Case Studies in Thermal Engineering 28 (2021) 101686

Contents lists available at ScienceDirect

Case Studies in Thermal Engineering

journal homepage: www.elsevier.com/locate/csite

The study on thermal management of magnetorheological fluid

retarder with thermoelectric cooling module
Philip K. Agyeman a, b, c, Gangfeng Tan a, b, *, Frimpong J. Alex a, b, d, Dengzhi Peng a, b,
Jamshid Valiev a, b, Jingning Tang e
a
School of Automotive Engineering, Wuhan University of Technology, Wuhan, 430070, China
b
Hubei Key Laboratory of Advanced Technology of Automotive Components, Wuhan University of Technology, Wuhan, 430070, China
c
Mechanical Engineering Department, College of Engineering, KNUST, Kumasi, Ghana
d
Mechanical Engineering Department, Faculty of Engineering, Kumasi Technical University, Kumasi, China
e
Hubei Center for Quality Inspection of Special Purpose Vehicle, China

A R T I C L E I N F O A B S T R A C T

Keywords: The operational temperature of the magnetorheological (MR) fluid retarder is usually over 80 ◦ C,
Thermoelectric cooling and it reaches 150 ◦ C when heat dissipation doesn’t function well. The high temperature will
MRF retarder seriously affect the working efficiency of the MR fluid by decreasing the viscosity of the fluid,
Braking torque which affects the generated braking torque. For the analysis of the heat dissipation behavior of
Thermal resistance MR fluid retarder, an experimental study of a thermal management system using thermoelectric
Cooling capacity
cooling (TEC) module with heat sink as the cooling unit was conducted. The variation tendencies
of temperature for the TEC module were investigated during the working process of the MR fluid
retarder using the TEC module and natural convection air-cooling. The performance of the MR
fluid at the desired temperature was analyzed by applying one-dimensional thermodynamic
model and a numerical calculation method. The influences of temperature on the performance of
MR fluid retarder were analyzed by comparing the performance characteristics of the TEC
modules with Peltier-Seebeck effect and natural convection air-cooling systems. The results
indicate that TEC has a better effect on the MR fluid retarder performance compared to the
natural convection air cooling systems. The TEC module coupled with heat sink achieved a higher
heat transfer efficiency, maintained the working temperature of the MR fluid in the retarder,
improved the retarder’s efficiency, and maximized the braking torque generated. The proposed
system will further enhance the longevity of the retarder and provide a platform for heat dissi­
pation and cooling of automobile systems.

* Corresponding author. School of Automotive Engineering, Wuhan University of Technology, Wuhan, 430070, China.
E-mail addresses: [email protected] (P.K. Agyeman), [email protected] (G. Tan), [email protected] (F.J. Alex), [email protected].
cn (D. Peng), [email protected] (J. Valiev), [email protected] (J. Tang).

https://doi.org/10.1016/j.csite.2021.101686
Received 19 September 2021; Received in revised form 15 November 2021; Accepted 3 December 2021
Available online 6 December 2021
2214-157X/© 2021 The Authors. 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/).
P.K. Agyeman et al. Case Studies in Thermal Engineering 28 (2021) 101686

Nomenclature

Acp Cross section of ceramic (m2 )

ATEC Cross section of thermoelectric pair (m2 )
Aw Surface area of retarder wall (m2 )
cp Specific heat capacity J/kg/K)
DR Diameter of retarder (m)
H Magnetic field
h Heat transfer coefficient
K Thermal conductivity (W /m /K)
lTEC Length of TEC thermocouple (m)
Lcp Length of ceramic plate (m)
m Mass flow rate (kg /s)
N Number of disc
Nu Nusselt number
I Applied current (A)
P Power consumption of TEC (W)
Pr Prandtl number
Q̇ Heat transferred per unit time (W)
Qc Cooling capacity (W)
Qh Heat rejection (W)
qx Heat flux (W /m2 )
R Electrical resistance (Ω)
r1 Inner radius of disc (m)
r2 Outer radius of disc (m)
T Temperature (K)
Kf Thermal conductivity of MR fluid (W /m /K)
t Time (s)
Tb Braking torque (Nm)
Tc Cold surface temperature of TEC (K)
Th Hot surface temperature of TEC (K)
Tm Mean temperature (K)
RT Total thermal resistance (K /W)
z Figure of merit
Re Reynolds number

Greek symbol
α Seebeck coefficient (V /K)
μ Viscosity of fluid
ρ Electrical resistivity
τ Shear stress (Mpa)
γ̇ Shear rate (s− 1 )
τy Yield stress (kPa)

1. Introduction
Magnetorheological fluid (MRF) retarder is an auxiliary brake system that assists the main braking system of a vehicle. The MR
fluid is an induced fluid that changes its properties under the influence of a magnetic field. During operation of the MR fluid retarder,
an increase in temperature causes the viscosity of the fluid to decrease. The heat generated in the MR fluid causes a significant
reduction in the output braking torque and subsequently affects the performance of the MRF retarder [1]. To prevent the MR fluid from
losing its viscosity and control the heat dissipation, a TEC system is proposed. The TEC exhibits an efficient way of cooling because of
the Peltier effect. Its high-level thermal performance is relevant to the regulation of the heat generated by the MR fluid retarder system.
To maximize the thermal performance and braking capability of the MRF retarder, it is vital to identify and understand the concepts
that guide the cooling performance and structural parameters of the TEC.
Several scholars have researched into the thermal management of MRF brakes and other cooling methods. Liu, W. et al. [2]
proposed the use of an independent two-phase evaporator to strengthen the coolant heat absorption capacity in the transmission
medium via oil outlets of the retarder. This was done to increase the thermostability of heavy-duty vehicles during working operations
by applying vacuum flow boiling heat transfer. They concluded that the two-phased warm exchanger fortifies the heat retention

2
P.K. Agyeman et al. Case Studies in Thermal Engineering 28 (2021) 101686

capacity of the cooling medium. Additionally, it retains the heat from the transmission medium within the water-powered retarder.
Xiong, S. et al. [3] studied the effect of temperature on the braking efficiency of MRF braking device to stabilize the braking torque by
compensating the magnetic field strength. The results showed that the braking torque is nonlinear and decreased with a rise in
temperature under the same magnetic field. Wang, D.M. et al. [4] experimented on a novel high-torque MR brake, in which a
water-cooling medium was adopted to assist in dissipating the heat generated during the braking process. Natural cooling of the MR
fluid’s temperature increased linearly with respect to time. However, the forced cooling (water cooling) medium effectively slowed the
increase of the fluid’s temperature. Weiss et al. [5] experimentally considered the effect of temperature increment on the phase
characteristics of an MR fluid and found that the plastic viscosity and dynamic yield stress diminished by 95 and 10%, as the tem­
perature rose from − 40 to 150 ◦ C.
Thermal effects under different schemes of the MR fluid are also studied [6–8], from which the temperature gradient distribution is
observed to relate to the heat transfer capabilities and output braking torque. Philip A.K. et al. [9] researched on the thermal man­
agement of magnetorheological fluid auxiliary braking device with a flat plate heat pipe (FPHP) as the cooling medium. The authors
optimized the coupling relationship between the heat generated at the wall of the retarder and FPHP. The results showed a decrease in
temperature at the retarder’s wall by 5 – 8 ◦ C. Erol, O., et al., and Nagaya, K., et al. [10,11], performed a study on the temperature
distribution and characteristics of the MR fluid. In their analysis, they noted that temperature has a great influence over the MR fluid,
hence, it curbs the implementation in phase-type engineering.
Thermoelectric (TEC) is a smart device that has temperature variations with voltage input. The integration of the Peltier device
with phase change materials was presented by Zhao et al. [12]. The phase change material acted as a capacity buffer to bolster space
cooling at a point when the Peltier device was inaccessible. In addition to single-stage Peltier devices, many researchers have also
studied two-stage Peltier devices [13–15], where it is commonly concluded that the two-stage Peltier device performs with a higher
coefficient of performance (COPs), particularly when the desired temperature differential is generally expansive. Ibáñez-Puy et al.
[16], further examined the effect of Peltier device for space warming of a little adiabatic box with the extreme point by applying the
information to the device in buildings.
Sangchabdr [17] adopted the approach of cooling with the combination of heat pipes and thermoelectric module. The proposed
cooling system had sufficient heat dissipation of 200W. The temperature of the proposed cooling system was found to be lower
compared to the existing systems. Xiaoqin [18] analyzed the performance of a thermoelectric cooling system with gravity-assisted heat
pipes for cooling electronics. The cooling capacity was improved by 64.8% and electricity consumption was reduced by 39.3%
compared to TEC with air-cooling heat sink. However, many reports revealed the possibility of using TEC modules as a promising
competitor to conventional air-cooling and heating systems. Tan et al. [19] display a TEC framework for air conditioning with inte­
gration of a phase change material as a substitute to heat sink. The expansion of the phase change material empowered them to reach
56% upgrade of the TEC module implementation and an increase in COP from 0.5 to 0.78. Cosnier et al. [20] presented a study on the
TEC cooling and heating framework with four TEC modules at different air flow rates and an input current ranging from 0 to 5A. The
results showed that in the heating mode, COP ranged from 0.7 to 4.5 at temperatures between 10 and 70 ◦ C, whereas in the cooling

Fig. 1. Physical model of thermoelectric unit.

3
P.K. Agyeman et al. Case Studies in Thermal Engineering 28 (2021) 101686

mode, the COP values varied between 0.3 and 4.5 at surface temperatures of 5–30 ◦ C.
Gokcek and Sahin [21] did also conduct their research on the performance of a mini-channel water-cooled thermoelectric fridge.
They assessed the performance of their model by considering the cold surface temperature, COP, and cooling rate at different cooling
water mass rates. Their results demonstrated that the integration of the mini-channel heat sinks increased the effectiveness of TEC,
with a temperature decrease of 2 ◦ C at 0.8L/min to − 0.1 ◦ C at 1.5L/min. Additionally, in a 2hr duration, the COP increased from 0.19
at 0.8L/min to 0.03 at 1.5L/min. Astrain et al. [22] attempted to improve the COP of thermoelectric fridge by optimizing the heat
dissipation of the hot surface of TEC module. Simultaneously, using phase-change materials and thermosiphon, COP of the framework
improved by 23% with the thermal resistance of the heat sink increasing up to 36%.
The literature presented is an indication of the extensive studies on thermoelectric cooling (TEC) have advanced. In this paper, a
MRF retarder is taken as the research object. Based on the experimental measurement of TEC performance parameters, the equivalent
thermal path calculation method is adopted to build a thermal resistance model with and without TEC assembly under natural con­
vection, and a reasonable mathematical formula is selected to describe its performance. Then, the variation trend of heat dissipation
performance of MRF retarder with the working current of TEC was calculated and analyzed. Therefore, this paper studies the cooling
application of thermoelectric devices to regulate, maintain the temperature of the MR fluid retarder, and to enhance the braking torque
of a vehicle.

2. System mathematical modeling

2.1. Mathematical modeling of thermoelectric cooling module
2.1.1. Physical model of thermoelectric element
In this study, the most commonly used numerical model is the thermal resistance model, which is based on the energy balance
equation. Fig. 1 illustrates a schematic diagram of a miniature thermoelectric unit.
The working performance of the TEC can be characterized by performance indexes. These includes the net heat absorption at the
cold surface, total heat released at the hot surface, maximum temperature differences and performance coefficient [23]. To facilitate
the analysis and calculation, the following assumptions were made [24,25] and more details can be found in Ref. [26]:
1. One-dimensional steady state heat transfer was considered, i.e., the heat flows only in the direction parallel to the electric arm.
2. Half of the Joule heat generated by the electric arm flows to the cold surface of the TEC and half to the hot surface.
3. The electric arm layer of the TEC is very thin and the temperature gradient in the electric arm is very small, hence, the Thomson
effect is ignored.
4. The physical properties of the TEC such as the Seebeck coefficient, electrical resistance, and thermal conductance are independent
of temperature.
5. The contact and electrical resistance of the ceramic plates, electrical soldering, and cable are negligible.
6. The thermal losses of the TEC and heat sink are not considered.

Based on the forementioned assumptions and analytical theories, both Peltier and Thomson effects are bulk heating effects.
Thomson effect describes the reversible heat or cooling in a homogenous semiconductor material. The Peltier effect shows the
endothermic and exothermic effects at the interface between the cold and hot surfaces. Hence, the cooling effect of the Peltier module
is vital in regulating the temperature of the MRF. This paper chose the TEC1 – 12706 module as the cooling device due to its per­
formance and commercial availability. The technical parameters of the module are shown in Table 1.
The performance parameters and energy equilibrium model have been verified by Refs. [25,27–30] and generally used in deter­
mining the heat transfer rates. According to the boundary condition, the energy equations for the cooling capacity, heat rejection,
power consumption, and performance efficiency are expressed in Table 2.
From the equations described in Table 2, the cooling capacity, current, and the temperature difference between the hot and cold
surfaces are not linear. Thus, it is essential to evaluate these parameters for different applications. Such evaluation was carried out in
this study. The authors considered two major conditions for understanding the effect of these variables for different applications. One
such condition is to maximize the cooling capacity while the other is to maximize the cooling efficiency [23].

Table 1
Technical Parameters of TEC module TEC1 – 12706.

Parameter Value Unit

Dimensions (LxWxH) 40x40x38 mm

No. of thermocouples 127
Seebeck Coefficient 0.05343 V/K
Electric Resistance 2.1–2.5 Ω
Thermal Conductivity 0.52 W/K
Maximum Rated Voltage 15.4 Volts
Maximum Rated Current 6 Amps
Maximum Cooling Capacity 55–65 Watts
o
Maximum Temperature Difference >65 C
o
Range of Operating Temperature − 50–85 C
Weight 22 g

4
P.K. Agyeman et al. Case Studies in Thermal Engineering 28 (2021) 101686

Table 2
Calculation of the TEC performance parameters.

Performance parameters Equation

Cooling capacity, (W) Qc = αITc − 0.5I2 R − KΔT (1)

Heat rejection, (W) Qh = αITh + 0.5I2 R − KΔT (2)
Power consumption, (W) P = I2 R + αIΔT (3)
Performance efficiency αITc − 0.5I2 R − KΔT Qc
ε= = (4)
I2 R + αIΔT P

2.1.2. Maximum cooling efficiency working condition

The performance metric for evaluating the maximum cooling efficiency are to establish the required optimum current. Hence, the
optimum current is calculated from the partial derivation of Equation (4) and setting the current to zero. The optimum current, op­
timum cooling capacity and maximum cooling efficiency are computed as shown in Equations 5–7, respectively.
( )
2KM M − Th/T
(5)
c
Qo = ( )
(M − 1) 1 + Th/T
c

( )
M − Th/T
Tc
(6)
c
εmax =
ΔT M+1

αΔT αΔT
Iopt = ( √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅) = (7)
K (1 + ZTm ) − 1 K(M − 1)

√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅ α2 . ZT is the figure of merit of the mean temperature, T .

where, M = 1 + ZTm , Z = KR m m

2.2. MR fluid retarder braking model

In MRF retarder, the braking characteristics of the MR fluid are manifested in the magnitude of the braking torque produced by the
brake. The MR fluid produces a rheological effect when acted upon by a magnetic field H, which results in the braking torque. The fluid
shear stress τ(γ̇, H) delivered by the fluid was considered as a contribution of both fluid-dependent dynamic yield stress τy (H) and a
viscous friction which depends on the shear rate γ̇ and the fluid viscosity μ. The constitutive equation can be described as:
{
τ = τy (H) + ηγ̇, τ ≥ τy (H)
(8)
γ̇ = 0, τ ≤ τy (H)

The brake torque varies with the applied current to the electromagnet coil, the angular velocity, and the radius of the rotating disc.
The MR fluid braking torque combines the torque generated by the applied magnetic field and the viscous friction of the MR fluid as
shown in Equation (9) [31].
2N πτ ( 3 ) N πωμ ( 4 )
Tb = r2 − r13 I + r2 − r14 (9)
3 2h

Where r2 and r1 are the outer and inner radius of the disc, N is the number of the disc and I is the applied current.

3. Heat transfer characteristics

3.1. Heat dissipated from MR fluid retarder
Analysis of the temperature distribution is fundamental in the operation of the MR fluid retarder as the fluid works within a limited
temperature range. The heat generated from MR fluid retarder exits the system through a convectional heat transfer caused by me­
chanical loss between sliding of the rotors and loss of electrical power, i.e., Joule effect due to the current in the coil [32,33]. The
amount of heat transferred per unit time, Q̇ is expressed as in Equation (10):
( )
Q̇ = − hAw T(t) − T∞ (10)

Where h is the convection coefficient, Aw is the surface area of the retarder wall affected by the convection flow, T(t) .and T∞ is the
function of temperature with respect to time and ambient temperature. Using the Dittus-Boelter correlation equation [33], the heat
convection coefficient is calculated as a function of the Nusselt number and expressed as:
( )
/
h = Kf D *Nu (11)
R

5
P.K. Agyeman et al. Case Studies in Thermal Engineering 28 (2021) 101686

cp *μ 2ro2 − 2ri2
where, Nu = 0.0265Re0.8 Pr0.3 , Re = m*D
μ*Aw , Pr =
R
Kf , and. DR = 2ro +2ri
The mechanical and electrical power entering the system is computed from Equation (12):

2N πωτ ( 3 ) N πω2 μ ( 4 )
Ẇ = r2 − r13 I + r2 − r14 + I 2 R (12)
3 2h
The thermal characteristics of the MRF retarder were analyzed in a steady state, hence, the temperature of the fluid was constant
under ideal conditions. Based on the energy conservation law, the rate of energy stored in the system is estimated from Equation (13):
dEst
Ėst ≡ = Ėin − Ėout + Ėg (13)
dt

Where Ėin and Ėout are the rate of thermal and mechanical energy entering and leaving the system. Ėg is the rate of energy generated in
the system. The kinetic energy from the transmission shaft of the MRF retarder were converted into the thermal energies during the
working process of the retarder. The energy generated is the product of the temperature gradient and the sum of the thermal masses of
components in the system. Hence, based on the energy conservation law, the overall energy equation of the MR fluid retarder is
expressed in Equation (14):
( ) ∑ dTi
− hAw T(t) − T∞ + Ẇ = cpi · mi · (14)
dt

3.2. Thermal resistance of system

The thermal resistance at the sides of the thermoelectric module is significant in determining the performance of the thermoelectric
cooling system. Fig. 2 shows a schematic diagram of the thermoelectric system, in which the MRF retarder and heat sink are situated on
either side of the thermoelectric module. The heat flux (qx ) absorbed by the thermoelectric cooling module is assumed to be of equal
′

amount to the flow through the cold surface Tc . Analyzing the heat flow in the thermoelectric cooling module and the flow of air
around the retarder and heat sink, the thermal resistance of the thermoelectric cooling module is shown in Equation (15) [34]. The
physical representations of the thermal resistance between the wall of the MRF retarder and the heat sink are shown in Table 3.
According to the cooling path shown in Fig. 2, temperatures are set to a series of nodes: temperature at the wall of the MRF retarder,
Tw , temperatures at the cold and hot surfaces of the thermoelectric cooling module Tc & Th , the temperatures at the base of the heat
sink and at the fins Tb hs & Tfin . The thermal resistance in series constitutes an equivalent thermal path model. The total thermal
resistance, RT , of the MRF retarder cooling system is the sum of the individual components thermal resistance as shown in Equation
(15):
( )
Lwall Lcp ρlTEC α Tb hs − Tfin
RT = RMRF− Wall + RWall− cp + Rcp− TEC + RTEC− HS = + + 1+ + (15)
hmrf Awall kcp Acp ATEC lTEC mCp (Tin − Tout )

4. Experimental set-up
To achieve the desired results, the thermal management system of the MRF retarder and thermoelectric cooling systems must be
closely related to the experimental set-up as shown in Fig. 3. The thermal performance of the system was evaluated by conducting an
experiment. The apparatus and procedure of the experiment are discussed as follows. A thorough check of the equipment was carried
out before the experiment. The ambient temperature was recorded at 23 ◦ C ± 0.5 ◦ C and set at 22 ◦ C during the simulation to account
for varying conditions. Four TEC modules were installed at specified measurement points on the wall of the MRF retarder. The TEC
module was connected in series and powered by a DC power supply. The initial voltage supply to the TEC module was set at 2V with an
increment of 1V.
Temperatures were measured by PT1000 sensors and recorded at the end of every experimental cycle. The rotating speed of the
retarder was set at 100 rpm with an increment of 50 rpm and controlled by an ABB-ACS510 drive. Table 4 shows the components of the
data acquisition system. A LabVIEW program was generated to acquire the temperature at the wall of the MRF retarder. Eight analog

Fig. 2. Thermal Resistance Analysis of the MRF retarder-TEC Cooling System.

6
P.K. Agyeman et al. Case Studies in Thermal Engineering 28 (2021) 101686

Table 3
Physical meaning of thermal resistance.

Symbol Meaning SI unit

R1 = Contact thermal resistance between the MRF retarder wall and cold surface of the TEC module. K/
W
RMRF− Wall

R2 = RWall− cp Thermal resistance between the ceramic plate and the metal core printed circuit board.
R3 = Rcp− TEC Thermal resistance between the cold and hot surface of the TEC.
R4 = RTEC− HS Thermal resistance between the hot side of the TEC and bottom of the heat sink.

Fig. 3. The MRF retarder and DAQ experimental set-up.

Table 4
Main components and models of the data acquisition system.

Name Model Parameter

Speed torque sensor LorenZDR-211 Range: ±30 N⋅m

Temperature Sensor PT1000 ±200 ◦ C
Data acquisition card 1 NIcDAQ-9174
Data acquisition card 2 NIPXIe-6361
Industrial computer NIPXIe-8840
Junction Box NISCB-68A
Gauss meter CH-hall3600
DC power supply EDKORSIT6500 Two-way
ABB ACS510

Fig. 4. Relationship between Cooling Efficiency and Heat Transfer Efficiency with varying Input Currents.

7
P.K. Agyeman et al. Case Studies in Thermal Engineering 28 (2021) 101686

input channels were used to measure the data from the set-up. The experimental data was recorded using NI cDAQ9174 device
interfaced with a computer and recorded at intervals of 60s. The temperature data observed were those of the retarder wall tem­
perature and thermoelectric cold surface temperature.

5. Results and discussion

5.1. Influence of heat transfer coefficient on cooling capacity and efficiency of the MRF retarder
The effect of the thermoelectric cooling module on the thermal management system of the MR fluid retarder is simulated in Figs. 4
and 5. The key parameters that were factored into the simulation process are the temperature variations, the speed of the retarder, the
heat dissipated within the retarder and the cooling performance of the TEC. In Fig. 4, the cooling efficiency at 0.5, 0.8, 1 and 1.2A had
an increment of 17.92, 19.61, 20.38 and 21.1% with reference to the heat transfer coefficient which shows the effect of the input
current on determining the cooling efficiency of system. A similar trend is shown in Fig. 5 with the cooling capacity with respect to the
input current and the heat transfer coefficient. The variations tendencies per each current were recorded as 30.65, 40.16, 49.58 and
59.33%. From the observed percentage increment in the cooling capacity and efficiency, it is safe to say that an increase in the input
current to the TEC, results in an increase in the cooling capacity and efficiency. Additionally, heat transfer characteristics between the
hot surface of the TEC and the heat sink was used to analyze the change of thermal cooling capacity and efficiency of the TEC.
Siahmargoi, M [35]. point out that, increasing the convective heat transfer coefficient at the cold and hot surfaces has a significant
effect on the cooling capacity and efficiency of a TEC, which is vital controlling the wall temperature of the MRF retarder. Also, Krakov,
and Nikiforov, Wen, Tsai, and Leong [36,37], studied the thermal and magnetic characteristics of the magnetic fluid and reported that
the heat transfer characteristics is related to the temperature gradient. From Figs. 4 and 5, the cooling efficiency had a range of
0.34–0.48% and the cooling capacity ranged between 10.09 and 15.98W per each current reading with regards to the heat transfer
coefficient at the hot surface of the TEC.

5.2. Influence of thermal resistance on MRF retarder-TEC/heat sink

With reference to Fig. 2, the thermal resistance between the MRF retarder wall and the cooling surface of the TEC, and the thermal
resistance between the hot surface of the TEC and heat sink is shown in Figs. 6 and 7. Under a constant value of cooling capacity, the
thermal resistance between the MRF retarder wall and the TEC’s cold surface decreases at the initial current of 0.5A and then increases
until it reaches the optimum current Iopt as shown in Fig. 6. In the figure, the optimum current Iopt varied from 0.67 to 0.84A at the
rotating speed 700 rpm and gradually increased as Iopt approached 1A. The experimental curve followed a similar trend with a per­
centage variation from the simulation curve of 11.7%. The thermal resistance between the two surfaces reaches near equilibrium at an
optimal current of 0.97A, which shows that the current input at the TEC has a significant effect on the thermal resistance and hence the
temperature at the MRF retarder wall.
In Fig. 7, the simulated and experimental curves have a similar trend due to the heat dissipated from the hot surface of the TEC and
the material properties of the heat sink. The percentage variation between the experimental and simulated curve is 10.1% at the
optimum current of 0.84A. It can be seen that the thermal resistance increased sharply within a short duration of time and gradually
decreased due to the fin efficiency of the heat sink.
Furthermore, the difference between the simulation and the experimental results mainly come from three sources:

Fig. 5. Relationship between Cooling Capacity and Heat Transfer Efficiency with varying Input Currents.

8
P.K. Agyeman et al. Case Studies in Thermal Engineering 28 (2021) 101686

Fig. 6. Thermal Resistance between the MRF retarder Wall and TEC’s Cold Surface.

Fig. 7. Thermal resistance between the TEC’s hot surface and heat sink.

1. In terms of material and structure, the simulation ignores the change of the thermal physical property of the TEC with temperature.
The additional heat transfer temperature difference, the weld resistance, and the miscellany heat dissipation exchange between the
cold and hot surfaces also had an impact on the cooling effect in the experiment, which was ignored.
2. In terms of temperature distribution processing, the numerical simulation assumes that any point in the heat sink has the same
temperature, which is impossible from an experimental perspective. In the experiment, the temperature distribution of heat sink is
a time-varying non-uniform field, meaning the heat transfer between the heat sink and the environment is much more complicated
than in the simulation.
3. In the experiment, due to the thermal resistance between the ceramic insulation layer, the filler metal layer, the thermal conductive
plate itself and the contact surfaces, additional heat transfer temperature difference is generated. This results in the actual tem­
perature difference between the two ends of the TEC decreasing with the increase of thermal resistance under the same conditions.

5.3. Influence of TEC on MRF retarder wall temperature and braking torque
The temperature characteristics of the MRF retarder with and without the TEC is simulated and experimented as shown in Fig. 8.
From Fig. 8a, it can be seen that there is a continuous increase in temperature at the wall of the MRF retarder with time. At the peak
speed of 1200 rpm, the temperature at each measurement points peaks and had an average of 129.4 ◦ C for the simulated curve and

9
P.K. Agyeman et al. Case Studies in Thermal Engineering 28 (2021) 101686

Fig. 8. Effect of TEC on MRF retarder.

117.3 ◦ C for the experimented curve with a variation of 9.3%.

In Fig. 8b, by observing the simulation and experimental data, it is can be seen that the temperature from the four points of the TEC
positions shows that the temperature of the MRF retarder is not completely distributed according to the radial distance of the retarder
wall. Due to gravity, the flow rate from the top quadrant is higher than the bottom quadrant, hence the friction inside the retarder is
faster and the temperature is higher under the same heat dissipation conditions. From Fig. 8b, the heat dissipation from the MRF
retarder is minimized from two aspects. First, the effect of the TEC with the heat sink which reduces the temperature and the second is
the speed of rotation. With the TEC installed at the wall of the MRF retarder, the temperature is uniform.
Likewise, from Fig. 9, the temperature effect at the retarder wall is plotted against the rotating speed of the retarder’s rotor. It can
be seen that the simulated and experimental curves with TEC attain equilibrium of 97.7 and 94.1 ◦ C with a variation of 3.75%. Without
the TEC, the simulated and experimented curves recorded 128.5 and 121.8 ◦ C with a variation of 5.2%.
The simulation and experimental curves show the effect of the TEC on the braking torque of the MRF retarder and temperature at
the surface of the retarder wall in Fig. 10. The braking torque was very much affected by the temperature as also presented by Limpert
[38]. In a related study, Zhou, L., et al. [39] identified that in order to ensure the braking safety of a vehicle, it is important to maintain
the temperature of the fluid within the working temperature of the fluid while braking. As can be seen from Fig. 10, the experimental
results follow a similar trend as the simulation results. At a rotating speed of 700 to 1200RPM, the braking torque for the experiment
reduced by 8.6% and 9.03% for the simulation with the TEC. In addition, at the same speed range, the braking torque decreased by

Fig. 9. Relationship between MRF wall temperature and speed.

10
P.K. Agyeman et al. Case Studies in Thermal Engineering 28 (2021) 101686

Fig. 10. The Effect of the TEC Module on Braking Torque of MRF retarder.

34.3% and 40.7% for the simulation and experiment within the working temperature. Hence, the simulation braking torque deviated
4.76% from the experimental braking with the TEC and 15.7% without the TEC.
In the simulation process without the TEC, the effect of the fluid viscosity and temperature of the fluid is related to the speed of the
retarder under a natural air-cooling process. This same principle is applied in the experimental process without the TEC. Due to the
speed of the retarder, the viscosity of fluid losses its magnetorheological properties and increases in temperature which causes the
braking torque to reduce. Whilst, in the simulation and experimental process with the TEC, the cooling effect plays an important role in
maintaining the magnetorheological and temperature effect of the MR fluid. Hence, due to small variations, the simulation results can
be used as a means to determine the brake torque with TEC, especially when the experimental results are not available. It is convenient
to conclude that the TEC module with high current input, high heat transfer coefficient at the hot surface of the TEC, and low thermal
resistance at the wall of the MRF retarder is able to maintain the temperature of the MR fluid during operation and achieve the desired
braking torque.

6. Conclusion
This paper presents the study on the thermoelectric cooling of an MRF retarder. The main aim was to compare the effect of the TEC
on MRF retarder over natural convection air-cooling and the following observations were made.

1. Based on the heat transfer model of the system, the thermoelectric cooling of the MRF retarder is simulated numerically. The results
show that at the optimal current, the cooling effect of the TEC and heating power from the MRF retarder reaches equilibrium faster.
Moreover, the effectiveness of the heat sink provides a better cooling effect.
2. With reference to Refs. [9,40] and analysis of data collected, the TEC had a higher rate of reducing the temperature at the wall of
the retarder. Quan, J. and B. Huang projected a temperature reduction between 5.2 and 7.57%, Philip, A.K., et al., had a tem­
perature reduction of the retarder of 7.2% and the TEC with the retarder reduced the temperature of 3.5% in comparison to the
other papers. Therefore, the TEC has a better cooling effect and a bigger influence on the braking torque of the MRF retarder.
3. The MRF retarder experiment system includes many parts, such as simulation, thermal interface material, thermoelectric module
and heat sink, etc., and there are errors in the temperature measurement of each part. In addition, there are also measurement
errors in the data collection and detection part of the experiment, and the errors are transferable. The errors of the experimental
system are usually magnified due to the accumulation of errors of various components. As there are many influencing factors in the
experiment and the single experiment lasts for a long time, the change of ambient temperature and air flow will have an impact on
the experimental results. Therefore, in the further study, this experiment should be carried out in an airtight space to minimize the
shadow of external factors.

Data availability statement

Not applicable.

11
P.K. Agyeman et al. Case Studies in Thermal Engineering 28 (2021) 101686

Funding
This research received no external funding.

CRedit author statement

Philip K. Agyeman: Conceptualization, Methodology, Software, Writing-Original draft preparation. Frimpong J. Alex: Resources,
Writing-Reviewing and Editing. Dengzhi Peng: Software, Visualization. Jamshid Valiev: Software, Investigation. Jingning Tang:
Project administration. Gangfeng Tan: Supervision, Validation.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to
influence the work reported in this paper.

Acknowledgments
The authors give acknowledgements to Hubei Key Laboratory of Advanced Technology for Automotive Components, Wuhan
University of Technology, Wuhan 430070, China; School of Automotive Engineering, Wuhan University of Technology, Wuhan
430070, China, and Suizhou-WUT Industry Research Institute, Suizhou 441300, China.

References
[1] W. Song, et al., Thermal and tribological characteristics of a disc-type magnetorheological brake operated by the shear mode, J. Intell. Mater. Syst. Struct. 30 (5)
(2019) 722–733, https://doi.org/10.1177/1045389x18770740.
[2] W. Liu, et al., Boiling coolant vapor fraction analysis for cooling the hydraulic retarder, SAE Int. J. Eng. 8 (4) (2015), https://doi.org/10.4271/2015-01-1611.
[3] S. Xiong, et al., Effect of Temperature on Braking Efficiency Stability of Magnetorheological Fluid Auxiliary Braking Devices. https://doi.org/10.4271/2017-01-
2510, 2017, 1.
[4] D.M. Wang, Y.F. Hou, Z.Z. Tian, A novel high-torque magnetorheological brake with a water cooling method for heat dissipation, Smart Mater. Struct. 22 (2)
(2013), 25019, https://doi.org/10.1088/0964-1726/22/2/025019.
[5] K.D. Weiss, T.G. Duclos, Controllable fluids: the temperature dependence of post-yield properties, Int. J. Mod. Phys. B 8 (20n21) (1994) 3015–3032.
[6] N. Wang, et al., Effect of temperature on the transmission characteristics of high-torque magnetorheological brakes, Smart Mater. Struct. 28 (5) (2019), 57002.
[7] M. McKee, F. Gordaninejad, X. Wang, Effects of temperature on performance of compressible magnetorheological fluid suspension systems, J. Intell. Mater. Syst.
Struct. 29 (1) (2018) 41–51.
[8] I.Y. Forero-Sandoval, et al., Study of thermal conductivity of magnetorheological fluids using the thermal-wave resonant cavity and its relationship with the
viscosity, Smart Mater. Struct. 26 (2017), 25010.
[9] A.K. Philip, et al., Research on the performance of magnetorheological fluid auxiliary braking devices thermal management system based on flat plate heat
pipes, SAE Int. (2020) 1–10.
[10] O. Erol, et al., Magnetic induction control with embedded sensor for elimination of hysteresis in magnetorheological brakes, J. Intell. Mater. Syst. Struct. 23 (4)
(2012) 427–440, https://doi.org/10.1177/1045389X11435432.
[11] K. Nagaya, et al., MR fluid viscous coupling and its torque delivery control, Tribol. Int. 40 (1) (2007) 89–97, https://doi.org/10.1016/j.triboint.2006.02.059.
[12] D. Zhao, G. Tan, Experimental evaluation of a prototype thermoelectric system integrated with PCM (phase change material) for space cooling, Energy 68
(2014) 658–666, https://doi.org/10.1016/j.energy.2014.01.090.
[13] S.C. Kaushik, S. Manikandan, The influence of Thomson effect in the performance optimization of a two stage thermoelectric cooler, Cryogenics 72 (2015)
57–64, https://doi.org/10.1016/j.cryogenics.2015.08.004.
[14] T.h. Wang, et al., Parameter analysis and optimal design for two-stage thermoelectric cooler, Appl. Energy 154 (2015) 1–12.
[15] H. Lv, et al., Enhancement of maximum temperature drop across thermoelectric cooler through two-stage design and transient supercooling effect, Appl. Energy
175 (2016) 285–292, https://doi.org/10.1016/j.apenergy.2016.05.035.
[16] E. Ibáñez-Puy, et al., Thermal and energy performance assessment of a thermoelectric heat pump integrated in an adiabatic box, Appl. Energy 228 (2018)
681–688, https://doi.org/10.1016/j.apenergy.2018.06.097.
[17] B. Sangchandr, V. Afzulpurkar, A novel approach for cooling electronics using a combined heat pipe and thermoelectric module, J. Eng. Appl. Sci. 2 (4) (2009)
603–610.
[18] X. Sun, et al., A thermoelectric cooler coupled with a gravity-assisted heat pipe: An analysis from heat pipe perspective, Energy Convers. Manag. 155 (2018)
230–242, https://doi.org/10.1016/j.enconman.2017.10.068.
[19] G. Tan, D. Zhao, Study of a thermoelectric space cooling system integrated with phase change material, Appl. Therm. Eng. 86 (2015) 187–198, https://doi.org/
10.1016/j.applthermaleng.2015.04.054.
[20] M. Cosnier, G. Fraisse, L. Luo, An experimental and numerical study of a thermoelectric air-cooling and air-heating system, Int. J. Refrig. 31 (6) (2008)
1051–1062, https://doi.org/10.1016/j.ijrefrig.2007.12.009.
[21] M. Gökçek, F. Şahin, Experimental performance investigation of minichannel water cooled-thermoelectric refrigerator, Case Studies Thermal Eng. 10 (2017)
54–62, https://doi.org/10.1016/j.csite.2017.03.004.
[22] D. Astrain, et al., Improvements in the cooling capacity and the COP of a transcritical CO2 refrigeration plant operating with a thermoelectric subcooling system,
Appl. Therm. Eng. 155 (2019) 110–122, https://doi.org/10.1016/j.applthermaleng.2019.03.123.
[23] M. Chen, G.J. Snyder, Analytical and numerical parameter extraction for compact modeling of thermoelectric coolers, Int. J. Heat Mass Tran. 60 (2013)
689–699, https://doi.org/10.1016/j.ijheatmasstransfer.2013.01.020.
[24] Y. Cai, et al., Thermal performance of an active thermoelectric ventilation system applied for built space cooling: network model and finite time thermodynamic
optimization, Energy 170 (2019) 915–930, https://doi.org/10.1016/j.energy.2018.12.186.
[25] Y. Cai, et al., Entropy generation minimization of thermoelectric systems applied for electronic cooling: parametric investigations and operation optimization,
Energy Convers. Manag. 186 (2019) 401–414, https://doi.org/10.1016/j.enconman.2019.02.064.
[26] P.E. Ruiz-Ortega, M.A. Olivares-Robles, Peltier supercooling in transient thermoelectrics: spatial temperature profile and characteristic cooling length, Entropy
21 (3) (2019), https://doi.org/10.3390/e21030226.
[27] A. Caglar, Design and experimental investigation of a novel thermoelectric water dispenser unit, Appl. Therm. Eng. 149 (2019) 822–828, https://doi.org/
10.1016/j.applthermaleng.2018.11.028.
[28] D.C. Guo, et al., Application of thermoelectric cooler in temperature control system of space science experiment, Appl. Therm. Eng. 168 (2020), https://doi.org/
10.1016/j.applthermaleng.2019.114888.

12
P.K. Agyeman et al. Case Studies in Thermal Engineering 28 (2021) 101686

[29] T.H. Kwan, X.F. Wu, Q.H. Yao, Complete implementation of the combined TEG-TEC temperature control and energy harvesting system, Control Eng. Pract. 95
(2020), https://doi.org/10.1016/j.conengprac.2019.104224.
[30] T.H. Kwan, et al., Performance analysis of the thermoelectric device as the internal heat exchanger of the trans-critical carbon dioxide cycle, Energy Convers.
Manag. 208 (2020), https://doi.org/10.1016/j.enconman.2020.112585.
[31] A.Z. bin Zainordin, et al., Design and characterisation of magnetorheological brake system, Int. J. Eng. Syst. Model Simulat. 7 (1) (2015) 62–70.
[32] G. Marannano, G.C. Duboka. Virzì Mariotti, Analysis of the behaviour of an optimized magnetorheological brake, Rev. Appl. Phys. 2 (2) (2013) 23–32.
[33] T.L. Bergman, et al., Fundamentals of Heat and Mass Transfer, John Wiley & Sons, 2016.
[34] X. Lu, et al., Thermal resistance matching for thermoelectric cooling systems, Energy Convers. Manag. 169 (2018) 186–193.
[35] M. Siahmargoi, et al., An experimental study on the performance evaluation and thermodynamic modeling of a thermoelectric cooler combined with two
heatsinks, Sci. Rep. 9 (2019), https://doi.org/10.1038/s41598-019-56672-9.
[36] C.-y. Wen, R. Tsai, K.-P. Leong, Natural convection of magnetic fluid in a rectangular Hel… Shaw cell of different aspect ratios, Phys. Procedia 9 (2010)
181–185, https://doi.org/10.1016/j.phpro.2010.11.041.
[37] M.S. Krakov, I.V. Nikiforov, To the influence of uniform magnetic field on thermomagnetic convection in square cavity, J. Magn. Magn Mater. 252 (2002)
209–211, https://doi.org/10.1016/S0304-8853(02)00653-4.
[38] R. Limpert, Brake design and safety in SAE, 2011.
[39] L. Zhou, et al., Study of Energy Recovery System Based on Organic Rankine Cycle for Hydraulic Retarder. https://doi.org/10.4271/2016-01-0239, 2016, 1.
[40] J. Quan, B. Huang, Research on thermal management of magnetorheological fluid retarder based on phase change principle, SAE Int. (2020).

13