You are on page 1of 25

Journal of Intelligent Material Systems and

Structures
http://jim.sagepub.com/

A review on the simulation and modeling of magnetorheological fluids


Ali Ghaffari, Seyed Hassan Hashemabadi and Mahshid Ashtiani
Journal of Intelligent Material Systems and Structures published online 21 August 2014
DOI: 10.1177/1045389X14546650

The online version of this article can be found at:


http://jim.sagepub.com/content/early/2014/08/21/1045389X14546650

Published by:

http://www.sagepublications.com

Additional services and information for Journal of Intelligent Material Systems and Structures can be found at:

Email Alerts: http://jim.sagepub.com/cgi/alerts

Subscriptions: http://jim.sagepub.com/subscriptions

Reprints: http://www.sagepub.com/journalsReprints.nav

Permissions: http://www.sagepub.com/journalsPermissions.nav

Citations: http://jim.sagepub.com/content/early/2014/08/21/1045389X14546650.refs.html

>> OnlineFirst Version of Record - Aug 21, 2014

What is This?

Downloaded from jim.sagepub.com at TULANE UNIV on September 1, 2014


Review Article

Journal of Intelligent Material Systems


and Structures
1–24
A review on the simulation and Ó The Author(s) 2014
Reprints and permissions:
modeling of magnetorheological fluids sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/1045389X14546650
jim.sagepub.com

Ali Ghaffari, Seyed Hassan Hashemabadi and Mahshid Ashtiani

Abstract
The magnetorheological fluids are classified as smart materials with controllable rheological properties. The fast growing
application of magnetorheological fluids in recent years has increased the demand for simulation and modeling of these
fluids. From the invention of magnetorheological fluids up to now, many experimental and also theoretical investigations
have been carried out to study these types of smart materials; also many attempts have been made to formulate and
simulate their behavior. The aim of this investigation is to present a review on the different models and simulation meth-
ods that were applied in the studying of magnetorheological fluids. In this study, the different simulation methods of mag-
netorheological fluid have been categorized into two general approaches: continuum and discrete phase approaches. The
different rheological and structural models of magnetorheological fluids in continuum approach have been summarized in
this study. The computational framework of discrete approach and the basic models for magnetorheological fluid in this
approach are also discussed.

Keywords
Smart fluid, magnetorheological fluid, magnetorheological modeling, continuum simulation, discrete approach

Introduction the value of yield stress and apparent viscosity of the


MRF. The value of enhancement depends on the
The magnetorheological fluids (MRFs) as the smart applied field magnitude and direction. The particle sizes
materials with varying yield stress were discovered in are around 1–20 mm (Bossis et al., 2003b). A simple
the late 1940s by Rabinow (1948). The variation in their and quite efficient MRF can be synthesized by disper-
rheological properties (due to their specific composition sing the iron powder in oil with the use of some surfac-
and structure) in less than few milliseconds, in the pres- tant, for instance, stearic acid, in order to prevent
ence of external magnetic field, makes them appropriate irreversible aggregation (Bossis et al., 2002; Omidbeygi
for many industrial applications. The interest in MRFs and Hashemabadi, 2012).
has increased in the last decades due to their promising The MRFs should be stable against the settling and
characteristics. have a high magnetic saturation (Claracq et al., 2004).
The MRFs are composed of susceptible magnetic Furthermore, the magnetic particles should have a
particles dispersed in a continuum fluid that is usually good resistance against the corrosion. Different types
oily fluid. While in the absence of magnetic field, the of magnetic particles have been examined by research-
MRFs behave like a Newtonian fluid, but with the ers: magnetic latex made of polystyrene particles with
application of an external magnetic field, the dispersed inclusions of magnetite (Lemaire et al., 1992; Volkova
magnetic particles make chain-like structures that resist et al., 1999b), polymer-coated nano-sized ferrite parti-
against the flow (Omidbeygi and Hashemabadi, 2013). cles (Kormann et al., 1996), meso-scale carbonyl iron
As a result, the rheological properties of MRFs change
with the application of magnetic field, and the fluid
Computational Fluid Dynamics (CFD) Research Laboratory, School of
shows a non-Newtonian behavior which is usually Chemical Engineering, Iran University of Science & Technology (IUST),
described with the viscoplastic rheological models Tehran, Iran
(Lange et al., 2001). The volume fraction of dispersed
magnetic particles in conventional MRFs is around Corresponding author:
Seyed Hassan Hashemabadi, Computational Fluid Dynamics (CFD)
20%–40% (Sahin et al., 2012; See et al., 2006; Ulicny Research Laboratory, School of Chemical Engineering, Iran University of
et al., 2005). Chain-like structure of particles in the Science & Technology (IUST), Narmak, Tehran 16846-13114, Iran.
presence of an external magnetic field causes to enhance Email: hashemabadi@iust.ac.ir

Downloaded from jim.sagepub.com at TULANE UNIV on September 1, 2014


2 Journal of Intelligent Material Systems and Structures

and nickel–zinc ferrites (Phulé and Ginder, 1999), parameters are as follows: volume fraction, size, type,
Fe3O4 nanoparticles (Zafarani-Moattar and Majdan- and shape of dispersed particles; physical properties of
Cegincara, 2013), and the most commonly used carbo- carrier fluids; direction and intensity of applied mag-
nyl iron particles (De Vicente and Berli, 2013; Ji Eun netic field; and operational mode. Variation in any of
et al., 2012; Park et al., 2011). Among all these, only these parameters leads to a change in rheological prop-
the carbonyl iron–based suspensions have already erties of MRF such as yield stress and apparent viscos-
enough performance to be used in MR devices ity. Depending on the operational conditions, there
(Claracq et al., 2004). may also be other effective parameters, but they are in
MRFs are used in many different industrial applica- second degree of importance.
tions from their invention until now. Linear dampers General methods of modeling and simulation of
(Goldasz and Sapinski, 2012; Parlak and Engin, 2012; MRFs can be classified into two categories: the conti-
Parlak et al., 2012; Zhu et al., 2012), rotor dampers nuum and discrete approaches. In the first one, the mix-
(Carmignani et al., 2006), rotary brakes (Farjoud et al., ture of dispersed particles and suspending fluid is
2008), rotary clutches (Olabi and Grunwald, 2007), and considered as a one homogeneous phase, and its beha-
directional control valves (Böse et al., 2012; Kostamo vior can be modeled by applying appropriate rheological
et al., 2012; Quoc-Hung et al., 2007; Salloom and equations. In this approach, the coefficients and para-
Samad, 2011, 2012) are some of the examples which meters of rheological models are defined as a function of
have found industrial applications for MRFs. Ma et al. suspension’s characteristics and applied magnetic field
(2012) have proposed a continuous variable transmis- intensity. The main problem in this approach is to find
sion system based on the MRF and shape memory an appropriate rheological model that can predict the
alloy for improving the operating efficiency of the cen- behavior of flow in a wide range of shear rates and
trifugal fans. Nosse and Dapino (2007) have presented applied magnetic field (Omidbeygi and Hashemabadi,
a hybrid actuator based on the rectification of magne- 2012). Only the macroscopic properties are evaluated in
tostrictive vibrations by means of MR flow control. A this method, and microscopic specifications such as
method of jet stabilization based on using MRF in jet chain-like structure of dispersed particles are not
fishing technology has been proposed by Kordonski tracked. In the second approach (discrete model), MR
and Shorey (2007). Recently, the MRFs were utilized suspension is considered as a two-phase particulate flow
in the haptic devices (Blake and Gurocak, 2009; Yang consisting of the dispersed particles and the surrounding
et al., 2009, 2010, 2012). The benefits of MRFs such as fluid. In this approach, the motion of particles and sus-
fast response time and low power requirement provide pending fluid is handled separately by appropriate mod-
the ability to size minimization and also elimination of els. Modeling in discrete approach is based on the basic
instabilities in such devices (Yang et al., 2012). The physical laws, and thus, it is more general in comparison
MRFs were also used for medical applications. Sheng to continuum approach. Discrete approach can also be
et al. (1999) have proposed a novel cancer therapeutic divided into two categories according to the methods,
method using MRFs. In this method, the MRF is used which are used to model the motion of dispersed parti-
to inhibit the blood supply to a tumor in order to cles. In one method, the motion of particles is tracked
starve it. The MR effect is also utilized for drug deliv- by implementation of Newton’s second law, and in
ery systems. Therefore, it can be said that MR technol- another one, the particles’ trajectories are calculated via
ogy has a wide range of applications in human life and kinetics-based models. The main problem in dispersed
it is increasingly growing. More details on the applica- model is complexity of equations and high computa-
tions of MRFs can be found in Bica et al. (2013). tional cost. The discrete and also continuum approaches
A broad range of possible applications is presum- are schematically shown in Figure 1.
ably the reason for the intense research on MRFs in There are some review articles dealing with the
recent years. Nowadays, modeling and simulation have synthesis, application, and technological aspects of
become inevitable part of engineering problems in MRFs (Bica et al., 2013; Bossis et al., 2002; Goncalves
order to find the optimum design. It can be expressed et al., 2006; Mrlik et al., 2013; Park et al., 2010; Zhu
that modeling and simulation have significant role in et al., 2012). There are also some articles that review
the development of science and engineering. In addition the characteristics and structures of MRFs (Bossis
to the fast increasing applications of MRFs, modeling et al., 2002, 2003b; De Vicente et al., 2011a). Even
and simulation of these fluids have received much inter- though many attempts have been made by numerous
est from researchers. Modeling and simulation of authors for simulation and modeling of MRFs, based
MRFs are so challenging due to their complex struc- on our knowledge, there is no comprehensive review on
tures. Researches in this field indicate that there are yet this field published yet. Therefore, the aim of this
no unique and explicit methodologies for simulation review article is to survey the MRF simulation and
and modeling of MRFs. It is necessary to determine modeling approaches and also to classify and categor-
the effective parameters for studying the behavior of ize these approaches. The main modeling approaches
MRFs for modeling and simulation. The main effective (continuum and discrete approaches) with their

Downloaded from jim.sagepub.com at TULANE UNIV on September 1, 2014


Ghaffari et al. 3

Figure 1. Schematic of the two main modeling approaches for MRFs. The suspension in continuum approach is considered as a
one-phase fluid with new density and viscosity. In discrete approach, particles and surrounding fluid will be tracked separately.
MRF: magnetorheological fluid.

computational frameworks and necessary correlations where L, g, and W are the geometric length, gap, and
and models are discussed in this article. The main sec- width of the flow rectangular channel between the fixed
tions and subsections of this article are listed in tree poles, respectively; Q is the flow rate; h is the fluid visc-
chart of Figure 2. osity in the absence of magnetic field; and t y is the yield
stress resulted from the applied magnetic field. The
parameter c is a coefficient depending on the ratio of
MRF operational modes
DPt =DPh . The minimum value of the c is 2 (for
Determination of the MRF operational mode is neces- DPt =DPh less than ;1) and its maximum value is 3 (for
sary because the rheological models in continuum mod- DPt =DPh greater than ;100) (Salloom and Samad,
eling approach are derived based on the flow mode 2012).
(Nicholas and Norman, 2004; Quoc-Hung et al., 2007). In shear mode, also referred to as direct shear mode,
MRFs may be operated in different ways depending on the MRF experiences a relative shearing force from one
the requirements of an application. The three common of the plates enclosing the fluid gap (Figure 3(b)).
operational modes include valve mode, shear mode, Clutches and rotary brakes are examples of devices
and squeeze mode, all shown in Figure 3. which utilize the MRFs in shear mode. The total force
The valve and shear modes are more common in in the shear mode can be split into two parts: a viscous
MR devices. The squeeze mode is less common, and term Fr which is only dependent on the rheological
hence, it is less studied in the literatures. MR devices in properties of fluid in off state and a magnetic field–
valve mode contain stationary poles, whereas in shear dependent component Fmr (Olabi and Grunwald, 2007).
mode these poles are moving. Servo valves, dampers, Accordingly, the total shear force can be approximated
and shock absorbers are examples of MR devices which
operate in valve mode (Ahmadkhanlou, 2008; Farjoud hSA
F = Fr + Fmr = + t mr A ð2Þ
et al., 2011; Grunwald and Olabi, 2008; Kostamo et al., g
2012). The valve operational mode of MRFs was exten-
sively studied by numerous researchers (Grunwald and where h, S, and A are the dynamic viscosity, the relative
Olabi, 2008; Nicholas and Norman, 2004; Quoc-Hung speed, and the working interface area, respectively. tmr
et al., 2007; Sahin et al., 2012; Salloom and Samad, is the yield stress developed in response to the applied
2011), to improve the efficiency and/or to find the opti- magnetic field.
mum geometry. In most of these studies, the Bingham Most of the studies on the rheometry of MR suspen-
plastic rheological equation with a variable yield stress, sions are in the direct shear flow mode (Joung and See,
which is dependent on the magnetic field strength, has 2008; Ramos et al., 2011; Sedlacik et al., 2013; Ulicny
been used to describe the flow behavior. The total pres- et al., 2005). This is perhaps due to the widespread use
sure drop in the MR devices operated in the valve of MRF in this operational mode and/or due to the fact
mode is evaluated by summing the viscous component that the results obtained in this mode can be generalized
DPh and the yield stress component DPt (Salloom and to find the constitutive equation of flow which can be
Samad, 2011) used in other operational modes.
The squeeze mode of operation (Figure 3(c)) is the
12hQL ct y L least common and maybe the least understood
DP = DPh + DPt = + ð1Þ
g3 W g (Farjoud et al., 2011). This mode places the acting force

Downloaded from jim.sagepub.com at TULANE UNIV on September 1, 2014


4 Journal of Intelligent Material Systems and Structures

Figure 2. The main sections and subsections of this article.


MRF: magnetorheological fluid.

Figure 3. MR fluid operational flow modes: (a) valve mode, (b) shear mode, and (c) squeeze mode.
MR: magnetorheological.
Adapted from Farjoud et al. (2011).

in line with the magnetic field direction and the particle and MRF–elastomer vibration isolators. Among the
chains (Carmignani et al., 2006; Wang et al., 2005). three modes, the squeeze flow mode provides the larg-
The squeeze mode has been explored for use in the est yield stress under the same field (Ruiz-López et al.,
small-amplitude vibration control, impact dampers, 2013). However, understanding the behavior of MRFs

Downloaded from jim.sagepub.com at TULANE UNIV on September 1, 2014


Ghaffari et al. 5

under squeeze flow mode is still far to be complete, id=57#Features) to investigate the influence of magnetic
mainly because of the lack of both a thorough under- field on the MRFs. A wide variety of modified versions
standing of the basic MR mechanisms and reliable of conventional stress- and strain-controlled rheometers
experimental data (Ruiz-López et al., 2012). A compre- can be found in the literature for studying MR suspen-
hensive review of squeeze flow magnetorheology has sions: concentric two cylinders (Klingenberg et al., 2007b;
been presented by De Vicente et al. (2011). Wang and Gordaninejad, 2006), two parallel plates
In practice, MR devices may utilize a combination (Bossis et al., 2003a; De Vicente et al., 2004; See, 2003),
of the three operational modes. In fact, conventional and cone-plate geometry (Claracq et al., 2004; Volkova
applications involve the presence of complex flows. et al., 1999a).
However, in order to get a full understanding of the Extensive studies have been performed to under-
complex flows, it is necessary to simplify the flow beha- stand the rheological properties of MR suspensions at
vior by considering the overall complex flow as a result different operational modes (Fang et al., 2009;
of a superposition of standard flows such as valve, Klingenberg and Ulicny, 2011; Laun et al., 1996;
shear, and squeeze flows (De Vicente et al., 2011). A Martin, 2001; Park et al., 2010; Rankin et al., 1999;
center MR bearing from a driveshaft line is an example Santiago-Quinones et al., 2013; Sim et al., 2013). The
of MR device which includes the combined squeeze and parameters of MRFs rheological models are found to
shear modes. Kulkarni et al. (2003) have investigated be dependent on the operational conditions, especially
the properties of MRFs under the squeeze combined the intensity of the applied magnetic field. The MRFs
with torsional mode. Their results showed that introdu- resist the flow by applying a magnetic field, and thus,
cing oscillatory shear (torsional mode) in pure squeeze the yield stress and apparent viscosity of MRF increase
mode can improve the damping force and performance with increase in the magnetic field intensity. Therefore,
of a typical damper. An investigation on the combined due to the yield behavior, it is convenient to use visco-
shear and squeeze mode of MRF is also reported by plastic rheological models in order to describe the
See and Tanner (2003). They observed that in the behavior of MRFs (Huang et al., 2012). Mitsoulis
absence of shear forces, the squeezing normal force was (2007) has presented a detailed review on the rheologi-
increased with the magnetic flux density. But it was cal models of viscoplastic fluids. Some of the most con-
decreased by applying the shear forces. This behavior venient reported rheological models for MRFs are
was referred to the breakage of the chain-like structure biviscous model, Bingham model, biplastic Bingham
of particles due to the shear forces. In a similar work, model, Herschel–Bulkley model, Casson model, and
See et al. (2006) investigated the behavior of MRF Papanastasiou model. Table 1 presents different rheo-
under a combination of squeeze and shear modes. They logical models that have been used for MRFs.
found that the compression effects are less significant Analytical and numerical simulations for viscoplastic
than shear strains. But, however, they have not pro- fluids under different operational modes can be found
posed any constitutive model. It should be declared in the literature (for instance, see Firouzi and
that there are a few reported works in the literature Hashemabadi, 2008, 2009). The viscoplastic models
concerning combined flows in spite of their importance presented in Table 1 are depicted in Figure 4(a) to (d).
in practical MR devices. The development of a general
constitutive framework for the combined operational Application of rheological models
modes will be very helpful for modeling of real MR
devices. Even though a wide variety of viscoplastic rheological
models were examined to describe the behavior of
MRFs (Table 1), it should be noted that the Bingham
Rheological models plastic and Herschel–Bulkley models are the most com-
monly used. This may be due to the simpler form of
As mentioned above, the main problem in continuum these two models in comparison with the other visco-
approach is to determine an appropriate rheological plastic models. Especially the linear form of Bingham
model for MRF (Omidbeygi and Hashemabadi, 2012). model makes it appropriate for high computational
Accordingly, experimental studies are needed to derive problems. A list of rheological models that are used for
the rheological model and its coefficients for different MRFs at different operational modes is tabulated in
operational modes of MR suspensions. There are few Table 2. The flow mode and the main obtained results
commercial rheometers to study the behavior of are summarized in this table. A review on this table will
MRFs, and consequently in most of the researches, the be useful for the selection of appropriate rheological
common commercial rheometers were modified model for different operational modes.
and developed to study the MRF behavior (Laun et al., Li et al. (2002) studied the rheological behavior of a
1996). An MR rheometer has been provided by standard MRF, 132LD MRF, at different operational
Anton Paar Company (http://www.anton-paar.com/ modes, and they have observed a linear viscoelastic
Magnetorheology/Rheometer/60_Corporate_en?product_ behavior at low values of stresses. They have

Downloaded from jim.sagepub.com at TULANE UNIV on September 1, 2014


6 Journal of Intelligent Material Systems and Structures

Table 1. Different viscoplastic rheological models that are used for magnetorheology.

Rheological model Reference Equation

Bingham plastic Mitsoulis (2007) t = t y + hg_ jt j.t y (3)


n
Herschel–Bulkley Mitsoulis (2007) t = t y + h(g)_ jt j.t y (4)
pffiffiffi pffiffiffiffi pffiffiffiffiffiffi
Casson Mitsoulis (2007) t = t y + hg_ jt j.t y (5)

hr g_ t<t1
Biviscous Goldasz and Sapinski (2012) t= (6)
t y + hg_ t.t 1

t 1 + hr g_ t<t 2
Biplastic Bingham Goldasz and Sapinski (2012) t= (7)
t y + hg_ t.t 2
Papanastasiou Papanastasiou (1987) _ = t N (g)W
t(g) _ 1 (g)
_ + t Bn (g)W
_ 2 (g)
_ (8)

Figure 4. Viscoplastic rheological models: (a) biviscous model; (b) biplastic Bingham model; (c) Bingham, Herschel–Bulkley, and
Casson models; and (d) Papanastasiou model.

demonstrated that nonlinear viscoelastic properties will may be described with Kelvin–Voigt model.
be dominated by increasing the applied stress. Gandhi Viscoelastic behavior of MRFs is also investigated by
and Bullough (2005) have investigated the behavior of some other researchers (Claracq et al., 2004; Segovia-
MRF at stresses less than the yield point and observed Gutiérrez et al., 2012). Numerous studies on the viscoe-
a solid-like behavior for fluid in this stress range which lastic fluids can be adapted for modeling and

Downloaded from jim.sagepub.com at TULANE UNIV on September 1, 2014


Ghaffari et al. 7

Table 2. Rheological models for MRFs in some investigations.

Reference Description Flow mode Rheological equation, Table 1

Berli and De Vicente A structural viscosity model on the Shear mode Equations (3) and (5)
(2012) basis of Bingham plastic and Casson
models was proposed
Goldasz and Sapinski A robust set of non-dimensional Valve mode in dampers Equations (3), (4), (6), and (7)
(2012) parameters for the purpose of
modeling of MR/ER dampers was
proposed
Omidbeygi and CFD simulation of rotational eccentric Shear mode Equation (3)
Hashemabadi (2012) cylinder in MRF
Kim et al. (2011) MR characteristics of the nano-sized Shear mode in rotational Equation (5)
magnetic particle–based MRF was rheometer
investigated
Farjoud et al. (2011) Analytical investigation of MRFs in Squeezing flow in Equation (8)
squeezing flows was performed low shear rate
Farjoud et al. (2008) At high shear rates, most of the MRFs MR brakes Equation (4)
are shear thinning and Bingham plastic
model is inappropriate to use.
Herschel–Bulkley model is more
appropriate in high shear rates
Wang and Gordaninejad Rheological properties of MR materials Squeeze flow Equation (3)
(2006) at high shear rates
See (2003) A comparison study on the behavior of Shear and squeezing flow Equation (4)
an MR suspension under steady shear
flow and constant velocity squeezing
flow

MRF: magnetorheological fluid; ER: electrorheological; CFD: computational fluid dynamics.

simulation of viscoelastic behavior of MRFs experience shear rates more than 10,000, it is necessary
(Hashemabadi et al., 2003; Hashemabadi and to find the appropriate rheological models to describe
Mirnajafizadeh, 2010) the MRFs’ behavior at such shear rates too. Farjoud
In fact, the viscoplastic behavior of MRFs is et al. (2008) have studied the behavior of MRFs at high
observed when the applied stress is more than the yield shear rates, up to 12,000 s21. They observed a shear
point. However, most of the studies on the MRF beha- thinning behavior for MRF at high shear rates.
vior deal with its viscoplastic behavior; that is, the Accordingly, they have concluded that the Herschel–
MRFs were investigated more under the stresses more Bulkley model is more appropriate than the Bingham
than the yield point. plastic model in the case of high shear rates. Omidbeygi
Lange et al. (2001) have studied the rheological and Hashemabadi (2013) have also used the Herschel–
behavior of MRF by a capillary rheometer at high val- Bulkley model to investigate the hydrodynamic charac-
ues of stresses in flow mode and observed a viscoplastic teristics of MRF flow within an eccentric annulus.
behavior. They have also used Bingham plastic equa- Their analytical and numerical simulations showed a
tion (equation (3)) to describe the flow of MRF. In a good consistency at high shear rates in a wide range of
research done by Choi et al. (2005), the flow behavior applied magnetic field intensities. Bingham plastic,
curve (shear stress vs shear rate) of MR and electro- biviscous, biplastic Bingham, and Herschel–Bulkley
rheological fluids were theoretically analyzed by means models are used by Goldasz and Sapinski (2012) to
of a rotational coaxial cylinder and a rotational parallel analyze the flow mode behavior of an MR damper.
disk viscometer. In their study, some relations between They have compared the analytical results with the
shear stress and torque as well as shear rate and angu- experimental results and observed a good consistency
lar velocity were derived on the basis of the Bingham between the results obtained from biplastic Bingham
plastic (equation (3)), Biviscous (equation (6)), and model and experimental data.
Herschel–Bulkley (equation (4)) constitutive models. Casson rheological equation (equation (5)) is also
In steady state, the MRF flow is commonly modeled used by some authors. A Casson-like structural viscos-
as a Bingham plastic fluid. However, most of the stud- ity model was proposed by Berli and De Vicente (2012).
ies deal with the investigation of rheological behaviors They have demonstrated that the Casson-like model
of MRF and work in approximately low shear rates can account the deviations from Bingham plastic model
(below 1000 s21). By knowing the fact that the MR predictions. Kim et al. (2011) have investigated the
devices in some applications such as rotary brakes characteristics of the nano-sized magnetic particle–

Downloaded from jim.sagepub.com at TULANE UNIV on September 1, 2014


8 Journal of Intelligent Material Systems and Structures

based MRF dispersed in a non-magnetic carrier fluid The value of n shows how fast or slow the transition
using a rotational rheometer under different external from low shear rate quasi-Newtonian behavior to a
magnetic field strengths. Based on the investigation of Bingham plastic behavior occurs (Farjoud et al., 2011).
flow behaviors at a steady shear mode and yield stress, Also, a Papanastasiou–Herschel–Bulkley model was
they found that the flow curve can be predicted well presented (Susan-Resiga, 2009) as a rheological model
with the Casson equation. for MRFs by blending a quasi-Newtonian behavior at
Claracq et al. (2004) have presented experimental very low shear rates with a Herschel–Bulkley model for
study of the viscoelastic behavior of MRFs under small high shear rates. This model has been validated with
deformations. They have observed that the yielded fluid experimental data for shear rates lower than 1000 s21
models such as Bingham plastic, Casson, and Herschel–
Bulkley have no acceptability at low shear rates. A _ = t N (g)W
t(g) _ 1 (g)
_ + t HB (g)W
_ 2 (g)
_ ð14Þ
cross model defined by the following equation has been
where
used to fit their experimental rheometry results
  t N (g)
_ = h0 g,
_ _ = t y + cg_ 1n
t Bn (g) ð15Þ
h0  h ‘
t = h‘ + g_ ð9Þ
1 + lg_ With the following weighting functions
   
where h0 is the viscosity at very low shear rates and h‘ g_ g_
is the viscosity at high shear rates (i.e. plastic viscosity). _ = 1  tanh 
W1 (g) W2 (g)
_ = tanh  ð16Þ
g_ g_
The relaxation time l is related to the transition from
the ‘‘solid state’’ to the ‘‘liquid state.’’ In low shear rates here, g_  is a shear rate value in the neighborhood of
(when g_ ! 0), the fluid has a Newtonian behavior, and the transition between the Newtonian and Herschel–
in high shear rates (when g_ ! ‘), the MRF behaves Bulkley models.
like a Bingham plastic fluid. Many of the MR devices such as MR dampers in
The cross models based on Papanastasiou’s (1987) automotive applications and also shock absorbers are
constitutive relation were also proposed for MRFs. accompanied by high values of shear rates and veloci-
The advantage of these cross models is their ability to ties. As illustrated in Table 2, some authors who stud-
capture the variation in the shear stress versus the ied the behavior of MRFs at high shear rates (up to
shear rate, within a large range of the shear rate 12,000 s21) (Farjoud et al., 2008) reported that the
(1026–103 s21) (Susan-Resiga, 2009). Herschel–Bulkley model could appropriately be used
A cross model known as Papanastasiou–Bingham for modeling the flow behavior of MR Brakes.
rheological equation was used by Farjoud et al. (2011) Goncalves and Ahmadian (2005) also investigated the
to model the MRFs at low shear rate squeezing flows. behavior of MRF subjected to high rates of shear and
This cross model includes the combination of high flow velocities. They used a custom-made high
Newtonian fluid behavior at low shear rates and shear slit flow rheometer to investigate the MRF
Bingham plastic behavior for high values of shear rates response at velocities ranging from 1 to 37 m/s, with
corresponding shear rates ranging from 0.14 3 105 to
_ = t N (g)W
t(g) _ 1 (g)
_ + t Bn (g)W
_ 2 (g)
_ ð10Þ 2.5 3 105 s21. A yield behavior of MRF was reported
in their observations in the presence of magnetic field,
t N (g)
_ = h0 g,
_ t Bn (g)
_ = t y + h0 g_ ð11Þ while a nearly Newtonian post-yield behavior was
reported for high shear rate tests in the absence of mag-
The weighting functions W1 (g)
_ and W2 (g)
_ in equa- netic field. In order to model the yield stress, a new
tion (10) should be chosen so that W1 (g)_  W2 (g)
_ at term called the ‘‘dwell time’’ was introduced and
_  W2 (g)
very low shear rates and W1 (g) _ at high shear defined as ‘‘the amount of time the fluid spends in the
rates, with W1 (g)
_ + W2 (g)
_ = 1. Some of the continuous presence of a magnetic field.’’ The authors indicated
functions that meet these general requirements can be that the normalized yield stress is predominantly a
found in Mitsoulis (2007). Farjoud et al. (2011) have function of dwell time. To date, little studies have been
used the following weighting functions reported to investigate the MRFs in such high velocity
and high shear applications. There is a need to more
_ = eng_
W1 (g) _ = 1  eng_
W2 (g) ð12Þ experimental and theoretical works to obtain a com-
plete model for MRFs at high values of shear rates and
After substitution of these functions in equation (10)
strains.
and rearrangement, the following stress–strain relation
In summary, it can be said that the Bingham plastic
is obtained
equation is the most common model which has been
   applied by the authors to describe the rheological beha-
t y 1  eng_
t = h0 + g_ ð13Þ vior of MRFs at moderate shear rates. Good consis-
g_ tency with experimental data and also the linear

Downloaded from jim.sagepub.com at TULANE UNIV on September 1, 2014


Ghaffari et al. 9

formulation of this model might be the reason of its


abundant use. But this model loses its validity at high
values of shear rates, due to the shear thickening beha-
vior of MRFs at high shear rates. Herschel–Bulkley
and Casson models have shown a good consistency
with experimental data at high shear rates. Also, it
should be noted that a Newtonian behavior is observed
for MRFs at low values of shear rates. Therefore, there
is a need for a rheological model that can describe the
behavior of MRF at a wide range of shear rates. The
cross models such as Papanastasiou models have been
used by the authors for this aim.
The main problem in all these rheological models is
to find the values of their parameters which are depen-
dent on the operational conditions and the magnitude
of applied magnetic field. To determine the coefficients
of rheological models for each application, some experi-
mental works are necessary. Accordingly, finding the
correlations to derive the functionality of coefficients
from the basic properties of MRF and its operational
conditions is an urgent need in studying of MRFs.

Figure 5. Forces acting on an MR fluid.


Rheological models based on the MRF structure MR: magnetorheological.
It should be noted that the rheology of MRFs is
strongly dependent on the state of structuration of the and Heyes, 1993). These dimensionless numbers are
suspension (Volkova et al., 2000). As mentioned in the discussed in the following subsection.
previous section, the main problem in the case of rheo-
logical models is to find the correlations for depen-
l number. The magnitude of the magnetic interaction
dency of coefficients on the operational conditions and
energy between the two magnetic moments in the linear
fluid structural properties. The aim of this section is to
regime relative to the thermal energy is the so-called l
review the different studies that have investigated the
parameter (De Vicente et al., 2011). When an isolated
dependency of rheological coefficients on the structural
particle that is surrounded by a fluid is exposed to an
properties of MRFs at different operational conditions.
external magnetic field H0 , this particle will acquire a
From the design perspective, it is important to estab-
magnetic moment as (Bossis et al., 2003b)
lish a quantitative relationship between the rheological
properties (viscosity, yield stress, etc.) and variables mp  mf
such as the volume fraction of the particles, their mag- m = 4pm0 mf ba3 H0 , b= ð17Þ
mp + 2mf
netic properties, and the intensity of the applied exter-
nal field. where mf , mp , and m0 are relative magnetic permeability
It is common to use the dimensionless numbers to of fluid, particle, and vacuum, respectively, and a is the
formulate the behavior of MRFs. In order to obtain radius of the particle. This magnetic moment causes
these numbers, by ignoring the inertial forces (Bossis aggregation of magnetic particles and their alignment
et al., 2003b), it can be started from the equation of one in the direction of applied magnetic field. On the other
particle motion according to Newton’s motion law. The hand, the thermal energy, which is proportional to the
procedure to obtain the dimensionless numbers was dis- product of Boltzmann constant (k) and temperature in
cussed by Bossis et al. (2003b). To find the dimension- Kelvin (T), leads to disruption of particle chains.
less parameters of system, knowledge of the acting Accordingly, the formation of aggregates of particles
forces will be essential. Hydrodynamic, magnetic, and depends on the ratio of magnetic interaction energy to
Brownian forces are the main acting forces in MRFs. kT. This ratio is known as l parameter
These forces are schematically shown in Figure 5.  
Mason number (Mn), Peclet number (Pe), and 1 m2 1 2 3 2 1
l= = 4pm0 mf b a H0 ð18Þ
Lamda (l) are the main dimensionless numbers which 4pm0 mf a3 kT kT
were used by many authors to formulate the rheologi-
cal behavior of MRFs based on the fluid structure For sufficiently large values of l, magneto-static par-
(Bossis and Cebers, 1999; Lemaire et al., 1993; Melrose ticle interactions dominate over thermal motion,

Downloaded from jim.sagepub.com at TULANE UNIV on September 1, 2014


10 Journal of Intelligent Material Systems and Structures

resulting in chain-like particle aggregates. Bossis et al. Accordingly, it can be concluded that for a given
(2002, 2003b) have declared that at l = 1, the inter- MR suspension with a specified concentration, two of
particle magnetic forces dominate Brownian forces. these three dimensionless parameters are sufficient to
For 1 mm particles with large permeability (b’1) at describe the dependency of structural properties. Some
moderate magnetic field intensity (H = 127 A=m) and researchers have developed theoretical and empirical
T = 300 K, the value of l obtained is equal to 1. correlations to describe the dependency of rheological
parameters as a function of the structural properties of
MRF. The viscosity and yield stress of MR suspensions
Mason number. The primary forces governing the beha-
are the most important rheological characteristics in
vior of MR suspensions are magneto-static polarization
magnetorheology.
forces induced by the applied magnetic field and the
Some investigations declared that for MR suspen-
hydrodynamic forces caused by the particle motion rela-
sions with large enough magnetic fields and in the case
tive to the continuous phase. The ratio of hydrodynamic
of negligible short-range inter-particle forces, all the
to magnetic forces is known as Mason dimensionless
rheological properties—and in particular the
number, Mn, which is defined as (Bossis et al., 2002)
viscosity—show the same magnitude for the same val-
_ 2
6phf ga 8hf g_
ues of Mn, l, and volume fraction of particles, f
Mn =  = ð19Þ (Bossis et al., 2002, 2003b; De Vicente et al., 2011). The
fp m0 mf b2 H02 equilibrium structure of MRFs in the absence of flow
also can be determined only by the particle volume
here, in this equation, hf is the viscosity of the sur-
fraction f and the l parameter (De Vicente et al., 2011;
rounding fluid. It should be noted that in different lit-
Furst and Gast, 2000).
eratures, the definition may differ by a multiplicative
Accordingly, some correlations and structural mod-
factor. Similar Mason numbers employing the particle
els were proposed to describe the dependency of rheo-
magnetization or magnetic moment have been defined
logical properties of MRF on its structural properties.
by other authors (De Gans et al., 1999, 2000).
Most of these structural models deal with determina-
Klingenberg et al. (2007a) have developed a relation
tion of viscosity and yield stress dependency on the
for Mason number that is defined in terms of the sus-
magnetic field intensity. The proposed models for visc-
pension magnetization as
osity and also yield stress have been reviewed in the fol-
_ 2
9hf gf
lowing subsections.
Mn[ ð20Þ
2m0 mf h M i2 Viscosity structural models
where f is the volume fraction of dispersed particles As mentioned above, the Bingham plastic model is the
and h M i represents the suspension magnetization. They most common equation used to describe the flow beha-
have declared the Mason number that is defined in vior of MR suspensions. An equation for the dimen-
terms of the suspension magnetization can be employed sionless viscosity can be derived by starting from
to produce a collapse of experimental data at various Bingham plastic model
magnetic field strengths and shear rates.
 
h Mn 1
Peclet number. The ratio of hydrodynamic to Brownian =1+ ð23Þ
h‘ Mn
forces is expressed as Peclet number. In fact, it is an
index to determine the significance of Brownian forces The shear viscosity is defined as the ratio of shear
in comparison with shear forces. The Brownian forces _ h‘ is the high shear
stress to shear rate (h = t=g);
can be neglected in the case of large values of Peclet viscosity (field strength–independent) in Bingham
number. This dimensionless parameter may be defined plastic model; and Mn is the critical Mason number
as (Bossis et al., 2003b) that determines the transition from magnetization to
hydrodynamic control of suspension structure (Berli
_ 3
6phga and De Vicente, 2012). The critical Mason number is
Pe = ð21Þ a function of the volume fraction f. Several models
kT
have been presented in the literature to describe this
In summary, it can be said that these three dimen- dependency as follows (De Vicente et al., 2004;
sionless parameters, l, Mn, and Pe, represent the ratios Martin and Anderson, 1996)
between the main forces acting on a conventional
MRF. Also, these parameters can be related to each Cfhf
other as Mn = ð24Þ
h‘
2Pe here, C is a coefficient which is derived according to
Mnl = ð22Þ
3 specific assumptions and/or simplifications in the

Downloaded from jim.sagepub.com at TULANE UNIV on September 1, 2014


Ghaffari et al. 11

Table 3. Different values of D in equation (25) and C in strength of the field-induced structure (Bossis et al.,
equation (24). 2002, 2003b).
Applied magnetic field intensity, volume fraction of
Reference. D C
particles, viscosity of carrier fluid in the absence of
Volkova et al. (2000) 20.87 \ D \ 20.74 1.91 external field, the magnetic particle size, and the type
De Gans et al. (1999) 20.9 \ D \ 20.8 5.25 are the main parameters that may influence the yield
Martin and Anderson (1996) – 8.82 stress of MRFs. Several researches have focused on the
De Vicente et al. (2004) – 8.485 determination of the dependency of yield stress on
Felt et al. (1996) 20.83 \ D \ 20.74 –
these effective parameters. Even though several differ-
ent models are presented in the literature to formulate
the dependency of yield stress, there are no comprehen-
mechanical stability conditions of MR suspensions. sive and unique correlations that can predict this
The range of C has been tabulated in Table 3. In order dependency. The effect of applied magnetic field was
to capture the experimental data trend, a more general given more attention in comparison to other influen-
form of equation (23) has been used in the literature as cing parameters. Numerous studies in the field of mag-
below netorheology have been focused on the effects of the
applied magnetic field (Agustı́n-Serrano et al., 2013;
h Bica, 2012; Bossis and Cebers, 1999; Chin et al., 2001;
= 1 + Mn MnD ð25Þ
h‘ Fang et al., 2009; Kim et al., 2011; Park et al., 2011;
Promislow and Gast, 1996; Santiago-Quinones et al.,
Table 3 presents different values for exponent D. But 2013). The effects of the other parameters such as vol-
it must be noted that this equation fails to describe the ume fraction, viscosity of surrounding fluid, and size of
low shear regime and also predicts a too sharp solid– particles have been less reported in comparison with
liquid transition (Berli and De Vicente, 2012; De the field strength (Kittipoomwong et al., 2005; Lemaire
Vicente et al., 2011). A modified more complex struc- et al., 1995; Ruiz-López et al., 2012; Zafarani-Moattar
tural model for viscosity of MRF was derived by Berli and Majdan-Cegincara, 2013). Table 4 presents some
and De Vicente (2012) based on the concept of the of these models and correlations which were proposed
effective volume fraction and viscosity function of non- based on the structure of MRFs. The operational
Newtonian suspensions. This structural model for visc- mode, MRF composition, and also the range of effec-
osity of MRFs is expressed as tive parameters are reported in this table.
!2 The rheological behavior of MR suspensions con-
h 1 + ðMn=Mn Þ1=2 taining magnetic micro-particles was investigated by
= ð26Þ Bossis et al. (2003a). They proposed a mechanism for
h‘ ðh‘ =h0 Þ1=2 + ðMn=Mn Þ1=2
determining the yield point of MRFs under shear load
where h0 is the low shear viscosity, and other para- and magnetic field. Based on their theory, the aggre-
meters are the same in equation (25). It was declared gates of particles which are formed due to the presence
(Berli and De Vicente, 2012) that this structural model of external magnetic field begin to rotate after some
is capable to explain deviations in the experimental friction on the walls and break at a strain smaller than
data from the Bingham model predictions. unity. This critical strain is the value when hydrody-
A simplified dimensionless form of the Casson plas- namic shear forces overcome the magnetic force and
tic equation can also be obtained by expanding equa- cause breakage of chain-like aggregates. In this theory
tion (26) for h0  h‘ as following expression a linear chain structure is supposed to the aggregates.
By neglecting the interactions between chains, the fol-
h lowing correlation is obtained to describe the depen-
= 1 + ðMn=Mn Þ1 + 2ðMn=Mn Þ1=2 ð27Þ
h‘ dency of Bingham yield stress

A good consistency between the predictions of this ty 9fr gc


structural model and experimental results has also been =   ð28Þ
ðm0 H 2 fÞ 8 1 + g2 2
c
reported.
here, g c and fr are the critical strain and a coefficient,
respectively, which are dependent on the magnetic per-
Yield stress dependency meability of the particles. Typical values for the particle
The yield stress, a critical value of shear stress when a with high permeability are g c = 6 and fr = 2 (Bossis
viscoplastic fluid begins to flow, is an important para- et al., 2003a).
meter of MRFs. The efficiency of an MRF is first In the case of high values of permeability
judged through its yield stress, ty , which measures the (a = mp =mf  1), an analytical equation for prediction

Downloaded from jim.sagepub.com at TULANE UNIV on September 1, 2014


12
Table 4. Different structural models proposed in the literatures.

Operational H (kA/m) Particle types (size, mm) Base fluid (viscosity, Pa s) u (vol. %) g_ (s21) Correlation Reference
mode

Squeeze flow 0–354 Carbonyl iron Silicone oil (0.02–0.5) 5–30 0.03–0.2a t CE }h0:33 f2:0 H2:0 Ruiz-López et al. (2012)
3=2
t y }H0
Shear flow Up to 0.8 Magnetite (;5), Silicone oil (0.1) 0.1–0.6 0.1–100 Chin et al. (2001)
and ;8–800 carbonyl iron (;4.5–5.2) t y }f (f\0:2)
h‘ }f0:45 H2:9
0
Shear flow 7–30 Carbonyl iron (;7) Silicone oil (13 and 606) 5–30 \2.5 Claracq et al. (2004)
t y }f0:5 H1:5
0
Shear mode 0–800 Polystyrene (;0.5–1) Distilled water and 0.014–0.12 \200 t y }f:H2 Felt et al. (1996)
excess SDS (0.001)
Shear flow 0–300 Carbonyl iron (1–2) Silicone oil 0–0.5 0.01–1000 D = 20.74, 20.81, 20.83, and 20.87b Volkova et al. (2000)

t y (H0 ) = aH20
pffiffiffiffiffiffiffiffiffiffiffiffiffi!
Shear flow 0–343 Carbonyl iron (4.25) Lubricant oil ;20 0.1–500 tanh H0 =Hc Fang et al. (2009)
3 pffiffiffiffiffiffiffiffiffiffiffiffiffi
H0 =Hc

SDS: sodium dodecyl sulfate.


a
Elongational rate range.
b
Din equation (25).

Downloaded from jim.sagepub.com at TULANE UNIV on September 1, 2014


Journal of Intelligent Material Systems and Structures
Ghaffari et al. 13

of the yield stress based on the standard structural of volume fraction of particles. They have deduced a
models has been presented (Bossis et al., 2002, 2003b) power law correlation, with exponent 1.5, for depen-
dency of yield stress as
3=2
t y = 2:31fm0 Ms1=2 H0 ð29Þ
t y }f0:5 H01:5 ð35Þ
here, Ms is the saturation magnetization. For the case
of low values of permeability (a = mp =mf \5), the yield Vereda et al. (2011) have used
 the
concept of the
stress can be calculated by the following relation average particle magnetization Mp for the scaling of
pffiffiffi the yield stress of dilute MRFs. They have shown that
ty 3 3   2 f(1  f) the average particle magnetization, as obtained from
= ms ð30Þ
m0 mf H 2 16 Cs + ms (1  f) the magnetization measurements of suspensions, can
ms be a good parameter for the scaling of the yield stress
ms = 1 ð31Þ (static as well as dynamic) of conventional MRFs at
mf
low particle concentration. A correlation for the
where Cs = 1 if the particles are gathered into stripes dependency of the static yield stress has been pro-
and Cs = 2 if they form cylinders. It is clear from this posed as
equation that the yield stress of materials with low per-  2:0
meability is proportional to the magnetic field strength t ystatic = 2:19 3 103 f Mp ð36Þ
squared. Also at low field strength, a similar depen-
dency has been reported (De Vicente et al., 2011; And also the correlation for the dynamic yield stress
Ramos et al., 2011). based on Bingham plastic model was obtained
A universal hybrid equation has been adopted from  2:1
the electrorheological structural models to find the field t ydynamic = 1:824 3 103 f Mp ð37Þ
dependency of the yield stress as (Fang et al., 2009;
Kim et al., 2011) The average particle magnetization can be obtained
by dividing the measured suspension magnetization by
pffiffiffiffiffiffiffiffiffiffiffiffiffi! the volume fraction as follows (Klingenberg et al.,
tanh H0 =Hc
t y (H0 ) = aH02 pffiffiffiffiffiffiffiffiffiffiffiffiffi ð32Þ 2007a; Vereda et al., 2011)
H0 =Hc
 MS (H)
here, a is a constant related to the susceptibility of the Mp (H) = ð38Þ
fluid, and volume fraction and other analogous physi- f
cal parameters; Hc is the critical magnetic field strength Using the average particle magnetization instead of
which satisfies H0 leads to validity of the correlations in a wide range
 3=2
of the field strengths. A good consistency has been
t y } H02 H\Hc ð33Þ reported between the predictions of proposed equations
H0 H.Hc (36) and (37) and the experimental results in a wide
Based on equation (32), two limiting behaviors with range of the magnetic field intensities (Vereda et al.,
respect to Hc can be obtained 2011).
In summary, it can be expressed that the field depen-
 dency of the yield stress is usually represented by a
aH02 H0  Hc
ty = pffiffiffiffiffiffi 3=2 ð34Þ power law relation (ty ;H n ) (Bossis et al., 2003b; De
a Hc H0 H0  Hc
Vicente et al., 2011). Due to the nonlinearity of the dis-
A good consistency between experimental results persed materials’ magnetization, different exponents (n)
and predictions of equation (34) was observed (Fang are reported for small, intermediate, and high values of
et al., 2009; Kim et al., 2011). Fang et al. (2009) in their applied magnetic field strength. The case of n = 2 is
experimental work on a carbonyl iron–based MR sus- expected for a linear magnetic material at low fields or
pension have obtained the value of Hc = 151:74 kA=m. for particulates of low permeability (Ginder et al., 1996;
Even though in most of the proposed correlations, a Vereda et al., 2011). For intermediate fields, n = 3/2
squared variation with the magnetic field strength is was reported (Bossis et al., 2002). The smaller value of
observed for yield point at relative small field strengths, n may be the result of the saturation of polar regions of
some observations do not confirm this relationship. each particle under the effect of intermediate fields
Claracq et al. (2004) have experimentally investigated (Ginder et al., 1996). The yield stress at the magnetic
the rheological properties of MRFs at low shear rates, saturation regime, which occurs at the strong external
less than 3 s21, and also at a wide range of magnetic fields, becomes field independent and scales with the
field strengths, 7 kA/m \ H0 \ 30 kA/m. Their square of saturation magnetization of particles (De
experimental work included 5%, 10%, 15%, and 30% Vicente et al., 2011; Fang et al., 2009).

Downloaded from jim.sagepub.com at TULANE UNIV on September 1, 2014


14 Journal of Intelligent Material Systems and Structures

Particle size effects nonlinear dependency, as it can be seen in Table 3, has


Lemaire et al. (1995) have investigated the effect of par- also been discussed. For instance, Ruiz-López et al.
ticle size on the rheological behavior of the MR suspen- (2012), who have investigated the performance of
sions composed of monodisperse and polydisperse MRFs in low compression rate, have proposed a quad-
spherical particles. Their observations revealed that for ratic dependency. Also, Tang and Conrad (2000) have
small particles where the parameter l is still larger than reported a nonlinear polynomial dependency on the
unity (l’102  103 ), the yield stress is strongly volume fraction for the yield stress.
enhanced with the increase in the particle size. Some
authors have demonstrated that only for very high val-
Medium viscosity
ues of l (l’109 ), there will be no dependency on the
size of the particles. De Gans et al. (2000) have also Even though the medium viscosity has influence on the
reported a strong increase in MR properties with the sedimentation rate of the dispersed particles, its effect
increase in the particle size for MR suspensions con- has not often been investigated in MR studies. Ruiz-
taining small particles. The decrease in the yield stress López et al. (2012) considered a power law dependency
for smaller particles may be due to the relatively for yield stress on the medium viscosity as
smaller magnetization of these particles (Genc x and ty ;h0:3360:008 . They have declared that the effect of
Phulé, 2002). particle concentration is found to be significantly more
Particle size effects on the MR properties can be important than the effect of surrounding fluid
related to the influence of the Brownian motions which viscosity.
are significant in the case of small particles. However,
at high values of l, Brownian forces can be neglected in
comparison with magnetic forces. Felt et al. (1996) have Summary
also investigated experimentally the effect of particle Yield stress is the most important parameter in the
size on a monodisperse MR suspension. Although their characterization and studying the magnetorheology so
investigations included dilute suspension with small that the MRF performance is judged through this para-
particle size range, they have demonstrated that both meter. Yield stress, ty , which is dependent on the
apparent viscosity and the yield stress show strong par- applied magnetic field intensity, H0, has been investi-
ticle size dependency. A linear relationship between gated experimentally in several researches. A power
apparent viscosity and particle diameter and also simi- law relation, ty ;H0n with 0 6 n 6 2, was usually pre-
larly for yield stress and particle diameter was reported. sented by researches to formulate this dependency. The
n = 2 was reported in the case of a linear magnetic
Particle volume fraction material at the low fields or particles with low perme-
ability. Yield stress becomes field independent when
Experimental study by Chin et al. (2001) on the rheol- the particles reach saturation magnetization, that is,
ogy behavior of MRFs composed of magnetite and car- n = 0. In the saturation regime, the yield stress scales
bonyl iron particles has revealed a linear dependence of with the square of the saturation magnetization of the
yield stress on the volume fraction of the particles, f. particles. For intermediate fields, n = 3=2 has been
They have demonstrated that at low concentration reported in most of the researches.
(f \ 0.2), the yield stress increases linearly with parti- The influence of other effective parameters such as
cle concentrations but it grows faster at high volume volume fraction, size of particles, and medium viscosity
fractions. The effect of magnetic field intensity was also has been less often discussed in the literatures. Hence,
investigated in this study, and they have found the fol- there is a need for more experimental works to investi-
lowing relation for considering the effect of parameters gate their effects more comprehensively. However, in
on the yield stress general, it can be concluded from the open literatures
that at low values of volume fraction, the yield stress
3=2
t y } fm0 Ms1=2 B0 ð39Þ shows linear dependency on the volume fraction. On
the contrary, a power law relation has been proposed
The magnetic induction B0 can be related with the
for this dependency at moderate and high volume frac-
external field H0 by the following equation
tions. The influence of particle size is characterized by l
B0 = m0 (H0 + Ms ) ð40Þ dimensionless number so that the particle size effect can
be neglected at high values of l number. Increasing the
In most of the studies implemented in the field of particle size causes enhancement in the yield stress at
magnetorheology, a linear dependency of the yield low l number. The medium viscosity effects have rarely
stress on the particle volume fraction has been reported been reported, but a power law dependency with the
(Bossis et al., 2003b; Chin et al., 2001; Felt et al., 1996; exponent 0.33 is more common, ty } h0:33 (Ruiz-López
Ginder et al., 1996). However, in some studies, a et al., 2012).

Downloaded from jim.sagepub.com at TULANE UNIV on September 1, 2014


Ghaffari et al. 15

A review on the different rheological and structural hydrodynamic interactions between carrier fluid and
models that are applied in the continuum approach of particles and also particle–particle interactions.
the simulation of MRFs was presented. All models in The computational framework and relevant models
this approach are macro-scale and consider the MR and correlations in discrete modeling approach of
suspension as a single continuum. The continuum MRFs will be discussed below, and also the general
approach suffers from some weaknesses; the rheologi- equation of motion for each particle in a typical MRF
cal constants are derived from fitting the experimental and different effective forces will be introduced.
measurements, consequently their application is limited
to a narrow range of applicability (Ahmadkhanlou
et al., 2010). Most of the rheological models are almost The particles’ motion equation
inaccurate at low values of stress because of assuming a The governing equation for the translational motion of
strict yield stress. particle i with mass mi, based on Newton’s second law,
can be written as
Discrete approach dvi X X
mi = Fijc + Fiknc + Fif + Fiext ð41Þ
The discrete approach, due to its accuracy and funda- dt j k
mental construction, is a growing simulation method
for particulate flows. The flow field of continuous dxi
ui = ð42Þ
phase, motion and interaction of particles in discrete dt
phase, and magnetic field distribution are the three
where Fijc is the contact force acting on particle i by par-
main phenomena which should be considered simulta-
ticle j and Fiknc is the non-contact force acting on parti-
neously in MRF simulation. The discrete approach
cle i by particle k. Indeed, the first two terms on the
based on the basic physical laws can be used to model
right-hand side of equation (41) are the summation of
these three phenomena and their interactions. Modeling
particle–particle interactions. The particle–fluid interac-
in this approach is mostly adopted from the numerical
tion forces are represented as the term Fif , and finally,
simulations of particulate flows and tends to a set of
the term Fiext is the representative of the acting forces
complex equations that need to be solved by the numer-
on particle i by external fields such as gravitational and
ical techniques.
magnetic or electric fields. Also, Brownian forces may
Discrete approach in the simulation of MRFs can
be included in this term (Li et al., 2011). Forces acting
be divided into two general categories according to the
simulation methods of particles’ motion. The first view on particles in discrete approach computational frame-
is adopted from the molecular dynamics method and is work are schematically presented in Figure 6.
mostly known as discrete element method (DEM). In
this vision, Newton’s second law is used to model the Fluid–particle interactions
discrete particles’ motion. The second approach is
based on the kinetic theory which considers the pair of Various forces can be generated due to the interaction
particles as dumbbells (Ahmadkhanlou et al., 2010). In between the particles and the carrier fluid. Some of
both these methods, the hydrodynamic of the sur- these forces are drag force, pressure gradient, virtual
rounding fluid and its interaction with discrete phase mass, Basset, and lift forces (Zhu et al., 2007). Among
can be obtained via solution of Navier–Stokes equa- these forces, the drag force is more significant and com-
tions. It should be noted that the first category is more monly the simple form of Stokes law which has been
convenient and also more reported in the literatures. implemented by the authors (Han et al., 2010b). But
Because of using the basic physical laws, there are
fewer assumptions in the discrete approach modeling,
and consequently, this approach will be more accurate
in comparison with the continuum approach. But high
computational cost is the main problem which has lim-
ited the use of this method.

Discrete approach based on Newton’s second law


In principle, the behavior of any suspension flow can
be obtained by the solution of Newton’s equations of
motion for discrete particles and the Navier–Stokes
equations for continuum fluid together with appropri-
ate boundary and initial conditions. Furthermore, there Figure 6. Forces acting on a particle in the discrete approach
is a need for appropriate relations in order to model computational framework.

Downloaded from jim.sagepub.com at TULANE UNIV on September 1, 2014


16 Journal of Intelligent Material Systems and Structures

other forces, in the spite of their importance in some particles should be included in the fluid flow equations.
specific particulate flows, are usually ignored in the Particle-induced forces can be included as source term
simulation of MRFs. in momentum equation, and also volume fractions can
The drag force coming from the hydrodynamic fric- be considered in Navier–Stokes equations for continu-
tion is the main force acting on particles by the sur- ous phase. Details of coupling schemes and related
rounding fluid. Several models have been developed to equations in particulate flows can be found in Li et al.
calculate accurately the drag force in different particu- (2011) and Zhu et al. (2007).
late flows. For an isolated particle in a fluid, the equa- In most of the studies in the discrete approach for
tion to determine the drag resistance force is well simulation of MRFs, the effect of particles on the flow
described by simple Stokes law as (Li et al., 2011) field has been ignored (Kittipoomwong et al., 2005;
Klingenberg et al., 2010; Mohebi et al., 1996; Ukai and
F drag = 6phaðu‘ (xi )  ui Þf ð43Þ Maekawa, 2004). Nevertheless, there are few studies
that have considered the effect of particle forces on the
here, u‘ (xi ) is the velocity of continuous phase at posi- flow field. Pappas and Klingenberg (2006) studied the
tion xi , ui is the particle’s velocity, and finally f is a cor- MRFs in plane Poiseuille flow with particle-level simu-
rection factor. lations. Indeed, they have used discrete approach
Due to the fact that computational costs must be method in their simulation. They considered the effec-
minimized and also in most cases the MRF flow field is tive forces on the flow field due to the particles as a
laminar and/or stationary (very close to Stokes flow source term in Stokes equation
Re \ 0.1), this straightforward relation ( f = 1) has
been implemented in most of the studies for the simula- hr2 u‘ (x) = rP  fbulk (x) ð45Þ
tion. In most of the reported studies, which have used
discrete approach as their simulation method, the main where fbulk is the body force. Analytical solution of
concern is the equilibrium state of MRF and chain-like equation (45) in the flow direction was presented by
structure of the particles (Han et al., 2010a). Even in Tamura (1992)
some cases where the dynamic behavior of particles
and the fluid flow field were investigated together, the t w 1 X nhyd
u‘x (z) = zðz  Lz Þ + F
flow was also considered in laminar regime; therefore, hLz hLx Ly k k, x

  ð46Þ
the assumption of Stokes drag law is reasonable for zk
these conditions. However, some modifications were 3 1 z  ðz  zk ÞHevðz  zk Þ
Lz
employed in the Stokes equation in order to meet pre-
dictions for different particulate flows well. To achieve where Lx, Ly, and Lz are the lengths of plane in x-, y-,
this aim, a correction factor f is multiplied to Stokes and z-directions, respectively; Hev(z) is the Heaviside
equation (equation (43)) step function; and Fk,nhyd
x denotes the x component of all
non-hydrodynamic forces acting on particle k. Shear
f = CI CC CF ð44Þ stress at the wall (tw ) is calculated as

where CI is the inertial correction factor which  


dp Lz
accounts the effects of inertial forces on the particles. tw =  ð47Þ
dx 2
CC is known as the Cunningham factor and is used to
modify the assumption of no-slip boundary condition Joung and See (2007) have used Rotne–Prager–
which has been implemented in the derivation of Yamakawa (Yamakawa, 1970) tensor to determine the
Stokes equation. The drag force is influenced by the long-range hydrodynamic velocity perturbation of
presence of other particles in the vicinity of the given spheres. Accordingly, in this way, they have considered
particle. This particle hindering effect is addressed the effect of particles on the flow field. Due to the large
through the correction factor CF . There are a wide vari- number of particles, enormous equations are required
ety of expressions for these correction factors for differ- to be solved for the motion of each particle. Therefore,
ent particulate flows under various conditions (Li the resolution of the flow field has to be fine enough to
et al., 2011). resolve the flow of continuum fluid through the pores
In order to quantify the fluid–particle interaction, it among closely spaced particles. As a result, depending
is necessary to solve the fluid flow field coupled with on the time and length scales of interest, simplifications
discrete particle motions. Therefore, understanding the have to be made when this theoretical approach is fol-
fundamentals of the flow, governing equations, and lowed. In most of the practical cases, simplifications in
constitutive relationships is very important. Just as the order to obtain an analytical solution for the flow field
fluid exerts forces on the particles in the flow field, the are not possible. In these cases, numerical methods
momentum is exchanged from the particles to the fluid. such as computational fluid dynamics (CFD) can be
So, these reaction forces and also volume fraction of used. Different coupling schemes such as direct

Downloaded from jim.sagepub.com at TULANE UNIV on September 1, 2014


Ghaffari et al. 17

numerical solution (DNS)–DEM, lattice Boltzmann where rij is the distance between the particle centers and
modeling (LBM)–DEM, CFD–DEM, and large eddy rmin is the minimum distance between the surfaces of
simulation (LES)–DEM have been used by many particles. The parameters k and F0rep should be adjusted
researchers to study the particulate flows (Zhu et al., to equilibrate the repulsive contact force with attractive
2007). CFD–DEM coupling is one of the useful methods magnetic interaction forces (Martin, 2001).
in discrete approach modeling that is used extensively in The classic Hertzian contact model was also used by
the simulation of particulate flows. But this method has some authors (Han et al., 2010a, 2010b)
been less reported in the case of MRF simulations. 0 1
Nevertheless, the other coupling methods like LBM– 4pE
DEM have been implemented to study the MRFs by Fijc =  Fjic = @ d3=2
ij
Aer ,
some authors (Han et al., 2010a; Han et al., 2010b). 3R 1  v2p ð50Þ
 
dij = 2R  rij , rij = xj  xi
Particle–particle interaction forces
where E and vp are Young’s modulus and Poisson’s
A particle in MRF maybe exposed to two types of inter- ratio of the particles, respectively.
action forces with other particles: contact forces and A quasi-hard-sphere model was also employed by
non-contact forces. Contact forces are referred to tan- Martin (2001) to determine the repulsive contact forces
gential and normal forces between two contacting parti- between particles in an electrorheological fluid
cles and are short-range forces. Non-contact forces,
known as long-range forces, include inter-particle forces A
that are applied from a distance. Inter-particle magnetic Fh, s (rij ) =  a , c = 0:97, a=6 ð51Þ
rij  cd
forces are the main non-contact forces acting in MRFs
X X where d is the particle diameter and parameter A is an
FI = Fijc + Fiknc ð48Þ adjustable coefficient.
j k

here, Fijc is the contact force acting on particle i by parti-


Inter-particle magnetic forces
cle j and Fiknc is the non-contact force acting on particle i
by particle k. A magnetic moment can be induced on each magnetiz-
able particle in the presence of external magnetic field.
Consequently, there will be a force acting on each parti-
Inter-particle contact forces
cle as a result of the magnetic fields of adjacent particles
When two particles are close enough to each other, a in the carrier liquid. Precise and also computationally
contact area can be found between them. Contact inter- efficient equations are needed to describe these inter-
actions over this area can be decomposed into a com- particle magnetic forces. Dipole and mutual dipole
ponent in tangential plane and one normal to the plane. models are the most common models in this area. Some
Due to its difficulty in accurately describing these con- relevant magnetic theories and different models for
tact interactions, it is necessary to use adopted simpli- simulation of MRFs are briefly summarized in this sec-
fied models or equations to determine the resulting tion; more details can also be found elsewhere in elec-
forces and torques. Various approaches have been pro- tromagnetic textbooks (Durney and Johnson, 1969;
posed for this aim. The models such as linear spring- Jackson, 1998). A comprehensive review of the existing
dashpot model (Cundall and Strack, 1979) are the most models for describing magnetic interactions in MRFs
intuitive and commonly used. More detailed discus- was also presented in Han et al. (2010b) and Keaveny
sions on the contact models that have been proposed and Maxey (2008).
for various kinds of particulate flows are addressed
elsewhere (Li et al., 2011; Zhu et al., 2007).
Tangential component of contact forces has been Accurate magnetic model. The general solution of poten-
ignored in all reported studies in the field of MRFs and tial f for Laplace equation in a system containing N
just normal component has been included in the mod- particles at a uniform magnetic field was presented in
els. In most investigations, an exponential form of hard Clercx and Bossis (1993). Within each domain of the
sphere interaction force was applied by the authors to particles, Oi , the potential fip takes the form of
describe the normal contact forces (Kittipoomwong X
et al., 2005; Mohebi et al., 1996; Pappas and fip ðri Þ = ci + f0 ðri Þ + bilm ril Ylm ðui , ui Þ
Klingenberg, 2006; Ukai and Maekawa, 2004) ð52Þ
r 2 Oi (i = 1, . . . , N )
 
rij  rmin
Fijrep = F0rep expk er ð49Þ And the potential in fluid domain
rmin

Downloaded from jim.sagepub.com at TULANE UNIV on September 1, 2014


18 Journal of Intelligent Material Systems and Structures

X 1
ff ðri Þ = ci + f0 ðri Þ + Qilm ri(l + 1) Ylm ðui , ui Þ where uij is the angle between particle centerlines and
l( > 0), m
2l + 1 magnetic field direction and H0 is the intensity of mag-
X X 1 netic field. The unit vectors in the r and u directions are
+ Qjlm bilm rj(l + 1) Ylm ðui , ui Þ er and eu , respectively; b is defined in equation (17).
j6¼i l( > 0), m
2l + 1
Multi-pole and multi-body interactions are neglected
ð53Þ in the point dipole approximation. However, it becomes
accurate in the magnetic saturation regime (Pappas and
where ri is the position vector related to the center of
Klingenberg, 2006) and also when the two particles are
particle i; f0 (ri ) is the potential due to the external
far apart. But it becomes inaccurate when the separa-
applied magnetic field; and ci = f0 (xi ) is the external
tion distance of the particles is less than particle dia-
potential at the center of particle i. Ylm (u, u) represents meter (Keaveny and Maxey, 2008). The error is mainly
the spherical harmonics defined with respect to the cen- due to the neglected strong interaction between the two
ter of particle i. bilm are unknown coefficients to be close magnetized particles and reaches maximum when
determined and Qilm are the so-called multi-pole two particles are in contact.
moments. In theory, after determination of bilm , the
inter-particle magnetic forces can then be resolved
according to the integration of magnetic tension over Mutual dipole model. In the mutual dipole model, the
each particle (Han et al., 2010a). induced magnetic fields by other neighbor particles are
In this analytical model, both the multiple particle considered in the magnetization of each spherical parti-
interactions and multi-poles effects are considered in the cle. Therefore, each sphere behaves as though isolated
calculation of the magnetic field. Resolving unknown and immersed in a field whose strength is equal to the
coefficients b and multi-poles Qilm in the series (equation sum of the external field and the fields due to the other
(53)) is the key problem in this approach. Higher multi- beads evaluated at the sphere’s center. An individual
dipole terms could be neglected due to the truncation of particle in mutual approach is still considered as a
series in numerical solutions. High computational cost point dipole, but it is subjected to a secondary magneti-
and tedious computational procedure are considered as zation from the other particles. The mutually magne-
the main limitations of this model. Therefore, accurate tized moment of particle i in a system containing N
magnetic modeling approach is difficult to apply for magnetizable particles in a uniform magnetic field H0 is
MRFs, particularly for cases with large number of parti- given by
cles. The simplified versions for the solution of problems
containing two particles were presented in Keaveny and mp + mf
mi = 4pa3 ðH0 + H(xi )Þ, (i = 1, 2, . . . , N )
Maxey (2008). The numerical solutions for some special mp + 2mf
configurations were also considered in Clercx and Bossis ð56Þ
(1993), Keaveny and Maxey (2008), and Klingenberg
et al. (1991). It should be noted that the computational where H(xi ) represents the total secondary magnetic
cost to obtain a numerical solution for a two-particle field generated by the other magnetized particles and
case is still substantial. can be expressed as (Han et al., 2010a, 2010b; Keaveny
and Maxey, 2008)
Fixed dipole model. The point dipole model is the most  
X
N   X N
1 3^rij mj :^rij  mj
common magnetic interaction model used in the simu- H(xi ) = Hj mj , rij =  3 ð57Þ
lation of MRFs. In this model, the magnetic field pro- j, j6¼i j, j6¼i
4p rij 
duced by other particles is neglected and only the effect
of externally applied magnetic field is considered. where ^rij = ^xi  ^xj . All the mi can be determined by sol-
External magnetic field induces dipole moments on ving a linear system of equations containing equations
each particle. In this approach, each particle is virtually (56) and (57) for N particles. These equations have to
replaced by a point dipole located at its center, thus be solved simultaneously for N particles at each time
inter-particle magnetic forces will be equal to magnetic steps. Unfortunately, the high computational cost asso-
forces between point dipoles. The force on sphere i at ciated with the solution of this linear system of equa-
the origin of a spherical coordinate due to sphere j, in tions has limited the application of this model (Han
the presence of the magnetic field H = H0ez, can be et al., 2010a, 2010b). Although the accuracy in the
expressed as (Joung and See, 2008) mutual dipole model is more improved in comparison
with fixed dipole model, there is still overestimations
 4
d    and/or underestimations in this model, particularly for
Fijd = F0 3 cos2 uij  1 er + sin 2uij eu ð54Þ the contacting particles in suspensions with high vol-
rij
ume fraction of magnetic particles.
3 Some improvements to the mutual dipole model
F0 = pm0 mf d 2 b2 H02 ð55Þ
16 have been presented by Keaveny and Maxey (2008). So

Downloaded from jim.sagepub.com at TULANE UNIV on September 1, 2014


Ghaffari et al. 19

that after the determination of total magnetized Conclusion


moments, the magnetic force between any two particles
The magnetic field, particles’ flow, and continuous fluid
is calculated by incorporating the two-body exact solu-
flow field should be simultaneously handled in model-
tion. Although some additional improvements have
ing and simulation of MRFs. Different modeling meth-
been achieved, the exact solution has still not been
ods have been used to study MRF which can be divided
obtained and the computational cost is also high in this
into two main approaches: continuum and discrete
method. An alternative approximation for improving
approaches. In continuum approach, the MR suspen-
the accuracy but retaining the computational simplicity
of fixed or mutual dipole model is to use some empiri- sion is considered as a one-phase homogeneous fluid,
cal correlations to describe the magnetic interactions and viscoplastic rheological models are used in this
when the particles are close to each other (Han et al., approach to model its motion. Bingham plastic and
2010b; Klingenberg et al., 1991). Herschel–Bulkley models are the most common rheolo-
However, this procedure needs substantial pre- gical models used in continuum approach. The relative
computations and have to be repeated for different sus- simplicity of these rheological models is considered as
ceptibility values, particularly for a large number of their advantages in modeling of MRF. However, their
particles it will be more tedious. Therefore, finding limitation in prediction of MRF behavior for a wide
computationally efficient models that can account for range of operational conditions causes limitation in use.
the accurate magnetic interaction is still open to chal- Many investigations have been carried out by
lenge and further works are needed in this area. researchers to find the relation between the coefficients
of rheological models and the effective parameters of
the MRF. Structural models have been proposed in the
Brownian forces literature to identify the dependency of yield stress and
Due to the large sizes of dispersed particles, Brownian also apparent viscosity of MRF on the applied mag-
forces are usually neglected in the simulation of MRFs. netic field intensity and structural specifications of fluid
Experimental observations show that for usual mag- such as concentration and size of particles. These struc-
netic fields, the magnetic forces dominate the Brownian tural models present good estimations of the fluid beha-
forces (Bossis et al., 2003b; De Vicente et al., 2011). vior in the specified range of applications. However,
Ocalan and McKinley (2013) have declared that the the application of these empirical models will be limited
magnetic forces in common flow situations of MRFs to the narrow range of operational conditions in experi-
dominate over Brownian forces even at the elevated mental measurements. Accordingly, finding the macro-
temperatures. However, in some studies, this force was scale structural models that can be validated for a wide
considered in spite of the large sizes of particles (Peng range of applications is still a challenging matter in the
et al., 2012; Segovia-Gutiérrez et al., 2013a, 2013b). continuum approach. Also, finding a rheological model
The Brownian force, FB , is a vector stochastic force, that can be applied for a wide range of shear rates is in
and the statistics of its component is formulated as demand.
(Drayton and Brady, 1996; Peng et al., 2012) Although the macro-scale models in the continuum
approach are relatively simple for application, they
E½FB (t) = 0, E½FB (t)FB (t + t) = 2Dd(t) ð58Þ have no particle-level insight, they are incapable to dif-
ferentiate between particles and carrier fluid, and they
where t is the correlation time, d is the Dirac delta func- have no explicit relation to fundamental first physical
tion, and D is the translational diffusion coefficient, principles. Therefore, the particle-level simulations (dis-
given by the Stokes–Einstein relation crete approach) will be needed for design and control
kB T of MRF-based devices. Unfortunately, the discrete
D= ð59Þ approach model suffers from high computational cost.
h
Accordingly, the investigation for finding models and
here, kB is Boltzmann’s constant correlations that decrease the computational cost while
(kB = 1:38 3 1023 JK1 ) and T is the absolute tem- increasing the accuracy in the discrete approach is an
perature in Kelvin. open area to research. Even though numerous models
In summary, in this section, we discussed on the dis- for calculation of magnetic forces are proposed, there
crete phase approach in the simulation of MRFs. The still is a demand for models that increase the accuracy
computational framework in this approach and the and applicability in the numerical solutions. In all the
necessary models were illustrated. Different forces’ reported MRF simulations that have used discrete
equations and models that have used by the authors approach, the classical Hertzian contact model was
for the prediction of different interactions in a common implemented. Using the other contact models may
MRF were also reviewed. A brief summary of different improve the accuracy of predictions. In coupling the
researches of MRFs that have implemented the discrete fluid flow and particle motions, only the drag force is
approach is tabulated in Table 5. considered. However, a sensitive analysis on the

Downloaded from jim.sagepub.com at TULANE UNIV on September 1, 2014


20 Journal of Intelligent Material Systems and Structures

Table 5. Different studies in the field of using discrete approach simulation for MRFs.

Reference Description of models Case study

Ahmadkhanlou et al. (2010) Particle dynamics: kinetic theory–based Shear mode, carbonyl iron powder in silicone
model; hydrodynamic: Maxwell velocity oil, comparison with two rheological models:
distribution; internal forces: linear spring Bingham and Herschel–Bulkley
model; magnetic forces: fixed dipole
Han et al. (2010b) Particle dynamics: Newton’s second law, Shear mode. Representative volume element
discrete element approach; hydrodynamic: of an MRF
LBM; contact forces: Hertzian contact model;
magnetic forces: mutual dipole model
Feng et al. (2010) Particle dynamics: Newton’s second law, Combined three-dimensional LBM and
discrete element approach; hydrodynamic: discrete element method for modeling fluid–
LBM; contact forces: Hertzian contact model particle interactions.
Ciocanel et al. (2006) Particle dynamics: kinetic theory–based Shear and elongational flows
model; hydrodynamic: Stokes drag law;
internal forces: linear; magnetic forces: fixed
dipole model
Tao (2006) Particle dynamics: Newton’s second law, Shear mode, representative volume element
discrete element approach; hydrodynamic:
stokes drag law; contact forces: not
mentioned; magnetic forces: fixed dipole
model
Brigadnov and Dorfmann (2005) Full system of equations for moving isotropic The flow of an MRF between two parallel
MRFs in an electromagnetic field is presented fixed plates under the influence of a constant
magnetic field perpendicular to the flow
direction was considered
Cook et al. (2004) Particle dynamics: Newton’s second law, A direct two-dimensional simulation method
discrete element approach; hydrodynamic: for particle–fluid systems without magnetic
LBM; internal forces: linear force
Ly et al. (1999) Particle dynamics: Newton’s second law; A two-dimensional rectangular framework in
hydrodynamic: Stokes drag; contact forces: valve mode was simulated
hard sphere rejection; magnetic forces: multi-
pole model

MRF: magnetorheological fluid; LBM: lattice Boltzmann modeling.

different forces acting on particles can declare the con- for Telerobotic Systems. Columbus, OH: Mechanical Engi-
tribution of each force. This will be helpful for the deci- neering Graduate Program, The Ohio State University, p.
sion to ignore or to include such forces. In most of the 207.
studies, the simple Stokes drag model is applied which Ahmadkhanlou F, Mahboob M, Bechtel S, et al. (2010) An
will not be so accurate in the flows with high gradients improved model for magnetorheological fluid-based actua-
tors and sensors. Journal of Intelligent Material Systems
of velocity.
and Structures 21: 3–18.
Berli CLA and De Vicente J (2012) A structural viscosity
Acknowledgements model for magnetorheology. Applied Physics Letters 101:
021903.
The authors gratefully acknowledge financial support from
Bica I (2012) The influence of the magnetic field on the elastic
the Iran National Science Foundation (INSF).
properties of anisotropic magnetorheological elastomers.
Journal of Industrial and Engineering Chemistry 18:
Declaration of conflicting interests 1666–1669.
The authors declared no potential conflicts of interest with Bica I, Liu YD and Choi HJ (2013) Physical characteristics of
respect to the research, authorship, and/or publication of this magnetorheological suspensions and their applications.
article. Journal of Industrial and Engineering Chemistry 19:
394–406.
Blake J and Gurocak HB (2009) Haptic glove with MR
References brakes for virtual reality. IEEE/ASME Transactions on
Agustı́n-Serrano R, Donado F and Rubio-Rosas E (2013) Mechatronics 14: 606–615.
Magnetorheological fluid based on submicrometric silica- Böse H, Rabindranath R and Ehrlich J (2012) Soft magnetor-
coated magnetite particles under an oscillatory magnetic heological elastomers as new actuators for valves. Journal
field. Journal of Magnetism and Magnetic Materials 335: of Intelligent Material Systems and Structures 23: 989–994.
149–158. Bossis G and Cebers A (1999) Effects of the magnetodipolar
Ahmadkhanlou F (2008) Design, Modeling and Control of interactions in the alternating magnetic fields. Journal of
Magnetorheological Fluid-Based Force Feedback Dampers Magnetism and Magnetic Materials 201: 218–221.

Downloaded from jim.sagepub.com at TULANE UNIV on September 1, 2014


Ghaffari et al. 21

Bossis G, Khuzir P, Lacis S, et al. (2003a) Yield behavior of carbon nanotube additive and its yield stress scaling func-
magnetorheological suspensions. Journal of Magnetism tion. Colloids and Surfaces A: Physicochemical and Engi-
and Magnetic Materials 258–259: 456–458. neering Aspects 351: 46–51.
Bossis G, Lacis S, Meunier A, et al. (2002) Magnetorheologi- Farjoud A, Ahmadian M, Mahmoodi N, et al. (2011) Non-
cal fluids. Journal of Magnetism and Magnetic Materials linear modeling and testing of magneto-rheological fluids
252: 224–228. in low shear rate squeezing flows. Smart Materials and
Bossis G, Volkova O, Lacis S, et al. (2003b) Magnetorheol- Structures 20: 085013 (14 pp.).
ogy: fluids, structures and rheology. In: Odenbach S (ed.) Farjoud A, Vahdati N and Fah YF (2008) Mathematical
Ferrofluids. Berlin, Heidelberg: Springer, pp. 202–230. model of drum-type MR brakes using Herschel-Bulkley
Brigadnov IA and Dorfmann A (2005) Mathematical model- shear model. Journal of Intelligent Material Systems and
ing of magnetorheological fluids. Continuum Mechanics Structures 19: 565–572.
and Thermodynamics 17: 29–42. Felt DW, Hagenbuchle M, Liu J, et al. (1996) Rheology of a
Carmignani C, Forte P and Rustighi E (2006) Design of a magnetorheological fluid. Journal of Intelligent Material
novel magneto-rheological squeeze-film damper. Smart Systems and Structures 7: 589–593.
Materials and Structures 15: 164. Feng YT, Han K and Owen DRJ (2010) Combined three-
Chin BD, Park JH, Kwon MH, et al. (2001) Rheological dimensional lattice Boltzmann method and discrete ele-
properties and dispersion stability of magnetorheological ment method for modelling fluid–particle interactions with
(MR) suspensions. Rheologica Acta 40: 211–219. experimental assessment. International Journal for Numeri-
Choi YT, Cho JU, Choi SB, et al. (2005) Constitutive models cal Methods in Engineering 81: 229–245.
of electrorheological and magnetorheological fluids using Firouzi M and Hashemabadi SH (2008) Analytical solution
viscometers. Smart Materials and Structures 14: 1025. for Newtonian–Bingham plastic two-phase pressure driven
Ciocanel C, Lipscomb G and Naganathan NG (2006) A par- stratified flow through the circular ducts. International
ticle pair model for magnetorheological fluids. Proceedings Communications in Heat and Mass Transfer 35: 666–673.
of SPIE 6170: 198–206. Firouzi M and Hashemabadi SH (2009) Exact solution of two
Claracq J, Sarrazin J and Montfort J-P (2004) Viscoelastic phase stratified flow through the pipes for non-Newtonian
properties of magnetorheological fluids. Rheologica Acta Herschel–Bulkley fluids. International Communications in
43: 38–49. Heat and Mass Transfer 36: 768–775.
Clercx HJH and Bossis G (1993) Many-body electrostatic Furst EM and Gast AP (2000) Micromechanics of magnetor-
interactions in electrorheological fluids. Physical Review E heological suspensions. Physical Review E 61: 6732–6739.
48: 2721–2738. Gandhi F and Bullough W (2005) On the phenomenological
Cook BK, Noble DR and Williams JR (2004) A direct simula- modeling of electrorheological and magnetorheological
tion method for particle-fluid systems. Engineering Compu- fluid preyield behaviour. Journal of Intelligent Material
tations 21: 151–168. Systems and Structures 16: 237–248.
Cundall PA and Strack ODL (1979) A discrete numerical Genc x S and Phulé PP (2002) Rheological properties of mag-
model for granular assemblies. Geotechnique 29: 47–65. netorheological fluids. Smart Materials and Structures 11:
De Gans BJ, Duin NJ, Ende Dvd, et al. (2000) The influence 140–146.
of particle size on the magnetorheological properties of an Ginder JM, Davis LC and Elie LD (1996) Rheology of mag-
inverse ferrofluid. Journal of Chemical Physics 113: netorheological fluids: models and measurements. Interna-
2032–2042. tional Journal of Modern Physics B 10: 3293–3303.
De Gans B-J, Hoekstra H and Mellema J (1999) Non-linear Goldasz J and Sapinski B (2012) Nondimensional characteri-
magnetorheological behaviour of an inverse ferrofluid. zation of flow-mode magnetorheological/electrorheologi-
Faraday Discussions 112: 209–224. cal fluid dampers. Journal of Intelligent Material Systems
De Vicente J and Berli CA (2013) Aging, rejuvenation, and and Structures 23: 1545–1562.
thixotropy in yielding magnetorheological fluids. Rheolo- Goncalves FD and Ahmadian M (2005) A study on MR
gica Acta 52: 467–483. fluids subject to high shear rates and high velocities. Smart
De Vicente J, Klingenberg DJ and Hidalgo-Alvarez R (2011a) structures and Materials 2005: Damping and Isolation 5760:
Magnetorheological fluids: a review. Soft Matter 7: 46–56.
3701–3710. Goncalves FD, Koo JH and Ahmadian M (2006) A review of
De Vicente J, López-López MT, Durán JDG, et al. (2004) the state of the art in magnetorheological fluid
Shear flow behavior of confined magnetorheological fluids technologies—part I: MR fluid and MR fluid models. The
at low magnetic field strengths. Rheologica Acta 44: Shock and Vibration Digest 38: 203–219.
94–103. Grunwald A and Olabi AG (2008) Design of magneto-
De Vicente J, Ruiz-López JA, Andablo-Reyes E, et al. rheological (MR) valve. Sensors and Actuators A: Physical
(2011b) Squeeze flow magnetorheology. Journal of Rheol- 148: 211–223.
ogy 55: 753–780. Han K, Feng YT and Owen DRJ (2010a) Modelling of mag-
Drayton YB and Brady JF (1996) Brownian electrorheologi- netorheological fluids with combined lattice Boltzmann
cal fluids as a model for flocculated dispersions. Journal of and discrete element approach. Communications in Compu-
Rheology 40: 1027–1057. tational Physics 7: 1095–1117.
Durney CH and Johnson CC (1969) Introduction to Modern Han K, Feng YT and Owen DRJ (2010b) Three-dimensional
Electromagnetics. New York: McGraw-Hill. modelling and simulation of magnetorheological fluids.
Fang FF, Choi HJ and Jhon MS (2009) Magnetorheology of International Journal for Numerical Methods in Engineering
soft magnetic carbonyl iron suspension with single-walled 84: 1273–1302.

Downloaded from jim.sagepub.com at TULANE UNIV on September 1, 2014


22 Journal of Intelligent Material Systems and Structures

Hashemabadi SH and Mirnajafizadeh SM (2010) Analysis of modes. Journal of Intelligent Material Systems and Struc-
viscoelastic fluid flow with temperature dependent proper- tures 14: 99–104.
ties in plane Couette flow and thin annuli. Applied Mathe- Lange U, Richter L and Zipser L (2001) Flow of magnetor-
matical Modelling 34: 919–930. heological fluids. Journal of Intelligent Material Systems
Hashemabadi SH, Etemad SG, Thibault J, et al. (2003) Ana- and Structures 12: 161–164.
lytical solution for dynamic pressurization of viscoelastic Laun HM, Kormann C and Willenbacher N (1996) Rheome-
fluids. International Journal of Heat and Fluid Flow 24: try on magnetorheological (MR) fluids. Rheologica Acta
137–144. 35: 417–432.
Huang J, Ma J and Wei S (2012) Viscoplastic flow of the mag- Lemaire E, Bossis G and Grasselli Y (1993) Yield stress and
netorheological fluid in a cylindrical-type clutch. Interna- structuration of magnetorheological suspensions. Journal
tional Journal of Advancements in Computing Technology of Magnetism and Magnetic Materials 122: 51–52.
(IJACT) 4: 458–465. Lemaire E, Grasselli Y and Bossis G (1992) Field induced
Jackson JD (1998) Classical Electrodynamics. New York: structure in magneto and electro-rheological fluids. Jour-
Wiley. nal de Physique II France 2: 359–369.
Ji Eun K, Jae-Do K, Ying Dan L, et al. (2012) Effect of Lemaire E, Meunier A, Bossis G, et al. (1995) Influence of the
medium oil on magnetorheology of soft carbonyl iron par- particle size on the rheology of magnetorheological fluids.
ticles. IEEE Transactions on Magnetics 48: 3442–3445. Journal of Rheology 39: 1011–1021.
Joung C and See H (2007) Simulation of magneto-rheological Li S, Marshall JS, Liu G, et al. (2011) Adhesive particulate
fluids incorporating hydrodynamic effects. Journal of Cen- flow: the discrete-element method and its application in
tral South University of Technology 14: 271–274. energy and environmental engineering. Progress in Energy
Joung CG and See H (2008) The influence of wall interaction and Combustion Science 37: 633–668.
on dynamic particle modelling of magneto-rheological sus- Li WH, Du H, Chen G, et al. (2002) Experimental investiga-
pensions between shearing plates. Rheologica Acta 47: tion of creep and recovery behaviors of magnetorheologi-
917–927. cal fluids. Materials Science and Engineering: A 333:
Keaveny EE and Maxey MR (2008) Modeling the magnetic 368–376.
interactions between paramagnetic beads in magnetorheo- Ly HV, Reitich F, Jolly MR, et al. (1999) Simulations of par-
logical fluids. Journal of Computational Physics 227: ticle dynamics in magnetorheological fluids. Journal of
9554–9571. Computational Physics 155: 160–177.
Kim I, Song K, Park B, et al. (2011) Nano-sized Fe soft- Ma J, Shu H and Huang J (2012) MR continuously variable
magnetic particle and its magnetorheology. Colloid and transmission driven by SMA for centrifugal fan in nuclear
Polymer Science 289: 79–83. power plant. Science and Technology of Nuclear Installa-
Kittipoomwong D, Klingenberg DJ and Ulicny JC (2005) tions 2012: 205675 (6 pp.).
Dynamic yield stress enhancement in bidisperse magnetor- Martin JE (2001) Field-induced rheology in uniaxial and biax-
heological fluids. Journal of Rheology 49: 1521–1538. ial fields. International Journal of Modern Physics B 15:
Klingenberg DJ and Ulicny JC (2011) Enhancing magnetor- 574–595.
heology. International Journal of Modern Physics B Martin JE and Anderson RA (1996) Chain model of electro-
(IJMPB) 25: 911–917. rheology. Journal of Chemical Physics 104: 4814–4827.
Klingenberg DJ, Olk CH, Golden MA, et al. (2010) Effects Melrose JR and Heyes DM (1993) Simulations of electro-
of nonmagnetic interparticle forces on magnetorheological rheological and particle mixture suspensions: agglomerate
fluids. Journal of Physics: Condensed Matter 22: 324101. and layer structures. Journal of Chemical Physics 98:
Klingenberg DJ, Swol Fv and Zukoski CF (1991) The small 5873–5886.
shear rate response of electro rheological suspensions. I. Mitsoulis E (2007) Flows of viscoplastic materials: models
Simulation in the point-dipole limit. Journal of Chemical and computations. In: Binding DM, Hudson NE, Keun-
Physics 94: 6160–6170. ings R (eds) Rheology reviews. Glasgow, Universities
Klingenberg DJ, Ulicny JC and Golden MA (2007a) Mason Design & Print, pp 135 –178.
numbers for magnetorheology. Journal of Rheology 51: Mohebi M, Jamasbi N and Liu J (1996) Simulation of the for-
883–893. mation of nonequilibrium structures in magnetorheologi-
Klingenberg DJ, Ulicny JC and Smith AL (2007b) Effects of cal fluids subject to an external magnetic field. Physical
body forces on the structure and rheology of ER and MR Review E 54: 5407–5413.
fluids. International Journal of Modern Physics B (IJMPB) Mrlik M, Sedlacik M, Pavlinek V, et al. (2013) Synthesis and
21: 4841–4848. magnetorheological characteristics of ribbon-like,
Kordonski W and Shorey A (2007) Magnetorheological
polypyrrole-coated carbonyl iron suspensions under oscil-
(MR) jet finishing technology. Journal of Intelligent Mate-
latory shear. Journal of Applied Polymer Science 128:
rial Systems and Structures 18: 1127–1130.
2977–2982.
Kormann C, Laun HM and Richter HJ (1996) MR fluids
Nicholas CR and Norman MW (2004) Volume-constrained
with nano-sized magnetic particles. International Journal
optimization of magnetorheological and electrorheological
of Modern Physics B 10: 3167–3172.
valves and dampers. Smart Materials and Structures 13:
Kostamo E, Kostamo J, Kajaste J, et al. (2012) Magnetor-
1303.
heological valve in servo applications. Journal of Intelli-
Nosse DT and Dapino MJ (2007) Magnetorheological valve
gent Material Systems and Structures 23: 1001–1010.
for hybrid electrohydrostatic actuation. Journal of Intelli-
Kulkarni P, Ciocanel C, Vieira SL, et al. (2003) Study of the
gent Material Systems and Structures 18: 1121–1136.
behavior of MR fluids in squeeze, torsional and valve

Downloaded from jim.sagepub.com at TULANE UNIV on September 1, 2014


Ghaffari et al. 23

Ocalan M and McKinley G (2013) High-flux magnetorheol- Sahin H, Gordaninejad F, Wang X, et al. (2012) Response
ogy at elevated temperatures. Rheologica Acta 52: time of magnetorheological fluids and magnetorheological
623–641. valves under various flow conditions. Journal of Intelligent
Olabi AG and Grunwald A (2007) Design and application of Material Systems and Structures 23: 949–957.
magneto-rheological fluid. Materials & Design 28: Salloom MY and Samad Z (2011) Finite element modeling
2658–2664. and simulation of proposed design magneto-rheological
Omidbeygi F and Hashemabadi SH (2012) Experimental valve. International Journal of Advanced Manufacturing
study and CFD simulation of rotational eccentric cylinder Technology 54: 421–429.
in a magnetorheological fluid. Journal of Magnetism and Salloom MY and Samad Z (2012) Design and modeling mag-
Magnetic Materials 324: 2062–2069. netorheological directional control valve. Journal of Intelli-
Omidbeygi F and Hashemabadi SH (2013) Exact solution gent Material Systems and Structures 23: 155–167.
and CFD simulation of magnetorheological fluid purely Santiago-Quinones D, Raj K and Rinaldi C (2013) A com-
tangential flow within an eccentric annulus. International parison of the magnetorheology of two ferrofluids with
Journal of Mechanical Sciences 75: 26–33. different magnetic field-dependent chaining behavior.
Papanastasiou TC (1987) Flows of materials with yield. Jour- Rheologica Acta 52: 719–726.
nal of Rheology 31: 385–405. Sedlacik M, Pavlinek V, Vyroubal R, et al. (2013) A
Pappas Y and Klingenberg DJ (2006) Simulations of magne- dimorphic magnetorheological fluid with improved oxida-
torheological suspensions in Poiseuille flow. Rheologica tion and chemical stability under oscillatory shear. Smart
Acta 45: 621–629. Materials and Structures 22: 035011.
Park BJ, Fang FF and Choi HJ (2010) Magnetorheology: See H (2003) Field dependence of the response of a magnetor-
materials and application. Soft Matter 6: 5246–5253. heological suspension under steady shear flow and squeez-
Park B, Park B, Hato M, et al. (2011) Soft magnetic carbonyl ing flow. Rheologica Acta 42: 86–92.
iron microsphere dispersed in grease and its rheological See H and Tanner R (2003) Shear rate dependence of the nor-
characteristics under magnetic field. Colloid and Polymer mal force of a magnetorheological suspension. Rheologica
Science 289: 381–386. Acta 42: 166–170.
Parlak Z and Engin T (2012) Time-dependent CFD and See H, Mackenzie S and Chua BT (2006) Effect of compres-
quasi-static analysis of magnetorheological fluid dampers sion on the response of a magneto-rheological suspension.
with experimental validation. International Journal of Korea-Australia Rheology Journal 18: 121–126.
Mechanical Sciences 64: 22–31. Segovia-Gutiérrez JP, Berli CLA and de Vicente J (2012)
Parlak Z, Engin T and C xallıI_ (2012) Optimal design of MR Nonlinear viscoelasticity and two-step yielding in magne-
damper via finite element analyses of fluid dynamic and torheology: a colloidal gel approach to understand the
magnetic field. Mechatronics 22: 890–903. effect of particle concentration. Journal of Rheology 56:
Peng Y-B, Ghanem R and Li J (2012) Investigations of micro- 1429–1449.
structured behaviors of magnetorheological suspensions. Segovia-Gutiérrez JP, De Vicente J, Hidalgo-Alvarez R, et al.
Journal of Intelligent Material Systems and Structures 23: (2013a) Brownian dynamic simulations and experiments
1351–1370. of MR fluids. Journal of Physics Conference Series 412:
Phulé PP and Ginder JM (1999) Synthesis and properties of 012056.
novel magnetorheological fluids having improved stability Segovia-Gutiérrez JP, De Vicente J, Hidalgo-Alvarez R, et al.
and redispersibility. International Journal of Modern Phy- (2013b) Brownian dynamics simulations in magnetorheol-
sics B 13: 2019–2027. ogy and comparison with experiments. Soft Matter 9:
Promislow JHE and Gast AP (1996) Magnetorheological 6970–6977.
fluid structure in a pulsed magnetic field. Langmuir 12: Sheng R, Flores GA and Liu J (1999) In vitro investigation of
4095–4102. a novel cancer therapeutic method using embolizing prop-
Quoc-Hung N, Young-Min H, Seung-Bok C, et al. (2007) erties of magnetorheological fluids. Journal of Magnetism
Geometry optimization of MR valves constrained in a spe- and Magnetic Materials 194: 167–175.
cific volume using the finite element method. Smart Mate- Sim H, Kwon S and Choi H (2013) Xanthan gum-coated soft
rials and Structures 16: 2242. magnetic carbonyl iron composite particles and their mag-
Rabinow J (1948) The magnetic fluid clutch. Transactions of netorheology. Colloid and Polymer Science 291: 963–969.
the American Institute of Electrical Engineers (AIEE) 67: Susan-Resiga D (2009) A rheological model for magneto-
1308–1315.
rheological fluids. Journal of Intelligent Material Systems
Ramos J, Klingenberg DJ, Hidalgo-Alvarez R, et al. (2011)
and Structures 20: 1001–1010.
Steady shear magnetorheology of inverse ferrofluids. Jour-
Tamura H (1992) Electro-rheological effects in a flow through
nal of Rheology 55: 127–152.
a narrow slit. Journal of the Physical Society of Japan 61:
Rankin PJ, Horvath AT and Klingenberg DJ (1999) Magnetor-
3984–3993.
heology in viscoplastic media. Rheologica Acta 38: 471–477.
Tang X and Conrad H (2000) An analytical model for magne-
Ruiz-López JA, Hidalgo-Alvarez R and de Vicente J (2012)
torheological fluids. Journal of Physics D: Applied Physics
On the validity of continuous media theory for plastic
33: 3026.
materials in magnetorheological fluids under slow com-
Tao R (2006) Structure and dynamics of dipolar fluids under
pression. Rheologica Acta 51: 595–602.
strong shear. Chemical Engineering Science 61: 2186–2190.
Ruiz-López JA, Hidalgo-Alvarez R and de Vicente J (2013)
Ukai T and Maekawa T (2004) Patterns formed by paramag-
Continuous media theory for MR fluids in non-shearing
netic particles in a horizontal layer of a
flows. Journal of Physics Conference Series 412: 012057.

Downloaded from jim.sagepub.com at TULANE UNIV on September 1, 2014


24 Journal of Intelligent Material Systems and Structures

magnetorheological fluid subjected to a dc magnetic field. Yamakawa H (1970) Transport properties of polymer chains
Physical Review E 69: 032501. in dilute solution: hydrodynamic interaction. Journal of
Ulicny JC, Golden MA, Namuduri CS, et al. (2005) Transient Chemical Physics 53: 436–444.
response of magnetorheological fluids: shear flow between Yang T-H, Koo J-H, Kim S-Y, et al. (2012) Application of
concentric cylinders. Journal of Rheology 49: 87–104. magnetorheological fluids for a miniature haptic button:
Vereda F, De Vicente J, Segovia-Gutiérrez JP, et al. (2011) experimental evaluation. Journal of Intelligent Material
Average particle magnetization as an experimental scaling Systems and Structures 23: 1025–1031.
parameter for the yield stress of dilute magnetorheological Yang T-H, Kwon H-J, Lee SS, et al. (2009) Conceptual design
fluids. Journal of Physics D: Applied Physics 44: 425002. of miniature tunable stiffness display using MR fluids. In:
Volkova O, Bossis G, Guyot M, et al. (2000) Magnetorheol- TRANSDUCERS 2009: international solid-state sensors,
ogy of magnetic holes compared to magnetic particles. actuators and microsystems conference, Denver, CO, 21–25
Journal of Rheology 44: 91–104. June, pp. 897–899. New York: IEEE.
Volkova O, Cutillas S and Bossis G (1999a) Shear banded Yang T-H, Kwon H-J, Lee SS, et al. (2010) Development of
flows and nematic-to-isotropic transition in ER and MR a miniature tunable stiffness display using MR fluids for
fluids. Physical Review Letters 82: 233–236. haptic application. Sensors and Actuators A: Physical 163:
Volkova O, Cutillas S, Carletto P, et al. (1999b) Flow-induced 180–190.
structures in magnetorheological suspensions. Journal of Zafarani-Moattar M and Majdan-Cegincara R (2013) Stabi-
Magnetism and Magnetic Materials 201: 66–69. lity, rheological, magnetorheological and volumetric char-
Wang J, Meng G, Feng N, et al. (2005) Dynamic performance acterizations of polymer based magnetic nanofluids.
and control of squeeze mode MR fluid damper–rotor sys- Colloid and Polymer Science 291: 1977–1987.
tem. Smart Materials and Structures 14: 529. Zhu HP, Zhou ZY, Yang RY, et al. (2007) Discrete particle
Wang X and Gordaninejad F (2006) Study of magnetorheo- simulation of particulate systems: theoretical develop-
logical fluids at high shear rates. Rheologica Acta 45: ments. Chemical Engineering Science 62: 3378–3396.
899–908. Zhu X, Jing X and Cheng L (2012) Magnetorheological fluid
dampers: a review on structure design and analysis. Journal
of Intelligent Material Systems and Structures 23: 839–873.

Downloaded from jim.sagepub.com at TULANE UNIV on September 1, 2014

You might also like