IOP Conference Series: Materials Science and Engineering

PAPER • OPEN ACCESS

Damage Evolution in Nakajima Tests of DP800 Dual Phase Steel

To cite this article: T Bergs et al 2018 IOP Conf. Ser.: Mater. Sci. Eng. 418 012048

View the article online for updates and enhancements.

This content was downloaded from IP address 139.81.158.93 on 22/09/2018 at 01:50

International Deep Drawing Research Group 37th Annual Conference IOP Publishing
IOP Conf. Series: Materials Science and Engineering 418 (2018) 012048 doi:10.1088/1757-899X/418/1/012048
1234567890‘’“”

Damage Evolution in Nakajima Tests of DP800 Dual Phase

Steel

T Bergs1), M Nick2),a), D Trauth2), F Klocke2)

1
Fraunhofer-Institut für Produktionstechnologie, Steinbachstr. 17, 52074 Aachen, Ger-
many
2
Laboratory for Machine Tools and Production Engineering (WZL) of RWTH Aachen
University, Chair for Manufacturing Technology, Campus-Boulevard 30, 52074
Aachen, Germany
a
M.Nick@wzl.rwth-aachen.de

Abstract. In order to extend the understanding of damage evolution in sheet metal forming,
standardized Nakajima tests are carried out on a DP800 dual phase steel. Sample geometries for
characteristic stress-strain states are drawn in incrementing stages and their damage states are
analyzed using light and scanning electron microscopy as well as micro hardness measurements.
Numerically analyzed load paths are correlated with the respective damage states to allow pre-
diction of damage evolution in deep drawing processes. The influence of anisotropy is investi-
gated by testing samples cut at various angles to the rolling direction of the sheet material. The
result of the conducted research is the understanding of interactions between load paths and
damage evolution in sheet forming. These results will later be used to optimize load paths in a
deep drawing process, taking into account Lode parameter and stress triaxiality to produce dam-
age controlled parts.

1. Introduction and state of the art

Human-made climate change has been identified as one of the most urgent threats to the environment
[1]. One way to reduce carbon dioxide emissions and thereby global warming is the reduction of vehicle
mass. This can be achieved by lightweight design using advanced high strength steel [2]. Vehicle body
parts are primarily made in forming processes. However, service properties of formed parts depend
strongly on the damage state after forming. Material discontinuities such as voids, inclusions, or precip-
itations, act as damage initiation locations and undergo evolution with respect to stress states during
forming processes. An increased damage state after forming facilitates damage evolution and, eventu-
ally, failure in service. [3]
Damage models are categorized into two distinct approaches. Cockcroft and Latham [4], and Bao
and Wierzbicki [5] proposed models that quantify damage through a scalar parameter and predict failure
at a critical value for that parameter. Lemaitre [6] proposed a microstructural approach which takes into
account softening and allows relation of damage to plastic properties. Both types of models are primarily
designed to predict failure in the manufacturing process. To be able to reliably model damage evolution
and thereby control damage during forming processes, closer examination in specific forming processes
is necessary. Li has shown that the investigation of damage evolution requires the regard of not only an
equivalent stress, as used in flow curves, but of the actual stress state, characterized by the von Mises
equivalent stress 𝜎eq, the Lode parameter 𝐿 and stress triaxiality 𝜂, which are invariants of the stress

Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd 1
International Deep Drawing Research Group 37th Annual Conference IOP Publishing
IOP Conf. Series: Materials Science and Engineering 418 (2018) 012048 doi:10.1088/1757-899X/418/1/012048
1234567890‘’“”

tensor [7]. The value of these invariants can be identified with certain stress states. The von Mises
equivalent describes the magnitude of the stress state. The Lode parameter 𝐿 is a normalized value and
takes values of 𝐿 = −1 … 1. A Lode parameter of 𝐿 = −1 describes biaxial tension with uniaxial com-
pression, 𝐿 = 0 describes pure shear stress and 𝐿 = 1 describes uniaxial tension. Stress triaxiality 𝜂
takes values of 𝜂 = −∞ … ∞ with 𝜂 = −2/3 describing plane compressive stress, 𝜂 = −1/3 uniaxial
compressive stress, 𝜂 = 0 pure shear stress, 𝜂 = 1/3 uniaxial tensile stress and 𝜂 = 2/3 plane tensile
stress [8]. Yin has identified positive values of stress triaxiality 𝜂 as an indication of damage critical
stress states in in-plane torsion tests of DP600 dual phase steel [9].

Figure 1. Nakajima test principle: (a) specimen geometry; (b) test setup.

Evolving stress states are described using load paths. These show the evolution of loads, in this case
of the three invariants, over time [10]. To achieve different load paths, Nakajima tests are executed.
These were originally developed to record forming limit diagrams, which help in predicting failure in
sheet metal forming. Nakajima tests use specimens with varying cutout widths (Figure 1 (a)) to achieve
different stress-strain states. These specimens are clamped using a circular blank holder and then drawn
until fracture through a drawing die using a hemispherical punch (Figure 1 (b)). Major and minor strains
on the specimen surface are recorded when fracture occurs and plotted in the forming limit diagram,
resulting in a curve representing the limit of formability of the sheet material [11].
Pathak [12] investigated damage evolution in DP780 dual phase steel during Edge Stretching. Holes
were alternatively reamed and sheared into a tensile test specimen and the specimen was then drawn.
Experiments were interrupted at different strains and the damage states were analyzed. Void nucleation,
void growth and void coalescence were investigated as damage mechanisms. Void density, as an indi-
cator for void nucleation, showed a strong increase with increasing strain. Both average void diameter
and void orientation did not change significantly with increased strain [12].

2. Methods and materials

In the context of this work, specific stress states are created in a DP800 dual phase steel using Nakajima
tests. These stress states are correlated to damage states in the material. The specimens were water jet
cut, which has been determined to prevent both excessively frayed edges and heat influence. Frayed
edges might lead to premature failure at the edges while heat influence would change the steel micro-
structure, especially the damage state.

2.1. Load paths

Specimens with cutout widths 𝑤C ∈ {11 mm; 47 mm; 83 mm; 107 mm} were produced. Finite Ele-
ment simulations of the Nakajima tests on these specimens were carried out in previous work using
Abaqus/Explicit with solid three-dimensional reduced integration 8-node hexahedral elements (type

2
International Deep Drawing Research Group 37th Annual Conference IOP Publishing
IOP Conf. Series: Materials Science and Engineering 418 (2018) 012048 doi:10.1088/1757-899X/418/1/012048
1234567890‘’“”

C3D8R). Convergence analysis has shown element lengths of 𝑙p = 0.3 mm in the sheet plane and 𝑙s =
0.5 mm in sheet thickness direction to give consistent results. The simulations have yielded load paths
for various cutout widths [12]. The simulations were validated with strain fields obtained from physical
Nakajima tests. While a significant drop-off in strain towards the specimen center could be observed in
the simulations that did not occur in physical experiments, strain results starting at 20 mm from the
center were consistent with the fields observed in experiment. Load paths were therefore extracted for
elements at the center line of the specimen, at the end of the cutout (see Figure 2 (a)). The extracted load
paths are shown in Figure 3. Specimens with cutout width 𝑤C = 47 mm were selected to use in exper-
iments since their load paths show gradually increasing values of the Lode parameters 𝐿 at the top sur-
face and a steep increase to the maximum value at the bottom surface.

Figure 2. Finite Element model of a 𝑤C = 95 mm Nakajima test: (a) location of investigated ele-
ments at 𝑡 = 10 s; (b) load paths at the upper specimen surface.

Figure 3. Load paths with respect to Lode parameter (a) at top surface, (b) at bottom surface for vari-
ous cutout widths.

3
International Deep Drawing Research Group 37th Annual Conference IOP Publishing
IOP Conf. Series: Materials Science and Engineering 418 (2018) 012048 doi:10.1088/1757-899X/418/1/012048
1234567890‘’“”

Figure 4. BUP 1000 Sheet Metal Testing Machine: (a) machine setup; (b) lubrication detail.

2.2. Experimental approach

The tests mentioned before were executed on a Zwick Roell BUP 1000 Sheet Metal Testing Machine
(see Figure 3). The machine provides a maximum force of 𝐹max = 1,000 kN, which is available for both
clamping and forming. A Nakajima tool was used with a punch diameter of 𝑑P = 100 mm, a drawing
die diameter of 𝑑d = 132 mm and a forming matrix diameter of 𝑑M = 200 mm. The testing machine
is fitted with a GOM ARAMIS digital image correlation (DIC) system, which can be used to record
trains during forming. However, since the focus of this work lies on the damage evolution in relation to
the specific stress state and stress history, strains were not recorded.
Nakajima tests are ideally executed without friction in the contact zone between punch and specimen.
To reduce friction as far as possible, a layered lubrication system was used (see Figure 3 (b)). Beruforge
170 from Bechem Lubrication Technology, a MoS2 based coating lubricant, was first applied to the
punch. The lubricant was then covered by a PVC disc with a diameter of 𝑑PVC = 90 mm and another
layer of lubricant was applied onto this disc. Experiments at the Institute of Forming Technology and
Lightweight Components (IUL) at TU Dortmund have proven this combination to be effective in reduc-
ing friction to near-zero.
To investigate damage evolution during forming, stopped tests are undertaken. Specimens are drawn to
depths of 𝑠d ∈ {7.5 mm; 12.5 mm; 17.5 mm}. These are then cut along the line at the end of the cutouts
and their microstructure is analyzed, specifically with regard to voids. Specimens are grinded and pol-
ished with grain sizes of 6, 3 and 1 µm. To remove caps from the voids that result from material being
smeared over them, the polished specimens are etched in a 3% Nital solution for 5 s. To remove the
etched structure from the metal matrix, a polishing step of 1.5 min with 1 µm grains is used. The extent
of damage in the specimens is identified through the area of voids that can be seen in the micrograph.

3. Results and Discussion

Sample micrographs of a 𝑠d = 17.5 mm specimen both near the top and the bottom surface are shown
in Figure 5 (a). The DP800 steel matrix is seen as a uniform, light gray area. All black areas are regarded
as voids, with their percentage used as a measure for damage. Light pink inclusions are not counted
towards the damage estimation.

3.1. Void percentage

Average void percentages 𝑣 for the investigated specimens are displayed in Figure 5 (b). While top
surface values increase from 𝑣 = 0.047 % at 𝑠d = 7.5 mm through 𝑣 = 0.052 % at 𝑠d = 12.5 mm
to 𝑣 = 0.075 % at 𝑠d = 17.5 mm, this is entirely within the standard deviation of the sample. For the

4
International Deep Drawing Research Group 37th Annual Conference IOP Publishing
IOP Conf. Series: Materials Science and Engineering 418 (2018) 012048 doi:10.1088/1757-899X/418/1/012048
1234567890‘’“”

bottom surface, average void percentage increases slightly from 𝑣 = 0.058 % at 𝑠d = 7.5 mm to
𝑣 = 0.060 % at 𝑠d = 12.5 mm and then decreases to 𝑣 = 0.053 % when the drawing depth is in-
creased to 𝑠d = 17.5 mm. Again, regarding standard deviation, no effect of larger drawing depths can
be seen.

Figure 5. Void percentage in specimens: (a) void identification in micrograph; (b) void percentage
averages for different drawing depths.

Possible reasons for the lack of effect on measured damage are issues with the evaluation procedure,
a lack of resolution with the light microscope, or an actual lack of detectable effects. Damage initiation
in the form of void generation happens through brittle fracture of the martensite phase, through ductile
deformation of the ferrite phase or through stress concentration at inclusions or precipitations [13].
These effects occur at a scale that cannot be investigated using light microscopy. Continued deformation
of the specimen does eventually lead to voids that can be detected with a light microscope, but these
deformations were not achieved in the experiments.
Due to symmetry, the direction of the maximum principal stress at the investigated location is ex-
pected to be perpendicular to the plane in which the micrograph was taken. Since voids will grow pri-
marily in the direction of the maximum principal stress, detection of growth is difficult.

4. Summary and Outlook

The drawing of the selected 𝑤C = 47 mm specimen to depths of 𝑠d ∈ {7.5 mm; 12.5 mm; 17.5 mm}
has proven to not produce sufficiently significant damage in the form of voids to analyse the influence
of load paths on damage evolution using light microscopy. Although an increase in the average void
percentage at the top surface can be seen, this is entirely within the standard deviation of the measure-
ment.
The cited study by Pathak [12] found void percentages of around 𝑣 = 0.2 % in reamed hole speci-
mens at equivalent strain 𝜀eq = 0.05. These grew to around 𝑣 = 0.8 % at 𝜀eq = 0.4. While potentially
different initial material states and different testing methods prevent a direct comparison, similar mag-
nitudes can be seen.
Further experiments need to be undertaken to analyse damage evolution. The drawing depths at fail-
ure are known from experiments executed in order to validate the numerical model of the Nakajima
tests. Drawing specimens to a depth just before failure will create a state of maximum damage. Further-
more, the numerical model must be refined in order to correctly predict strain not only at the location
investigated here, but specifically at the center of the specimen, where failure does eventually occur.
This will allow the investigation of load paths at a location where damage is expected to be more pro-
nounced.

5
International Deep Drawing Research Group 37th Annual Conference IOP Publishing
IOP Conf. Series: Materials Science and Engineering 418 (2018) 012048 doi:10.1088/1757-899X/418/1/012048
1234567890‘’“”

The results of the presented work and further investigations will be used in subprojects within the
Collaborative Research Center CRC/Transregio 188 Damage Controlled Forming Processes that are
specifically concerned with damage modelling to extend and calibrate various damage models for the
specific dual-phase steel used within the CRC.

Acknowledgements
The project presented in this paper is funded by the German Research Foundation (DFG) in the Collab-
orative Research Center CRC/Transregio 188 Damage Controlled Forming Processes, subproject Influ-
encing Damage in Deep Drawing, and in the Individual Grant Program Investigation of the Influence of
Reversing Strokes on Work Piece Properties in Deep Drawing with Servo-Press Technology. The au-
thors would like to also thank Prof. Dr.-Ing. A. Erman Tekkaya and Alexander Schowtjak, M.Sc. of the
Institute of Forming Technology and Lightweight Construction at TU Dortmund for support in the ex-
ecution of the experimental work.

References
[1] J. Rogelj et al.: Paris Agreement climate proposals need a boost to keep warming well below 2 C.
Nature 534.7609 (2016), pp. 631-639.
[2] H.J. Kim, C. McMillan, G.A. Keolian and S.J. Skerlos: Greenhouse Gas emissions payback for
lightweighted vehicles using aluminum and high-strength steel. Journal of Industrial Ecology
14 (2010) No. 6, pp. 929-946
[3] F. Klocke and D. Breuer: Using the finite element method and artificial neural networks to predict
ductile fracture in cold forming processes. Materials Processing and Design: Modeling,
Simulation and Applications, Columbus 2004, pp. 1944-1949
[4] M. Cockcroft, D. Latham: A simple Criterion Of Fracture For Ductile Metals. National
Engineering Laboratory, 1966
[5] Y. Bao, T. Wierzbicki: On fracture locus in the equivalent strain and stress triaxiality space.
International Journal of Mechanical Sciences 46.1 (2004): 81-98.
[6] J. Lemaitre: A Continous Damage Mechanics Model for Ductile Fracture. Journal of Engineering
Material and Technology, 107 (1985), pp. 83-89
[7] Y. Li; M. Luo, J. Gerlach; T. Wierzbicki: Prediction of shear-induced fracture in sheet metal
forming. Journal of Materials Processing Technology 210 (2010) No. 14, 1858-1869
[8] M. Dunant, D. Mohr: Effect of Lode parameter on plastic flow localization after proportional
loading at low stress triaxialities. Journal of the Mechanics and Physics of Solids 66 (2014),
pp. 133-153
[9] Q. Yin: Verfestigung- und Schädigungsverhalten von Blechwerkstoffen im ebenen
Torsionsversuch. Dissertation TU Dortmund, 2014
[10] N. Shamsaei, M. Gladskyi, K. Panasovskyi, S. Shukaev, A. Fatemi: Multiaxial fatigue of titanium
including step loading and load path alteration and sequence effects. International Journal of
Fatigue 32 (2010), No. 11, pp. 1862-1874
[11] ISO 12004-2:2008: Metallic materials – Sheet and strip – Determination of forming-limit curves
– Part 2: Determination of forming-limit curves in the laboratory
[12] N. Pathak, C. Butcher, M.J. Worswick, E. Bellhouse, J. Gao: Damage Evolution In Complex-
Phase and Dual-Phase Steels during Edge Stretching. Materials 10 (2017), No. 4, pp. 1-29
[13] T. Bergs, M. Nick, A. Feuerhack, D. Trauth, F. Klocke: Numerical-experimental Investigation Of
Load Paths in DP800 Dual Phase Steel During Nakajima Test. ESAFORM 2018, April 2018,
Palermo. Unpublished conference paper. AIP Publishing, 2018
[14] M. Calcagnotto, Y. Adachi, D. Ponge, D. Raabe: Deformation and fracture mechanisms in fine-
and ultrafine-grained ferrite/martensite dual-phase steels and the effect of aging. Acta
Materialia 59 (2011), pp. 658–670