic fluid, evidence has accumulated that Carbopol is far from being an
ideal Herschel-Bulkley fluid; in particular, elasticity cannot be neglected at low
shear stress. This has led to a substantial renewal of interest in elasto/viscoplasticity and plastic compressibility. Recent models and applications were
presented by Saramito, Forterre, Georgiou, and de Souza Mendes. In parallel,
the quest for a genuinely Herschel-Bulkley fluid is launched again. Graners
experiments revealed the value of two-dimensional foams to investigate plastic
flow problems, but elasticity still plays a role in the bulk-stress generation.
From the presentations, it appeared that Carbopol requires relatively long
mixing times in order to achieve reproducible results. There seems to be some
need for guidelines in preparing Carbopol (and other viscoplastic candidates),
4

a database listing standard Herschel-Bulkley fits as functions of concentration

and pH, and more sophisticaed viscoplastic models incorporating elastic effects
and slip.
A number of particular applications were also presented. As shown by Moyers, Nouar and Rust, interfacial, thermal, and/or convective instabilities in
industrial ducts or in magmatic chambers lead to complex problems in stability analysis. Fester and Slatter explained why determining energy losses in
turbulent/laminar viscoplastic flows is of great importance for pipeline design
and operation. Geomaterials involved in natural gravity-driven flows have often been modeled as plastic fluids. Bartelt reviewed some key questions in
the rheology of snow and recast rheological questions within a thermodynamics framework by highlighting the role of energy and mass transfers in snow
avalanches. As shown by Bonnoit, McElwaine, and Kowalski, there is still a
host of questions (e.g., segregation in granular flows, behavior of multicomponent fluids) that have received partial answers so far. Today, in the realm of
geophysical flows, it is still unclear how to forge a symbiotic relation between
engineers who use crude shallow-layer theories coupled with Bingham-like drag
parameterizations (amongst others) and physicists who are developing more
physically-based theories on the basis of laboratory data. Can we help to better
constrain those former models with the latter type of approaches? Turbulence
in viscoplastic seems to be an open field for research: when does transition
occur, and how are friction factors affected by viscoplasticity? What is the
phenomenology of viscoplastic turbulence? Can it be characterized by relatively simple dimensional scaling theories and ideas about the effect of yield
stress on the energy containing scales, the inertial range and the dissipation
scales?
The meeting was scientifically rewarding and enjoyable, which encouraged the
organization of subsequent workshops with the same format. Monte Verit`a was
a striking venue for a focused workshop of this nature. We are very grateful
to Centro Steafano Franscini for their financial and technical support.