Available online at www.sciencedirect.

com

ScienceDirect
Procedia CIRP 40 (2016) 653 – 658

13th Global Conference on Sustainable Manufacturing - Decoupling Growth from Resource Use

Dry rotary swaging with structured tools

Marius Herrmanna,*, Christian Schencka, Bernd Kuhfussa
a
bime, Bremen Institute for Mechanical Engineering, University of Bremen, 28359, Bremen, Germany
* Corresponding author. Tel.: +49 421 21864822; fax: +49 421 21864821. E-mail address: [email protected]

Abstract

Rotary swaging is a cold bulk forming process. The diameter of the workpiece is reduced incrementally by oscillating movement of the tools.
The established processes use lubricants which fulfill necessary functions such as lubricating, cooling and cleaning of the tools. Disadvantages
caused by the use of lubricant are costs of recycling, replacement of lost coolant and for the cleaning of the workpiece. To eliminate the
lubricant it is necessary to substitute the functions of the lubricant in other ways. For example by means of coating and structuring of the tools.
In this study infeed rotary swaging with structured tools is investigated using finite element simulations. Different structures are implemented
in the reduction zone of the tools. The influence is investigated by the radial and axial process forces. Additionally first structured tools are
manufactured and experimentally tested. The results of modeling and testing are discussed.
© 2016 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
© 2016 The Authors. Published by Elsevier B.V.
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the International Scientific Committee of the 13th Global Conference on Sustainable Manufacturing.
Peer-review under responsibility of the International Scientific Committee of the 13th Global Conference on Sustainable Manufacturing
Keywords: cold forging; dry metal forming; structured tools

1. Introduction Commonly, the rotary swaging process is studied using

finite element method (FEM) [2,3,4]. The investigations are
Rotary swaging is an incremental cold forming process and conducted to examine strain and stress distributions with
has an important field of application in the automotive regard to the influence of axial feed velocity [3] or friction and
industry for the production of axes, steering spindles and gear tool angle [4]. Also, different designs of the reduction zone of
shafts. It allows the reduction of the diameter of rods and the tools are investigated like convex, concave and hybrid
hollow shafts and features advantages like improved material contours [5]. Simulation of the process is of great potential to
properties as increased tensile strength, undisturbed fiber flow determine the influence of tool geometry, friction etc. for the
and for hollow shafts the adjustable wall thickness in case of rotary swaging process.
an optimal use of material resources. Thus rotary swaging
belongs to the near net shape and suitable production
techniques, with great potential for lightweight construction. F RI FR
The forming of the workpiece is achieved incrementally by d0 F RII d1
the oscillating tool stroke hT. The principle setup of the infeed
rotary swaging process is shown in Fig. 1. The workpiece is
axially fed into the swaging head with the feeding force Ff. µ
Ff FA
Due to the radial forming force FRI in the reduction zone (I)
and the tool angle α an axial reaction force FA counteracts to
?
Ff. To reduce the axial reaction force the effective friction in
I II FR III hT
the reduction zone of conventional tools is increased by a
tungsten carbide layer [1]. At the same time the friction in the
calibration zone is low to prevent wear due to manually L cal
polishing.
Fig. 1. Principle setup of infeed rotary swaging.

2212-8271 © 2016 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the International Scientific Committee of the 13th Global Conference on Sustainable Manufacturing
doi:10.1016/j.procir.2016.01.150
654 Marius Herrmann et al. / Procedia CIRP 40 (2016) 653 – 658

Nomenclature The proceeding to cope with this challenge is to control the

material flow and thus the required forces by a combination of
α ° tool angle coating and structuring of the reduction zone of the tools. The
β ° exit tool angle first approach is a coating on the whole tool with a low
µ - friction coefficient value friction coefficient to reduce wear and tear as well as cold
λ mm wavelength shuts [9]. However this creates a conflict, due to the reduction
A mm amplitude of the friction coefficient the axial reaction force increases.
d0 mm initial diameter The common tungsten carbide layer is not applicable any
d1 mm final diameter longer because in dry swaging the flushing effect of the
fst Hz stroke frequency lubricant with cleaning the tools from swarf is missing.
FA N axial reaction force Therefore a deterministic geometric structure in the reduction
Ff N feed force zone of the tools in contrast to a smooth zone is intended to
FR N radial force increase the effective friction. In this work simple contours are
FRI N radial force in zone I investigated by numerical simulations to obtain a basic
FRII N radial force in zone II understanding of the impact on the process. A qualitative
H mm overall height comparison of the simulation results reveals promising
hT mm tool stroke structures for the reduction zone and first structured tools
L mm overall length which are identified by the simulations are manufactured and
lcal mm length of zone II experimentally tested.
ltool mm tool length
r0 mm initial radius 2. Simulation
r1 mm final radius
Ra µm roughness 2.1. Modeling
s mm wall thickness
x mm feed length First investigations with the simulation model which is
I - reduction zone used for this study are conducted by Moumi [2]. This 2D-
II - calibration zone axisymmetric model is necessary to study a great variety of
III - exit zone different structures. The model is implemented with the
software ABAQUS 6.13-1 and consists of two parts the tool
and the workpiece (see Fig. 2). The tool is implemented as
From both the economical as well as the ecological point of rigid body. As friction model the penalty formulation and the
view the interest of dry metal forming becomes larger. In case coulomb friction coefficient is used. This model is used due to
of rotary swaging a large amount of lubricant based on the simplicity and the good results in cold metal forming
mineral oil is still used. The elimination of the use of simulations [10]. The numerical method is based on
lubricants results in three significant advantages [6], reduction Abaqus/Explicit due to the fast dynamic rotary swaging
of financial costs, of environmental impacts and of possible process. With an Intel® Core™ i5-4670 processor the FEM-
health burden. simulation of one process for one structured tool needs with 4
The reduction of the costs is based on the elimination of cores about 4 hours.
process steps like the recycling of lubricant and cleaning of It is not necessary to provide the complete length of the
the workpiece. Also no replacement is necessary of lost workpiece with a fine mesh but only the first 35 mm. After
lubricant due to the release during the process. Especially for the feeding of the first 35 mm into the swaging unit the tool is
the forming of tubes a lot of lubricant is removed with the completely filled and a quasi-steady state is reached in the
workpiece from the process. Furthermore the design of the process. The elements are reduced integrated A 4-node
machine is simplified. But the lubricant fulfills essential bilinear axisymmetric quadrilateral with hourglass control.
functions like the reduction of friction and thus the reduction The mesh size in the deformed area of the workpiece is
of tool load and wear. Furthermore it serves as separation chosen as small as possible and as large as possible to limit
layer to minimize cold welding processes. In addition it cools the computation time. Preliminary investigations showed that
the process and flushes the working zone to remove particle these restricting constraints do not falsify the results. The
abrasion. workpiece is modeled as an elastic-plastic isotropic material
The basic feasibility of dry rotary swaging in the micro with parameters from literature [11]. Additional settings for
range is presented by Moumi [7]. It is shown the process the simulation are summarized in table 1.
window, i.e. the applicable feed rate could be extended in
swaging of aluminium but is reduced in swaging of steel.
Furthermore the dry rotary swaging in the macro range is
tested. The analysis of the recorded process parameters and
the formed geometry of the workpiece reveal the potential of
dry rotary swaging, but also the difficulties that arise. Dry
rotary swaging needs a modification of the process and system
parameters as well as an adjustment of the tools [8]. Also the
function of the lubricant must be realized in another way. Fig. 2. 2D-axisymmetric-FE-model with process parameters.
 Marius Herrmann et al. / Procedia CIRP 40 (2016) 653 – 658 655

Table 1. Process settings for the simulation.

parameter value parameter value
feed velocity vf 30 mm/s feed length 35 mm
zone II length lcal 20 mm tool length ltool 100 mm
stroke frequency fst 37,5 Hz element type CAX3R
material AISI304 elements size (fine) 0.5x0.29 mm
friction coefficient 0.1 elements size (tall) 2.5x12.75 mm
A cosine structure was chosen in order to allow both, the
Fig. 4. Radial force.
provision of a surface structures with distinct impact on the
frictional properties and within highlighted geometrical range
the application of ball-end milling tools of larger diameter
(d ı 0.5 mm) to distinctly reduce the machining time when
transferring structures to the surfaces of rotary swaging tools.
The varied parameters are the wavelength λ and the amplitude
A, see Fig. 3. The parameters values are λ = 0.9, 1.1, 1.3 mm
and A = 10, 30, 40, 50, 100, 150, 200 μm. The overall height
H = 8.28 mm and overall length L = 47 mm are fixed which
leads to α = 10 °. The friction coefficient in the reduction and
calibration zone is fixed to μ = 0.1 for all simulations derived
from the intended coating (without lubrication) and
lubrication (without coating) of the tool [12].

A amplitude
H overall heigth Fig. 5. Axial force.
L overall length
λ
α toolangle 3. Experimental procedure
λ wavelength
H 3.1. Setup
A
α
L To compare the results of the forming process with
structured tools as well as the workpiece results also
Fig. 3. Cosine geometry in the reduction zone of a structured tool. investigations with other tools were carried out. At first
conventional tools with a tungsten carbide layer (T1) in the
2.2. Results reduction zone are used (see Fig. 6 a)). Furthermore tools with
a smooth reduction zone (T2) were used to form tubes with
The simulation results are analyzed to observe the effects and without lubricant (see Fig. 6 b)).
of the geometrical parameters on the radial and axial force.
The considered axial and radial forces are the average of the
maximum axial and radial forces per stroke over the complete
quasi-steady state of the process, the radial force is considered
for one tool. The results of the comparative simulation
without steps are set to 100 %.
The radial and axial forces are represented versus the
amplitude A with the three different wavelengths. For the
radial process force no clear effect of the amplitude is
noticeable, see Fig. 4. But a taller wavelength leads to a lower
radial force. Only a reduction to FR = 96.8 % in comparison to
the simulation without structured reduction zone is possible a) b)
for a cosine shape with λ = 1.3 and A = 0.05 mm. A stronger
effect on the axial process force is shown in Fig. 5. While the Fig. 6. Rotary swaging tools; a) conventional tools with a tungsten carbide
amplitude rises the axial force decreases. Also a taller layer (T1), b) smooth tools (T2).
wavelength leads to a reduction of the axial force. A reduction
to FR = 34 % in comparison to the simulation without A first experimental implementation of a structure in the
structured reduction zone is achievable however for reduction zone for rotary swaging tools are realized.
λ = 1.1 mm and A = 0.2 mm. The axial force is reduced by the Therefore a cosine shape with wavelength of λ = 1.3 mm and
benefited material flow. This is due to the geometry since the a amplitude of A = 150 μm is chosen. The tool T3 consisting
flow is less impeded due to the space generated by the cosine of the material 1.2379 is visualized in Fig. 7.
geometry as well as the altered distribution of strain.
656 Marius Herrmann et al. / Procedia CIRP 40 (2016) 653 – 658

error. The lubrication condition has a very small influence but

for dry rotary swaging the tracking error is slightly lower. For
the forming with tools with high friction coefficient value the
increasing of the tracking error with increased feed velocity is
higher than for the forming without lubricant respectively
with structured tools. In summary it can be said, that
structured tools reduce the tracking error and thus also the
axial reaction force.

Fig. 7. Rotary swaging tool with cosine geometry (T3).

The experiments were conducted with a swaging unit

HE32 with four tools. The workpiece was fed into the
swaging unit by a linear direct drive over a feed length of
x = 130 mm. Four different feed velocities were used with as
well as without lubricant. For the forming without lubricant
the tools were cleaned each time. The steel tubes were formed
from an initial diameter of d0 = 20 mm with a wall thickness
of s = 2 mm to a final diameter of d1 = 15 mm. The
experimental settings are summarized in Table 2.

Table 2. Process settings for the experiments.

parameter value parameter value
feed velocities vf 500, 1000, 1500, workpiece 1.0038
2000 mm/min material
tool material 1.2379, ASP 2023 feed length x 130 mm
tool angle α 10 ° initial diameter d0 20 mm
tool stroke hT 1 mm final diameter d1 15 mm Fig. 8. Tracking error for different feed velocities, wet and dry.
stroke frequency fst 37,5 Hz zone II length lcal 20 mm
3.3 Results - workpiece
3.2. Results - process
The work quality was investigated by measuring the
The process parameters were measured and the mean of roundness deviation and surface roughness. Therefore the
the maximal value per stroke is calculated. The data were roughness was measured for all three tools for the forming
analyzed for the different phases separately but only the with and without lubricant with a feed velocity of
results for the stable process phase are discussed because they vf = 500 mm/min. The roundness deviation was only
allow the best statement. The actual and the set value of the measured for the workpieces formed with the tools T1. It is a
feeding drive position were measured continuously. Thus, the value for the deviation of the profile form from the perfect
tracking error Δx were calculated by the difference between circular form and indicates the distance between the maximal
actual and set value of the feeding drive position. This positive and the maximal negative deviation of the perfect
difference is due to the axial reaction force FA of the forming circular form. The roundness deviation increases with higher
process and thus allows a statement about this force. The feed velocities irrespective of the lubricant condition. The
increase of the tracking error with increasing velocity is also roundness value for dry formed workpieces is worse and the
determined by the gain factor value in the position closed loop variance is higher than for the wet forming (see Fig. 9).
controller. The plot of the theoretical tracking error Δxtheoretical
is sketched in Fig. 8.
The tracking error is lower for all feed velocities for rotary
swaging without lubricant in comparison to rotary swaging
with lubricant (see Fig. 8). For the tools T2 rotary swaging
with lubricant generates the highest tracking errors for all feed
velocities due to the lowest friction coefficient value. The
tracking error for the forming with lubricant and tools T1 is
lower but also high due to the still low friction coefficient
value. For the forming without lubricant and the tools T1 and
T2 the tracking error is lower due to the higher friction
coefficient value. The tools T3 generate the lowest tracking
Fig. 9. roundness value for different feed velocities, T1 wet and dry.
 Marius Herrmann et al. / Procedia CIRP 40 (2016) 653 – 658 657

Furthermore the roughness is worse for the dry formed

workpieces formed with the tools T1 and T2 compared to the
lubricated forming (see Fig. 10). For the forming with the
tools T3 the roughness is the highest for wet as well as dry
forming due to the structure in the reduction zone of the tools.
But the roughness after wet forming is worse than after dry
forming.
After the forming of one tube without lubricant a large
amount of particle abrasion adhered at the surface of the tools
T1 and the particle abrasion covered the tungsten carbide
layer. The higher amount of particles can be explained with
the higher friction coefficient during dry forming. Due to the
missing of the flushing function of the lubricant the particles
Fig. 11. Structured tools after forming with vf = 500 /mm/min; a) wet, b) dry.
adhered to the tools. The same was seen at the workpiece
where a lot of particles are on the surface after forming
without lubricant. The particles abrasion was struck into the
workpiece surface what could be observed when cleaning
these. This is one of the reasons for the high roughness. For
the same experiment with the tools T1 with lubricant much
less amount of particle abrasion was observed.
The particle abrasion after dry forming with the tool T2
were much less compared to dry forming with the tools T1. 0.5 mm
The difference of the amount of particles was slightly higher
for dry forming. Summarized the quality for tubes after the
forming with the tools T1 and T2 is better with lubricant. For Fig. 12. Workpiece surface after forming with T1wet (vf = 500 mm/min).
the forming without lubricant the quality for tubes is better
with the tools T2.

0.5 mm

Fig. 13. Workpiece surface after forming with T1dry (vf = 500 mm/min).

Fig. 10. roughness of the workpieces with a feed velocity of vf = 500 mm/min
for T1, T2 and T3, wet and dry.

After the forming with structured tools (T3) the surfaces

had the worst quality compared to the forming with tools T1
and T2. In the valley of the structure of the tools T3 gather the
particle abrasion of both processes, dry and wet (see Fig. 11). 0.5 mm
Pictures of the surfaces of the workpieces are shown in Fig.
12-17 for all tools with wet and dry conditions. Fig. 16 and 17
show the surface of a workpiece formed wet or dry with the Fig. 14. Workpiece surface after forming with T2wet (vf = 500 mm/min).
tools T3. There are valleys to recognize in peripheral direction
produced by the structure in the reduction zone of the tools.
They occur for wet and dry forming. This valleys return at
regular intervals, which can be seen in the profile line in Fig.
18. The valleys are generated by the structure of the reduction
zone of the tools. The surface finishing in the calibration zone
of the tools is not possible anymore due to the strong
disruption. The figures of the surface of the workpieces
formed with the tools T1 (see Fig. 12 and 13) and T2 (see Fig. 0.5 mm
14 and 15) show a better surface quality without valleys
irrespective of the lubricant condition. Fig. 15. Workpiece surface after forming with T2dry (vf = 500 mm/min).
658 Marius Herrmann et al. / Procedia CIRP 40 (2016) 653 – 658

A first application of structured tools for rotary swaging is

examined. It influences positively the process parameter
tracking error. Due to the particles abrasion and the
disadvantageous surface generated by the missing lubricant
and the structure in the reduction zone the tools need an
additional modification. The structure in the reduction zone is
too high. Therefore a smaller value for the amplitude A has to
0.5 mm be found which leads still to a decrease of the tracking error
while the negative influence on the surface quality is
minimized. One idea for further investigations is a changing
Fig. 16. Workpiece surface after forming with T3wet (vf = 500 mm/min).
amplitude A which decreases to the calibration zone. In
addition a skew cosine should be of interest due to the
potential to have the same positive influence on the tracking
error while smaller amplitude A and less impact on the
surface quality. In a next step coated tools have to be
investigated to identify the decrease of the particle abrasion.
Furthermore other structures for example with a smaller
amplitude or a skew cosine have to be examined to get a
better surface quality. Finally both tool modifications,
0.5 mm structuring and coating, have to be combined to show the
complete potential of dry rotary swaging. The change of the
Fig. 17. Workpiece surface after forming with T3dry (vf = 500 mm/min). temperature conditions due to the missing cooling of the
lubricant needs also further investigations

Acknowledgements

The authors gratefully acknowledge the support by the

German Research Foundation DFG for the sub-project
" Potentials of Dry Rotary Swaging " within the priority
Fig. 18. Axial profile line of the workpiece surface after forming with T3dry. program 1676 “ Dry metal forming - sustainable production
through dry processing in metal forming “.
4. Summary
References
Forming with structured tools under different lubrication
conditions in comparison to forming with conventional and [1] Kuhfuss B, Moumi E. Incremental Forming. In: Vollertsen F, editors.
Micro Metal Forming. Springer Berlin; 2013. p. 104-113.
smooth tools is studied during infeed rotary swaging. [2] Moumi E, Ishkina S, Kuhfuss B, Hochrainer T, Struss A, Hunkel M. 2D-
Therefore steel tubes are reduced in diameter and the process Simulation of material flow during feed rotary swaging using finite
parameters as well as the formed geometry are examined. The element method. Procedia Eng. 2014; 81: 2342–2347.
feasibility is shown in particle and also the difficulties that [3] Rong L, Nie ZR, Zuo TY. FEA modeling of effect of axial feeding
arise. Following conclusions are drawn: velocity on strain field of rotary swaging process of pure magnesium. T.
Nonferr. Metal. Soc. 2006; 16: 1015-1020.
x Dry rotary swaging features modified recorded process [4] Ameli A, Movahhedy MR. A parametric study on residual stresses and
parameters in comparison to forming with lubricant. The forging lead in cold forming process. Int. J. Adv. Manuf. Tech. 2007; 33:
tracking error decreases due to the increased friction 7-17.
coefficient. [5] Ghaei A, Movahhedy MR, Karimi Taheri A. Study of the effects of die
x Forming with structured tools also leads to a decrease of geometry on deformation in the radial forging process. J. Mat. Proc.
Tech. 2005; 170: 156-163.
the tracking error. Whereas the use of tools with smooth [6] Vollertsen F, Schmidt F. Dry metal forming: Definition, chances and
reduction zone shows the highest tracking error for the challenges. Int. J. of Pre. Eng. and Manufac.-Green Tech. 2014. 1: 59-62
forming with lubricant. [7] Moumi E, Influences of dry machining on micro rotary swaging. 4th
x During dry forming with conventional tools which have a International Conference on Nanomanufacturing 2014.
tungsten carbide layer a high amount of particle abrasion [8] Herrmann M, Schenck C, Kuhfuss B. Dry rotary swaging - tube forming.
Key Engineering Mat. 2015. 651-653: 1042-1047
adhere to the tools and the workpiece. Also during dry [9] Herrmann M, Hasselbruch H, Boehmermann F, Kuhfuss B, Zoch HW,
rotary swaging with smooth and structured tools particle Mehner A, Riemer O. Potentials of Dry Rotary Swaging. Dry Metal
abrasion adhere to the tools and the workpieces. Forming OAJFMT. 2015; 1:63-71.
x The workpiece quality of the tubes formed by dry rotary [10] Xincai T. Comparisons of friction models in bulk metal forming.
swaging is disadvantageous due to the worse roundness Tribology I. 2002; 35: 385-393.
[11] Doege E, Meyer-Nolkemper H, Saeed I. Fließkurvenatlas metallischer
and roughness. But especially tubes formed with structured Werkstoffe. Hanser Verlag. Wien: 1986.
tools exhibit strong disruption at the surface independent [12] Czichos H, Habig KH, Celis JP. Tribologie-Handbuch: Tribometrie,
from the lubricant condition. Tribo-materialien, Tribotechnik. Vieweg + Teubner. Wiesbaden