See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/320966011

Design Optimization of a Bi-Fold Magnetorheological Damper Subject to Impact

Loads

Conference Paper · September 2017

DOI: 10.1115/SMASIS2017-3753

CITATIONS READS
2 605

3 authors, including:

Muftah Saleh Rama B. Bhat

Algonquin College Concordia University Montreal
5 PUBLICATIONS   21 CITATIONS    373 PUBLICATIONS   3,836 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Rotor dynamics and tribology View project

Crashworthiness of helicopters View project

All content following this page was uploaded by Muftah Saleh on 17 November 2017.

The user has requested enhancement of the downloaded file.

 The ASME 2017 Conference on Smart Materials, Adaptive Structures and Intelligent Systems
DESIGN OPTIMIZATION OF A BI-FOLD MAGNETORHEOLOGICAL DAMPER
SUBJECT TO IMPACT LOADS
Muftah Saleh Ramin Sedaghati Rama Bhat
Department of Mechanical, Industrial & Aerospace Engineering, Sir George William Campus,
1455 De Maisonneuve Blvd. W., Montreal, Quebec, Canada, H3G 1M8
ABSTRACT
Dynamic range (D) (ratio of the maximum on-state damping
force to the off-state damping force), is an important index
characteristic of the performance of the Magnetorheological
Energy Absorbers (MREAs). In high speed impact, the dynamic
range may fall into the uncontrollable zone ( 1) due to the
increase in the off-state damping force which is associated with
the transition of the flow from laminar to turbulent condition.
Therefore, it is of paramount importance to design optimize the
MREA in order to increase its dynamic range while
accommodating the geometry, MR fluid flow and magnetic field
constraints. In this study, a design optimization problem has been
formulated to optimally design a bi-fold MREA to comply with
the helicopter crashworthiness specifications for lightweight
civilian helicopters. It is required to have a minimum dynamic
range of 2 at 5 m/s impact velocity while satisfying the
constraints imposed due to the geometry, volume of the device,
magnetic field and the flow of the magnetorheological fluid in
the MR valve. Meanwhile in order to comply with the helicopter
crashworthiness requirement, the MREA device should be
designed to generate 15 kN field-off damping force at the design
impact velocity if the MREA is to be integrated with skid landing
gear systems. The magneto-static analysis of the MREA valve
has been conducted analytically using simplified assumptions in
order to obtain the relation between induced magnetic flux in the
MR fluid gaps in active regions versus the applied current and
MREA valve geometrical parameters.
Both Bingham plastic models, with and without minor loss
factors, have been utilized to derive the dynamic range and the
results are compared in terms of the generated off-state damping
force, on-state damping force, and dynamic range. The Bingham
plastic model with minor loss coefficients was found to be more
accurate due to the turbulent condition in the MREA caused by
the impact. Finally, the performance of the optimized bi-fold
MREA has been evaluated under different impact speeds.
SMASIS
September 18-20, 2017, Snowbird, Utah, USA
SMASIS2017-3753
1. INTRODUCTION
Surviving the death and casualties due to vertical crashes in
helicopters has grown to be a major concern in helicopter
crashworthiness. In these events, the helicopter occupants might
become vulnerable to high abrupt accelerative forces of 15 g’s
or more due to inadequate provision to absorb the impact energy
[1].
To protect the helicopter and its occupants in the event of
hard landing or crash, several agencies have established
guidelines for designing, testing, and certification of each
helicopter crashworthy system, i.e. landing gear, sub-floor
structure, and occupant seat, see for example [2-5]. The crash
design conditions for civil helicopters are different from those
for military ones. For instance, the skid landing gears (SLGs) of
the civilian helicopters, which are mainly used with lightweight
helicopters, should be successfully drop tested at vertical
velocity of 2.44 m/s and level attitude [4].
In recent decades, the emerging of smart materials
technology such as magnetorheological fluids has led to the
development of innovative adaptive energy absorption systems
which could be effectively utilized to attenuate vibration and
shock. To enhance the performance of the conventional SLGs in
attenuating impact loads at sink rates higher than 2.44 m/s, they
can be integrated with MR based dampers. These fail-safe semi-
active devices are capable of generating variable and large
damping forces rapidly based on impact conditions while
requiring low power consumption compared with fully-active
devices.
The single and multi-coil conventional (single-flow path)
MR dampers have been extensively studied, optimized and
validated for use in landing gears under different impact load
conditions. Batterbee et al. [6, 7] optimized and experimentally
validated a single-flow path MR damper retrofitted from a
passive damper. The study was conducted at a sink rate of 2.43
m/s and a drop mass of 473 kg. Hengbo et al. [8] proposed a new
1 Copyright © 2017 by ASME
phenomenological model based on the Bouc-Wen model to the devices were drop tested to evaluate their performance in
predict the MR response under shock loads. The study showed predicting the peak on-state damping force and dynamic range
that the acceleration and damping force estimated by the for a drop speed of up to 6.75 m/s; the unsteady models take into
proposed model were in good agreement with the impact test consideration the effect of fluid inertia at high speed velocity
results. Choi et al. [9] designed and drop tested a multi-coil impacts while the HM model considers the effect of fluid
single-flow mode MR damper for the use in a landing gear compressibility in addition to the fluid inertia. By comparing the
system of a helicopter. The proposed MR damper was developed simulation and test results, the authors concluded that the
based on Bingham plastic (BP) model and it also accounts for developed HM model was more accurate in predicting the
minor loss factors which otherwise were neglected in basic BP response of the bi-fold MREA because it accounts for fluid
model. The new model has been given the acronym “BPM”. The compressibility in addition to the unsteady behavior of the
system demonstrated its capability of attenuating the impact MREA.
loads for a sink rate range of 1.83-3.65 m/s. Mao et al. [10, 11]
In the present study, a bi-fold MREA will be geometrically
designed and successfully drop tested a MR energy absorber
optimized for an impact velocity of up to 5 m/s. The objective is
(MREA) at sink rate as high as 5 m/s. In their study, Bingham
to maximize the dynamic range to maintain good controllability
plastic model with minor losses (BPM model) could predict the
of the MREA without violating the constraints.
off-state damping force more accurately than the basic BP model
because the BPM model took into consideration the effect of NOMENCLATURE
non-linear velocity squared damping effect along with the minor
losses due to the geometry profile and the type of flow inside the A1 area of the bobbin core
MR valve. Singh et al. [12] performed an optimization study on A2 area of the bobbin external flange
a conventional MREA to maximize its dynamic range at impact A3 area of the 3rd linkage of the bobbin
velocity of up to 8 m/s. Using BPM model in the design of the Ad effective area of the MR valve
damper, the damper achieved dynamic range of 1.73 at impact Ap effective area of the piston
velocity of 8 m/s. In skid gear arrangement, the common A surface area of the equivalent single MR valve
drawback of using the multi-coil conventional MREAs is that B magnetic flux density
they might not fit the space between the SLG and the helicopter Bavg average magnetic flux density
fuselage due to their extended length, even though they are Bs,sat saturation magnetic flux density of the bobbin
capable of generating the required damping force at high piston material
velocities. This prompted the engineers to design compact bi- B magnetic flux density in the MR valve volume
fold MREAs which can produce the required damping force B,sat saturation magnetic flux density of MR fluid
without violating the volume constraints. c passive viscous damping of MR fluid
Mao et al. [13-15] developed and experimentally validated d MR valve thickness
various analytical models to predict the response of bi-fold D dynamic range
MREA under different impact velocities. In [13], a hydro- Db equivalent diameter of single MR valve
mechanical (HM) model was developed to account for fluid Db1 diameter of inner MR valve
compressibility and inertia as well as the losses due to minor Db2 diameter of outer MR valve
factors. The HM model was used to characterize the MREA Dh hydraulic diameter of the MR valve
under shock and impact loads for a piston velocity of up to 6.71 Di diameter of bobbin core
m/s. Based on this model, the MR damper was fabricated and Din inner diameter of damper cylinder
tested using drop test rig. The device performance was evaluated Do inner diameter of the outer flange
by dropping a mass of 45.5 kg on the damper shaft at speeds 1, Dp piston head diameter
2, and 3 m/s. The results showed that the HM model could Dr piston rod diameter
accurately predict the behavior of the MREA device. In [14] the Ds MREA cylindrical shell diameter
authors designed and constructed a bi-fold flow-mode MR f Darcy friction factor
damper. The device performance was evaluated under both F damping force
shock and vibration loads. The speed envelop was ranging from Ff friction force
high velocity of 6.75 m/s for shock energy dissipation to low Fml damping force due to minor losses
velocity of 0.1 m/s for vibration mitigation. Two analytical Foff off-state damping force
models were developed; the Bingham plastic (BP) model for Fon on-state damping force
high velocity impacts and BP model coupled with low speed F controllable damper force
hysteresis (BPH). The later was found more accurate in F Darcy friction viscous damping force
predicting the behavior of the device under high and low H bobbin length
velocities. Mao and co-authors [15] designed three MREAs on Kml-i minor loss coefficient
the basis of the following three analytical models; namely: I applied current
Unsteady-BP, unsteady-BPM, and HM dynamic models. Then, l active length of the MR valve

2 Copyright © 2017 by ASME

 l3 length of the rear flange of the bobbin cannot be neglected at higher piston speed. As the field-off
lc length of the coil damping force increases with the increase in impact speed, the
lj length of the j-th link of the magnetic circuit controllability of the MREA may significantly reduce until it
L total active length of the MR valve diminishes at D  1. Thus, maintaining laminar flow inside the
N number of wire turns per coil MR valve requires having large MR valve thickness. However,
Re Reynolds number this adversely affects controllability which requires small gap
Rt total reluctance thickness because larger thicknesses weaken the intensity of the
sgn (.) signum function magnetic flux in the MR valve.
Vb volume of one bobbin
Vd fluid velocity 3. BINGHAM-PLASTIC MODEL WITH MINOR LOSSES
Vp piston velocity (BPM)
w coil width In this study the Bingham plastic model incorporating minor
xi design variable loss factors (BPM) will be used as the analytical model to
 surface roughness optimize and evaluate the device performance. The performance
 field-independent viscosity of MR fluid
 scale factor
 scale factor
 density of the MR fluid
 magnetic flux
 permeability
s permeability of the bobbin material
 permeability of the MR fluid
y yield stress
y,max maximum yield stress
Pml viscous pressure drop due to minor losses Fig. 1. Schematic diagram of bi-fold MREA.
P viscous pressure drop due to Darcy friction
P pressure drop due to MR effect
under this model will be compared to that evaluated for the basic
Bingham plastic model (BP). The basic BP model does not
2. BI-FOLD MREA DEVICE account for the minor loss factors and used extensively in
The bi-fold MREA device, shown in Fig. 1, consists of two modeling the dynamic behavior of MR dampers in low velocity
tubes. An outer tube serves as a protection shell and an inner tube vibration applications. During the impact and because of the
forms the bi-fold MR valve geometry between the ends of the motion of the piston, the flow of the MR fluid inside the MR
tube and the bobbin. It serves as a guide of the back-forth motion valve results into three components of MREA damping forces.
of the piston. The electromagnetic coils are imbedded in the They are due to the viscous pressure drop (Darcy friction), minor
bobbins next to the bi-fold valve on either side of the damper. losses and due to the controllable pressure drop under the
The MREA devices may suffer drastic drop in performance activation of magnetic field. The minor losses are calculated
under high impact velocities (> 1 m/s). This is attributed to the using semi-empirical relations that approximate the bi-fold MR
substantial increase in the off-state damper force in proportion to valve as a tube [16]. The damping force in the BPM model is
the increase in the non-linear velocity squared. The high velocity expressed as [9, 15]
impact environment prompts counteracting the applied force by 𝐹 = (Δ𝑃𝜂 + Δ𝑃𝑚𝑙 ) 𝐴𝑝 + (Δ𝑃𝜏 𝐴𝑝 + 𝐹𝑓 ) 𝑠𝑔𝑛(𝑉𝑝 ) (1)
producing a proportional damping force in a very short duration
impact. For example, to protect helicopter occupants in the event where the Darcy friction pressure drop is given by
of hard impact or crash, the MREA should be capable of 𝜌𝐿𝑉𝑑2
generating and maintaining an average magnitude of 15 kN Δ𝑃𝜂 = 𝑓 (2)
2𝐷ℎ
stroking load in less than 200 milliseconds impact duration [9,
16]. Therefore, we should be cautious so that the stroking load here the hydraulic diameter Dh = 2 d, the total active length of
does not exceed the threshold stroking load. In high velocity the annular MR gaps L = 4l, and the Darcy friction factor (f) is a
impact loading the flow inside the MR valve may no longer be piecewise function of Reynolds number, Re, of the flow in the
laminar as characterized by Reynolds number, Re. It may fall in MR gap, d, and surface roughness,  [16, 17]. For laminar flow
the transition or turbulent regime where Re is greater than 2000
[15]. Another factor that contributes to the sharp decrease in the 96
performance is the loss due the entrance, exit, sudden expansion, 𝑓= , Re  2000 (3)
𝑅𝑒
sudden contraction and bend inside the gap region, etc. These For transition flow
losses are collectively known as the minor loss factors and

3 Copyright © 2017 by ASME

 96 1
𝑓 = (1 − 𝛼) + 𝛼 2
2000 𝜀 ⁄𝐷ℎ 1.11 6.9
{1.8 𝑙𝑜𝑔10 [( ) + ]}
3.7 4000

for 2000 < Re  4000 (4)

and for turbulent flow condition

1 𝜀 ⁄𝐷ℎ 1.11 6.9

≈ 1.8 𝑙𝑜𝑔10 [( ) + ]
𝑓 1/2 3.7 𝑅𝑒
for Re > 4000 (5)
where Fig. 2. Equivalent single MR damper [13].
2 𝜌 𝑑 𝑉𝑑 𝑅𝑒−2000
𝑅𝑒 = and 𝛼= (6)
𝜂 4000−2000
𝐹𝜏 = Δ𝑃𝜏 𝐴𝑝 (15)
The total pressure drop from minor losses is given by
𝑉𝑑2 The field-on damping force (Fon) is calculated as the summation
Δ𝑃𝑚𝑙 = 𝜌 ∑ 𝐾𝑚𝑙_𝑖 (7)
2 of field-off and MR yield forces.
The pressure drop due to MR yield stress is expressed as

∆𝑃𝜏 =
4𝐿𝜏𝑦
(8) 𝐹𝑜𝑛 = (Δ𝑃𝜂 + Δ𝑃𝑚𝑙 + Δ𝑃𝜏 ) 𝐴𝑝 (16)
𝐷ℎ
The dynamic range (D) is expressed as the ratio of the field-on
and the signum function is defined as force at maximum applied current to the field-off force
1 𝑓𝑜𝑟 𝑉𝑝 > 0 𝐹𝑜𝑛
𝐷= (17)
𝑠𝑔𝑛(𝑉𝑝 ) = { 0 𝑓𝑜𝑟 𝑉𝑝 = 0 (9) 𝐹𝑜𝑓𝑓

−1 𝑓𝑜𝑟 𝑉𝑝 < 0
3.1 MINOR LOSS COEFFICIENTS
Applying the continuity equation between the flow in the inner The accurate prediction of the damping force in the high
tube bore and the flow inside the bi-fold MR valve, the velocity velocity impact/crash events require consideration of the earlier
of the fluid flowing in the valve (Vd) can be written as mentioned minor loss factors [9, 15, 16]. The regions of such
𝐴𝑝 losses are shown on a quarter MREA in Fig. 3. Based on Ref.
𝑉𝑑 = 𝑉 (10) [16], the coefficients of the minor losses are assumed as
𝐴𝑑 𝑝
where
𝜋 Kexit = 0.6, Kentrance = 0.4, and Kbend = 0.18
𝐴𝑝 = (𝐷𝑝2 − 𝐷𝑟2 ) (11)
4
and
𝐴𝑑 = 𝜋 𝐷𝑏 𝑑 (12)
To calculate the effective area of MR valve, the bi-fold valve is
transformed to a mathematically equivalent conventional MR
single valve as shown in Figure 2 [13]. Then, the effective valve
diameter is calculated as

2𝐷𝑏1 𝐷𝑏2
𝐷𝑏 = (13) Fig. 3. Coefficients of minor loss: (a): Exit, (b):
𝐷𝑏1 + 𝐷𝑏2
Bend, and (c): Entrance. On the other side: (d):
Entrance, and (e): Exit. (piston moving to the right).
The friction force is usually produced from the contact of the
seals in the piston rod and the piston head with the bobbin and
the inner tube, respectively. In the preliminary design stage, 3.2 MAGNETIC CIRCUIT ANALYSIS
however, this force can be neglected. Expression (1) can now be The effectiveness of the MREA in generating the desired
separated to the following components controllable damping force (F) is not merely dependent on the
strength of the magnetic field but also on the magnetic properties
𝐹𝑜𝑓𝑓 = (Δ𝑃𝜂 + Δ𝑃𝑚𝑙 ) 𝐴𝑝 (14) of the candidate flux return material (Bobbin), the MR fluid, and
and the MR yield force is the dimensions of the MR valve and the bobbin. The magnetic
circuit of the bi-fold MREA is shown in Fig. 4.

4 Copyright © 2017 by ASME

 The purpose of designing and performing the magneto- 𝑁𝐼 =
𝐵𝜏 𝐴𝜏 𝑙1
+
𝐵𝜏 𝐴𝜏 𝑙2
+
𝐵𝜏 𝐴𝜏 𝑙3
+
𝐵𝜏 𝐴𝜏 𝑙4
+
𝐵𝜏 𝐴𝜏 𝑙5
+
𝜇𝑠 𝐴1 𝜇𝑠 𝐴2 𝜇𝑠 𝐴3 𝜇𝑠 𝐴4 𝜇𝑠 𝐴5
static analysis of a magnetic circuit are: a) Maximize the 𝐵𝜏 𝐴𝜏 𝑙6 𝐵𝜏 𝑙𝜏
magnetic intensity applied to the MR fluid gap in the valve. This +2 (23)
𝜇𝑠 𝐴6 𝜇𝜏
ensures generating the required controllable damping force; b)
𝑙𝜏 𝑙1 𝑙2 𝑙3 𝑙4
Avoid magnetic lock off in the bobbin by assuring that the MR 𝑁 𝐼 = 𝐵𝜏 (2 + 𝐴𝜏 ( + + + +
𝜇𝜏 𝜇𝑠 𝐴1 𝜇𝑠 𝐴2 𝜇𝑠 𝐴3 𝜇𝑠 𝐴4
fluid in the valve saturates before the bobbin. The analytical 𝑙5 𝑙6
solution assumes a linear relationship between the + )) (24)
𝜇𝑠 𝐴5 𝜇𝑠 𝐴6
magnetomotive force and the magnetic intensity. The magnetic
substituting 𝑙𝜏 = 𝑑 in the above Eq., B is computed as
circuit is analyzed using Kirchhoff law and the magnetic flux
conservation rule [17]. According to the Kirchoff law the 𝐵𝜏 =
𝑁𝐼
(25)
2𝑑
magnetomotive force of a magnetic circuit is expressed as 𝜇𝜏
+𝐴𝜏 𝑅𝑡

where Rt is the total reluctance of the bobbin links and calculated

𝑁 𝐼 = ∑ 𝐻𝑗 𝑙𝑗 j = 1, 2,…,6 (18)
from
𝑙𝑗
where N is the number of wire turns per coil, I is the current 𝑅𝑡 = ∑61 (26)
𝜇𝑠 𝐴𝑗
passing through the coil, H is the magnetic intensity, and l is the
link length. For the circuit shown in Fig. 4, Eq. (18) can be For the circuit in Fig. 4, the reluctance of different link is
written as tabulated in Table 1. To numerically confirm and validate the
analytical approach of analyzing the characteristics of the
magnetic circuit, an open Finite Element Magnetic Method
(FEMM) software was used. In the FEMM package, the active
regime of the MREA and the surrounding bobbin are modeled as
axisymmetric problem, meshed automatically using triangle, and
the flux density is calculated in the active MR gap and the
bobbin.
Table 1. Parameters of bi-fold MREA magnetic circuit.
Link Length (lj) Area (Aj) Reluctance
𝜋 2 𝑙1
1 0.5(𝑙 + 𝑙3 ) + 𝑙𝑐 (𝐷 − 𝐷𝑟2 ) 𝑅1 =
4 𝑖 𝜇𝑠 𝐴1
Fig. 4. Magnetic circuit of the bi-fold MREA.
𝜋 2 𝑙2
2 0.5(𝑙 + 𝑙3 ) + 𝑙𝑐 (𝐷 − 𝐷𝑜2 ) 𝑅2 =
4 𝑖𝑛 𝜇 𝑠 𝐴2
𝑁 𝐼 = 𝐻1 𝑙1 + 𝐻2 𝑙2 + 𝐻3 𝑙3 + 𝐻4 𝑙4 + 𝐻5 𝑙5 + 𝐻6 𝑙6 +
𝐷𝑖𝑛 + 𝐷𝑜 − 𝐷𝑖 − 𝐷𝑟 𝜋(𝐷𝑖𝑛 + 𝐷𝑜 + 𝐷𝑖 + 𝐷𝑟 )𝑙3 𝑙3
2𝐻𝜏 𝑙𝜏 (19) 3 𝑅3 =
4 4 𝜇 𝑠 𝐴3

Assuming that the magnetic flux (∅) passing through the cross 4
𝐷𝑖 − 𝐷𝑟 𝜋(3𝐷𝑖 + 𝐷𝑟 )𝑙
𝑅4 =
𝑙4
sectional areas of the links is the same, 4 4 𝜇 𝑠 𝐴4

∅ = 𝐵𝜏 𝐴𝜏 = 𝐵𝑗 𝐴𝑗 ; 𝑗 = 1, 2, … , 6 (20) 𝜋(𝐷𝑖 + 𝑤)𝑙 𝑙5

5 w-2d 𝑅5 =
4 𝜇 𝑠 𝐴5
here B is the magnetic flux density in the MR fluid in the gap, 𝐷𝑖𝑛 − 𝐷𝑜 𝜋(𝐷𝑖𝑛 + 3𝐷𝑜 )𝑙 𝑙6
6 𝑅6 =
Bj is the magnetic density of the j-th link of the magnetized 4 4 𝜇 𝑠 𝐴6
metallic material, and 𝐴𝜏 and 𝐴𝑗 are the areas normal to the path
of the magnetic field lines of the MR valve and the j-th linkage, In order to calculate the yield force, the field dependent yield
respectively. The magnetic flux density can now be expressed in stress (y) is expressed in the magnetic flux density of the MR
terms of magnetic intensity as fluid. The commercial MR fluid type MRF-132DG from Lord
Corporation [18] is used in the current study to generate the
𝐵= 𝜇𝐻 (21) required damping force. The induced yield stress is adopted from
Ref. [19] and the y-B curve is shown in Fig. 5. The function of
where  is the permeability of the magnetic linkage. Substituting 𝜏𝑦 is given in kPa as
expression (21) into (19) yields
𝐵1 𝑙1 𝐵2 𝑙2 𝐵3 𝑙3 𝐵4 𝑙4 𝐵5 𝑙5 𝐵6 𝑙6 𝐵𝜏 𝑙𝜏 𝜏𝑦 = 52.962 𝐵𝜏4 − 176.51 𝐵𝜏3 + 158.79 𝐵𝜏2 +
𝑁𝐼 = + + + + + +2 (22) 13.708 𝐵𝜏 + 0.1442 (27)
𝜇𝑠 𝜇𝑠 𝜇𝑠 𝜇𝑠 𝜇𝑠 𝜇𝑠 𝜇𝜏

The substitution of (20) into (22) gives

4. DESIGN OPTIMIZATION OF THE BI-FOLD MREA

5 Copyright © 2017 by ASME

 The objective of the current study is to maximize the Table 2. The upper and lower bounds of design variables.
dynamic range of the proposed MREA device over the impact Variable Description Bounds Unit
velocity range of up to 5 m/s. x1 = d MR valve 0.5-2.0 mm
x2 = w Coil width 3.5-8.0 mm
4.1 METHODS USED TO SOLVE THE OPTIMIZATION x3 = l Active length of MR valve 8.0-35.0 mm
PROBLEM x4 = lc Coil length 5.0-20.0 mm
To solve the optimization problem and find the optimal x5 =  Scale factor 1.02-1.07 -
solution, two optimization methodologies have been formulated x6 =  Scale factor 1.43-1.50 -
and implemented in MATLAB. The first is the stochastic based
where  and  in the above table are defined respectively as
60
Yield stress (y), kPa

𝐵𝑠,𝑠𝑎𝑡 𝐴1
50 𝛾= (28)
𝐵𝜏,𝑠𝑎𝑡 𝐴𝜏
40
𝐴1 𝐷𝑖
30 𝛽= = (29)
𝐴𝑝 𝐷𝑝
20
The factor  is defined as the flux in the bobbin core to the flux
10 in the MR gap both at their saturation state, while the factor  is
0 the ratio of the external diameter of bobbin core to the diameter
0 0.3 0.6 0.9 1.2 1.5 of the piston head.
Magnetic flux density (B), T
Fig. 5. Yield stress of MRF-132DG [19].

algorithm based on the Genetic Algorithm (GA) optimization

technique and the second is the gradient based nonlinear
mathematical technique based on the Sequential Quadratic
Programming (SQP). GA algorithms are general optimization
techniques because they do not require/use the gradient functions
in calculations. Therefore, they are used to solve all kinds of
problems. GA can provide the global optimal solution, however
it requires large number of calculations even for reasonable
problem size. Therefore, we used GA to identify the best fitness Fig. 6. Optimized design variables of the bi-fold MREA.
values and the corresponding optimal values of design variables
in the vicinity of the global optimum solution. On the other hand,
the SQP numerical technique has been proven as a powerful and Table 3. MREA parameters.
robust algorithm in solving the nonlinear optimization problems. Parameter Value Unit
SQP is a local optimizer which may trap in a local optimum. To
MR fluid density 3050 Kg/m3
ensure catching the global optimum, different initial design
points are selected randomly. These design points are taken as MR fluid viscosity 0.112 Pa-s
the optimal design points obtained earlier using GA algorithm. Relative permeability of MR fluid 6.0 
In this case, the calculation time is reduced substantially. The Relative permeability of steel 1006 1404 
technique in this case is known as “GA+SQP” technique. Two Saturation flux density of MR fluid
1.5 Tesla
codes were written in the MATLAB (R2015b) environment to (B,sat)
solve the optimization problem. Saturation flux density of Steel 1006
3.0 Tesla
(Bs,sat)
4.2 DESIGN VARIABLES Maximum permissible MR yield stress
Six control variables (illustrated in Figure 6) that determine 48.0 kPa
(y)
the geometry of the MREA are defined. These parameters and Piston rod diameter (Dr) 12.0 mm
their lower and upper bounds are given in Table 2. The other Maximum permissible bobbin diameter
design parameters of the MREA are tabulated in Table 3. The B- 90.0 mm
(Din)
H curve of the bobbin material (alloy steel 1006) is shown in Fig.
Design impact velocity (Vp) 5.0 m/s
7 [20]. The MR gap size bounds of 0.5 and 2.0 mm are selected
as per Ref. [21]. 4.3 CONSTRAINTS

6 Copyright © 2017 by ASME

 The constraints on the bi-fold MREA can be categorized
into: Geometrical, field-off damping, and the last one is due to
the magnetic circuit constraints. In the subsequent sections, the
aforementioned constraints will be described briefly.
4.3.1 CONSTRAINTS ON THE GEOMETRICAL
PARAMETERS
The MREA in this study will be assembled with a SLG
system. Therefore, the MREA exterior dimensions (diameter and
full extended length) should be constrained to fit the MREA in
the designated space between the fuselage sub-floor structure
and the SLG. In addition, the variation in the thickness of the MR
active annular channels has significant impact on the amount of
the produced damping force. Smaller valve thickness offers
superior controllability of the damping force. However, in this
case the flow inside the valve may no longer be laminar which
increases the off-field damping force due to the velocity squared
nature of the flow. Thus, it decreases the performance sharply.
Therefore, there should be a compromise between these two
requirements. In addition, the design variable () is an important
variable in maximization of the generated yield force
simultaneously with the minimization of the amount of material
in the inner tube the viscous damping force as Eq. (1) depicts.
3 F  14.0 kN
Magnetic flux density, B (T)
2.5
2 Ds  0.095 m
1.5
1 0.0035  x2  0.008 m
0.5
0 1.02  x5  1.07
0 100 200 300 400 500 600
Magnetic flux intensity, H (kA/m)
Fig. 7. B-H curve for Steel 1006 [20].
4.3.2 CONSTRAINTS ON THE FIELD-OFF DAMPING
FORCE
For the bi-fold MREA to be incorporated in a skid landing
gear, it should be capable of generating a stroking load of about
15 kN at the peak impact velocity of 5.0 m/s when the magnetic
circuit is turned off. This force is corresponding to the maximum
permissible impact load that an occupant/payload can sustain
with minor or no injuries/damage and is approximately
equivalent to sudden acceleration of 15 g’s [1, 12]. The reason
of turning off the magnetic circuit at the high-end velocity is that
the MREA can generate the required damping force passively
due to its squared relationship with the piston velocity [22].
4.3.3 CONSTRAINTS BASED ON THE MAGNETIC
CIRCUIT
The rheology of the fluid is controlled by alternating the
magnetic field which in turn is controlled by applying current to
the coil buried into the bobbin. Therefore, it is crucial to analyze
the magnetic circuit carefully to ensure that the bobbin core
saturates after the MR in the gap does. However, this might result
in a massive damper size. To satisfy the above condition at
minimum bobbin volume, a design variable () was introduced.
Factor  relates the saturation flux constraint of the effective area
of the MR link (B.A) to the saturation flux constraint of the core
material link (B1.A1). In this study a coil made of copper wire of
size 22 AWG was used. The wire diameter is 0.64262 mm and
can transmit current of up to 7 amperes. To avoid overheating the
embedded coil, a maximum current of 3.0 A is used in the design
optimization of the magnetic circuit.
4.4 FORMULATION OF THE OPTIMIZATION
PROBLEM
The optimization problem can now be formally formulated
in the standard form as
Find Design Variables X*:
𝐹
To maximize 𝐷 = 𝑜𝑛
𝐹𝑜𝑓𝑓
Subject to the following equality and inequality constraints:
Foff = 15.0 kN
B  1.5 T
Re  1600
Vb  3.7  10-4 m3
0.0005  x1  0.002 m
0.008  x3  0.035 m
0.005  x4  0.002 m
700 1.43  x6  1.45
5. RESULTS AND DISCUSSION
The optimization problem was solved using the genetic
algorithm (GA) and sequential quadratic programming (SQP)
methods. GA method was first executed to identify the near
global optimum solution. The results from GA were then used as
initial design points for SQP to find the true global optimum
solution. In all SQP simulations the different initial points
yielded the same optimum solution (dynamic range). This
confirmed that the global solution was successfully caught.
Table 4 shows the convergence of the SQP to the global optimum
solution for different design points found by GA. The
dimensions of the optimized bi-fold MREA and the other
relevant performance parameters are tabulated in Tables 5 and 6,
respectively. The geometrical dimensions are given in mm. The
iteration history of the SQP solution is presented in Fig. 8 and
shows that the objective function converges to the global optimal
solution in only three iterations.
7 Copyright © 2017 by ASME
Table 4. Convergence of SQP to the global maximum solution. quadratically. The error increases to 22% with respect to MBP
Initial design variables (X0)* Optimum design variables (X)* D model at 5 m/s impact velocity. This means that the viscous
(0.81,5.2,25.3,10.5,1.02,1.45) damping force due to the minor losses is significantly high and
(0.81,5.2,25.3,10.5,1.02,1.45) should not be neglected.
(0.81,5.3,25.3,10.3,1.02,1.45)
(0.81,5.3,25.3,10.3,1.02,1.45) (0.82,5.3,25.3,10.3,1.02,1.45) 1.93 The dynamic range (D) for the bi-fold MREA is shown in
(0.82,5.2,25.3,10.4,1.02,1.45) Fig. 10. In contrast to the damping force, the BP model
(0.82,5.7,25.1,9.6,1.03,1.45) overestimates the dynamic range of the MREA because of the
(0.82,5.3,25.3,10.3,1.02,1.45) absence of the minor loss component. The attained dynamic
* Variables x1 x4 in mm, x5 and x6 are dimensionless range was 9 and 8 for BP and BPM models, respectively, at
Table 5. Dimensions of the optimized MREA. BP model BPM model Error
Parameter Value 30 25

Damping force (kN)

MREA Shell diameter (Ds) 95.0 25 20
Bobbin length (H) 51.64 20
15

Error %
MR Valve thickness (d) 0.82
15
Coil width (w) 5.30 10
MR valve diameter (Db) 63.68 10
Piston head diameter (Dp) 40.50 5 5
MR valve active length (l) 25.30
0 0
Coil length (lc) 10.30 0 1 2 3 4 5
Impact velocity(m/s)
Table 6. Relevant parameters of the optimized MREA. Fig. 9. Comparison of damping force predictions.
Parameter Value Unit
Re 1600  impact velocity of 1 m/s. As the velocity increases, D decreases
Foff (BPM model) 15.0 kN non-linearly with a quadratic increase in the error between D for
N 168 Turn BP and BPM models. The maximum error is 35% at high-end
B 0.937 Tesla impact velocity. Fig. 11 depicted the prediction of the field-on
damping force at different current inputs for the impact velocity
y, max 48.0 kPa
range zero to 5 m/s. Here, it is observed that all the curves have
F 14.0 kN
identical pattern regardless of the value of the applied current.
This indicates that the pattern of the profile is dominated by the
off-field force. To employ the optimized bi-fold MREA in the
1.94 semi-active control strategy, the yield force is defined as a
function of the applied current as depicted in Fig. 12. The curve
1.9375 indicates that the relationship between the yield force and the
Obj Fun (D)

applied current is non-linearly (expressed as cubic function).

1.935
BP model BPM model
1.9325 10 40
35
Dynamic range (D)

1.93
0 0.5 1 1.5 2 2.5 3 6
No. of iterations
Fig. 8. Iteration history of the objective function.
Fig. 9 compares the generated on-state damping force by
BP and BPM models. It can be seen from the figure that when
the minor losses are neglected, then the damping force is linearly
proportional to the impact velocity. On the other hand, when the
minor loss coefficients are taken into account, then the curve is
non-linear and is more accurate [15]. As it can be seen the BP
underestimates the damping force, and the estimation error rises
8 30
25
Error %
20
4 15
2 10
5
0 0
1 2 3 4 5
Impact velocity (m/s)
Fig. 10. Dynamic range vs. impact velocity.
When the current supply to the MR fluid approaches 3 A, then
the fluid is magnetically saturated and no further increase in the
yield force is expected. The analytical approach obtains the
magnetic flux in the MR valve as 0.937 Tesla. The numerical

8 Copyright © 2017 by ASME

solution results in 0.8 T which is very close to that of the
analytical solution.
I=0A
30 I=2A
Damping force (kN)
25
20
15
10
5
0 Department of Commerce, Phoinex, AZ, U.S.A., 1989.
0 1 2 3
Impact velocity (m/s)
Fig. 11. Damping force generated at different
current inputs.
50
y = -2.3583 I3 + 9.9162 I2 + 7.4285 I
Yield stress, y (kPa)
40
30
20
10
0 2. Validation using Experimental Data," Smart Mater.
0 0.5 1 1.5 2 2.5
Current, I (A)
Fig. 12. Yield force vs. applied current.
6. CONCLUSION
The present optimization study was conducted to maximize
the dynamic range of the bi-fold MREA at peak impact velocity
of 5 m/s. The constraints were applied based on the practical
fabrication and size requirements of a specific skid landing gear
configuration. All the problem constraints were met. With these
constraints, the MREA was optimized to generate a peak
stroking load of 15 kN to meet the crashworthiness specifications
for vertical rotorcraft impacts as the optimized device is intended
to be incorporated in a skid landing gear of lightweight
helicopter. Two optimization methodologies were used in this
study: The GA and the SQP. The former method could capture
near global optimum solution. This solution was then used as an
initial point for SQP algorithm to refine the search of the global
optimal solution of the problem. A maximum global could be
obtained at 5 m/s peak impact velocity. Two versions of Bingham
plastic model were compared in this study, the basic Bingham
plastic, BP, and the Bingham plastic model incorporating the
minor loss coefficients, BPM. The geometry of the bi-fold
MREA could be optimized to generate a maximum damping
force of 29 kN in the field-on while ensuring the flow is laminar
at the high-end impact speed value. At this velocity a minimum
dynamic range of about two was achieved. This ensures a
wideband controllability of the device to be employed in a semi-
active system. The study emphasizes that the use of BPM model
is more accurate than the BP model since the pressure drop
estimation by the BP model leads to erroneous results.
I=1A
I=3A
REFERENCES
[1] D. F. Shanahan, Human Tolerance and Crash
Survivability, Madrid: North Atlantic Treaty Organisation
(NATO), 2004, pp. 1-16.
[2] R. E. Zimmermann, et al.,"Aircraft Crash Survival Design
Guide, Vol. 3-Aircraft Structural Crash Resistance," U.S.
4 5
[3] R. G. Fox,"Hilecopter Crashworthiness-Part One,"
Helicopter Safety, vol. 15, no. 6, pp. 1-6, Nov./Dec. 1989.
[4] P. Minderhoud,"Development of Bell Helicopter's Model
429 Sleigh Type Skid Landing Gear," in AHS 64th Annual
Forum, Montreal, Canada, 2008.
[5] "MIL-STD-1290A(AV)," Department of Defense,
Washington, D. C., 1989.
[6] D. C. Batterbee, et al.,"Magnetorheological Landing Gear:
1. a Design Methodology," Smart Mater. Sturct., vol. 16,
pp. 2429-2440, 2007.
[7] D. C. Batterbee, et al.,"Magnetorheological Landing Gear:
3 Sturct., vol. 16, pp. 2441-2452, 2007.
[8] X. Hengbo, et al.,"Experimental Investigation into
Magnetorheological Damper Subject to Impact Loads,"
Trans. Tianjin Univ., vol. 14, pp. 540-544, 2008.
[9] Y.-T. Choi, et al.,"Analysis and Control of a
Magnetorheological Landing Gear System for a
Helicopter," J. AHS, vol. 61, pp. 032006-1 - 032006-8,
2016.
[10] M. Mao, et al.,"A Nonlinear Analytical Model for
Magnetorheological Energy Absorbers under Impact
Conditions," in Proceedings of ASME 2009 Conference on
Smart Materials, Adaptive Structures and Intelligent
Systems (SMASIS2009), Oxnard, California, U.S.A., 2009.
[11] M. Mao, et al.,"Experimental Validation of a
Magnetorheological Energy Absorber Design Analysis," J.
Intel. Mater. Syst. Struct., vol. 25, no. 3, pp. 352-363, 2013.
[12] H. J. Singh, et al.,"Experimental Validation of a
Magnetorheological Energy Absorber Design Optimized
for Shock and Impact Loads," Smart Mater. Struct., vol.
23, pp. 1-17, 2014.
[13] M. Mao, et al.,"Shock Load Mitogation using
Magnetorheological Energy Absorber with Bifold Valves,"
in Proc. of SPIE Conf. on Industrial and Commercial
Applications of Smart Strcturces Technologies 2007, San
Diego, California, U.S.A., 2007.
9 Copyright © 2017 by ASME
[14] M. Mao, et al.,"A Magnetorheological Damper with Bifold
Valves for Shock and Vibration Mitigation," J. Intel.
Mater. Syst. Struct., vol. 18, pp. 1227-1232, 2007.
[15] M. Mao, et al.,"Nonlinear Modeling of
Magnetorheological Energy Absorbers under Impact
Conditions," Smart Mater. Struct., vol. 22, pp. 1-12, 2013.
[16] F. M. White, Fluid Mechanics, 4 ed., Boston, 2003.
[17] Y.-M. H. S.-B. Choi, Magnetorheological Fluid
Technology Applications in Vehicle Systems, Boca Raton:
Taylor & Francis Group, LLC., 2013.
[18] "Data Sheet of MRF-132DG Magnetorheological Fluid,"
Lord Corporation, 2011. [Online]. Available:
www.lord.com. [Accessed 25 March 2017].
[19] Q.-H. Nguyen, et al.,"Geometry Optimization of MR
Valves Constrained in a Specific Volume Using the Finite
View publication stats
Element Method," Smart Mater. Struct., vol. 16, p. 242–
2252, 2007.
[20] T. Planche, "Design Note TRI-DN-13-31 Ariel Dogleg
Vertical Dipoles EHBT:MBO & EHBT:MB5A,"
TRIUMF, 2013.
[21] Y.-J. Nam and M.-K. Park, "Electromagnetic Design of a
Magnetorheological Damper," J. Intel. Mater. Syst. Struct.,
vol. 20, pp. 181-191, 2009.
[22] Y.T. Choi, et al.,"Analysis and Control of a
Magnetorheological Landing Gear System for a
Helicopter," in AHS 68th Annual Forum & Technology
Display, Fort Worth, Texas, U.S.A., 2012.
10 Copyright © 2017 by ASME