Journal of Materials Processing Technology 143144 (2003) 807813

Mild steel (En8) rod tests under combined tensiontorsion loading

N.M. Zarroug

, R. Padmanabhan, B.J. MacDonald, P. Young, M.S.J. Hashmi

School of Mechanical and Manufacturing Engineering, Centre for Design and Manufacturing, Dublin City University, Dublin, Ireland
Abstract
This paper examines the results obtained from combined tensiontorsion loading tests carried out on Mild steel (En8) specimen. The
loading of the specimen was carried out in different modes: (i) maintaining tensile force or axial displacement constant and increasing
torque or angle of twist; (ii) maintaining torque or angle of twist constant and increasing load or axial displacement. A nite element
solution of the problem was obtained to gain further insight into the effects of the loading modes. The results from the nite element
analysis (FEA) were validated against experimental results.
2003 Published by Elsevier B.V.
Keywords: Torsion; Tension; Constant angle of twist; Constant axial displacement; ANSYS
1. Introduction
In recent times, more attention has been paid to the ex-
perimental investigation and analysis of results using sim-
ulation packages to determine the mechanical behaviour of
material, under complex loading conditions since uniaxial
experiments have been found to be inadequate for reveal-
ing the material behaviour completely. Experiments carried
out by Meguid et al. [1], to determine the behavior of thin
walled tubular specimen made of Mild steel (En8) under
non-proportional straining gave results in good agreement
with the von Mises yield condition. It was possible to ob-
tain almost entire positive quadrant of the yield curve from
a single test, without unloading and reloading the specimen.
Further it was said that at constant angle of twist and in-
creasing the axial load, the experimental shear stress trajec-
tory follows the von Mises yield curve upto a point and then
onwards deviate crossing the von Mises yield curve. This
can be attributed to the deformation prior and during the on-
set of the necking of the specimen [2]. Similar experiments
carried out by Ali and Hashmi [3] on circular steel rod sub-
jected to combined torque and tension loading reported that
the shear stress and axial stress relationship was found to
be non-linear. This paper investigates the behaviour of Mild
steel (En8) rod under combined tension and torsion. The
material is subjected to different combined loading modes
using a purpose built torquetension machine to establish
the behaviour in the elastic-plastic region until failure. Four

Corresponding author.
E-mail address: zarroug99@yahoo.com (N.M. Zarroug).
different modes of loading the specimen were considered
for a xed (i) axial load, torque was increased, (ii) axial dis-
placement, torque was increased, (iii) torque, axial load was
increased and (iv) angular twist, axial load was increased.
The results from the experiments were analysed and found
to be consistent with the theoretical analysis of the com-
bined load test based on the Ideal Plasticity theory, namely:
that there is no one to one relationship between stress and
strain over the yield plateau.
2. Experimental setup
The test machine is a purpose built machine capable of ap-
plying tension and torsion loads simultaneously at different
strain rates. Two servomotors control the loading of speci-
men, one for tension and one for torsion. The load, torque,
axial displacement and angular twist can be maintained at
constant independently. This is possible by controlling the
speed and torque of individual motors, through servo con-
trollers. The machine is controlled from a remote computer
using LabVIEW software and data acquisition devices. Four
transducers were used to measure load, torque, axial dis-
placement and angular twist and the measured units were
amplied with a modular amplier and fed to LabVIEW for
control and display. A closed loop is created with LabVIEW,
as the program takes action according to the signals from
the transducers. These inputs from the transducers are saved
as database for analysis. A schematic of the experimental
setup is shown in Fig. 1.
Apart from the transducers used in the machine, strain
gauges are attached to the specimen to obtain the strain
0924-0136/$ see front matter 2003 Published by Elsevier B.V.
doi:10.1016/S0924-0136(03)00362-5
808 N.M. Zarroug et al. / Journal of Materials Processing Technology 143144 (2003) 807813
Fig. 1. Schematic diagram of the test rig setup.
information of the specimen under different loads. The
strain gauges are properly balanced for resistance and the
output voltage is supplied to the modular amplier for fur-
ther amplication and then to the LabVIEW system. Two
types of strain gauges, one for the axial strain measurement
and other for the shear strain measurement, are used in
the experiments, and are attached within the gauge length
of the specimen. The test specimen used in the experi-
ments is Mild steel designated as En8 (AISI 1039 Steel)
with the following composition of components (wt.%):
C Fe Mn P S
0.360.44 98.4798.94 0.71 0.04 (max.) 0.05 (max.)
The specimen is prepared in standard dimensions
and properly t in the grippers of the purpose built
torquetension machine. The dimensions of the specimen
are shown in Fig. 2. The machine is controlled by the
LabVIEW software. A virtual machine control panel, con-
sisting of variable knobs, boolean switches and transducer
signal indicators, was programmed in the computer and all
the controls and indicators in the front panel were properly
wired in the block diagram which runs in the background.
The input signals from the transducer amplier is analysed
and according to the test condition and front panel inputs,
the motors are controlled. Four different programmes for
four loading modes are prepared.
3. Experimental programme
Combined tensiontorsion load tests have been carried out
at normal room temperature in four different loading modes.
In the rst set of experiments, the specimen was loaded un-
til an extension corresponding to 50% of axial yield load
was achieved and the load motor was set to HOLD mode
to keep the axial displacement constant. The extension for
50 and 75% axial yield load was found out from simple
tension tests carried out on steel specimens. Two tension
Fig. 2. Geometry of specimen.
tests were carried out on virgin specimens under same load-
ing conditions and the proof axial stress at 0.02% offset,
Fig. 3a, and corresponding axial yield load was averaged to
be 30.5 kN, Fig. 3b. After loading the specimen to the above
mentioned extension, a torque is applied to the specimen.
For this test, strain gauges are not used as there is no further
extension of the specimen and the torque applied is at a pre-
determined shear strain rate. The specimen was loaded until
yield point was reached. The same procedure was adopted
for tests on samples initially subjected to 75% of yield
load.
In the second set of experiments, the specimen was sub-
jected to an initial torque corresponding to 50% of yield
torque and axial load was increased by maintaining the an-
gle of twist constant. This was achieved by setting the torque
motor to HOLD position and running the load motor. The
yield torque was found out from pure torque tests on steel
specimens. Two pure torsion tests were carried out on vir-
gin specimen under same loading conditions and the proof
yield torque was averaged to be 67.9 Nm, Fig. 4. In this
case, the axial load was applied until the specimen failed.
The procedure was repeated by applying an initial torque
corresponding to 75% yield torque and increasing the axial
load, but keeping the angle of twist constant.
In the third set of experiments, a predetermined load of
50% axial yield load was applied to the specimen and that
load was maintained in the specimen throughout the test.
With the presence of axial load, a torque was applied to the
specimen until the ultimate point. The same procedure was
repeated for 75% axial yield load also.
In the fourth set of experiments, 50% yield torque was
applied on the specimen and the axial load was increased
until failure. The torque was maintained constant throughout
N.M. Zarroug et al. / Journal of Materials Processing Technology 143144 (2003) 807813 809
Fig. 3. (a) Axial stress vs. axial strain; (b) uniaxial tensile load vs. axial strain.
Fig. 4. Pure torque vs. shear strain.
the experiment. The same procedure was carried out for 75%
yield torque.
Four transducers, namely, an axial load cell, a torque
load cell, an angle measuring device and a linear variable
displacement transducer (LVDT) were used to measure the
axial load, torque, angle of twist and axial displacement,
respectively. The modular amplier excites the transducers
and the signal are received and amplied and then fed to
LabVIEW. The signal, in volts, from the amplier is scaled
properly and shown on the indicator in the front panel as
well as stored as a Microsoft Excel le, for future analysis.
4. Results and discussion
4.1. Axial displacement held constant
Initially the specimen was extended to a displacement
corresponding to 50% yield load and then a torque was ap-
plied. It was observed from Fig. 5, that as the torque was
increased gradually, the initially applied axial load remains
constant for certain level of torque. It then decreases very
marginally until the combined axial load and torque pro-
duces von Mises yield condition. The axial load in the spec-
imen then decreases rapidly as the normalised torque ex-
ceeds 1.5 times the yield torque. When the specimen was
subjected to an initial axial load corresponding to 75% yield
load and then torque is applied, the axial load starts to de-
crease at a greater rate at the initial stage of loading and
tends to drop rapidly after about 1.5 times the yield torque.
On comparison of these two cases, when the initial ax-
ial load is smaller, the subsequent torque holding capacity
of the specimen is extended upto 1.75 times yield torque
after which yielding occurs. The specimen sustains larger
torque because of the nonlinear, inelastic behaviour across
the cross-section in shear. This increases the torque carrying
behaviour of the specimen. From the yield criteria for com-
bined loading according to von Mises, the material starts
yielding at lower torque when the initially applied axial load
is increased. This follows the yield criteria
2
+ 3
2
= Y
2
,
where is the axial stress, the shear stress and Y the ax-
ial yield stress. After reaching the yield plateau the com-
bined stresses in the specimen follows the yield plateau as
shown in Fig. 5. The normalised torque exceeds unity due
to the strain hardening property of the material as shown in
Fig. 4.
810 N.M. Zarroug et al. / Journal of Materials Processing Technology 143144 (2003) 807813
Fig. 5. Normalised axial load vs. normalised torque (constant axial displacement).
4.2. Angular twist held constant
In this loading mode, the specimen was subjected to 50%
yield torque and the angle of twist was held constant. An
axial load was then applied to the specimen already under
torque. The torque remains constant at the initial level until
the applied axial load reaches a value which when combined
with the torque causes the material to yield. Thereafter, for
any increase in axial load the torque reduces rapidly till
the axial load reaches a maximum of about 1.2 times the
yield axial load as shown in Fig. 6. At this stage any at-
tempt to increase the axial load causes the onset of plas-
tic instability and necking. The axial load decreases rapidly
at a constant torque until the specimen fails through frac-
ture. Similar behavior is repeated for an initial torque of
about 73% yield torque. The results plotted in terms of nor-
malised shear stress and axial stress for specimens initially
subjected to initial torque of 29 and 79% yield torque, give
a clearer indication of when plastic yielding commences un-
der combined torsion and tension loads. These results shown
in Fig. 7 indicate that the specimen subjected to higher ini-
Fig. 6. Normalised axial load vs. normalised torque (constant angle of twist).
tial torque followed by increasing axial load, yields plas-
tically at an early stage than the specimen subjected to a
lower initial torque. This is due to the fact that the combined
stress level meets the strain hardening yield plateau at an
early stage and then follows the yield plateau until necking
starts. The combined stress in the specimen with the lower
initial torque meets the yield plateau drawn using maximum
stresses just prior to the onset of necking. The combined
stress levels in both cases do not follow the yield plateau
drawn with the values of initial yield stresses in tension and
torsion.
4.3. Constant axial load
In the third loading mode, the specimen was subjected to
an initially applied axial load of 50% of the uniaxial proof
yield load and subsequent application of torque, maintaining
the axial load constant throughout the test. This procedure
was repeated for an initially applied axial load of 75% of the
uniaxial yield load. As the torque was gradually increased,
the shear strain as well as the axial strain was observed to
N.M. Zarroug et al. / Journal of Materials Processing Technology 143144 (2003) 807813 811
Fig. 7. Normalised axial stress vs. normalised shear stress (constant angle of twist).
Fig. 8. Normalised shear strain vs. normalised axial strain (constant load).
increase so as to maintain the axial load constant. To observe
the effect of initially applied axial load on the strength of
the solid rod with the subsequently applied torque, the test
results are plotted in terms of normalised axial strain versus
normalised shear strain and are shown in Fig. 8. It was
observed from this gure that at higher initial axial load, the
normalised shear strain does not reach more than about 2
Fig. 9. Normalised axial strain vs. normalised shear strain (constant torque).
but it exceeds 3.5 for the lower initially applied axial load.
This shows that, since the combined stress is lower in the
latter case it can accommodate more shear strain. The rate
of increase in the axial strain rate is signicantly higher for
the higher initial axial load compared to that for low initial
axial load, for a given increase in the shear strain and hence
shear stress.
812 N.M. Zarroug et al. / Journal of Materials Processing Technology 143144 (2003) 807813
4.4. Constant torque
In this loading case, the specimen was subjected to an
initial torque of 50% yield torque of the steel specimen, and
subsequently an increasing axial load was applied. The same
procedure was repeated for an initially applied torque of 75%
yield torque as well. Normalised axial strain and shear strains
are plotted as in Fig. 9. The gure shows that with increasing
axial strain, due to the increase in the axial load the shear
strain increases to maintain the initially applied torque. This
increase in shear strain is much higher for initially applied
higher torque values. As the axial load and hence the axial
strain increases, the outer layers of the specimen begin to
yield and to maintain the torque constant the shear strain
increases.
5. Results of nite element analysis
ANSYS version 5.6 was used for preliminary analysis of
the combined load cases. Combinations of 50 and 75% of
yield axial loads and 50 and 75% of yield torque values
were used for the analyses. The element type chosen was
Fig. 10. Images from ANSYS showing deformation of specimen at 50% yield torque and 118% yield axial load.
SOLID73, a 3D eight-node structural solid element with
rotation. The element has six degrees of freedom at each
node. Youngs modulus used was 210 GPa based on material
tests and the Poisson ratio was set to 0.3. Both the loads
(torque and tension) were applied in a single load step. The
specimen geometry is supplied to ANSYS and meshed. The
sparse direct solver was used to solve all the combination
of loads.
Images from the general post processor for axial displace-
ment, angular twist, shear stress and von Mises equivalent
stress for one load case (T = 0.5 yield torque and F = 1.18
axial yield load) are presented in Fig. 10. The axial displace-
ment of the gauge length is calculated from the cumulative
values of displacements of all the sections as deformation in
other sections was found to be negligible, whereas the angle
of twist is calculated from the extreme section value at the
load end. As shown in Fig. 11 for axial displacement and
Fig. 12 for angle of twist, the closeness of the curves shows
good agreement of the nite element analysis (FEA) and ex-
perimental test results. The same procedure was adopted for
three other loading cases. All the results revealed the same
trend of good agreement between the FEA and experimental
test results, hence they are not presented individually.
N.M. Zarroug et al. / Journal of Materials Processing Technology 143144 (2003) 807813 813
Fig. 11. Comparison of axial displacement in FEA and experiment.
Fig. 12. Comparison of angular twist in FEA and experiment.
6. Conclusion
Experimental results fromcombined torsiontension load-
ing indicates that the initially applied torque or axial load
begins to decrease with the subsequently applied load or
torque respectively, when the combined stress reaches a
critical value which is governed by the yield criteria.
When the axial displacement was maintained constant and
torque was increased, the axial load carrying capacity of
the specimen decreases. A rapid drop in the axial load
carrying capacity is observed once yield torque is reached.
When the angle of twist was maintained constant and the
axial load was increased, the torque carrying capacity of
the specimen drops rapidly once the yield load is reached.
When the initially applied torque is closer to the yield
torque, the maximum axial load that can be applied to the
specimen decreases signicantly. After the specimen be-
comes fully plastic, even without any increase in load the
axial stress and shear stress increases rapidly as necking
commences and fails rapidly.
To maintain the initially applied load on subsequent ap-
plication of torque, the axial strain increases along with
the shear strain. The axial strain was found to increase
rapidly for initially applied higher load case.
Similarly, to maintain the initially applied torque on sub-
sequent application of axial load, the shear strain increases
along with the axial strain.
The results from preliminary FEA were found to be in
close agreement with the experimental results.
A renement in analysis of load cases is carried out to
simulate the exact loading path of the actual experiment.
References
[1] S.A. Meguid, L.E. Malvern, J.D. Campbell, Plastic ow of mild steel
under proportional and non-proportional straining at a controlled rate,
Trans. ASME 101 (1979) 248253.
[2] S.A. Meguid, M.S. Klair, L.E. Malvern, Theoretical and experimen-
tal results of the plastic and strain-hardening behavior of En8 at a
controlled rate, Int. J. Mech. Sci. 26 (11/12) (1984) 607616.
[3] A.R.M. Ali, M.S.J. Hashmi, Theoretical and experimental results of the
elasticplastic response of a circular rod subjected to non-proportional
combined torque and tension loadings, in: Proceedings of the Insti-
tution of Mechanical Engineers, Part C, vol. 213, 1999, pp. 251
261.