Journal Pre-proof

Evaluation of Nonlinear Dynamic Phenomena in the Hysteretic

Behaviour of Magnetorheological Dampers

Wael Elsaady , S Olutunde Oyadiji , Adel Nasser

PII: S2666-4968(20)30019-4
DOI: https://doi.org/10.1016/j.apples.2020.100019
Reference: APPLES 100019

To appear in: Applications in Engineering Science

Received date: 17 June 2020

Revised date: 23 September 2020
Accepted date: 24 September 2020

Please cite this article as: Wael Elsaady , S Olutunde Oyadiji , Adel Nasser , Evaluation of Nonlinear
Dynamic Phenomena in the Hysteretic Behaviour of Magnetorheological Dampers, Applications in
Engineering Science (2020), doi: https://doi.org/10.1016/j.apples.2020.100019

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.

© 2020 Published by Elsevier Ltd.

This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/)
 1

Evaluation of Nonlinear Dynamic Phenomena in the Hysteretic

Behaviour of Magnetorheological Dampers
Wael Elsaady1, 2, *, S Olutunde Oyadiji1, Adel Nasser1

1
Department of Mechanical, Aerospace and Civil Engineering, Faculty of Science and
Engineering, The University of Manchester, M13 9PL, Manchester, UK
2
Mechanical Engineering Branch, Military Technical College, Cairo 11838, Egypt
*
Author for correspondence; Email: [email protected],
[email protected]
 2

Evaluation of Nonlinear Dynamic Phenomena in the Hysteretic

Behaviour of Magnetorheological Dampers

Abstract

The performance of a commercially-available magnetorheological (MR) damper is modelled

via a one-way coupled numerical viscoelastic-viscoplastic approach. The approach adopts a
Finite Element Analysis (FEA) of the magnetic circuit of the damper and a transient
Computational Fluid Dynamics (CFD) analysis of fluid flow. The apparent viscosity of the
MR fluid is defined as a function of the magnetic field intensity and local shear rate. The
effects of different sources of the hysteretic behaviour of MR dampers, namely: fluid
compressibility, fluid inertia, viscoelasticity and friction are investigated. Moreover, the
effect of employing different rheological models in the numerical approach is investigated.
The results indicate that that the effects of fluid compressibility and dynamic friction are the
main sources of the hysteretic behaviour of the damper. Also, the results show the proper
selection of the rheological model employed in the numerical approach is critical, as it leads
to major differences in the predictions of the numerical approach.

Keywords

MR fluids, MR dampers, finite element analysis, finite volume analysis, computational fluid
dynamics, rheology.
 3

1 Introduction

Magnetorheological (MR) dampers are a class of semi-active smart devices that manifest
highly nonlinear behaviours due to the interactions between multi-physics phenomena that
combine structural mechanics, magnetism and rheological fluid flow [1]. MR dampers are
employed in different mechanical systems such as luxury cars, buildings, bridges and railway
systems, prosthesis, and aircraft mechanisms [2-5].

Modelling of the rheological flow in MR fluid applications is thought to be complicated, as it

involves interactions between magnetic and dynamic flow characteristics that are affected by
different sources of nonlinearity, such as magnetic saturation, non-Newtonian viscoelastic-
viscoplastic flow, temperature, fluid inertia, and compressibility [6-8]. The rheological flow
is affected by the magnetic field, which is not homogenous in the whole fluid domain in the
damper. Rather, it is localised in the throttling area of the damper. This inhomogeneity of the
distribution of magnetic field brings an additional difficulty in the definition of the
characteristics of rheological flow in modelling approaches [9, 10]. Therefore, most of the
existing models of MR dampers in the literature adopt analytical models and solutions that do
not depend on modelling of that rheological flow. Rather, most of these models depend on
damper representation by a set of springs, dashpots, and other elements that represent
nonlinearity [11]. Studies on the numerical modelling of MR fluid devices are uncommon
compared to those that employ analytical methods [1, 10].

Regarding numerical modelling approaches, Parlak et al. [12] presented a numerical approach
for modelling the viscoplastic fluid flow in MR dampers. Gurubasavaraju et al. [6] employed
a similar technique for the same purpose. Both techniques employed a Finite Element (FE)
model for the solution of magnetic field in conjunction with a Computational Fluid Dynamics
(CFD) model that investigates the characteristics of rheological flow. In those two papers, the
effect of the magnetic field was only defined at the locations of magnetic poles that impose a
uniform magnetic field. However, the effect of the non-uniform distribution of the magnetic
field along the throttling area of the damper was recently considered in [13], in which the
enhanced model shows a better accuracy relative to the experimental data.

Case et al. [14] developed a two-way coupled approach for a small-scale MR damper used in
an orthotics system. For a large-scale MR damper, Zheng et al. [15] used a two-way coupled
technique in which they included the effects of fluid temperature. The flow of MR fluid in
these studies was assumed as incompressible viscoplastic flow. Therefore, the models were
shown to be incapable of representing the hysteretic behaviours of the dampers, which is
 4

mainly caused by the effects of fluid compressibility, friction, inertia, viscoelasticity, and
presence of a gas chamber in common designs of MR dampers [16, 17]. Recently, an FE
approach has been presented in [7, 8], in which the compressibility and inertial effects were
modelled and found to be mainly responsible for the hysteretic behaviours of the dampers.
However, not many studies investigated other sources of nonlinearity such as the viscoelastic
effects, presence of gas pocket, friction, and other sources of inertia.

Unlike common analytical models of MR dampers, this paper presents a one-way coupled
numerical viscoelastic-viscoplastic approach for the analysis of a commercially-available MR
damper. The results of the numerical approach have been validated by experimental
measurements. The paper aims to model and evaluate the different phenomena that are
reported to cause the hysteretic behaviour of MR dampers such as fluid compressibility,
inertia, friction and viscoelasticity by using the numerical approach. In most studies that
employ numerical techniques, the effect of viscoelasticity and fluid compressibility are
neglected [3, 6, 12, 14, 18-21], whereas the current study accounts for these effects. Besides,
it accounts for the other sources of nonlinearity caused by the presence of gas pocket in the
damper and inertial effects of fluid due to piston motion with higher velocity. Moreover, the
numerical approach has been used to evaluate the hysteretic behaviour of MR dampers due to
dynamic friction force.

The paper is organised as follows: the construction and experimental characterisation of the
MR damper under investigation in this study are presented in Section 2. Then, the method of
the numerical viscoelastic-viscoplastic approach and the coupling technique are illustrated in
Section 3, followed by the validation of the model and the study of the effects of the variation
of different parameters in Section 4. Afterwards, the effect of dynamic friction caused by the
sealing glands of the damper is evaluated using the current theoretical and experimental
methods, as shown in Section 5. Finally, the conclusions are drawn in Section 6.

2 Construction and experimental characterisation of the MR damper

The MR damper under investigation in this study is the short-stroke MR damper, model RD-
8040-1, produced by LORD Corporation. It is a mono-tube MR damper with an air chamber
appended to the damper to work as an accumulator. The chamber contains air under certain
pressure, and is separated from the fluid domain by means of a diaphragm. The gas pre-
charge pressure is required to reduce the effects of cavitation [22]. The construction and main
dimensions of the damper are presented in [23, 24], and also shown in Figure 1. The piston
 5

provides the housing for an electromagnetic coil that generates the magnetic field in the
piston, as shown in Figure 1(b).

(a) (b)

Figure 1: Construction of the RD-8040-1 MR damper: (a) pictorial view [3], and (b)
schematic cross-sectional view [24].

The dynamic characteristics of the damper were measured by the Electro-Servo-Hydraulic

(ESH) machine, which applies sinusoidal excitations to one end of the damper and measures
the output force on the other blocked end. Therefore, the kinetics of the damper in terms of
force, velocity and displacement were determined at different excitation currents applied to
the damper, as shown in Figure 2. It is seen that the output force of the damper increases
significantly with the increase of input current. While the piston is operated at 1 A, which is
the maximum operating current of the damper according to the damper datasheet [22], the
maximum force of the damper was found to be approximately 20 times greater than the off-
state force of the damper. The current experimental data were found to match those measured
for the same damper in [25], and they were used to validate the mathematical approach
presented in the current study.
 6

Figure 2: Dynamic characteristics of the RD-8040-1 MR damper: (up) force-time histories in

a complete cycle at different currents, = 1 (Hz), = 3 (mm), (down-left) Force-
displacement diagrams, and (down-right) force-velocity diagrams.

It is worth mentioning that the measurements of the dynamic characteristics of the damper
shown in Figure 2 do not include the static force of the damper caused by the pre-charge
pressure in the air chamber. The damper was installed on the ESH test machine so that the
piston was initially fixed at a certain level at which the force monitored by the ESH machine
was set as zero. Therefore, the force measured by the ESH machine is the dynamic force of
the damper. That is why the force diagrams shown in Figure 2 are seen to be nearly
symmetric about the -axis, unlike those presented in [22, 26] in which the force diagrams
are seen to be asymmetric about the -axis. That is because the force diagrams measured in
these studies [22, 26] represent the total force of the damper which includes the static force
due to the pre-charge pressure in the gas chamber. The sole contribution of gas pressure in
the current measurements is due to the variation of gas pressure caused by the motion of the
piston. That variation is due to the change of volume caused by the immersed/protruded part
of the piston rod. Therefore, the output force of the damper according to the current
measurements can be expressed as follows:

( ) ( ⁄| |) (1)

where and are the cross-sectional areas of the piston head and piston rod, respectively,
and are the pressures in the rebound and the compression chambers, respectively,
is the variation of gas pressure due to the change of air volume caused by the piston
motion, and ( ⁄| |) is the summation of all friction resistances multiplied by the
velocity sign to interpret its direction as a resisting force.
 7

Equation (1) has been used to validate the current numerical approach by the determination of
different parameters via the numerical approach, and the direct comparison between the
theoretical results and experimental data, as will be shown in Section 4.2. The current
numerical approach accounts for the pressure and viscous forces, and the friction force
caused by the shear stresses on piston walls caused by the flow MR fluid, denoted as the “wet
friction” [27, 28], as will be shown in Section 3.3. The “dry” friction force caused by the
gland sealing applied to the piston was neglected. However, the current numerical approach
has been used to evaluate the dry friction, as will be shown in Section 5. The variation of the
gas force, , is determined assuming adiabatic expansion of gas according to the
following equation [23]:

[( ) ] (2)

where is the pre-charge pressure of the air chamber whose value is approximately 3.44
bar, based on some data provided to the first author from Lord Corporation [29]. is the
initial volume of the gas (nitrogen) chamber, is the piston displacement, and is the
polytropic exponent of nitrogen taken as 1.4.

3 Mathematical approach

The current method adopts the same numerical approach presented by the current authors in
[30], in which a one-way coupled numerical approach was developed to model the
performance of MR dampers. The numerical method couples an FE model developed for the
magnetic circuit and a transient numerical model of fluid flow that is implemented using a
CFD model. The coupling between the two solvers was implemented by the definition of
fluid viscosity as a function of the magnetic field density and local shear rate, as in [9, 13,
30]. Different numerical modelling approaches that employ either decoupled or one- or two-
way coupled techniques between the magnetic field and fluid flow solvers in MR fluid
research have been recently reviewed in [10].

The former study, [30], presented a theoretical study of the viscoplastic flow in an MR
damper that was previously-tested in [31, 32]. The approach presented in the current study
expands the former approach by the inclusion of the viscoelastic-viscoplastic behaviour of
the fluid. Moreover, the current study shows the roles of compressibility, viscoelasticity, fluid
inertia and friction in producing the hysteretic behaviour of a commercially-available MR
damper. The motion conditions of the damper involve very high shear rates of the fluid
compared to those conditions studied in [30]. That is because the thickness of the throttling
 8

area of the current damper is smaller than that considered in [30]. Therefore, the rheological
model used in the current study is different from the one presented in [30]. Also, the effects
of magnetic saturation and fluid inertia are included in the current study, whereas they are not
presented in [30]. The dynamic characteristics measured for the current MR damper were
compared with the theoretical results predicted by the numerical approach.

3.1 Generation of computational domains and model conditions

The FE analysis of the electromagnetic circuit is performed using COMSOL/Multiphysics

software, whereas the fluid flow analysis is applied by ANSYS/Fluent software. The CFD
model is a two-phase flow analysis based on the Volume-of-Fluid (VOF) model in
ANSYS/Fluent, to account for the existence of the air chamber in the damper. The
computational domains of both solvers are shown in Figure 3(a) and (b), respectively. Figure
3(a) shows a two-dimensional axisymmetric grid established by COMSOL/Multiphysics for
the magnetic circuit of the MR damper (piston head), in which magnetic insulation was
assumed at the boundaries of the piston. The mesh is refined in the MR fluid region, as
shown in the figure to determine the magnetic field distribution accurately. The direction of
current in the coil and the number of coil turns are defined in the FE model based on
Maxwell’s equations. Also, the nonlinear magnetic properties due to magnetic saturation are
accounted for in the model by the definition of magnetic permeability of materials according
to the corresponding – curve.

Figure 3(b) shows the fluid domain of the damper described by a two-dimensional
axisymmetric grid that is solved in ANSYS/Fluent using a transient two-phase flow model.
All boundaries are defined as walls except for the axis of symmetry. The fluid is excited by
the motion of piston walls causing it to flow in the throttling area of the piston. To simulate
piston motion in the model, the dynamic mesh layering technique available in ANSYS/Fluent
was used. Dynamic mesh layering technique adds or removes cells adjacent to moving walls
or fluid zones in a computational domain that can be only represented by quadrilateral cell
shapes for 2D problems, or wedge and hexahedron cell shapes for 3D problems [33]. A
predefined velocity was assigned to a part of the computational domain, termed as “fluid-
movement zone”. To do so, two stationary interior boundaries were assigned so that the
layering (adding or removing cells from the mesh) is performed.
 9

(a)

(b)
Figure 3: Computational domains of the current numerical approach: (a) the magnetic circuit
domain established for the piston using COMSOL/Multiphysics, and (b) the two-phase flow
fluid domain solved by ANSYS/Fluent.

The dynamic layering technique adopted to simulate the piston movement can be performed
by another method, which seems to be more straightforward. That is to assign the motion to
the walls of the piston rather than assigning a “fluid-movement zone” and two stationary
interior boundaries, as shown in Figure 4 and also presented in [34]. In that case, there is no
need to assign the stationary interior boundaries shown in Figure 3(b) because the moving
piston walls perform the layering on the stationary mesh. However, this approach is not
employed in the current study, as it is thought to be problematic in terms of solution setup
and stability. The problems arise from the need to assign “interface” boundary conditions
perpendicular to the moving walls if the motion is assigned to walls rather than fluid zones.
 10

This means that at least two interfaces should be defined in the way shown in Figure 4, which
will cause ANSYS/Fluent to read the mesh as three separate meshes split by the two
interfaces. This may lead to convergence problems that can be imagined as fluid separation at
the interfaces. Therefore, interfaces are not recommended in CFD simulations if they do not
represent the real problem description such as an interface between two rigid bodies.

Figure 4: Mesh interfaces required to assign the motion to solid walls rather than a fluid zone
(not applied in the current study).

3.2 Coupling technique

A User-Defined Function (UDF) performs the coupling between the FE and CFD models by
the definition of fluid viscosity. The viscosity is defined as a function of the magnetic-field-
dependent fluid yield stress, ( ), and the local shear rate, ̇ , computed in the CFD model.
Several functions are used in the literature to define the viscosity such as Papanastasiou’s
model [35], Eyring model [36], hyperbolic tangent function [20], or bi-viscous Bingham
model [37]. However, the best results were obtained by using a blending equation that was
developed by Susan-Resiga [38]. That is:

̇ ̇ ̇ (3)
[ ( )] [ ( ) ( ) ] ( ) ̇
̇ ̇ ̇

where, , ̇ , , and are rheological parameters that control the growth of viscosity at
very low shear rates, whose values are shown in Table 1 at the different input currents
investigated in the current study. That function was found to fit the experimental
measurements of the viscosity of the MR fluid used in the damper over a wide range of
applied shear rates. It should be noted that the rheological shear stress-shear rate diagram is
highly nonlinear and it differs for each type of MR fluid device, and also for the same MR
fluid device at different magnetic fields. So, fitting an equation that describes the
 11

characteristic shear stress-shear rate diagram of an MR fluid requires equation parameters to

be determined at different magnetic fields and fit a wide range of shear rates, as achieved by
Susan-Resiga [38]. The bi-viscous Bingham model has been also investigated in the current
study, as will be shown in Section 4.6. The apparent viscosity according to the bi-viscous
Bingham model is defined as:

̇ ̇
{ (4)
̇ ̇
̇

where is the maximum viscosity at the critical shear rate ̇ , and is the plastic viscosity
whose value is reported to be 100 to 1000 times less than the maximum viscosity [12, 39].
The maximum viscosity, which occurs at zero shear rate, is often referred to as the solid
viscosity to denote the shear between the fluid and the solid plug formed. The values adopted
for these parameters are also shown in Table 1.

The definition of fluid viscosity by a steady-state relation, ( ( ) ̇ ) as in equation (3)

and (4), includes modelling the viscoplastic behaviour of the fluid. To express the
viscoelastic behaviour exerted by the fluid in the pre-yield zone, the yield stress should be
expressed as a function of time. To do this, the following equation was used [39]:

( ̇ ⁄ ( ))
( ) ( )[ ] (5)

where, is time and is the fluid shear modulus, which can be defined as the ratio of shear
stress to the shear strain [40].

The magnetic field strength, , is determined by the FE model, then the corresponding yield
stress, ( ), is defined in the CFD model according to the information obtained from the
fluid data sheet [41]. The fluid yield stress is fed as a text file that is read and processed by
the UDF based on cell coordinates. Therefore, the UDF uses the fluid yield stress in
conjunction with the local shear rate determined by the CFD solver to define the viscosity
according to equations (3) and (5).
 12

Table 1: Values of the rheological parameters of Susan-Resiga [38] and the bi-viscous
Bingham model [37].

Susan-Resiga model Bi-viscous Bingham model

= 0.2 A = 0.4 A = 1.0 A = 0.2 A

(Pa.s) 84 x 105 55.78 x 105 55.78 x 105 (Pa.s) 1420

̇ (s-1) 7.556 x 10-5 2.673 x 10-4 2.673 x 10-4 (Pa.s) 3

(Pa) 258.7 757.5 1100 ̇ (s-1) 2

0.855 0.855 0.855

3.3 Governing equations

The main governing equations of the magnetic field analysis are the well-known Maxwell’s
equations developed in the weak form [42], given by the following equations:

⃑ (6)

⃑ (7)

⃑ (8)
⃑

⃑ (9)
⃑

where ⃑ is the electric flux density, is the electric charge density, ⃑ is the magnetic flux
density, ⃑ is the electric field intensity, ⃑ is the magnetic field intensity (strength), and is
the current density. On the other hand, the main governing equations used in the fluid flow
analysis are the continuity and momentum equations that are described in the VOF model,
respectively, as follows:

(10)
[ ( ) ( ⃑⃑⃑ ) ̇ ̇ ]

(11)
( ⃑ ) ( ⃑ ⃑ ) ⃑⃑ ⃑⃑⃑
 13

where and represent the corresponding phase, is the volume fraction of the
corresponding phase, is the density of each phase, ⃑⃑⃑ is the flow velocity vector of the
corresponding phase, ̇ is the mass transfer from phase to phase , ⃑ and are the
mass-averaged velocity and density of the mixture, respectively, is the fluid pressure,
⃑⃑ ( ( ) ̇ ) is the second-order deviatoric stress tensor, and is the gravity
acceleration.

It is worth mentioning that there are two approaches for numerical calculation of multiphase
flows, namely: the Euler-Euler approach and the Euler-Lagrange approach [33]. In the Euler-
Euler approach, to which the VOF model belongs, the different phases are treated as
interpenetrating/immiscible continua. On the other hand, the Euler-Lagrange approach treats
the fluid phase as a continuum, while the dispersed phase has the form of a large number of
particles, bubbles or droplets, whose positions are tracked through the computational domain.
There are three multiphase models available in the Euler-Euler approach, namely: VOF,
Mixture model, and the Eulerian model. The VOF model is the model preferred when
modelling a flow field that is represented by immiscible continua, in which the interface
between the phases is initially known [33, 43, 44].

The density of the MR fluid is defined in ANSYS/Fluent by the simplified Tait equation
assuming a linear growth of fluid bulk modulus, , as a function of pressure. That is:

( ) (12)

where, is the reference bulk modulus taken as: = 6 x 106 (Pa), and is the density
exponent taken as = 14. The MR fluid density at the reference pressure is = 3050
3
(kg/m ).

Therefore, the governing equations of the CFD model are solved within the computational
fluid domain shown in Figure 3(b) so that the different flow parameters are determined. The
fluid pressure in each chamber was used to determine the hydraulic force of the damper
according to equation (1). The wet friction force was determined by the calculation of the
total shear stress on piston walls multiplied by the side area of the piston, as in [17, 30].

To summarise the current computational method, the FE and CFD models were established in
the same manner presented in [30]. The FE model studies the flow of magnetic flux in the
MR piston, shown in Figure 1(b), based on Maxwell’s equations using
 14

COMSOL/Multiphysics. The CFD model studies the MR fluid flow in the fluid domain of
the damper using a two-phase flow model developed by the VOF model that is available in
ANSYS/Fluent. The UDF couples both solvers by the definition of fluid viscosity in the CFD
model according to equation (3) or (4), and equation (5). The value of the fluid yield stress,
, in the equation is determined according to the magnetic field intensity determined by the
FE model. The main parameters of the numerical approach are shown in Table 2, which also
presents the main dimensions of the damper shown in Figure 3(a) and (b).

Table 2: Parameters of the numerical approach.

Radius of the air gap, = 1.5 mm

Radius of the inner surface of the throttling area, = 15 mm

Radius of the outer surface of the throttling area, = 16 mm

Radius of piston core, = 10 mm

Outer radius of the, = 20 mm

Total length of the piston, = 21 mm

Length of piston shoulders (magnetic poles) , = 5 mm

Type of fluid used MRF-132 DG [23]

Material of the piston SAE-1020 [23]

Electric conductivity of the MR fluid, = 10-11 S/m

Electric conductivity of SAE-1020, = 8.41  106 S/m

Relative permeability of the MR fluid, – curve [41]

Relative permeability of Vacoflux-50, – curve [45]

Number of turns of each coil, = 150

MR fluid density at reference pressure, = 3050 kg/m3

Reference bulk modulus, = 6 x 106 Pa

density exponent, = 14

Height of the air pocket = 10 mm

Total length of the damper = 110 mm

 15

4 Results and discussions

4.1 Results of the FE model

The steady-state contours of magnetic field density in the piston head of the MR damper at
= 0.2 A are shown in Figure 5(a). It is seen that the main flow of the magnetic field in the
MR fluid region occurs at the farthermost ends of the piston. The distributions of magnetic
field intensity in the MR fluid region at different input currents is shown in Figure 5(b),
whereas the variation of the corresponding yield stress of the fluid is shown in Figure 5(c).
The fluid yield stress was determined by applying curve fitting of the diagram
available in [41].

(a)
 16

(b)

(c)
Figure 5: Distribution of magnetic field in the MR piston and corresponding yield stress of
the fluid: (a) contours of magnetic field density at = 0.2 (A), (b) distribution of magnetic
field intensity in the middle of the MR region at different input currents, and (c)
corresponding yield stress of the fluid determined from the fluid data sheet [41].

4.2 Model validation

The current numerical approach has been validated by the direct comparison with
experimental data presented in Section 2. The theoretical and experimental dynamic
characteristics of the damper performance at different conditions are compared as shown in
Figure 6 to Figure 8. The figures depict the dynamic characteristics of the damper at = 0.2,
0.4, and 1 A, respectively, and the values of = 1 Hz and = 3 mm for the sinusoidal
motion of the piston. It is seen that the theoretical results nearly coincide with the
experimental data, except for the sudden decrease of the force in the experimental data. These
sudden cusps are reported in different studies such as [19, 25, 46-49] and recently in [26] for
the same MR damper under investigation in the current study. However, the reason for their
occurrence is proposed in this paper for the first time. The occurrence of these sudden cusps
is thought to be mainly due to the dynamic effect of yield stress. In Figure 6 to Figure 8, at
= 0 and 0.5 s, the piston has instantaneous stops and starts to move either in the compression
or rebound strokes, respectively. Therefore, the fluid is totally plugged in the MR fluid region
at those moments, as also reported in [30]. When the piston starts to move, the plugged fluid
in the MR fluid region moves with the piston as one body until a sufficient pressure
 17

difference of the fluid in both chambers is built up. Once this pressure difference occurs, the
plug starts to break and this allows the fluid to flow in the region. However, this sudden flow
causes an instantaneous drop in the pressure difference, which is indicated by the cusps seen
on the force diagrams. It is thought that this instantaneous drop in the pressure difference
between the chambers enables the reformation of fluid plug in the MR fluid zone. Hence, the
pressure difference between the chambers is built up again until the plug is broken up again
at the beginning of the plastic zones ( 0.15 and 0.65 s). The occurrence of these sudden
peaks is thought to be more reported with relatively low values of piston velocities. These
peaks were not predicted by the current numerical approach as it is thought that more refining
of the mesh and the time step is required such that this phenomenon can be modelled
accurately.

Figure 6: Theoretical and experimental results of the dynamic characteristics of the RD-8040-
1 MR damper: (up) force-time histories in a complete cycle at = 0.2 A, = 1 Hz, =3
mm, (down-left) force-displacement diagrams, and (down-right) force-velocity diagrams.
 18

Figure 7: Theoretical and experimental results of the dynamic characteristics of the RD-8040-
1 MR damper at = 0.4 A.

Figure 8: Theoretical and experimental results of the dynamic characteristics of the RD-8040-
1 MR damper at = 1.0 A.

4.3 Effect of fluid compressibility

Fluid compressibility is reported to cause the hysteretic behaviour of MR dampers, which can
be described by the non-zero values of damper force at instantaneous locations of zero
velocity, as shown in the force-velocity diagrams in Figure 2 and Figure 6 to Figure 8. The
compressibility effects are due to the presence of air/gas chambers in most of the common
designs of MR dampers. The relative volume of these air chambers lowers the effective bulk
 19

modulus of the fluid considerably. To investigate the effect of fluid compressibility in the
current MR damper, the dynamic characteristics of the damper were determined based on the
coupled numerical approach employing different values of fluid bulk modulus, as shown in
Figure 9. It is seen that the lower value of fluid bulk modulus leads to a wider hysteretic zone
that is mostly noted in the characteristic force-velocity diagram. Therefore, it can be
concluded that the compressibility effects are the main source of hysteretic behaviour of MR
dampers, as also found by Guo et al. [7].

Figure 9: Theoretical force-displacement diagrams (left), and force-velocity diagrams (right)

for different values of the fluid bulk modulus.

4.4 Effect of viscoelasticity

The viscoelastic characteristics are mostly neglected in the study of MR dampers due to their
minor effects and the difficulty presented in numerical simulations by their constitutive
equations [17]. However, it should be noted that the viscoelastic effects are remarkable and
should not be neglected within MR dampers employing fluids with super high viscosity or
high velocity of the piston [17]. The viscoelastic characteristics are also reported to cause
further hysteretic behaviour of MR dampers. The viscoelastic characteristics of MR fluids
depend on the value of the shear modulus, , shown in equation (5). The higher value of
indicates higher elastic properties, therefore the fluid nearly deforms plastically. The
viscoelastic characteristics of MR fluids are measured by the determination of the value of
at different input currents, as in [50, 51]. The shear modulus is measured by applying an
oscillatory strain to a sample of the MR fluid and measuring the resulting stress [52]. The
values of the shear modulus of MR fluids are reported to be in the range of = 105 to 107 Pa
[50].
 20

To evaluate the viscoelastic characteristics in the current MR damper, the variation of

pressure in the rebound chamber was predicted by the current numerical approach using
different values of , as shown in Figure 10. The pressure is plotted only on a time interval of
⁄ , where = 1 s, and it represents the period of oscillation. This time interval represents
the piston motion in the compression stroke starting from rest at = 0, until maximum
velocity at ⁄ . The pressure in the rebound chamber mainly constitutes the damper
output force. That is because the variation of pressure in the compression chamber is
negligible due to the presence of the accumulator, as shown in Figure 1(b). It is seen in
Figure 10 that the variation of pressure in the rebound chamber is achieved faster at higher
values of . The retardation of pressure decrease, which indicates wider hysteretic behaviour
of the damper, is only seen at values of lower than 100 Pa, which is considerably lower
than the values reported for most MR fluids ( = 105 to 107 Pa). It was found that assuming
with higher values than 105 Pa leads to the same results as = 105 Pa. Therefore, it can be
concluded that the flow of MR fluids in MR dampers can be considered as a pure viscoplastic
flow, as it is seen that the inclusion of viscoelastic characteristics nearly leads to the same
results for high values of . So, from the results presented in Figure 9 and Figure 10, it can
be deduced that the prediction of the hysteretic behaviour of the current MR damper is
mainly caused by the inclusion of compressibility effects of the fluid. In other words, the
hysteretic behaviour of the current MR damper is due to fluid compressibility which is
nonlinear elasticity, but not viscoelasticity. The definition of liquids as incompressible media
is a quite common and acceptable assumption. However, in closed systems, the
compressibility effects should be accounted for as the effective bulk modulus of the liquid-
gas mixture is considerably low compared to the value of the bulk modulus of the fluid itself
[53]. Also, it should be noted that the viscoelastic characteristics may be remarkable at higher
frequencies and higher piston velocities and amplitudes.
 21

Figure 10: Variation of pressure in the rebound chamber due to employing different values of
in the numerical approach, = 1 x 107 Pa, = 1 Hz.

To show the viscoelastic effect in the current MR damper, the variation of pressure in the
rebound chamber was investigated at a high frequency of = 40 Hz, as shown in Figure 11.
The pressures are also plotted in the time interval of ⁄ for two values of , and the
8
reference bulk modulus was selected as a high value of 1 x 10 Pa, to minimise the effect of
fluid compressibility. = 1 x 108 Pa is the approximate value reported for silicone oil,
which is the most common carrier fluid employed in MR fluids [54]. It is seen in Figure 11
that employing different values of leads to a considerable difference in the profile of
pressure in the rebound chamber. The maximum difference between both curves was found to
be 15% relative to the pressure at = 100 Pa.
 22

Figure 11: Variation of pressure in the rebound chamber due to employing different values of
in the numerical approach, = 1 x 108 Pa, = 40 Hz.

4.5 Effect of fluid inertia

The hysteretic behaviour of MR dampers is also attributed to the effects of fluid inertia. High
piston velocity and high MR fluid density lead to a higher effect of fluid inertia. To evaluate
the role of fluid inertia in the current damper, three cases with different values of motion
frequency ( = 1, 10, and 20 Hz) were compared, as shown in Figure 12. The values of
motion amplitude and the applied current were kept constant in all cases. It is seen that the
higher inertia effect of fluid caused by the piston motion with higher frequency leads to a
wider hysteretic zone. Also, the higher frequency leads to an increase in the maximum force
of the damper and causes variations of the curve at the beginning of the plastic zone (around
= 0.1 and 0.6 s) in the same manner presented in [7]. However, the inertial effects due to the
variation of motion frequency are investigated in the current study, unlike the investigated
inertial effects due to the variation of fluid density presented in [7].

Figure 12: Theoretical dynamic characteristics of the RD-8040-1 MR damper due to different
motion frequencies, = 0.2 A.

4.6 Effect of different rheological models employed in the numerical approach

There is a plethora of models that are used in numerical simulations to define the apparent
viscosity of MR fluids, as shown earlier. So, it may be necessary to compare the theoretical
predictions of the current approach according to different models of rheology. The theoretical
output force of the damper due to the current model described by equations (3) and (5) is
compared with that force due to the bi-viscous Bingham model, defined by equations (4) and
(5), as shown in Figure 13. It is seen that the current model describes the damper behaviour
 23

fairly accurately relative to the experimental data compared to the bi-viscous Bingham
model. That is because the model developed by Susan-Resiga [38] defined by equations (3)
and (5) is shown to describe the viscosity measurements for the MRF-132DG MR fluid,
which is the fluid type employed in the damper, up to very high values of shear rate.

Figure 13: Theoretical output forces of the damper due to the current model described by
equations (3) and (5), and the bi-viscous Bingham model in comparison with the
experimental data.

4.7 Visualisations of flow parameters

One of the benefits of the investigation of fluid flow by CFD techniques is the availability of
visualisation of flow parameters. This helps in the better interpretation of flow characteristics,
which is necessary especially in complicated flows, and developing new methods to improve
system behaviours. To fully understand the characteristics of the rheological flow in the
damper, the shear rate and viscosity contours in the throttling area of the damper are shown
in Figure 14(a) and (b), respectively, whereas the velocity vectors at two locations in the
throttling area are shown in Figure 15. The relation between the different parameters in both
figures, namely: fluid shear rate, viscosity and velocity can be analysed. Figure 14(a) shows
the non-uniform distribution of the fluid viscosity due to the non-uniform distribution of the
magnetic field. Also, the fluid viscosity is seen to be low at the outer and inner cylindrical
surfaces of the throttling area. This is due to the high shear rate at these cylindrical surfaces,
as seen in the zoomed areas in Figure 14(b). It is also seen in Figure 14(b) that the shear rate
in the throttling area is approximately zero, since a plug flow profile is manifested, except for
the outer and inner cylindrical surfaces of the throttling area. The shear rate is seen in those
 24

areas to be very high, which heightens the need for the accurate definition of the rheological
model.

The variations of fluid viscosity and shear rate shown in Figure 14 lead to a slight variation
of the velocity profile along the MR fluid region, as shown in Figure 15(a). The figure shows
the velocity vectors at two locations in the throttling area: one of them is in the middle of the
region adjacent to the left magnetic pole, Zone (A), whereas the other is in the middle of the
throttling area (adjacent to the coil), Zone (B). It is seen that both velocity profiles are
slightly different, as can be seen also in Figure 15(b) which shows schematic drawings of the
velocity profiles at the two locations. The velocity magnitude is slightly higher in the centre
of Zone (B) compared to Zone (A). Also, the velocity magnitude is slightly lower in the outer
and inner parts of Zone (B) compared to Zone (A). This difference is seen clearly in Figure
15(a) in the zoomed areas that show the velocity vectors near the outer surface of the
throttling area in the two zones. This indicates that the pre-yield region in the middle of the
throttling area of the damper, Zone (B), is slightly smaller than that in the regions of
magnetic poles, Zone (A), which is due to the relatively lower viscosity in Zone (B), as
shown schematically in Figure 15(b). On the other hand, the post-yield region is slightly
larger in Zone (B) compared to Zone (A). It is seen that both of the velocity profiles in the
two regions are slightly different, although the magnetic field intensity is considerably high
in Zone (A) compared to Zone (B). The reason for that slight difference is due to the fact that
the fluid shear rate in the whole MR fluid region is approximately zero as shown earlier.
Therefore, the MR effect is manifested in the MR fluid region adjacent to the coil, even if the
applied magnetic field is small. Also, the zoomed areas in Figure 15(a) also indicate that the
velocity profile in the throttling area has common characteristics between Poiseuille and
Couette flows, as also illustrated in [17, 39, 55]. The main flow is a Poiseuille flow due to the
pressure difference along the throttling area, whereas the Couette flow is seen at the edges of
the throttling area at which a tiny portion of the fluid moves with the walls in the opposite
direction of the main flow due to the non-slip boundary condition.
 25

(a)

(b)
Figure 14: Viscosity, (a), and shear rate, (b), contours in the throttling arear of the damper at
⁄ = 0.25 s (maximum velocity of the piston in the right-to-left direction).
 26

(a)

(b)
Figure 15: Velocity vectors at two locations in the throttling area of the damper; one of them
is in the middle of the region of the left magnetic pole, whereas the other is in the middle of
the throttling area: (a) velocity vectors shown in the CFD solver, and (b) schematic drawings
of the velocity profiles.

Referring to Figure 14 and Figure 15, it is worth mentioning that the prediction of the
variation of fluid viscosity and shear rate, Figure 14(a) and (b), respectively, and the variation
of velocity profile, Figure 15, along the MR fluid region is due to the non-uniform
distribution of magnetic field intensity in the MR fluid region shown in Figure 5(b). The
effect of that non-uniform distribution has been accounted for in the current CFD model,
using the same technique presented in [9, 30]. Therefore, the fluid yield stress along the MR
fluid region has also a non-uniform distribution, as shown in Figure 5(c). In some previous
 27

studies such as [7, 12], the magnetic field intensity/density was only defined at the locations
of magnetic poles, therefore, the variation of the magnetic field in the middle region was not
accounted for. Alternatively, a homogenous magnetic field intensity/density in the whole MR
fluid region was considered in [6, 8], therefore, the whole MR fluid region is affected by a
uniform yield stress, whose value is determined according to the average magnetic field
intensity/density. However, the current numerical approach accounts for the non-uniform
distribution of fluid yield stress, as performed in [9, 13]. That is because the profile of yield
stress shown in Figure 5(c) has been defined as a 2D array based on the coordinates of the
MR fluid region using a text file that is compiled by the UDF.

5 Investigation of the off-state dry friction force

In the preceding modelling approach presented in Sections 3 and 4, it has been shown that the
fluid compressibility and inertia are the main reasons for the hysteretic behaviour of the
current MR damper, as shown in Figure 9 and Figure 12, respectively. The effect of
viscoelasticity was found to be minor. The friction forces are also reported to contribute to
the hysteretic behaviours of MR dampers. The effect of wet friction has been included in the
current numerical approach by the determination of shear stresses on piston walls. On the
other hand, the effect of dry friction of the piston rod with the sealing glands was neglected,
as in many numerical and multi-physics analyses that are presented in [6, 8, 17, 56, 57].
However, because the dry friction is reported to produce the hysteretic behaviours of
hydraulic dampers [58-60], therefore, it merits to be investigated in this study, as the study
investigates the nonlinear phenomena that lead to the hysteretic behaviour of MR dampers.

It should be noted that the dry friction resistance in hydraulic dampers subjected to reciprocal
loads is also a nonlinear dynamic behaviour, as shown by Heipl and Murrenhoff [58]. In that
paper, the dynamic dry friction resistance to the reciprocal motion of the piston rod in the
hydraulic cylinder was measured for different kinds of seals. They illustrated the difference
between dynamic and steady-state friction resistance, as shown in Figure 16. It is seen that
the dynamic resistance manifests a hysteretic loop, whereas the steady-state behaviour does
not. That hysteretic loop is also reported in MR dampers in both off- and on-states, as shown
in [60], in which the hysteresis of MR dampers was modelled taking into consideration the
Stribeck curve. The Stribeck curve is a fundamental concept in tribology which defines the
friction between sliding surfaces in relative motion [58].
 28

Figure 16: Difference between dynamic and steady-state friction resistance for hydraulic
dampers subjected to reciprocal loads [58].

The hysteretic loops are also seen in the dynamic characterisation of the current MR damper
even in the off-state, as shown in Figure 2. The off-state hysteretic loops are also reported in
[26] for the same MR damper under investigation in the current study. Therefore, that off-
state hysteretic behaviour can be attributed to viscous, inertia, compressibility and friction
effects. In order to evaluate the role of each element in producing the damper force and the
corresponding hysteretic behaviour, the damper behaviour in the off-state has been modelled
using the current CFD model. The CFD model was used to determine the damper behaviour
due to the effects of pressure and viscous forces, gas compression, and fluid inertia.
Therefore, the effects of dry friction and piston inertia can be evaluated by the difference
between the experimental dynamic characteristics of the damper in the off-state shown in
Figure 2, and the theoretical results of the preceding forces.

Hence, the CFD model presented in Section 3 has been modified to study the performance of
the damper in the off-state. The only modification was to the definition of fluid viscosity.
Rather than the localised definition of fluid viscosity based on cell coordinates using the UDF
to assign the magnetically activated and non-activated regions, as shown in Section 3.2,
another homogenous definition of fluid viscosity was adopted. The fluid viscosity has been
defined in the CFD model using the built-in library for Herschel-Bulkley fluids that is
available in ANSYS/Fluent. The fluid was considered as a non-Newtonian fluid, even in the
 29

off-state, in the light of the off-state shear stress – shear rate diagram that is available in the
fluid datasheet [41]. The diagram shows a nonlinear profile, in which the fluid manifests a
non-zero value of yield stress, whose magnitude is approximately 10 Pa, at zero shear rate.
Therefore, the built-in Herschel-Bulkley model in ANSYS/Fluent was adopted, and the
parameters of the model were assigned so that the viscosity equation according to the model
fits the experimental flow curve of the fluid in the off-state that is available in the fluid
datasheet [41]. The built-in Herschel-Bulkley model in ANSYS/Fluent employs a bi-viscous
definition of fluid viscosity that can be described by the following equations [33]:

̇⁄ ̇ ̇
( ) [( ) ( ) ] ̇ ̇
̇ ̇
(13)
( )
̇
( ) ̇ ̇
{ ̇ ̇

where is the consistency index and is the flow index. The values of the constants
shown in the preceding equation that were found to fit the experimental flow curve of the
fluid [41] were as follows: = 0.41 Pa.s, = 0.88, = 10 Pa, and ̇ = 0.01 Pa.s.

The CFD simulation has been performed and the same conditions of piston motion and the
other solver settings were also applied. Hence, the theoretical output force of the damper due
to pressure and viscous forces have been determined using equations (1) and (2). The
different elements of damper force, namely: the pressure and viscous forces according to the
CFD model and the dynamic force due to gas compression are compared with the total force
of the damper, as shown in Figure 17. The difference between the experimentally-determined
total force of the damper (black) and the total fluid forces predicted (blue) represents the role
of the dry friction and the inertia force of the piston. It is seen that the total force of the
damper is dominated by the dry friction and inertia force of the piston, rather than the
pressure, viscous and dynamic gas forces. The maximum value of the pressure, viscous and
dynamic gas forces represents about 21 % of the experimentally-determined total force of the
damper.

A similar analysis to that described by Figure 17 was presented in [60], in which the forces
due to pressure, viscous and dynamic gas forces were found to represent about 18 % of the
total damper force. Therefore, it can be concluded that the main output force of MR dampers
operated in the off-state under low values of frequency and amplitude is dominated by the dry
friction and inertial effects. The magnitude of the dry friction is influenced by the magnitude
of the high pre-charge pressure that is required to avoid cavitation problems in the damper.
The gauge value of pressure in the damper at the locations of sealing controls the static and
 30

dynamic friction resistance to the piston motion. Also, it can be deduced that the hysteretic
behaviour due to these effects is remarkable in comparison with those manifested by the
compressibility effects, as can be seen in the characteristic diagrams shown in Figure 17. It is
seen that a small hysteresis loop is manifested due to the pressure and viscous forces as
determined by the CFD model, whereas the hysteresis loop manifested due to the
experimental measurements is considerably wider.

In fact, compressibility effects are reported to cause nonlinear behaviour of different closed
systems, including hysteretic behaviours of MR dampers [1, 7, 53, 61, 62]. The
compressibility effects are related to the presence of air/gas chamber in MR dampers.
However, it is thought that the yield stress of the MR fluid itself causes further effects of fluid
compressibility. That is because there is a minimum pressure difference between the
chambers of the damper required to initiate flow between the chambers. While the piston is
moving and that threshold value of pressure difference is not achieved, the compressibility
effects are manifested by the fluid-gas mixture. For instance in the compression stroke, the
MR fluid compresses the air chamber and a gas pocket is formed in the rebound stroke, until
a sufficient pressure difference between the fluid in both chambers is achieved, therefore the
flow from the compression to rebound chamber starts in response to that pressure difference.
As this pressure difference is directly proportional to the applied current, the time lag of the
flow of the MR fluid in the throttling area increases. Therefore, wider hysteretic loops are
manifested as the applied current increases, as seen in the characteristic diagrams shown in
Figure 2.

Figure 17: Predictions of the theoretical pressure and viscous forces according to the CFD
model, ( ) , and gas forces computed by equation (2), , in
comparison with the total experimental force of the damper. The difference between the total
 31

experimental force (black) and the total theoretical forces (blue) represents the total dry
friction and inertia force of the damper.

6 Conclusions

The performance of a commercially-available MR damper has been studied using theoretical

and experimental methods. The theoretical method incorporates a coupled numerical
approach between FE and CFD analyses. The CFD model allows better assessment of the
characteristics of fluid flow which helps in design optimisation. It can be concluded that the
frictional, compressibility and inertial and effects are found to mainly form the hysteretic
behaviour of the MR damper, whereas the viscoelastic effects are minor. Also, the selection
of the rheological model used to define the fluid viscosity is critical, as it leads to major
differences in the predictions of the numerical approach.

The current numerical approach is thought to be useful not only to model the performance of
MR dampers but also to model nonlinear performances of different MR fluid devices. That is
because the current numerical approach presents a coupled analysis between magnetic and
fluid flow analyses, in which different sources of nonlinearity are accounted for, namely:
magnetic saturation, non-uniform distribution of magnetic field, and effects of fluid
compressibility, viscoelasticity, inertia, friction, and presence of an air pocket. Modelling the
performance of other MR fluid devices can be performed by changing the boundary
conditions according to the problem description in each device, and follow the same strategy
of the current numerical approach. Future work includes applying the same method on other
MR fluid devices and investigations of two-way coupling methods.

7 References

[1] J. Goldasz, Chapter 4 - Developments in modeling of magnetorheological actuators, in: S.-B. Choi,
W. Li (Eds.) Magnetorheological Materials and their Applications. The Institution of Engineering and
Technology, UK (2019), pp. 63-91.
[2] M. Ahmadian, J. A. Norris, Experimental analysis of magnetorheological dampers when subjected
to impact and shock loading, Communications in Nonlinear Science and Numerical Simulation, 13 (9)
(2008), pp. 1978-1985. https://doi.org/10.1016/j.cnsns.2007.03.028.
[3] R. Ahamed, S.-B. Choi, M. M. Ferdaus, A state of art on magneto-rheological materials and their
potential applications, Journal of Intelligent Material Systems and Structures, 29 (10) (2018), pp.
2051-2095. https://doi.org/10.1177/1045389X18754350.
[4] X. Yuan, T. Tian, H. Ling, et al., A Review on Structural Development of Magnetorheological Fluid
Damper Shock and Vibration, 2019 (2019), p. 33. https://doi.org/10.1155/2019/1498962.
[5] A. Dominguez, R. Sedaghati, I. Stiharu, Modeling and application of MR dampers in semi-adaptive
structures, Computers & Structures, 86 (3-5) (2008), pp. 407-415.
https://doi.org/10.1016/j.compstruc.2007.02.010.
 32

[6] M. Gurubasavaraju, H. Kumar, A. Mahalingam, An approach for characterizing twin-tube shear-

mode magnetorheological damper through coupled FE and CFD analysis, Journal of the Brazilian
Society of Mechanical Sciences and Engineering, 40 (3) (2018), pp. 1-14.
https://doi.org/10.1007/s40430-018-1066-z.
[7] P. Guo, J. Xie, X. Dong, et al., A Two-Dimensional Axisymmetric Finite Element Analysis of Coupled
Inertial-Viscous-Frictional-Elastic Transients in Magnetorheological Dampers Using the Compressible
Herschel-Bulkley Fluid Model, Frontiers in Materials, 6 (2019), p. 293.
https://doi.org/10.3389/fmats.2019.00293.
[8] P. Guo, J. Xie, Two-Dimensional CFD Modeling of Hysteresis Behavior of MR Dampers Shock and
Vibration, 2019 (2019), p. 14. 10.1155/2019/9383047.
[9] M. Kemerli, T. Engin, Z. Parlak A New Rheological Model of Magnetorheological Fluids for CFD:
Comparison and Validation. In: Proceedings of the ASME 2018 Conference on Smart Materials,
Adaptive Structures and Intelligent Systems, San Antonio, USA, 10-12 September 2018, p. 7984.
ASME, USA. 10.1115/smasis2018-7984.
[10] W. Elsaady, S. O. Oyadiji, A. Nasser, A review on multi-physics numerical modelling in different
applications of magnetorheological fluids, Journal of Intelligent Material Systems and Structures, 31
(16) (2020), pp. 1855-1897. https://doi.org/10.1177/1045389x20935632.
[11] D. H. Wang, W. H. Liao, Magnetorheological fluid dampers: a review of parametric modelling,
Smart Materials and Structures, 20 (2) (2011), p. 023001. https://doi.org/10.1088/0964-
1726/20/2/023001.
[12] Z. Parlak, T. Engin, İ. Çallı, Optimal design of MR damper via finite element analyses of fluid
dynamic and magnetic field, Mechatronics, 22 (6) (2012), pp. 890-903.
https://doi.org/10.1016/j.mechatronics.2012.05.007.
[13] M. Kemerli, T. Engin, Z. Parlak, Coupled Magnetic and CFD Modelling of a Structural
Magnetorheological Vibration Absorber with Experimental Validation, Mechanism, Machine,
Robotics and Mechatronics Sciences, 58 (2019), pp. 115-125. https://doi.org/10.1007/978-3-319-
89911-4_9.
[14] D. Case, B. Taheri, E. Richer, Dynamical Modeling and Experimental Study of a Small-Scale
Magnetorheological Damper, IEEE/ASME Transactions on Mechatronics, 19 (3) (2014), pp. 1015-
1024. https://doi.org/10.1109/TMECH.2013.2265701.
[15] J. Zheng, Z. Li, J. Koo, et al., Magnetic Circuit Design and Multiphysics Analysis of a Novel MR
Damper for Applications under High Velocity, Advances in Mechanical Engineering, 6 (2014), p.
402501. https://doi.org/10.1155/2014/402501.
[16] J. Gołdasz, B. Sapinski, Ł. Jastrzębski, Assessment of the magnetic hysteretic behaviour of MR
dampers through sensorless measurements, Shock and Vibration, 2018 (2018), p. 21.
https://doi.org/10.1155/2018/3740208.
[17] A. Syrakos, Y. Dimakopoulos, J. Tsamopoulos, Theoretical study of the flow in a fluid damper
containing high viscosity silicone oil: Effects of shear-thinning and viscoelasticity, Physics of Fluids, 30
(3) (2018), p. 030708. https://doi.org/10.1063/1.5011755.
[18] T. C. Bulea, R. Kobetic, C. S. To, et al., A Variable Impedance Knee Mechanism for Controlled
Stance Flexion During Pathological Gait, IEEE/ASME Transactions on Mechatronics, 17 (5) (2012), pp.
822-832. https://doi.org/10.1109/TMECH.2011.2131148.
[19] Z. Parlak, T. Engin, Time-dependent CFD and quasi-static analysis of magnetorheological fluid
dampers with experimental validation, International Journal of Mechanical Sciences, 64 (1) (2012),
pp. 22-31. https://doi.org/10.1016/j.ijmecsci.2012.08.006.
[20] D. Case, B. Taheri, E. Richer, Multiphysics modeling of magnetorheological dampers,
International Journal of Multiphysics, 7 (1) (2013), pp. 61-76. https://doi.org/10.1260/1750-
9548.7.1.61.
 33

[21] M. Gurubasavaraju, H. Kumar, A. Mahalingam, Performance analysis of a semi-active suspension

system using coupled CFD-FEA based non-parametric modeling of low capacity shear mode
monotube MR damper, Proceedings of the Institution of Mechanical Engineers, Part D: Journal of
Automobile Engineering, 233 (5) (2018), pp. 1214-1231.
https://doi.org/10.1177/0954407018765899.
[22] Lord Corporation, RD-8040-1 and RD-8041-1 Dampers datasheet. Available at:
http://www.lordmrstore.com/lord-mr-products/rd-8040-1-mr-damper-short-stroke (accessed on 02
October 2019).
[23] S. Seid, S. Chandramohan, S. Sujatha, Optimal design of an MR damper valve for prosthetic knee
application, Journal of Mechanical Science and Technology, 32 (6) (2018), pp. 2959-2965.
https://doi.org/10.1007/s12206-018-0552-7.
[24] S. Purandare, H. Zambare, A. Razban, Analysis of magnetic flux in magneto-rheological damper,
Journal of Physics Communications, 3 (7) (2019), p. 075012. https://doi.org/10.1088/2399-
6528/ab33d7.
[25] R. M. Desai, M. E. H. Jamadar, H. Kumar, et al., Evaluation of a commercial MR damper for
application in semi-active suspension, SN Applied Sciences, 1 (9) (2019), p. 993.
https://doi.org/10.1007/s42452-019-1026-y.
[26] B. Sapiński, Ł. Jastrzębski, J. Gołdasz, Electrical harmonic oscillator with MR damper and energy
harvester operating as TMD: Experimental study, Mechatronics, 66 (2020), p. 102324.
https://doi.org/10.1016/j.mechatronics.2020.102324.
[27] G. Chen, Chapter 3 - Fundamentals of contact mechanics and friction, in: G. Chen (Ed.)
Handbook of Friction-Vibration Interactions. Woodhead Publishing, Cambridge, UK (2014), 71-152.
[28] A. Cantelli, Uniform Flow of Modified Bingham Fluids in Narrow Cross Sections, Journal of
Hydraulic Engineering, 135 (8) (2009), pp. 640-650.
[29] Lord Corporation, Tecnical data for the RD-8040-1 damper. Data supplied by the company on 2
October 2019.
[30] W. Elsaady, S. O. Oyadiji, A. Nasser, A one-way coupled numerical magnetic field and CFD
simulation of viscoplastic compressible fluids in MR dampers, International Journal of Mechanical
Sciences, 167 (2020), p. 105265. https://doi.org/10.1016/j.ijmecsci.2019.105265.
[31] W. W. Chooi, Experimental characterisation of the properties of magnetorheological (MR) fluids
and MR damper, PhD Thesis (2005), The University of Manchester, UK.
[32] W. W. Chooi, S. O. Oyadiji, Experimental Testing and Validation of a Magnetorheological (MR)
Damper Model, Journal of Vibration and Acoustics, 131 (6) (2009), p. 061003.
https://doi.org/10.1115/1.3142885.
[33] Ansys, ANSYS FLUENT 12.0 Theory Guide. ANSYS, Inc., United States (2009).
[34] W. Elsaady, A. Ibrahim, A. Abdalla, Numerical Simulation of Flow Field in Coaxial Tank Gun Recoil
Damper, Advances in Military Technology, 14 (1) (2019), pp. 139-150.
[35] T. C. Papanastasiou, Flows of Materials with Yield, Journal of Rheology, 31 (5) (1987), pp. 385-
404. https://doi.org/10.1122/1.549926.
[36] Y.-T. Choi, L. Bitman, N. M. Wereley, Nondimensional Analysis of Electrorheological Dampers
using an Eyring Constitutive Relationship, Journal of Intelligent Material Systems and Structures, 16
(5) (2005), pp. 383-394. https://doi.org/10.1177/1045389x05050529.
[37] D. J. Ellam, R. J. Atkin, W. A. Bullough, Analysis of a smart clutch with cooling flow using two-
dimensional bingham plastic analysis and computational fluid dynamics, Proceedings of the
Institution of Mechanical Engineers, Part A: Journal of Power and Energy, 219 (8) (2006), pp. 639-
652. https://doi.org/10.1243/095765005x31306.
 34

[38] D. Susan-Resiga, A Rheological Model for Magneto-rheological Fluids, Journal of Intelligent

Material Systems and Structures, 20 (8) (2009), pp. 1001-1010.
https://doi.org/10.1177/1045389X08100979.
[39] W. Bullough, D. Ellam, A. Wong, et al., Computational fluid dynamics in the flow of ERF/MRF in
control devices and of oil through piezo-hydraulic valves, Computers & Structures, 86 (3-5) (2008),
pp. 266-280. 10.1016/j.compstruc.2007.01.043.
[40] E. M. Lifshitz, A. M. Kosevich, L. P. Pitaevskii, Chapter I - Fundamental Equations, in: E. Grabbe,
S. Ramo, D. Wooldridge (Eds.) Theory of Elasticity (Third Edition). Butterworth-Heinemann, Oxford,
UK (1986), 1-37.
[41] Lord Corporation, MRF-132DG Magneto-Rheological Fluid datasheet. Available at:
https://www.lord.com/sites/default/files/Documents/TechnicalDataSheet/DS7015_MRF-
132DGMRFluid.pdf (accessed on 11 December 2017).
[42] E. M. Purcell, D. Morin, Electricity and magnetism, Third ed. Cambridge University Press,
Cambridge, UK (2013).
[43] E. Adaze, A. Al-Sarkhi, H. M. Badr, et al., Current status of CFD modeling of liquid loading
phenomena in gas wells: a literature review, Journal of Petroleum Exploration and Production
Technology, 9 (2) (2019), pp. 1397-1411. 10.1007/s13202-018-0534-4.
[44] C.-D. Ho, H. Chang, H.-J. Chen, et al., CFD simulation of the two-phase flow for a falling film
microreactor, International Journal of Heat and Mass Transfer, 54 (15-16) (2011), pp. 3740-3748.
10.1016/j.ijheatmasstransfer.2011.03.015.
[45] Comsol, COMSOL Multiphysics Reference Manual. COMSOL Inc, Sweden (2018).
[46] K. Hudha, H. Jamaluddin, P. Samin, et al., Effects of control techniques and damper constraint
on the performance of a semi-active magnetorheological damper, International Journal of Vehicle
Autonomous Systems, 3 (2005), pp. 230-252. 10.1504/IJVAS.2005.008235.
[47] J. Yu, X. Dong, S. Sun, et al., Comparison of dynamic models based on backbone curve for rotary
magneto-rheological damper, proceedings of The Institution of Mechanical Engineers Part C: Journal
of Mechanical Engineering Science, 234 (14) (2020), pp. 2732-2740. 10.1177/0954406219856392.
[48] H. Ji, Y. Huang, S. Nie, et al., Research on Semi-Active Vibration Control of Pipeline Based on
Magneto-Rheological Damper, 10 (7) (2020), p. 2541.
[49] H. Metered, P. Bonello, S. O. Oyadiji, The experimental identification of magnetorheological
dampers and evaluation of their controllers, Mechanical Systems and Signal Processing, 24 (4)
(2010), pp. 976-994. https://doi.org/10.1016/j.ymssp.2009.09.005.
[50] W. W. Chooi, S. O. Oyadiji, Characterizing the effect of temperature and magnetic field
strengths on the complex shear modulus properties of magnetorheological (MR) fluids, International
Journal of Modern Physics B, 19 (07n09) (2005), pp. 1318-1324.
https://doi.org/10.1142/S0217979205030244.
[51] W. Li, G. Chen, S. Yeo, Viscoelastic properties of MR fluids, Journal of Non-Newtonian Fluid
Mechanics, 57 (1) (1999), pp. 1-25.
[52] M. Meyers, K. Chawla, Mechanical behavior of materials. Cambridge University Press, UK (2008).
[53] M. Jelali, Chapter 3 - Physical Properties of Fluids, in: A. Kroll (Ed.) Hydraulic servo-systems:
modelling, identification, and control. Springer, London, UK (2003), pp. 30-51.
[54] Dow Corning, Compressibility data. Data supplied by the company on 22 November 2018.
[55] R. M. Bhatnagar, Transient effect modelling of magnetorheological fluid–based damper at low
speed and its comparison with the existing magnetorheological damper models, Journal of
Intelligent Material Systems and Structures, 24 (12) (2013), pp. 1506-1523.
10.1177/1045389x12474355.
 35

[56] W. J. Meng, S. M. Wu, B. L. Liu, et al., Multi-Physics Analysis of a Magnetorheological Brake with
Double Coils Placed on Side Housing, Key Engineering Materials, 739 (2017), pp. 252-263.
10.4028/www.scientific.net/KEM.739.252.
[57] Y. Wang, S. Li, W. Meng, Strong coupling analysis of fluid–solid for magnetorheological fluid
braking system, Journal of Intelligent Material Systems and Structures, 29 (8) (2017), pp. 1586-1599.
10.1177/1045389X17742730.
[58] O. Heipl, H. Murrenhoff, Friction of hydraulic rod seals at high velocities, Tribology International,
85 (2015), pp. 66-73. https://doi.org/10.1016/j.triboint.2015.01.002.
[59] L. Gagnon, M. Morandini, G. L. Ghiringhelli, A review of friction damping modeling and testing,
Archive of Applied Mechanics, 90 (1) (2020), pp. 107-126. 10.1007/s00419-019-01600-6.
[60] Y.-L. Zhao, Z.-D. Xu, A hysteretic model considering Stribeck effect for small-scale
magnetorheological damper, Smart Materials and Structures, 27 (6) (2018), p. 065021.
10.1088/1361-665x/aabc2c.
[61] R. Stanway, Smart fluids: current and future developments, Materials Science and Technology,
20 (8) (2004), pp. 931-939. 10.1179/026708304225019867.
[62] D. H. Wang, X. X. Bai, W. H. Liao, An integrated relative displacement self-sensing
magnetorheological damper: prototyping and testing, Smart Materials and Structures, 19 (10)
(2010), p. 105008. 10.1088/0964-1726/19/10/105008.