Experimental Thermal and Fluid Science 125 (2021) 110381

Contents lists available at ScienceDirect

Experimental Thermal and Fluid Science

journal homepage: [Link]/locate/etfs

An accurate calorimeter-based method for the thermal characterization of

heat pipes
Joseph P. Mooney a, b, *, Jeff Punch a, b, Nick Jeffers c, Vanessa Egan a, b
a
CONNECT, Stokes Laboratories, Bernal Institute, University of Limerick, V94 T9PX, Ireland
b
School of Engineering, University of Limerick, V94 T9PX, Ireland
c
Enovus Labs, Woodstock Rd, Newtown Mount Kennedy, Wicklow, Ireland

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

Keywords: This study presents a method that can be used to accurately determine the thermal performance of a cylindrical
Heat pipes heat pipe. In the method, the heat pipe is placed between two stainless steel 304 cylindrical blocks, configured as
Thermal characteristics radial calorimeters that achieve thermal contact with the evaporator and condenser sections of the pipe. A
Thermal resistance
flexible isothermal electrical heater mat surrounds the evaporator block, and a liquid-cooled copper pipe
Effective thermal conductivity
Calorimeters
wrapped around the condenser block is used to remove heat. High precision thermistors (±0.01 K) positioned at
Novel experiment fixed radial locations within the calorimeters are used to measure the heat supplied to the evaporator and the
heat extracted from the condenser. One-dimensional radial conduction is assumed to occur within each calo­
rimeter, and this enables the quantification of heat flows from the temperature readings. This assumption is
verified by a steady-state analysis of the radial, axial and circumferential temperature differences within the
evaporator calorimeter, based on data recorded for the lowest and highest heat inputs. Furthermore, a numerical
model is used to confirm that end effects have a negligible influence on radial conduction within each calo­
rimeter. This study concludes that the most commonly used characterization techniques for heat pipes can
greatly overestimate thermal performance (15–32% for input powers of 7.5–25 W respectively) due to inaccurate
quantification of heat flows into the evaporator and from the condenser. The calorimetric technique reported
here achieves uncertainties in thermal resistance of <7.5% for low thermal loads (<12.5 W) and <6% for higher
loads (>12.5 W). Moreover, the method achieves a significant improvement in the experimental thermal effi­
ciency, with values of >75% recorded for all heat inputs in this study. The use of radial calorimeters in the
current study obviates the requirement for calculating the losses from the heater to ambient, hence achieving low
uncertainties in thermal resistance and effective thermal conductivity for a range of heat inputs.

channel. A temperature driven pressure between the evaporator and

condenser ends of the heat pipe causes the vapor to flow in the direction
1. Introduction of the heat sink [2]. At the condenser end, heat is removed, and the
vapor condenses into the wick, normally saturating it. Due to the
Heat pipes are passive heat transfer devices that utilise the latent vacated pore space at the evaporator and the saturated pores in the
heat of vaporisation of a working fluid in order to transport heat from a condenser, capillary pressures force the liquid back to the evaporator.
high temperature source to a colder sink [1]. A conventional heat pipe Heat pipes feature effective thermal conductivities of order 102–104
consists of a hollow metal capsule lined with an internal porous struc­ times superior to those of good heat-conducting materials (copper, sil­
ture, known as a wick, which is saturated with a working liquid. The ver, etc.) [3]. Heat pipes come in various types, depending on their
remaining internal volume of the heat pipe is occupied by the vapor of application, including: concentric, micro heat pipes, vapour chambers,
the liquid. Due to the stable coexistence of liquid and vapor, the internal flat heat pipes, and rotating heat pipes, amongst others. They are used in
pressure is equal to the vapor pressure that corresponds to saturation a multitude of applications such as: electronics cooling, energy con­
conditions. This ensures effective use of the latent heat of the liquid for version systems, nuclear and isotope reactor cooling, high-performance
quick evaporation and heat transfer. When heat is applied to one end of space infrastructure, and cooling devices for the leading edges of nose
the heat pipe, the liquid evaporates from the wick into the hollow vapor

* Corresponding author.
E-mail address: [Link]@[Link] (J.P. Mooney).

[Link]
Received 5 June 2020; Received in revised form 28 January 2021; Accepted 18 February 2021
Available online 25 February 2021
0894-1777/© 2021 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY license ([Link]
J.P. Mooney et al. Experimental Thermal and Fluid Science 125 (2021) 110381

Nomenclature Greek Symbols

Δ Uncertainty
Roman Letters η Efficiency, %
A Area, m2 θ Angle (◦ )
cp Specific Heat Capacity, J/kg.K
d Diameter, mm Subscripts
I Current, A average Average
k Thermal Conductivity, W/m.K c Condenser
L Length, mm e Evaporator
P Power, W i Inner
Q Heat Flow, W in in
r Radius, mm o Outer
Rth Thermal Resistance, K/W out out
T Temperature, ◦ C th Thermal
U Uncertainty, % s supplied
V Voltage, V
x distance, mm

cones of re-entry vehicles [1,4,5]. One of the more commonly used types with cartridge heaters, a copper coil, or direct contact cartridge heaters.
of heat pipe is a concentric tube, sintered copper powder heat pipe with Heat is typically supplied to the pipe via the power inputted by the
water as a working fluid. This heat pipe is known as a homogeneous power supply units (i.e. Joule heating), and Ohm’s law (Qs = P = IV,
wicked heat pipe because the wick comprises a single material with an where Qs ≥ Qin) is typically used to quantify this heat [1,7–9,11–15]. At
axially-invariant cross sectional structure [5]. the condenser end, a range of techniques can be used to remove heat: a
Thermal resistance (Rth), effective thermal conductivity (keff), and direct water jacket, a liquid cooled block, forced air-cooling, or natural
average temperatures along the axis of the pipe (Tx) are the most convection [1,7–9,11–15]. A popular method of heat removal is a water
commonly used thermal characteristics for heat pipes [1,6–9]. Effective jacket, as shown in Fig. 1 [1,7–9,12–15]. In this case, the heat rejected is
thermal conductivity is inversely proportional to thermal resistance, calculated using the temperature difference of the coolant entering and
hence the better performing heat pipes feature superior effective ther­ leaving the cooling jacket (i.e. a thermal energy balance, Qout = ṁcpΔT).
mal conductivity values. In the literature, there is an abundance of The heat supplied (Qs) is often compared to the heat rejected (Qout)
experimental and numerical analyses on heat pipes of varying geome­ in order establish the experimental quality or amount of heat lost from
tries, lengths and materials with reported values for effective conduc­ the experiment to ambient, i.e experimental thermal efficiency (ηth =
tivity and thermal resistance [10]. Qout/Qs) [7–9,14–16]. However, most of these studies [8,9,14,15] use
Across a wide range of applications, greater thermal performance is thermal efficiency as a measure of the performance of a heat pipe and
demanded of heat pipes in order to meet increasing stringent design assume that all the environmental losses are from the heat pipe to
requirements. Fig. 1 illustrates a frequently-used technique for experi­ ambient. This assumption neglects any losses from the experimental
mentally characterizing the performance of a heat pipe. A heat source is apparatus, and as shown in Table 1, the resulting Qin/Qout values imply
thermally attached to one end of the heat pipe (the evaporator end), and that heat pipes are highly inefficient heat transfer devices which
the other end of the pipe thermally connects to a heat sink (the conventionally is not the case. Table 1 also details the methods used for
condenser). Measures are implemented to minimize the thermal contact heat pipe characterization from recent literature and the range of
resistance between the pipe and the heat source and sink. Fig. 1 also measured Rth values for the heat pipe studies. It can be seen that for all
illustrates some of the most common means of supplying heat to the referenced studies, heat pipes tend to have low thermal resistances of
evaporator section which include: electrical heater mats, a heating block ~2.0–0.01 K/W. However, inconsistencies arise when evaluating Rth

Fig. 1. Conventional experimental set up for heat pipe characterization illustrating the methods of supplying and removing heat [17].

2
J.P. Mooney et al. Experimental Thermal and Fluid Science 125 (2021) 110381

Table 1
Previously published experimental analyses of heat pipe performance in a horizontal configuration. Note that all except two studies quantified Qin using Ohm’s Law
and only six of the studies quantified Qout (via an energy balance in the cooling fluid).
Qin (P = I.V) Heating Technique Qout Cooling Technique Thermal Resistance ηth (Qout/Qin)
K/W %

Mahdavi et al. 2018 Copper block with six electrical cartridge ṁcp (Tin -Tout ) 0.3–0.15 45–79%
[7] heaters Chilled water jacket with flowmeter (125 ml/m @
(Qin = 13–100 W) 25 ◦ C)
Sözen et al. 2019 Electrical heater ṁcp (Tin -Tout ) ~0.1 83–86%
[8] (Qin = 200–400 W) Cooling jacket
Sadeghinezhad Block heating units ṁcp (Tin -Tout ) 2–0.3 20–40%
et al. 2020 (Qin = 20–120 W) Refrigerated bath
[9] circulator cooling box
Naphon et al. 2009 Electrical heater ṁcp (Tin -Tout ) – 40–83%
[16] Cooling jacket
Water
Venkatachalapathy et al. 2015 Circumferential ceramic heater ṁcp (Tin -Tout ) 1.3–0.38 40–70%
[15] (Qin = 2.5–38 W) Cooling jacket
Water
Kumaresan et al. 2014 Circumferential ceramic heater ṁcp (Tin -Tout ) 1.4–0.4 39–74%
[14] (Qin = 10–150 W) Cooling jacket
Water
Khalili et al. 2016 [1] Flexible resistance heater Cooling water jacket 1.6–0.18 –
(Qin = 5–150 W)
Yong Li et al. 2013 Two copper blocks heated by a DC power Cooling block cooled by a water cooling bath 0.56–0.02 –
supply
[11] (Qin = 10–60 W)
Shung-Wen Kang et al. 2009 Electrical resistance heater powered by a DC ṁcp (Tin -Tout ) 0.215–0.17 –
[12] power supply Constant temperature thermal bath and a cooling
(Qin = 30–70 W) chamber (40 ◦ C)
Wenjie Zhou et al. 2020 [13] Heating block Isothermal water cooled block 0.24–0.01 –
(Qin = 10–20 W)
Kempers et al. 2008 Aluminum block with electric band heaters Acrylic water jacket ~0.6–0.17 –
[18] Qin = Qs–Qloss ṁcp (Tin -Tout )
(20–180 W)
Kempers et al. 2006 Brass annular block with electrical band heater Acrylic water jacket ~2–0.3 –
[19] Qin = Qs–Qloss ṁcp (Tin -Tout )
(10–300 W)

and ηth which can be explained with reference to the equation which thermal characterization of a test sample by calculating Qloss from the
defines the thermal resistance of a heat pipe: experimental apparatus. These include, but are not limited to, thermal
guarding [20,21], the insertion of thermocouples into the insulation to
Te − Tc
Rth = (1) measure Qloss, or the use of the thermal energy balance for evaporator
Qin
and condenser water jackets [22]. These techniques are all limited by
where, Te and Tc are the surface temperatures of the evaporator and the uncertainties associated with the quantification of heat loss, spe­
condenser respectively, and Qin is the heat flow into the evaporator. By cifically: the accuracy and spatial location of the thermocouples used to
definition, thermal resistance is dependent on Qin [6]. For the studies quantify Qloss; the thermal conductivity of the insulation; ambient con­
presented in Table 1, Qin is assumed to be Qs from the power supply. ditions; the geometric and thermal properties of the heaters; and the
However, the values of ηth, reveal that only a fraction (~20–83%) of Qin losses due to Joule heating.
is being rejected by the condenser. This implies that although low Similarly, there is also uncertainty when considering the thermal
thermal resistance values were recorded for the pipes, significant energy balance for the direct water jackets. As previously stated (see
amounts of heat were lost from the pipe to ambient. It is evident that Table 1), water is a commonly used condenser coolant for these exper­
these studies generally did not account for the heat losses from the iments [1,7–9,11–16,18,19]. For water, its high specific heat capacity
experimental apparatus to ambient. As a consequence, the reported (cp ~ 4,180 J/kg.K) results in low temperature differences between the
thermal resistances can be considered lower-bound values, and the ηth inlet and outlet flows of the cooling jacket, since most of the experiments
values (if given) are unreliable. Therefore, is it impossible to accurately have investigated power inputs of 20–100 W [1,7,9,11–16,18,19]. Small
evaluate Rth on the basis of Qs if there are significant uncertainties in the temperature differences in the condenser fluid require precision ther­
heat flowing through the heat pipe or lost from the apparatus. mometry in order to reduce the experimental uncertainties in Qout,
Two studies accounted for the losses, Qloss, from the evaporator to which were ~26–73.2% for one study [7]. These uncertainties could be
ambient [18,19]. In both studies, the losses are used to adjust the heater improved if very small mass flow rates were used, but this would result
power to obtain the heat input to the evaporator (i.e. Qin = Qs − Qloss), in increased uncertainties in flow rate. Additionally, at low flow rates,
and any losses between the evaporator and condenser end of the heat the flow around the pipe can feature vortex shedding that can induce
pipe are calculated using the thermal energy balance of the condenser. fluctuations in the rate at which the pipe is cooled [23,24], and transient
These studies achieved improved uncertainty in the measured thermal variations in the surface temperature of heat pipe.
characteristics of a heat pipe. The uncertainties associated with Rth and In this study, a novel method is presented to accurately characterize
keff of these methods were up to ±20% for lower thermal loads (below the thermal performance of a heat pipe. The method uses radial calo­
~10 W) and ±10% for loads between 10− 30 W. Precise knowledge of rimeters to quantify the heat flow into the evaporator and the heat
the experimental environment is required, however, and extra compu­ removed from the condenser. The use of radial calorimeters obviates the
tational effort is involved. Other methods can be used to improve the requirement for calculating the losses to ambient from the heater and

3
J.P. Mooney et al. Experimental Thermal and Fluid Science 125 (2021) 110381

the use of the thermal energy balance for a cooling water jacket. The calorimeter is formed from a cylindrical section of a solid material of
method achieves low uncertainties in Rth (≤7.5%), Keff (≤8%) and ηth known thermal conductivity. The heat pipe is tightly fit within a hole
(≤7%) for all thermal inputs. Section 2 details the operation of the drilled axially through the center of each calorimeter, as shown in Fig. 2.
calorimeters which are the key elements of the method. In Section 3, an An isoflux heat source is wrapped around the outer radius of the calo­
experimental apparatus that implements the method is described, with rimeter at the evaporator end of the pipe, and some form of cooling
data reduction and results presented in Sections 4 and 5 respectively. technique is used to achieve isoflux heat removal from the calorimeter at
the condenser end. If it is assumed that heat losses from the end caps of
2. Calorimetry the calorimeters are negligible (as shown in Sections 4.2 and 4.3), then
one-dimensional radial conduction will occur within each calorimeter,
Methods to measure the thermal conductivity of a homogeneous similar to ASTM D5470. Hence, the radial heat flows through each
material of unknown thermal properties, such as solid flat plate homo­ calorimeter can be quantified from point temperature readings at the
geneous materials, are well established in the literature. The well-known embedded locations shown in the figure. In this manner, the calorime­
implementation of the ASTM D5470 standard [25], depicted in Fig. 2(a), ters yield accurate values of the heat flows Qin and Qout. Eqs. (2) and (3)
is a popular approach for characterizing such materials. Although ASTM quantify one-dimensional radial conduction, in the cylindrical co-
D5470 is conventionally used for solid homogeneous materials, some ordinate system, into the evaporator (Qin) and out of the condenser
studies have applied this standard to achieve higher precision thermal (Qout), respectively:
calorimetry measurements of non-solid materials. For example, one
T1 − T2
study [26] used ASTM D5470 to measure the thermal properties of Qin = ( ) (2)
thermal interface materials (TIMs) with high precision and sensitivity. 2πklln rr21
This study managed to minimize heat leakage through the TIMs, and
maintain minimal temperature gradients through the meter bars while
T3 − T4
following ASTM D5470, in particular for lower heat powers. For the Qout = ( ) (3)
TIMs, a minimum thermal resistance of ~2.93E-3 K/W was measured 2πklln rr43
with an uncertainty of 2.7% using a heat transfer rate of~ 16.8 W. The
study achieved calorimetric measurements with an order of magnitude where, k is the known conductivity of the calorimeter blocks, l is the
improvement in precision and sensitivity over any previous investiga­ length of the blocks, r1, r2, r3 and r4 are the radial locations of the
tion in its field. measured temperatures T1, T2, T3 and T4 respectively, as illustrated in
The objective of the current study is to establish an accurate method Fig. 2(b). It should be noted that these radial calorimeters behave in an
of characterizing the thermal performance of a heat pipe with similar analogous manner to calorimeters based on one-dimensional flow in
attributes to those shown by previous analyses of solid materials [25] Cartesian geometries, ASTM D5470 [25], as illustrated in Fig. 2(a).
and TIMs [26]. The key element of the method is the implementation of Finally, the average surface temperatures of the evaporator and
a variation of the ASTM D5470 standard via the use of radial calorim­ condenser sections of the pipe, shown as Te and Tc in the figure, can
eters to obtain accurate values of the heat supplied to the evaporator either be recorded directly using sensors, or by extrapolation from the
section of the pipe, and the heat removed from the condenser section. sensors embedded within the calorimeters. An experimental imple­
Fig. 2 (b) illustrates the configuration of the calorimeters located at the mentation of this method will be detailed in the following section.
evaporator and condenser sections of a sample heat pipe. Each

Fig. 2. A schematic of the calorimeter-based method for the thermal characterization of a flat plate test sample (a: ASTM D5470) and a heat pipe (b: current study),
illustrating the radial calorimeters, and their embedded temperature sensors, at the evaporator and condenser ends of a sample heat pipe.

4
J.P. Mooney et al. Experimental Thermal and Fluid Science 125 (2021) 110381

Fig. 3. (a) Schematic of the experimental apparatus used [17,27,28]; (b) photograph of the setup used with the top layer of insulation removed; and (c) illustration of
calorimeter blocks with dimensional locations of thermocouples relative to the center of the heat pipe. Note that the thermistors and thermocouples are located
within 15 mm deep × 1 mm ∅ holes from both faces of the calorimeter blocks and can be identified in the figure from the legend at the bottom.

3. Experimentation electrical heat source into the evaporator section of the heat pipe, and
from the condenser section of the pipe to a liquid-based cooler. This
An experimental apparatus was realized to demonstrate the section presents details of the experimentation apparatus (Section 3.1)
calorimeter-based thermal characterization method, and to record the and procedure (Section 3.2). Further information relating to data
characteristics for a sample heat pipe. The apparatus, illustrated in reduction, calibration and uncertainty analysis is presented in Section 4.
Fig. 3, featured radial calorimeters to quantify heat flow from an

5
J.P. Mooney et al. Experimental Thermal and Fluid Science 125 (2021) 110381

3.1. Apparatus switched on. Input power levels (Qs) were tested over a range of 7.5–25
W. For each power input setting, the system was allowed to reach
The experimental apparatus included (Fig. 3(a), (b) and (c)): steady-state, defined as a change of <0.5 K over a 30 min period.
Furthermore, the thermocouples at the same radial locations as the
• A cylindrical heat pipe of 300 mm length × 6 mm diameter from thermistors had to be within 0.5 K so that steady-state radial conduction
Wakefield-Vette (WKV part number: 124656) with sintered Cu could be assumed (Fig. 3(c)). Once each test run was complete, the data
powder wicks and water as the working fluid. logger was stopped and the data stored. Steady-state temperature
• Thermal paste (heat sink compound, Dowsil/Dow Corning, thermal readings were averaged over the final 15 min of each power level. As the
conductivity 0.59 W/m.K as per the supplier’s data sheet). insulation used had a 96 ◦ C transition to burn temperature the experi­
• Two 100 mm diameter × 100 mm long stainless steel 304 cylinders, ment was stopped if any temperatures exceeded 90 ◦ C. At temperatures
configured as calorimeters. The cylinders featured fully drilled above this value, there was potential for insulation deformation.
through 6.1 mm diameter holes, axially along their centerlines, for
snug fitting of the heat pipe. The thermal paste was used to reduce 4. Data reduction
the thermal contact resistance at the interfaces between the calo­
rimeters and the pipe. The calorimeter at the evaporator end was This section presents the relations applied to achieve data reduction
wrapped within a 315 mm × 100 mm × 2 mm Keenova silicon for the calorimeter and heat pipe characterization [1,7–9,11–13,29,30].
flexible heater mat which was powered by an EL302RT Aim-TTi The full set of relations is used to obtain thermal characteristics of a
bench power supply (Fig. 3(a) and (b)). The calorimeter at the sample heat pipe (i.e. Rth, and keff) from the temperature data recorded
condenser end was surrounded by a 6 mm copper coil which carried in the experimentation. To this end, the calorimeters were used to
a coolant, silicone oil, at a controlled temperature (25 ◦ C) that was quantify the heat flow going into the evaporator section (Qin) and out of
regulated using a Huber Ministat 125 water bath. An M500 TTS the condenser section (Qout) of the heat pipe. Steady state temperature
Micro Series pump was used to circulate the coolant from the measurements of the calorimeter blocks, data reduction and an uncer­
isothermal bath through plastic tubing to the copper coil. tainty analysis are also presented.
• Two split blocks made from poured fit polyurethane expandable
insulation surrounded the calorimeters (Fig. 3(b)). Slots were 4.1. Thermal conductivity
machined for snug fitting of insulation around the copper coil and
heater mat. This ensured that the evaporator and condenser calo­ The thermal conductivity of the stainless steel 304 used to form the
rimeters were surrounded with insulation (Note: the polyurethane calorimeters is shown as a function of temperature in Eq. (4) [29], which
material had a quoted thermal conductivity of <0.05 W/m.K). is quoted to less than ±2% uncertainty [8]:
• Eight calibrated negative temperature coefficient (NTC) precision
2
thermistors (Amphenol Advanced Sensors, 0.768 mm diameter log10 k = − 1.4087 + 1.3982(log10 T) + 0.2543(log10 T)
probe, Manufacture No:MA100BF103A) were embedded into two 1 − 0.6260(log10 T) 3 + 0.2334(log10 T) 4
(4)
mm diameter × 15 mm deep holes on both faces of each of the + 0.4256(log10 T) 5 − 0.4658(log10 T) 6
+ 0.1650(log10 T) 7 − 0.0199(log10 T) 8
calorimeter blocks, see Fig. 3(c). The thermistors were used to
accurately characterize the heat flow through the calorimeter blocks,
where, k is the thermal conductivity of the material, and T is the average
into and out of the heat pipe. Thermal paste was used to minimize
absolute temperature of the object. The k value was calculated at a fixed,
potential uncertainties associated with interfacial thermal resistance
radially-averaged temperature within each of the evaporator and
between the thermistors and the calorimeter blocks.
condenser calorimeters, for each power setting. This value ranged from
• A total of 32 K-type thermocouples (Omega PFA-insulated K type
15.2 to 16.3 W/m.K over a temperature range of 25–60 ◦ C.
thermocouples, Stock No. 5TC-TT-KI-40–1 M) were used to record
the following: the average surface temperatures at the evaporator,
4.2. Quantification of heat flow
adiabatic and condenser sections of the heat pipe (at axial distances
of 5 mm, 95 mm , 115 mm, 185 mm, 205 mm and 295 mm from the
In designing the experiment, it was intended that the temperature
outer end of the evaporator, see Fig. 3(c)); the internal temperatures
field in the calorimeter blocks would only vary radially and therefore be
within the SS304 calorimeter blocks; the inlet and outlet tempera­
invariant in the circumferential and axial directions. It can be seen in
tures of the coolant; and the ambient air temperature; see Fig. 3(a),
Fig. 3(c) that the temperature measurements for calorimetry were
(b) and (c). Thermal paste was used, in a similar manner as the
recorded at known radial locations and axial depths. Axially, the
thermistors, for the thermocouples that were attached to the SS304
thermistor temperatures were averaged, in the evaporator and
blocks and the heat pipe.
condenser, to give TeA outer ,TeA inner , TcA inner and TcA outer respectively,
• Two National Instruments 4 channel thermistor NI-9219 modules
inner representing the thermistors located closest to the heat pipe. These
with USB Hi-Speed USB-9162 carrier chassis and a National In­
temperatures were used in Eqs. (5) and (6) to evaluate the average flow
struments 32 channel thermocouple NI TB-4353 module were used
of heat in (Qin) and out (Qout) of the heat pipe. It is important to note that
to wire the thermistors and thermocouples to a National Instruments
the power supplied (Qs) was also recorded for all experiments, in order
NI PXIe-1082 data recorder respectively, for measurement via Lab­
to quantify the difference between Qs and Qin.
View software (see Section 4.5 for thermistor calibration).
TeAouter − TeAinner
Qin = ( ) (5)
3.2. Procedure
2πkLe ln rrOi
The experimental apparatus was assembled and connected to all
ancillary equipment as shown in Fig. 3(a–c). For further information TcA − TcA outer
(6)
inner
Qout = ( )
relating to the details of calibration see Section 4.5. To achieve steady-
2π kLc ln rr0i
state conditions, the Huber Ministate 125 was used to circulate the
coolant through the copper coils. When a steady-state temperature of
~25 ◦ C was achieved on the thermocouples at the outer radius of the where, Le and Lc are the evaporator and condenser lengths, ri and ro are
condenser calorimeter block (i.e. to ensure constant thermal conditions), the radial locations of the thermocouples, and k is the thermal con­
the data recorder was started and the power supplied to the heater mat ductivity of the calorimeter blocks which was calculated using Eq. (4).

6
J.P. Mooney et al. Experimental Thermal and Fluid Science 125 (2021) 110381

In order to confirm the validity of the assumption that end effects due 4.3. Experimental verification of 1-D radial calorimetry for low Qin
to losses from the faces of the calorimeters are negligible (i.e. that the
temperature field within each calorimeter varies radially, and is largely In order to confirm 1-D radial conduction, an analysis of calorimeter
invariant in the circumferential and axial directions), a heat transfer temperatures at the lowest and highest thermal loads (6 W and 22 W)
analysis was carried out on the experimental apparatus using SOLID­ was performed. These configurations relate to the highest experimental
WORKS® 2018 (x64 SP3.0 Flow Simulation, version 26.3.0.0063), a uncertainty and temperature gradient respectively. This test measured
thermal simulation package based on a finite element solver. The the difference in the circumferential and axial point temperatures, with
boundary conditions and material selection were set to represent the ±0.01 K uncertainty in the thermistors and ±0.1 K in the thermocouples,
experimental test rig as follows: on both faces of the evaporator and condenser calorimeter blocks as
presented in Fig. 5(a). The steady state temperature measurements for
• the calorimeters were modelled as stainless steel AISI 304 cylinders the 6 W and 22 W configurations, used to verify that end losses were
of conductivity dependent on Eq. (4); negligible, are presented in Fig. 5(b) and (c) respectively. To confirm
• an isoflux condition totaling to 20 W heat was applied around the that the circumferential temperature variation was negligible and the
circumference of the evaporator calorimeter; radial temperature gradients followed the same slope (i.e. constant
• a fixed temperature boundary condition of 25 ◦ C was applied around circumferential Qin), four circumferential temperature measurements
the circumference of the condenser calorimeter; were recorded at both the inner and outer radii in the calorimeter blocks
• the heat pipe was modelled as a 300 mm long hollow copper pipe (with maximum positional uncertainty of ±0.4%). For verification of 1-
(wall thickness 0.8 mm, k = 395 W/m.K) with a highly conductive D axial conduction, a comparison was made between both faces of the
core (k = 55,000–60,000 W/m.K, to ensure that the thermal resis­ calorimeter blocks. Fig. 5(b) and (c) illustrate that, for the 6 W and 22 W
tance of the heat pipe ranged from 0.15 to 0.2 K/W); configuration, the differences in the radial and axial temperatures were
• 25 mm thick thermal insulation of 0.05 W/m.K conductivity sur­ ≤ 0.5 K and the temperature differences (i.e. slope of the curves in Fig. 5
rounded the calorimeters and adiabatic section of the heat pipe, and (b) and (c) range from − 0.0001 K/◦ to − 0.0003 K/◦ for the evaporator
to account for any radiative losses an emissivity value of 0.7 was and from 0.0003 K/◦ to 0.0005 K/◦ for the condenser) that define Qin
applied to the insulation; and Qout were constant/parallel. This verifies the assumption that con­
• external cooling by natural convection and radiation was applied duction within the calorimeter blocks is predominantly radial in
from all exposed surfaces into an ambient set to 23 ◦ C. direction.

Fig. 4 below illustrates the temperature gradients, axially and radi­

ally, for a cross section of the simulation. It was found that the heat lost, 4.4. Thermal characteristics
from 1-D conduction through the insulation, at the evaporator and
condenser end caps was <2% of the overall power inputted to the Using the values of Qin and Qout obtained from the calorimeters, the
evaporator. This served to confirm the assumption that end effects due thermal performance of the heat pipe was analyzed in a similar manner
to losses from the faces of the calorimeters are negligible. It was also to that presented in previous studies [1,7–9,11–13] in terms of thermal
found that for 15 mm deep temperature readings, in both faces of the resistance (Rth), effective thermal conductivity (keff) [6], and thermal
blocks (i.e. similar to the experimental conditions), the axial tempera­ efficiency (ηth) [7–9]. Thermal resistance is calculated using Eq. (7), and
ture differences were small (~0.1–0.4 K for 20 W) and considered to be it is used in Eq. (8) to evaluate the effective thermal conductivity:
negligible; and, moreover, temperatures were invariant circum­ Te − Tc
(7)
average average
ferentially. This confirmed that uniform heating and cooling of the heat Rth =
Qin
pipe occurred in the calorimeter blocks. To put this in context, the
temperature difference was ~2 K in the radial direction, between the L
keff = (8)
inner and outer radii in both faces of the calorimeters. Furthermore, the Rth .A
simulation showed that for a supplied heat of 20 W (i.e. Qs), the heat
input to the evaporator (Qin) was ≈16 W, and the heat transferred from where, Te_average and Tc_average are the averages of the point surface
the condenser (Qout) was ≈15 W. Thus, the simulation was consistent temperatures of the heat pipe in the evaporator and condenser sections
with the findings of the experimentation (see Section 5) that a significant respectively (see Fig. 3(c)), L is the length of the heat pipe, A is the cross
proportion of the supplied heat (Qs) is lost to the surroundings, and is not sectional area of the heat pipe, and Qin is the heat supplied to the
transported into the condenser (as Qin). evaporator section of the heat pipe (Eq. (5)), as quantified using the
evaporator calorimeter.
Thermal efficiency, ηth, is a measure of the quality of the

Fig. 4. Cross-sectional temperature fields from a numerical thermal simulation of the experimental apparatus: heat supplied was 20 W.

7
J.P. Mooney et al. Experimental Thermal and Fluid Science 125 (2021) 110381

Fig. 5. (a) Thermistor and thermocouple location and steady state temperature data for the (b) lowest thermal load (Qin = 6 W) and (c) highest thermal load (Qin =
22 W) on the inner and outer faces of the calorimeter blocks at the evaporator and condenser. Where, the data points at 0◦ locations are the thermistors and the points
at 90◦ , 180◦ and 270◦ locations are the thermocouples, inner face 1 represents the side of the calorimeter closest to the adiabatic section of the heat pipe and outer
face 2 is the corresponding outer face.

experimental apparatus. It describes the proportion of the heat received [31] for constant odds results, using the root-sum-square contribution of
by the evaporator of the heat pipe that is captured by the condenser the individual contributions of the variables (Eq. (10)). Where ΔZ was
block: the uncertainty (±) in the derived quantity Z = f (x1, x2, x3 … xN), and
Δxi was the uncertainty (±) in the primary variable xi. The uncertainty
Qout
ηth = (9) ΔZ was then calculated as:
Qin
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
√ N [ ]2
√∑ δZ
ΔZ = √ Δxi (10)
4.5. Calibration and uncertainty analysis i=1
δxi

Further relations capturing the uncertainties associated with the

A rigorous uncertainty analysis was employed in order to quantify
radial calorimeter apparatus are represented in Eqs. (11)–(15) below.
how the uncertainties of each measured quantity propagate to the
Although Qs was not used in the analysis of thermal resistance and
overall uncertainties in the thermal resistance and effective thermal
effective conductivity in this paper, it is assessed here to provide a
conductivity of the heat pipes under test. Each measured quantity and its
comparison to reported techniques that utilize Qs. In this context, the
associated uncertainty are listed in Table 2.
voltages and currents supplied to the heat source were calibrated using a
The maximum uncertainties in each quantity calculated in Eqs. (10)–
Fluke 175 True-RMS digital multimeter against a EL302RT Aim-TTi
(16) were obtained using the method of Kline and McClintock (1953)

8
J.P. Mooney et al. Experimental Thermal and Fluid Science 125 (2021) 110381

Table 2 [( )2 (√̅̅̅ )2 ]1/2

Uncertainty analysis for measured variables and derived quantities (Qin = ΔRth ΔQin 2 (ΔT)
5–22.5 W). = + (13)
Rth Qin δT
Measured Variable Uncertainty Derived Quantities Uncertainty
(%) The highest uncertainty associated with the effective thermal con­
ductivity of the heat pipe was thermal resistance. There were, however,
Temperature for Qin, ±0.01 K Power, Qs ±4.0
T additional uncertainties associated with the length and diameter of the
Heat Pipe ±0.1 K Heat In, Qin ±4.6–4.9 heat pipe. The manufacturer specification and Vernier precision of the
Temperature, Te, Tc length and diameter are presented in Table 2:
Length, l ±0.05 mm Heat Out, Qout ±4.7–5.1
[( )2 ( )2 ( )2 ]1/2
Voltage, V ±100 mV Thermal Resistance, Rth ±5.8–7.5 Δkeff ΔRth Δl Δd
Current, I ±10 mA Effective Thermal ±6.0–8.0 = + + (14)
keff Rth l d
Conductivity, keff
Radius, r ±0.05 mm Thermal Efficiency, ηth
Uncertainties in Qout do not influence the uncertainties in Rth and
±6.5–7.0
Diameter, d ±0.1 mm
keff. However, Qout was used to evaluate the losses from the apparatus to
Conductivity, ksteel ±0.32 W/m.
K ambient and to quantify the experimental thermal efficiency (ηth ), Eq.
(15). Hence, a similar calorimeter configuration was used for the
bench power supply to minimize any potential uncertainty in Qs, see condenser as the evaporator, and Eq. (12) applies. However, the tem­
Table 2. perature gradients were slightly smaller in the condenser (due to
ambient losses, Qin > Qout), leading to marginally larger percentage
[( ) ( )2 ]1/2
ΔQs ΔV
2
ΔI uncertainties in Qout. An isothermal Huber Ministat 125 water bath was
= + (11) used to hold the condenser coolant that circulated around the calorim­
Qs V I
eter bar at a constant temperature, Δx = ± 0.1 K. However, this un­
Accurate quantification of the heat transferred into (Qin) and out of certainty does not contribute to the uncertainty in Qout because the
(Qout) the heat pipe, as per Eqs. (5) and (6), is the main feature of this temperature measurements were read directly inside the calorimeter
study. In this regard, a set of precision thermistors was used to record block.
point temperature measurements in order to determine the radial heat [( )2 ( )2 ]1/2
flux into the test sample. The radial temperature distributions were Δηth ΔQin ΔQout
= + (15)
measured using the thermistors, and applied in Eqs. (5) and (6) to find ηth Qin Qout
Qin and Qout. Uncertainties in Qin and Qout depend on both the thermal
and spatial uncertainties of each of the thermistors. Thermistors were The maximum uncertainties for Qs, Qin, Rth, keff Qout and ηth at the
calibrated against a Fluke 1504 temperature reference probe (uncer­ lowest and highest thermal loads (5–22.5 W) are presented in Table 2.
tainty ± 0.002 K at 298 K), and a Fluke 5610 secondary reference Where, the maximum uncertainties are associated with the lowest
thermistor probe (uncertainty ± 0.01 K). The thermistors were cali­ thermal load. It is evident that the calorimeter method produced thermal
brated in a Huber Ministat 125 controlled water bath against the sec­ characterization data with low uncertainties compared to the leading
ondary reference probe using six points over a range from 25 to 60 ◦ C experimental techniques reported in the literature (±20% for ≤ 10 W
and curve fitted to a generalized Arrhenius form using a fourth order and ±10% for loads between 10–30 W [18,19]). Moreover, note that in
polynomial [26]. Accounting for any temperature instabilities in the this study the uncertainty for all measured variables was within ±6.5%
calibration environment and curve fitting errors, the relative tempera­ for Qin ≥ 12.5 W.
ture uncertainty in each thermistor was reduced to ±0.01 K accuracy
with a confidence interval of two standard deviations. 5. Results and discussion
The calorimeter blocks were made with high precision in the radial
(Δr) and longitudinal (Δl) axis; where spatial location uncertainty was This section presents the measured thermal characteristics, Rth, keff
within ±0.05 mm accuracy. For the stainless steel 304 blocks used to (Section 5.1) and ηth (Section 5.2) of a heat pipe tested over a power
quantify the flow of heat within the calorimeters, the thermal conduc­ input range of 7.5–25 W. Specific focus is placed on the values of the
tivity (k) was calculated according to Eq. (4) (see Section 4), which was characteristics that are calculated when Qin is quantified using radial
quoted with an uncertainty of less than ±2% for all temperatures used in calorimetry instead of Qs, as per other reported techniques
this study. The steady state measurements for thermal conductivity of [1,7–9,11–13].
the calorimeter blocks were recorded and used for the accurate quan­
tification of Qin. As evident in Eqs. (13)–(16), the uncertainties in the
5.1. Thermal resistance & thermal conductivity
thermal characteristics (Rth, keff and ηth ) are functions of the un­
certainties in Qin, which can be expressed as:
Fig. 6(a) and (b) below respectively present the thermal resistance
[( ) ( ) (
ΔQin Δk
2
Δl
2
1
(
Δr1
))2 (
− 1
(
Δr2
))2 (Rth) and effective thermal conductivity (keff) of the heat pipe as a
Qin
=
k
+
l
+
ln(r2/r1) r1
+
ln(r2/r1) r2 function of nominal power supplied, when the electrical heat input, Qs,
] and the calorimeter, Qin, are used to evaluate the heat supplied to the
( √ ̅̅
̅ )2 1/2
2 (ΔT) evaporator (Eq. (7)). From Fig. 6(a) and (b), it is evident that the heat
+ (12) pipe behaved in a manner that is consistent with previous studies, since
δT
an increase in the nominal heat input led to a decrease in Rth (~0.3–0.19
The equation for the experimental uncertainty in thermal resistance K/W), with a corresponding increase in keff (~35–54 kW/m.K). It is also
of the heat pipe [1,18], that was established using Eq. (7), is presented in evident that the thermal resistance value of the heat pipe is higher when
Eq. (13). It is evident that the uncertainty in the thermal resistance of the the evaporator calorimeter is used to evaluate Qin (Eq. (7)), rather than
heat pipe depends on the uncertainty in Qin and the uncertainty of the by assuming that Qin = Qs (as conventionally used in the previous
thermocouples attached to the surface of the heat pipe, which are pre­ techniques, see Table 1). It was found that the Rth values of the heat pipe
sented in Table 2: were 15–32% higher when calorimetry was used (corresponding to
effective thermal conductivity values 13–30%, lower). Arguably, due to
the calorimeter blocks employed in this method, it is likely that a lower

9
J.P. Mooney et al. Experimental Thermal and Fluid Science 125 (2021) 110381

Fig. 6. Thermal resistance, Rth (a) and effective thermal conductivity, keff (b) of the heat pipe as a function of the nominal heat inputs, Qnominal, where Qs denotes the
peformance of the heat pipe when Ohm’s law is assumed and Qin relates to the peformance when the radial calorimetry method is used. Note: the Qs data are
presented for illustrative purposes only, so uncertainty bars are not included.

proportion of Qs is transferred to Qin than in the previous techniques assume that all of the (electrical) heat supplied to the apparatus, Qs, is
because of the greater losses to ambient associated with the heat source transferred to the evaporator section of the heat pipe (i.e., Qin = Qs). In
and the larger calorimeter block. (i.e., the calorimeter imposes a higher this context, the calorimeter block method aims to illustrate the
thermal resistance between the heat source and heat pipe evaporator discrepancy between Qs and Qin and, in addition, the quality of the
than the resistances inherent in the previous techniques). However, the experimental test rig (ηth = Qout/Qin).
objective of this study is to obtain accurate measurements of Qin, and Fig. 7 illustrates the discrepancies in assuming that all of the heat
Qout. In other characterization techniques reported in the literature supplied to the apparatus, Qs, is extracted from the condenser section (i.
[7,9,12,14–16], the actual value of Qin at the evaporator is not known – e., Qout = Qin = Qs). It can be seen that over the power input range of
only an upper bound, Qs. The comparison of Rth calculated on the basis 7.5–25 W, Qin was only 70–90% of Qs with the largest difference
of both Qin and Qs is included in Fig. 6(a) and (b) as an indication of the occurring at lower power inputs (7.5–15 W). This illustrates that heat is
discrepancies in the literature that assume Qin is equal to Qs. From this lost from the source to ambient and Qs ∕ = Qin. The figure also includes the
study, it is established that inadequate quantification of Qin can lead to relevant data for the published studies that recorded ηth as a function of
large inaccuracies in Rth and keff. In the previous techniques, the losses to Qs [7,9,14,15]. Only one of these studies [7] features credible un­
ambient may not have been as large (in the absence of a calorimeter), certainties in ηth due to quantification of the environmental losses from
but this study concludes that the difference in Rth between using Qs and the heat source. Hence, data from [7] are presented in the figure. For the
Qin can be significant. Hence, the use of a calorimeter-based method to referenced studies in Fig. 7 there is a large discrepancy between Qs and
quantify the heat input to the evaporator section, Qin, is essential for Qout with ηth varying between 20% [9] to 79% [7]. For the calorimeter-
accurate thermal characterization of a heat pipe. Moreover, the method based method ηth is much higher giving values of 70–80% for the lower
can achieve measurements of Rth and keff with an upper bound uncer­ thermal loads (~6–12.5 W), rising to 85–91% for the higher loads (i.e.
tainty of ±7.5% for low thermal loads (~6–12.5 W) and ≤ ±6.5% un­ ~16.5–22 W). Hence, in the current study the difference between Qin
certainty for higher loads (≥12.5 W), as per Section 4.5. and Qout, or ηth, should predominantly represent losses from the ‘adia­
batic’ section of the pipe if calorimeter end cap losses are considered
negligible, Section 4.2. If the electrical heat input, Qs, is substituted for
5.2. Experimental thermal efficiency the calorimeter-based value of Qin, the ηth value of the experimental
apparatus dropped to 54–80% for all thermal loads – an indication of the
The primary focus of this study was to examine the limitations of magnitude of the errors incurred in previously published characteriza­
previously published characterization techniques [7,9,14–16] which tion techniques [1,7–9,11–16]. It is evident that the calorimeter-based
thermal characterization method presented in this paper yields greatly
improved quantification of heat transferred through the pipe.

6. Conclusions

In this study, an accurate calorimeter-based method for the thermal

characterization of a heat pipe has been demonstrated. The key element
of the method is the use of radial calorimeters to quantify the heat flow
into the evaporator section, and from the condenser section, of the heat
pipe, with low levels of uncertainty, ≤±4.9% for thermal loads ~6–12.5
W and ≤±4.6% for loads ≥12.5 W. In this regard, the method is greatly
superior to other characterization techniques, reported in the literature,
that do not quantify heat losses. To demonstrate the method, a sintered
Fig. 7. Experimental thermal efficiency, ηth, as a function of nominal heat copper heat pipe sample was characterized in terms of thermal resis­
input, Qnominal. Both conditions, Qin and Qs, used the same value of Qout to tance, effective thermal conductivity and thermal efficiency, for an input
illustrate the relatively low ηth values associated with prevously published power range of ~7.5–25 W and a maximum experimental uncertainty of
characterization techniques [7,9,14,15]. Note: the ‘Qs data’ and literature are 8%. The conclusions are as follows:
presented for illustrative purposes only, so uncertainty bars are not relevant
except for [7] who presented uncertainty in ηth.

10
J.P. Mooney et al. Experimental Thermal and Fluid Science 125 (2021) 110381

• Thermal resistances in the range of 0.3–0.19 K/W were recorded for [5] D.A. Reay, P.A. Kew, Introduction, in: D.A. Reay, P.A. Kew (Eds.), Heat Pipes (Fifth
Edition), Butterworth-Heinemann, Oxford, 2007, pp. 1–8.
heat inputs of approximately 6–22.5 W, with corresponding effective
[6] D.A. Reay, P.A. Kew, 2 - Heat Transfer and Fluid Flow Theory, in: D.A. Reay, P.
thermal conductivity values ranging from ~35–54 kW/m.K. These A. Kew (Eds.), Heat Pipes (Fifth Edition), Butterworth-Heinemann, Oxford, 2007,
resistance values are 15–32% higher than those obtained using the pp. 29–106.
techniques reported in the literature, because the method features a [7] M. Mahdavi, S. Tiari, S. De Schampheleire, S. Qiu, Experimental study of the
thermal characteristics of a heat pipe, Exp. Therm Fluid Sci. 93 (2018) 292–304,
more accurate means of quantifying the heat flow into the evapo­ [Link]
rator section of the heat pipe. [8] A. Sözen, M. Gürü, A. Khanlari, E. Çiftçi, Experimental and numerical study on
• Experimental thermal efficiency values of ~70–80% and 85–91% enhancement of heat transfer characteristics of a heat pipe utilizing aqueous
clinoptilolite nanofluid, Appl. Therm. Eng. 160 (2019), 114001, [Link]
were recorded for the lower (6–12.5 W) and higher (16.5–22 W) 10.1016/[Link].2019.114001.
ranges of input power, respectively. These values of experimental [9] E. Sadeghinezhad, A.R. Akhiani, H.S.C. Metselaar, S. Tahan Latibari, M. Mehrali,
thermal efficiency are far superior to those reported in the literature, M. Mehrali, Parametric study on the thermal performance enhancement of a
thermosyphon heat pipe using covalent functionalized graphene nanofluids, Appl.
which range from 45 to 79% due to inaccuracies in the methods used Therm. Eng. 175 (2020) 115385, [Link]
to quantify heat flows into and out of heat pipes. applthermaleng.2020.115385.
[10] A. Faghri, Heat pipes, review, opportunities and challenges, Front. Heat Pipes
(FHP) 5 (1) (2014).
The calorimeter-based thermal characterization method demon­ [11] Y. Li, H.-F. He, Z.-X. Zeng, Evaporation and condensation heat transfer in a heat
strated in this paper resolves two critical limitations in previously- pipe with a sintered-grooved composite wick, Appl. Therm. Eng. 50 (1) (2013)
reported techniques: the assumption that the heat flow into the evapo­ 342–351, [Link]
[12] S.-W. Kang, W.-C. Wei, S.-H. Tsai, C.-C. Huang, Experimental investigation of
rator section of the heat pipe can be quantified from the electrical heat
nanofluids on sintered heat pipe thermal performance, Appl. Therm. Eng. 29 (5)
supplied to the test apparatus (in fact, it can be significantly lower (2009) 973–979, [Link]
depending on experimental setup); and the very high uncertainties (up [13] W. Zhou, Y. Li, Z. Chen, L. Deng, B. Li, Experimental study on the heat transfer
to ~80%) in the measured values of the heat flow from the condenser performance of ultra-thin flattened heat pipe with hybrid spiral woven mesh wick
structure, Appl. Therm. Eng. (2020) 115009, [Link]
section of the pipe. In this regard, the calorimeter-based method pro­ applthermaleng.2020.115009.
vides accurate thermal characteristics for heat pipes and, as such, rep­ [14] G. Kumaresan, S. Venkatachalapathy, L.G. Asirvatham, Experimental investigation
resents a potential basis for a standard thermal characterization on enhancement in thermal characteristics of sintered wick heat pipe using CuO
nanofluids, Int. J. Heat Mass Transf. 72 (2014) 507–516, [Link]
technique for heat pipes. [Link].2014.01.029.
[15] S. Venkatachalapathy, G. Kumaresan, S. Suresh, Performance analysis of
CRediT authorship contribution statement cylindrical heat pipe using nanofluids – An experimental study, Int. J. Multiph.
Flow 72 (2015) 188–197, [Link]
ijmultiphaseflow.2015.02.006.
Joseph P. Mooney: Conceptualization, Methodology, Data curation, [16] P. Naphon, D. Thongkum, P. Assadamongkol, Heat pipe efficiency enhancement
Formal analysis, Software, Investigation, Writing - original draft, with refrigerant–nanoparticles mixtures, Energy Convers. Manage. 50 (3) (2009)
772–776, [Link]
Writing - review & editing. Jeff Punch: Conceptualization, Methodol­ [17] A.-T. B. P. Supply. [Link]
ogy, Validation, Formal analysis, Writing - review & editing, Supervi­ 6654906/ (accessed: 07/04/20).
sion, Project administration, Funding acquisition. Nick Jeffers: [18] R. Kempers, A.J. Robinson, D. Ewing, C.Y. Ching, Characterization of evaporator
and condenser thermal resistances of a screen mesh wicked heat pipe, Int. J. Heat
Conceptualization, Methodology, Supervision. Vanessa Egan: Concep­
Mass Transf. 51 (25) (2008) 6039–6046, [Link]
tualization, Methodology, Validation, Formal analysis, Writing - review ijheatmasstransfer.2008.04.001.
& editing, Supervision. [19] R. Kempers, D. Ewing, C.Y. Ching, Effect of number of mesh layers and fluid
loading on the performance of screen mesh wicked heat pipes, Appl. Therm. Eng.
26 (5) (2006) 589–595, [Link]
Declaration of Competing Interest [20] P. Kolodner, High-precision thermal and electrical characterization of
thermoelectric modules, Rev. Sci. Instrum. 85 (5) (2014) 054901, [Link]
The authors declare that they have no known competing financial 10.1063/1.4872276.
[21] C. Xing, C. Jensen, C. Folsom, H. Ban, D.W. Marshall, An optimal guarding scheme
interests or personal relationships that could have appeared to influence for thermal conductivity measurement using a guarded cut-bar technique, part 1
the work reported in this paper. experimental study, Appl. Therm. Eng. 62 (2) (2014) 850–857, [Link]
10.1016/[Link].2013.03.040.
[22] M.M. Sarafraz, O. Pourmehran, B. Yang, M. Arjomandi, Assessment of the thermal
Acknowledgements performance of a thermosyphon heat pipe using zirconia-acetone nanofluids,
Renewable Energy 136 (2019/06) 884–895, [Link]
This research is conducted with the financial support of Science renene.2019.01.035.
[23] P. Sellappan, T. Pottebaum, Vortex shedding and heat transfer in rotationally
Foundation Ireland (SFI) under Grant Number 13/RC/2077 and has oscillating cylinders, J. Fluid Mech. 748 (2014) 549–579, [Link]
been part funded by the European Regional Development Fund through 10.1017/jfm.2014.191.
the SFI Research Centres Programme. [24] E. Izadpanah, A. Ashouri, M. Liravi, Y. Amini, Effect of vortex-induced vibration of
finned cylinders on heat transfer enhancement, Phys. Fluids 31 (7) (2019) 073604,
[Link]
References [25] Standard test method for thermal transmission properties of thermally conductive
electrical insulation materials, A. S. D5470-06, Copyright ASTM International,
[1] M. Khalili, M.B. Shafii, Experimental and numerical investigation of the thermal Conshohocken, PA, 2007.
performance of a novel sintered-wick heat pipe, Appl. Therm. Eng. 94 (2016) [26] R. Kempers, P. Kolodner, A. Lyons, A.J. Robinson, A high-precision apparatus for
59–75, [Link] the characterization of thermal interface materials, Rev. Sci. Instrum. 80 (9) (2009)
[2] T. Dobre, O.C. Pârvulescu, A. Stoica, G. Iavorschi, Characterization of cooling 095111, [Link]
systems based on heat pipe principle to control operation temperature of high-tech [27] H. M. 125. [Link]
electronic components, Appl. Therm. Eng. 30 (16) (2010) 2435–2441, [Link] 4.0011.01 (accessed: 11/04/20).
org/10.1016/[Link].2010.06.014. [28] N. I. P. 1082. [Link]
[3] H. Zhang, J. Zhuang, Research, development and industrial application of heat pipe m-developers (accessed: 07/04/20).
technology in China, Appl. Therm. Eng. 23 (9) (2003) 1067–1083, [Link] [29] R. Xu, L. Xu, An experimental investigation of thermal contact conductance of
10.1016/S1359-4311(03)00037-1. stainless steel at low temperatures, Cryogenics 45 (10) (2005) 694–704, https://
[4] A. Huminic, G. Huminic,“Numerical study of the heat transfer and flow [Link]/10.1016/[Link].2005.09.002.
characteristics of heat pipes, in: IOP Conference Series: Materials Science and [30] T.M. Tritt, Thermal Conductivity : Theory, Properties, and Applications, Kluwer
Engineering, vol. 595, 2019, pp. 012013, doi: 10.1088/1757-899X/595/1/ Academic/Plenum Publishers, New York: New York, 2004.
012013. [31] S J. Kline, F.A. McClintock, Mech. Eng. (Am. Soc. Mech. Eng. 75, 3) 1953.

11