Journal of Materials Processing Technology 135 (2003) 204–210

Measurements of thermal contact conductance

M. Rosochowska, R. Balendra*, K. Chodnikiewicz
Department of Design, Manufacture and Engineering Management, University of Strathclyde, James Weir Building,
75 Montrose Street, Glasgow G1 1XJ, UK

Abstract

During forging, the transfer of heat between the component, the tools and the environment has an impact on tool-life and the accuracy of the
formed component. Consequently, the measurement of thermal contact conductance is of increasing interest to researchers and industrial
engineers participating in the manufacture of high-precision components by plastic deformation. It is recognised that thermal contact
conductance is a function of several parameters, the dominant ones being the type of contacting materials, the macro- and micro-geometry of
the contacting surfaces, temperature, the interfacial pressure, the type of lubricant or contaminant and its thickness. A new steady-state
method and measurement equipment are proposed in which the measurements are conducted on thin cylindrical specimens, which are retained
under pressure between two tools. A clear advantage of this method is the ability to measure the thermal contact conductance under precisely
controlled conditions. Due to the small aspect ratio of the specimen, the applied pressure may be of the same magnitude as that prevailing in
industrial bulk-metal forming processes. In the present paper some experimental results on the dependence of h on the pressure and the
specimen texture are presented.
# 2002 Elsevier Science B.V. All rights reserved.

Keywords: Heat; Transfer; Coefficient

1. Introduction The reliability of analytical approaches depends on the

The energy expended to plastically deform materials in
metal-forming processes is converted almost entirely into
heat, this energy increasing the temperature of the formed
component and the tools while some of it is dissipated to the
environment. The transfer of heat to the component and tools
has an impact on the accuracy of the formed component.
Consequently, heat transfer from the work-material to the
tool and the environment is of increasing interest to
researchers and engineers participating in the manufacture
of high-precision components by the plastic deformation of
engineering materials. At the initiation of manufacture, the
thermal conditions at the component–tool interface are in a
transitional state: cycles of temperature change occur over
both the individual forming operation and the working day.
Many forming operations have to be completed for the tools
to acquire their saturation temperature. At this stage, each
successive operation provides a quantum of energy, which
equals that which is dissipated to the environment. A quasi-
steady-state is then maintained so long as the forming
parameters and the environmental conditions remain
unchanged [1].
*
Corresponding author.
E-mail address: r.balendra@dmem.strath.ac.uk (R. Balendra).
accuracy of the material properties and the physical para-
meters that influence heat transfer between solids [1–3]. The
latter refers to values for specific heat, thermal conductance
and coefficient of thermal expansion of both the component
and the tool materials; whilst a further consideration is the
allocation of a value for thermal contact conductance, since
this determines the thermal balance in the component/tool/
environment system.
Thermal contact conductance h, also known as the heat-
transfer coefficient, is defined as follows [4]:
q
h¼ (1)
DT
in which DT is the temperature difference at the contacting
surfaces and q, the heat flux, defined as


d dQ
q¼ (2)
dA dt
It is recognised that thermal contact conductance is a func-
tion of several parameters, the dominant ones being the type
of contacting materials, the macro- and micro-geometry of
the contacting surfaces, the temperature, the interfacial
pressure, the type of lubricant or contaminant and its thick-
ness. The interfacial pressure between contacting surfaces
during plastic deformation (component/die, component/

0924-0136/02/$ – see front matter # 2002 Elsevier Science B.V. All rights reserved.
PII: S 0 9 2 4 - 0 1 3 6 ( 0 2 ) 0 0 8 9 7 - X
 M. Rosochowska et al. / Journal of Materials Processing Technology 135 (2003) 204–210
Nomenclature
A area of contact surface (m2)
h thermal contact conductance (W/(m2 K))
ks thermal conductivity of test specimen (W/
(m K))
kt thermal conductivity of tools (W/(m K))
m temperature gradient in tool
q heat flux (W/m2)
Q heat (J)
tp process time
TC1 temperature at the upper tool surface (8C)
TC2 temperature at the lower tool surface (8C)
TS1 temperature at the upper surface of specimen
(8C)
TS2 temperature at the lower surface of specimen
(8C)
DTts temperature difference at tool–specimen inter-
face (D 8C)
Greek symbols
d specimen thickness
y process relative time
punch and component/ejector) is often of an order higher
than the yield strength of the work-material. Further, the
variation of this interfacial pressure with time has a sig-
nificant influence on the thermal contact conductance. An
increasing trend in cold metal forming is to attempt to
manufacture the net shape of the component, which requires
a more complete understanding of the plastic deformation of
engineering materials. To-date, several different thermody-
namic models have been used to compute the thermal
contact conductance [5–7].
Measurements of the coefficient h have been carried out
while heat transfer was either in a steady-state [8] or
transient condition [9,10]. Experiments were conducted
using devices that contained two tools [9,10] or two tools
with a specimen sandwiched between them [9,11,12]. These
experiments were followed by an assessment of the thermal
contact conductance while the test specimen was deformed
plastically [9,11,12]: a further development involved the
incorporation of thermocouples in the specimen [13,14]. In
the simplest case of steady-state heat transfer, h may be
determined using Eq. (1). Another method is based on
matching the measured temperature distribution to analy-
tical or numerical solutions for various values of h [9–13].
While using this method, experiments are conducted to
establish temperature contours, which are then compared
to the results of numerical simulations. The thermal contact
conductance is assumed to be the value that provides the best
match between simulation and experimental results. A
further method is based on the solution of an inverse
problem. The sequential inverse method [14] has been used
to determine the thermal contact conductance in metal
205
forming processes. Review of publications [8–15] suggests
that values of h vary substantially, perhaps due to the fact
that these were derived using different experimental
approaches. Published values were derived from experi-
ments of different configurations, e.g. different materials,
surface preparation, lubricant, pressure and temperature,
thus disabling comparison. Values of h ¼ 7:5 and 17 kW/
m2 K were derived for an aluminium alloy (2024-0) by
compressing ring specimens between dies made of IN-
100 for a deformation speed of 1 mm/s and 56 m/s, respec-
tively, the lubricant used being Renite S52 [9]. Another
study, in which the dependence of h on the interfacial
pressure was taken into account, defined h ¼ 10 and
40 kW/m2 K under contact pressures of 6.9 and 110 MPa,
respectively. These tests were conducted using specimens
similar to those used in previous research, on dies made of
H-12 tool steel, lubricated with MoS2 [8]. Further, the
variation of thermal contact conductance with pressure
showed trends which appeared to depend on the work-
material and experimental conditions. From a more detailed
analysis [15] the thermal contact conductance was found to
be between 50 and 120 kW/m2 K for the cold upsetting of
non-lubricated aluminium, 80–220 kW/m2 K for the cold
upsetting of lubricated aluminium and 15–30 kW/m2 K for
the hot pressing of aluminium (billet at 200 8C) with tools at
room temperature. The reduction ranged from 14 to 50%. It
was proposed that under non-lubricated conditions an
approximate value for thermal contact conductance could
be 100 kW/m2 K for cold forming and 50 kW/m2 K for hot
forming operations.
It may be concluded that experiments used to determine
thermal contact conductance include ‘‘method-related
errors’’. Results depend on the measuring devices used in
the experiments and on the method of processing the experi-
mental data.
A new approach for deriving values of thermal contact
conductance under differing interfacial conditions is pre-
sented, together with results on the dependence of h on
pressure and specimen texture. A clear advantage of this
method is the ability to measure the thermal contact con-
ductance under precisely controlled and continuously sus-
tained conditions.
2. Equipment and procedure
2.1. Test equipment
The proposed approach is based on steady-state heat flow
along two cylindrical tools and through a specimen [16]. The
distinct characteristic of this approach is the use of a thin
(d ¼ 2 mm) specimen of the same diameter as the tool. Due
to the small aspect ratio of the specimen, pressures of the
same magnitude as those prevailing in bulk-metal forming
processes may be applied. The experimental equipment is
shown schematically in Fig. 1. The lower part of the device
206 M. Rosochowska et al. / Journal of Materials Processing Technology 135 (2003) 204–210

applied pressures of 30, 60, 120, 180, 240, 300 and

420 MPa, whilst the second set was subjected to a similar
range of pressure with the main heater temperature set at
300 8C. These temperature settings resulted in interfacial
temperatures of approximately 90 and 130 8C, respectively.

2.3. Test procedure

The test specimens were located accurately between the

tools. The temperature settings of the main heater and the
cooling water were used to define the thermal conditions of
the experiments. The continuous change in temperature
along the tools was monitored until a steady-state condition
was achieved. At the point that the temperature distributions
along the tools were stable and the difference of heat fluxes
in both tools was within a specified limit, readings from the
thermocouples were used to compute the thermal contact
conductance. All experiments were conducted with dry
interfaces.

2.4. Computational considerations

As it has been already mentioned, heat flow was achieved

Fig. 1. Layout of research equipment. by applying a heat source to the upper tool and a heat sink to
the lower tool. The temperature distribution along the tools,
measured along their axis by four 0.5 mm type-K thermo-
is attached to the press platen while the upper is attached to couples in each tool, was used to calculate the heat flux and
the ram. The upper tool (1) is equipped with a heater (2) temperature difference at the interface (Fig. 2). As both
while the lower tool (3) is fitted with a heat-sink (4). The test punches were made from the same material and had the
specimen (marked black in Fig. 1) is located between these same surface finish and hardness, it could be assumed that
tools. Two thermal barriers (passive and active) minimise
lateral heat-losses. A solid ceramic insulator (5), insulating
wool (6) and a reflective screen which surround these
elements (not shown in the figure) form the passive barrier.
The active barrier comprises a sleeve (7), compensating
heaters (8) and (9) and a heat-sink (10). These thermal-
compensation arrangements ensure a uniform and equal
temperature gradient along both tools. An insulating element
(11) insulates the load-cell (12). The heat-sinks (4) and (10)
are supplied from bath circulators.
Temperature control of the system in the range of 30–
300 8C was achieved using a bespoke control device. The
press possesses the capacity to exert a maximum interfacial
pressure of 800 MPa.

2.2. Materials

Tools were manufactured from tool steel, N1019, the

thermal conductance of which was measured at a National
Standards Laboratory. Cylindrical specimens (18 mm dia-
meter and 2 mm thickness) were made from Ma8 steel. Test
surfaces of all specimens were ground before imparting a
texture using electro-discharge machining. Two sets of
specimens with surface textures in the range Ra ¼
0:275:95 mm were used in the trials. One set was used
in experiments with the main heater set at 200 8C and Fig. 2. Schematic diagram of equipment (a), temperature distribution (b).
 M. Rosochowska et al. / Journal of Materials Processing Technology 135 (2003) 204–210 207

the temperature drop on both contact surfaces of the speci-

men with the punches was the same and could be defined by
ðTC1  TC2 Þ  ðTS1  TS2 Þ
DTts ¼ (3)
2
Thus, using (1), the thermal contact conductance, as an
average of that in the two specimen–tool interfaces, may
be determined using the following equation:
2q
h¼ (4)
ðTC1  TC2 Þ  ðTS1  TS2 Þ
in which the heat flux q was obtained from
q ¼ kt m (5)
Temperatures TC1 and TC2 were estimated by extrapolating
the temperature gradients in punches, which were computed
using the least-squares method. Using the relationship
d
TS1  TS2 ¼ q (6)
ks
the temperature difference ðTS1  TS2 Þ between the speci-
men surfaces was computed.

3. Results and discussion

3.1. Experimental results

The values of thermal contact conductance were com-
puted using Eq. (4). For the measured value of surface
texture, the variance of h with the applied pressure is shown
in Fig. 3. The thermal contact conductance is a function of
pressure and surface roughness. Thermal contact conduc-
tance bears an exponential relationship to interfacial pres-
sure. This finding agrees with that reported [8] for
experiments with aluminium (Al 1100-O and Al 6061-O)
specimens under dry conditions. Further, it was found that
the magnitude of the thermal contact conductance increased
more rapidly with pressure for a finer textured specimen.
It should be stressed that data on contact conductance
published in the literature refers mainly to contact between
steel dies and specimens made from aluminium alloys.
Thermal contact conductance, for steel-to-steel contact,
was obtained from tool-to-tool experiments. Values ranging
from 0.75 to 9 kW/m2 K were established for H-13 tool steel
dies (temperatures of dies were 420 and 50 8C) subjected to
pressures ranging from 0 to 150 MPa [12]. Similar values
were obtained for IN-100 die [9]. For both cases, h attained a
saturation value. A value of h ¼ 100 kW/m2 K proposed
[15] for cold forming appears to be the nearest to the results
presented in Fig. 3.
3.2. FE application
The compression of a cylindrical specimen was simulated
to assess the influence of changes in the thermal contact
Fig. 3. Variation of thermal contact conductance with contact pressure.
conductance on the temperature at the specimen and tool
contact surfaces. A 6 mm diameter specimen of 7.2 mm
height with thermo-mechanical properties of Ma8 and
cylindrical tools of 40 mm diameter and 25 mm height with
the properties of N1019 tool steel were used in simulations.
Friction between the specimen and the tool was assumed to
be zero to eliminate heat generation due to friction. The
lateral cylindrical surfaces of the tools and specimen dif-
fused heat by convection, the ambient temperature being
20 8C. The top surface of the upper tool and the bottom
surface of the lower were retained at a constant temperature
of 20 8C. Simulations were performed for the following
thermal contact conductance:
(i) 20 kW/m2 K, which was used for coupled thermo-
mechanical analyses [17],
(ii) 100 kW/m2 K, which was recommended for cold
forming [15],
(iii) the pressure-dependent value of h (Table 1) derived
from presented experiments for an initial surface
roughness of Ra ¼ 2:0 mm was used as an alternative.
Ram velocities of 42, 2.1 and 0.42 mm/s were assumed to
reduce the height of the specimen by 58%. The temperature
at the end of the deformation process on both the tool and
specimen surfaces for nodes located on the axis of symmetry
are shown in Table 2. The table contains also the temperature
208 M. Rosochowska et al. / Journal of Materials Processing Technology 135 (2003) 204–210

Table 1
Pressure-dependent heat contact conductance

P (MPa) h (kW/m2 K)

0 5.0
30 18.3
120 58.8
180 87.0
300 222
420 410

Table 2
Temperature at the die and specimen surface at the end of processes

h ¼ 20 kW/m2 K

0.42 mm/s 2.1 mm/s 42 mm/s

Specimen (8C) 45.9 79.7 100.6
Die (8C) 42.1 70.9 89.8
DT (8C) 3.8 8.8 10.8
h ¼ 100 kW/m2 K
Specimen (8C) 41.8 76.0 101.7
Die (8C) 41.2 74.5 99.7
DT (8C) 0.6 1.5 2.0
h ¼ f ðpÞ kW/m2 K
Specimen (8C) 40.94 75.2 101.9
Die (8C) 40.77 74.7 101.3
DT (8C) 0.17 0.5 0.6 Fig. 4. Variation of temperature with time at specimen and die surfaces for
ram velocity 42 mm/s.

drop DT between the above-mentioned surfaces. Temperature

variations with time at these nodes are shown in Figs. 4–6 for
a velocity of 42.0, 2.1 and 0.42 mm/s, respectively. It follows
from Table 2 and Figs. 4–6 that
(a) The final temperatures at the specimen and die surface
are a function of deformation velocity; higher velocities
resulting in higher temperatures.
(b) For the same value of h, the final temperature
difference DT is higher for shorter process time tp.
(c) The final temperature drop DT across the interface is
higher for h ¼ 20 kW/m2 K than for h ¼ 100 kW/
m2 K.
(d) The temperature drop, DT, when the pressure-depen-
dent h is used is greater than when h ¼ 100 kW/m2 K:
this difference being even greater when h ¼ 20 kW/
m2 K.
The explanation for the first observation follows from this
FE model used, in which heat transfer to the environment
was taken into account. For lower deformation velocity, the
process time tp is longer and for this reason a greater amount
of heat is dissipated to the environment, resulting in a lower
final temperature.
The explanation of the remaining observations requires
the introduction of the relative time of the process:
t
y¼ (7) Fig. 5. Variation of temperature with time at specimen and die surfaces for
tp press velocity 2.1 mm/s.
 M. Rosochowska et al. / Journal of Materials Processing Technology 135 (2003) 204–210
Fig. 6. Variation of temperature with time at specimen and tool surfaces
for ram velocity 0.42 mm/s.
where t is the actual time. Introducing (7) into (2), after a
simple modification of (1), the following is derived:
1 q
DT ¼
tp h
where


 d dQ
q ¼
dA dy
Observations (b) and (c) follow directly from (8). The
diagram shown in Fig. 7, which illustrates the relationship
between h and the relative time y, can be used to explain
observation (d). An average pressure-dependent value of h,
which was used in simulations, may be obtained from the
following equation:
Z 1
hav ¼ h dy (10)
0 [1] Y. Qin, R. Balendra, K. Chodnikiewicz, A method for the simulation
The magnitude of this parameter is independent of process
time as long as the deformed material is not strain rate or
temperature sensitive: this is the case in the discussed model.
From the diagram (Fig. 7), it can be observed that
hav > 100 kW/m2 K and hav @ 20 kW/m2 K. Thus using
(8) it can be shown that DThðpÞ < DTh¼100 as well as
DThðpÞ ! DTh¼20 . The influence of the pressure-dependent
h would be greater in processes where higher contact
pressures are encountered.
209
Fig. 7. Relationship between thermal contact conductance and relative
time.
4. Conclusions
The conducted research enables the following conclu-
sions to be made:
1. A new method of determining thermal contact con-
ductance has been successfully used to measure the
dependence of thermal contact conductance on pressure
and surface texture.
2. Thermal contact conductance increases with pressure and
decreases with the surface roughness of the work material.
3. Increase of the value of thermal contact conductance
with pressure was more pronounced for specimens with
smooth surface finishes.
4. The temperature difference across the specimen–die
(8) interface was smaller for the pressure-dependent h than
for 100 kW/m2 K, which has been recommended for
cold forming.
5. Differences in temperature distribution, resulting from
(9) using different values of h in FE simulations, are more
noticeable for rapid plastic deformation processes.
6. The influence of the pressure-dependent thermal contact
conductance on the die and specimen surface tempera-
tures would be more prominent in processes where high
interface pressures occur.
References
of temperature stabilisation in the tools during multi-cycle cold-
forging operation, Proceedings of the 15th International Conference
on Computer Aided Production Engineering, 1999.
[2] M.P. Miles, L. Fourment, J.L. Chenot, Calculation of tool
temperature during periodic non-steady-state metal forming forging,
J. Mater. Process. Technol. 45 (1994).
[3] G. Shen, S.L. Semiatin, E. Kropp, T. Altan, A technique to
compensate for temperature history effect in the simulation of non-
isothermal forging process, J. Mater. Process. Technol. 33 (1992).
[4] C.V. Madhusudana, Thermal Contact Conductance, Springer, Berlin.
ISBN 0-387-94534-2 (1996).
210 M. Rosochowska et al. / Journal of Materials Processing Technology 135 (2003) 204–210
[5] A. Degiovanni, A.S. Lamine, C.H. Moyne, Thermal contact in
transient state: a new model and two experiments, J. Thermophys.
Heat Transfer 6 (2) (1992).
[6] T. Jurkowski, Y. Jarny, D. Delaunay, Simultaneous identification of
thermal conductivity and thermal contact resistance without internal
temperature measurements, Proceedings of the Third UK Conference
on Heat Transfer, Birmingham, 1992.
[7] A. Degiovanni, A.S. Lamine, Ch. Moyne, Thermal contact in
transient state: a new model and two experiments, J. Thermophys.
Heat Transfer 6 (2) (1992).
[8] V.K. Jain, Determination of heat transfer coefficient for forging
application, J. Mater. Shaping Technol. (1990).
[9] S.L. Semiatin, E.W. Collings, V.E. Wood, T. Altan, Determination of
the interface heat transfer coefficient for non-isothermal bulk-
forming processes, J. Eng. Ind. 109 (1987) 49–57.
[10] Z. Malinowski, J.G. Lenard, M.E. Davies, A study of the heat-
transfer coefficient as a function of temperature and pressure, J.
Mater. Process. Technol. 41 (1994) 125–142.
[11] W. Nshama, J. Jeswiet, Evaluation of temperature and heat transfer
conditions at the metal-forming interface, Ann. CIRP 44 (1) (1995)
201–205.
[12] P.R. Burte, Im Yong-Taek, T. Altan, S.L. Semiatin, Measurements
and analysis of the heat transfer and friction during hot forging, J.
Eng. Ind. 112 (1990) 332–339.
[13] P. Dadras, W.R. Wells, Heat transfer aspects of non-
isothermal axisymmetric upset forging, J. Eng. Ind. 106 (1984)
187–195.
[14] V. Goizet, B. Bourouga, J.P. Bardon, Experimental study of the
thermal boundary conditions at the workpiece–die interface during hot
forging, Proceedings of the 11th IHTC, Korea, vol. 5, August 1998.
[15] J. Jewiet, W. Nshama, P.H. Oasthuizen, Evaluation of temperature
and heat transfer conditions in metal forming, Proceedings of the
11th International Heat Transfer Conference, Kyongju, Korea, 1998.
[16] T.S. Wisniewski, M. Rosochowska, K. Chodnikiewicz, A new
method of measurement of thermal contact conductance in metal
forming, Progress in Engineering Heat transfer, Proceedings of the
Third Baltic Heat Transfer Conference, IFFM Publishers, Gdansk,
1999.
[17] H. Long, R. Balendra, Evaluation of elasticity and temperature effect
on the dimensional accuracy of back-extruded components using
finite element simulation, J. Mater. Process. Technol. 81 (1998) 665–
670.