Escolar Documentos
Profissional Documentos
Cultura Documentos
Programa de Pósgraduação em Engenharia Metalúrgica
Escola de Engenharia Industrial Metalúrgica de Volta Redonda
Universidade Federal Fluminense
DOCTORAL THESIS
Experimental characterization and modeling of the
plastic behavior of a thirdgeneration
advanced high strength steel sheet
Rafael Oliveira Santos
Advisors:
Prof. D. Sc. Luciano Pessanha Moreira
Universidade Federal Fluminense, Volta Redonda, RJ, Brazil
Prof. D. Sc. António Manuel de Bastos Pereira
Universidade de Aveiro, Aveiro, Portugal
D. Sc. Marilena Carmen Butuc
Universidade de Aveiro, Aveiro, Portugal
Volta Redonda, RJ, Brasil
2020
RAFAEL OLIVEIRA SANTOS
EXPERIMENTAL CHARACTERIZATION AND MODELING OF THE PLASTIC BEHAVIOR OF
A THIRDGENERATION ADVANCED HIGH STRENGTH STEEL SHEET
A Doctoral Thesis submitted to the
Graduate Program on Metallurgical
Engineering of Universidade Federal
Fluminense in partial fulfillment of the
requirements for the degree of Doctor in
Metallurgical Engineering.
ADVISORS:
Prof. D. Sc. Luciano Pessanha Moreira
Universidade Federal Fluminense, Volta Redonda, RJ, Brasil
Prof. D. Sc. António Manuel de Bastos Pereira
Universidade de Aveiro, Aveiro, Portugal
D. Sc. Marilena Carmen Butuc
Universidade de Aveiro, Aveiro, Portugal
Volta Redonda, RJ, Brasil
2020
RESUMO
A terceira geração dos aços avançados de alta resistência (AHSS) têm atraído a atenção da
indústria automotiva devido a sua boa relação entre a conformabilidade e o custo de produção.
O estágio atual de desenvolvimento desses aços ainda conta com poucos produtos disponíveis
no mercado, assim como poucas pesquisas voltadas para o estudo da sua conformabilidade.
Com um limite de resistência à tração superior a 1000 MPa e alongamento total acima de 20%,
o AHSS de terceira geração com grau 980 MPa foi a primeira chapa desta classe de aço a ser
comercializada. Nesse trabalho, diversos ensaios mecânicos foram realizados para analisar o
comportamento plástico do aço Gen3 980T laminado a frio com espessura nominal de 1,58 mm.
As propriedades mecânicas e os coeficientes de anisotropia plástica de Lankford foram obtidos
a partir de ensaios de tração uniaxial realizados em 7 orientações angulares no plano da chapa.
O encruamento da chapa aço Gen3 980T também foi avaliado por meio do ensaio de expansão
hidráulica e o coeficiente de anisotropia biaxial foi obtido pelo ensaio de compressão de disco.
A conformabilidade foi analisada com auxílio dos ensaios de expansão de furo, estiramento
(Erichsen) e curva limite de conformação (CLC). As deformações limites foram definidas pela
norma ISO 120042 a partir de ensaios no plano (tração uniaxial e deformação plana por tração)
e fora do plano da chapa (Nakajima). As técnicas de difração de raios X e microscopia eletrônica
de varredura foram adotadas para a caracterização microestrutural do aço Gen3 980T. O modelo
GursonTvergaardNeedleman (GTN) foi adotado para descrever o comportamento de dano.
Foi proposta uma metodologia simples para identificação dos parâmetros do modelo GTN. O
procedimento de calibração foi validado comparando modelos de elementos finitos com
medidas experimentais dos ensaios mecânicos. O aço Gen3 980T tem uma microestrutura
formada por martensita, ferrita e austenita retida. No estado como recebido, a análise de
difração de raios X forneceu 12,2% de fração volumétrica de austenita retida. Na direção de
laminação da chapa, os valores médios dos limites de escoamento (𝑆𝑦 ) e resistência (𝑆𝑢 ) em
tração uniaxial são iguais a 604 e 1040 MPa, respectivamente, com um alongamento total igual
a 23,4% (𝑒𝑡 ). Deste modo, a chapa de aço Gen3 980T apresenta uma conformabilidade global,
definida por 𝑆𝑢 𝑥 𝑒𝑡 = 24.3 GPa%, valor condizente com o esperado para a terceira geração de
aços avançados de alta resistência. Foram encontrados valores muito próximos para os
coeficientes de Lankford determinados nas orientações angulares 0, 45 e 90 graus em relação a
direção de laminação, os quais forneceram coeficientes de anisotropia plástica planar e normal
iguais a − 0,079 e 0,917, respectivamente. A razão de expansão de furo determinada para o aço
Gen3 980T foi igual a 10,9% e o índice correspondente ao ensaio Erichsen igual a 10,62 mm.
As simulações numéricas foram capazes de descrever o aumento de conformabilidade em
decorrência da redução do coeficiente de atrito e devido aos efeitos inerentes a curvatura da
amostra do ensaio Nakajima. Os parâmetros do modelo de dano GTN, identificados a partir dos
resultados de tração uniaxial, forneceram boa previsão para os resultados experimentais
situados no lado esquerdo da CLC. Em relação ao lado direito da CLC, os resultados obtidos
para as deformações limites foram conservativos, principalmente para o modo de deformação
de tração biaxial simétrica.
Palavraschave: Aços avançados de alta resistência, classe 980, Ensaios mecânicos,
Deformações limites, Modelamento de dano, Conformação de chapas.
ABSTRACT
The third generation of the advanced highstrength steels (AHSS) has attracted the automotive
industry attention owing to its good compromise between the formability and production cost.
The current development stage of these steels so far has few products available on the market,
as well as little research aimed at studying their formability. With an ultimate tensile strength
greater than 1000 MPa and total elongation above 20%, the thirdgeneration AHSS with
980 MPa grade was the first sheet of this class of steel to be commercialized. In this work,
several mechanical tests were performed to analyze the plastic behavior of a coldrolled
Gen3 980T steel with a nominal thickness of 1.58 mm. The mechanical properties and the
plastic anisotropy Lankford coefficients were obtained from uniaxial tensile tests performed in
7 angular orientations in the plane of the sheet. The workhardening of the Gen3 980T steel
sheet was also assessed by means of the hydraulic bulge test, and the biaxial anisotropy
coefficient was obtained using the disc compression test. The formability was investigated by
means of the hole expansion, Erichsen cupping, and forming limit curve (FLC) test procedures.
The limit strains were defined by ISO 120042 standard from inplane tests (uniaxial tension
and plane strain in tension) and out of plane hemisphericalpunching tests (Nakajima
procedure). The GursonTvergaardNeedleman (GTN) damage model was adopted to describe
the fracture behavior of the Gen3 980T steel. A simple methodology for identifying the
parameters of the GTN model was proposed. This calibration procedure was validated by
comparing finite element predictions with the experimental measurements obtained from the
mechanical tests. The Gen3 980T steel has a microstructure composed of martensite, ferrite,
and retained austenite. In the asreceived condition, the Xray diffraction analysis provided
12.2% of the retained austenite volume fraction. In the sheet rolling direction, the average
values of the yield stress (𝑆𝑦 ) and ultimate tensile strength (𝑆𝑢 ) are equal to 604 and 1040 MPa,
respectively, along with a total elongation (𝑒𝑡 ) of 23.4%. In this way, Gen3 980T steel has global
formability, defined by 𝑆𝑢 𝑥 𝑒𝑡 = 24.3 GPa%, a value which is consistent with that expected for
the thirdgeneration of advanced highstrength steels. The Lankford coefficients determined in
the angular orientations 0, 45, and 90 degrees with respect to the rolling direction are very close,
which provided planar and normal anisotropy coefficients of − 0,079 and 0.917, respectively.
The hole expansion ratio (HER) determined for the Gen3 980T steel sheet was equal to 10.9%,
whereas the corresponding average Erichsen cupping test index (EI) was equal to 10.62 mm.
The numerical simulations were able to describe the increase in formability due to the reduction
of the friction coefficient and the effects inherent to the Nakajima test sample's curvature. The
GTN damage model parameters, identified from the experimental uniaxial tensile data,
provided a good forecast of the experimental results located on the lefthand side of the FLC.
Regarding the right side of the FLC, the results obtained for the limit strains were conservative,
mainly for the equal biaxial stretching.
Keywords: Advanced high strength steels, 980 class, Mechanical Testing, Limit strains,
Damage modeling, Sheet metal forming.
ACKNOWLEDGEMENTS
First, I would like to express sincere gratitude to my supervisors Prof. Luciano Pessanha
Moreira, Prof. António Manuel de Bastos Pereira, and Prof. Marilena Carmen Butuc, for their
guidance, encouragement, and support. I would also like to thank Prof. Gabriela Tamara Vincze
for all the shared knowledge and partnership in the development of this thesis.
I want to thank the support of Centro Federal de Educação Tecnológica Celso Suckow
da Fonseca – CEFET/RJ and express my acknowledgments to the Mechanical Engineering
Department, which allowed me to dedicate myself fully to this work for one sabbatical year.
Special thanks to my coworkers Prof. Alexandre Luiz Pereira, Prof. Marcelo dos Reis Farias,
and Prof. Thiago de Carvalho Silva, who covered my classes during this period.
I wish to express my gratitude to the Graduate Program on Metallurgical Engineering
of Universidade Federal Fluminense. I am equally grateful to all colleagues of TEMA (Centre
of Mechanical Technology and Automation) and the Department of Mechanical Engineering at
the University of Aveiro, who welcomed very well in Portugal and helped me develop my work.
This research work was financed in part by the “Coordenação de Aperfeiçoamento de
Pessoal de Nível Superior” (CAPES) Finance Code 001. The author also acknowledges the
research funding from the Operational Program for Competitiveness and Internationalization,
in its FEDER/FNR component, and the Portuguese Foundation of Science and Technology
(FCT), in its Budget component (OE), through projects POCI010145FEDER032466,
UIDB/00481/2020 and UIDP/00481/2020, and CENTRO010145FEDER022083 Centro
Portugal Regional Operational Programme (Centro2020), under the PORTUGAL 2020
Partnership Agreement, through the European Regional Development Fund. The support of
General Motors for supplying the material is gratefully acknowledged.
Finally, I wish to express my love and gratitude to my beloved family. Without their
endless encouragement and assistance, I would not have finished this thesis.
LIST OF FIGURES
Figure 2.1 – Global formability diagram for various automotive steels: the future opportunity and
definitions of the third generation AHSS. Adapted from Matlock et al. (2008). .................................... 8
Figure 2.2 – Schematic thermal cycles to produce the (a) TBF and (b) Q&P steels. Adapted from
Ebner et al. (2018). ................................................................................................................................ 11
Figure 2.3 – Global formability diagram of the 3rd generation of AHSS from the literature. Reference
number according to Table 2.1. ............................................................................................................. 13
Figure 2.4 – Engineering stressstrain curves of Q&P 980 steel obtained by several authors. ............. 15
Figure 2.5 – Results of hole expansion test for Q&P 980 steel: (a) punch loaddisplacement curve and
(b) appearance of cracks on the rim of an expanded hole (Li et al. 2018). ........................................... 16
Figure 2.6 – Results of Erichsen cupping test of Q&P 980 steel: (a) punch loaddisplacement curve
and (b) failure location (Li et al. 2018). ................................................................................................ 16
Figure 2.7 – Forming limit curve for a Q&P 980 steel sheet (Chen et al. 2017). ................................. 17
Figure 2.8 – (a) OM and (b) SEM images of a Q&P 980 steel (Guo et al., 2018). ............................... 18
Figure 2.9 – Optical micrographs of a Q&P 980 steel (a) in black and white, (b) photographed with
polarized light, and (c) color thresholding overlay (red), Finfrock et al. (2020). ................................. 19
Figure 2.10 – (a) XRD patterns and (b) correlation between the strain rate and the volume fraction of
retained austenite (Guo et al. 2018). ..................................................................................................... 19
Figure 2.11 – The volume fraction of retained austenite of a Q&P 980 steel as a function of the
longitudinal tensile strain (Chen et al. 2017). ....................................................................................... 20
Figure 2.12 – 3D Xray microtomography applied to void volume fraction determination for different
references in the literature (a) Maire et al. (2008), Requena et al. (2014), and (c) Cao et al. (2014). .. 23
Figure 2.13 – SEM images highlighting voids, inclusions, ferrite, and martensite phases for the (a)
uniaxial tension, and (b) plane strain deformation modes, Samei et al. (2016). ................................... 24
Figure 2.14 – (a) EDS analysis of the inclusion in DP800 steel and (b) SEM image with elongated
void areas formed around the inclusions in uniaxial tension at a strain level of 18%........................... 24
Figure 2.15 – Voids in the DP600 steel microstructure resulting from deformation modes under (ab)
uniaxial tension, (cd) plane strain, and (ef) biaxial tension specimens. After Samei et al. (2016)..... 25
Figure 2.16 – Microvoid quantities determined for the DP600 steel sheet deformed under uniaxial
tension, planestrain, and biaxial stretching modes: (a) void density, (b) void area fraction, (c) void
aspect ratio, and (d) mean void size. Adapted from Samei et al. (2016)............................................... 26
Figure 2.17 – Schematic representation of the specimens cutting method used by Saeidi et al. (2014).
............................................................................................................................................................... 27
Figure 2.18 – (a) Void area fraction and (b) void density as a function of the true thickness strain for
the DP780 steel, adapted from Saeidi et al. (2014). .............................................................................. 27
Figure 2.19 – (a) Specimen cutting sections for microvoid analysis, (b) volume fraction of strain
induced martensite as a function of the longitudinal truestrain in uniaxial tension, (c) void density and
void area fraction in the parallel, and (d) transverse crosssections. From Castanheira et al. (2019). .. 28
Figure 2.20 – DP600 steel (a) evolution of the void area fraction with thickness strain for tensile
specimen, and (b) variation of the void area fraction with the thickness strain for two categories of
void size. After Cingara et al. (2009). ................................................................................................... 29
Figure 2.21 –Morphology of voids at the crack initiation zone for DP600, Zhao et al. (2016). ........... 30
Figure 2.22 –(a) Specimen with the ring notch during a tensile test, (b) fracture surface of specimen (c)
microscopic photograph of the fracture surface, and (d) surface fracture photograph after binarization,
Wcislik (2016). ...................................................................................................................................... 30
Figure 3.1 – SEM micrographs of the Gen3 980T steel sheet with (a) 1,000x, and (b) 10,000x
highlighting the ferrite (F), martensiteaustenite (MA), and retained austenite (RA). .......................... 32
Figure 3.2 – XRay diffractogram patterns of Gen3 980T steel sheet determined from the asreceived
condition................................................................................................................................................ 34
Figure 3.3 – Nomenclature of the main dimensions of the uniaxial tensile test specimen. .................. 35
Figure 3.4 – Young’s modulus determination from a uniaxial tensile test. .......................................... 36
Figure 3.5 – The Poisson’s effect during a uniaxial tensile loading. .................................................... 37
Figure 3.6 – Engineering stressstrain curve showing the definitions of yield strength, ultimate tensile
strength, uniform elongation, and total elongation................................................................................ 37
Figure 3.7 Geometry and dimensions [mm] of the uniaxial tensile test specimen. ............................ 38
Figure 3.8 Spray pattern reference with 100 mm × 80 mm (ARAMIS, 2009). ................................. 39
Figure 3.9 – (a) Engineering stressstrain curve and (b) true stressstrain curve of Gen3 980T steel
sheet in the RD. ..................................................................................................................................... 39
Figure 3.10 – Position of Gen3 980T steel in the global formability diagram. ..................................... 40
Figure 3.11 – Angular orientations of the specimens with respect to the rolling direction................... 42
Figure 3.12 – (a) Comparison between engineering stressstrain curves for 0º, 45º, and 90º RD, and (b)
Lankford coefficient of the Gen3 980T steel as a function of orientation angle with respect to the
rolling direction. .................................................................................................................................... 43
Figure 3.13 – Normalized yield stress and ultimate tensile strength as a function of the orientation
angle with respect to the rolling direction. ............................................................................................ 44
Figure 3.14 – Uniform and total elongation as a function of the orientation angle with respect to the
rolling direction. .................................................................................................................................... 44
Figure 3.15 – Experimental true stress and true plasticstrain data of the Gen3 980T steel sheet at the
rolling direction and Hollomon’s predicted workhardening curve. ..................................................... 45
Figure 3.16 – Experimental true stress vs. true plastic strain and the curve fitted by the Ludwik
hardening equation. ............................................................................................................................... 46
Figure 3.17 – Experimental true stress and true plasticstrain data of the Gen3 980T steel sheet at the
rolling direction and Swift’s predicted workhardening curve.............................................................. 47
Figure 3.18 – Experimental true stress and true plasticstrain data of the Gen3 980T steel sheet at the
rolling direction and Voce’s predicted workhardening curve. ............................................................. 48
Figure 3.19 – The true stressstrain curves obtained with strain rates of 103 and 102s1. .................... 49
Figure 3.20 – Calculated and average strainrate sensitivity parameter 𝑚 of the Gen3 980T steel sheet
plotted as a function of the uniaxial tensile true plasticstrain. ............................................................. 50
Figure 3.21 – Experimental and fitted truestress and true plasticstrain along the rolling direction as a
function of the strainrate. ..................................................................................................................... 50
Figure 3.22 – (a) Sheet strip cut by wire electrical discharge machining and (b) disc compression
sample dimensions. ............................................................................................................................... 52
Figure 3.23 – Universal testing machine in compression mode and sample lubrication with graphite
grease (in detail). ................................................................................................................................... 52
Figure 3.24 – Measurements of the disc compression specimens of the Gen3 980T steel sheet: (a)
undeformed and (b) tested with a maximum load of 50 kN. ................................................................. 53
Figure 3.25 – Transverse strain vs. longitudinal strain in the disc compression test with graphite grease
lubricant................................................................................................................................................. 54
Figure 3.26 – (a) Specimen dimensions in [mm] for HET and (b) hole sheared edge produced by the
punching process. .................................................................................................................................. 55
Figure 3.27 – (a) Schematic setup of the device for HET and (b) detailed view of HET parts. ........... 56
Figure 3.28 – Hole sheared edge at (a) 0º and (b) 90º, according to the sheet rolling direction. .......... 57
Figure 3.29 – Punch reaction force curves determined from the HET of the Gen3 980T steel. ........... 58
Figure 3.30 – Final hole diameter with a fracture through the full thickness........................................ 59
Figure 3.31 – Specimen dimensions in [mm] adopted for the Erichsen cupping test. .......................... 60
Figure 3.32 – (a) Universal testing machine ECT setup and (b) detailed view of the tooling. ............. 61
Figure 3.33 – Deformed specimens and fracture locations obtained from Erichsen cupping tests of the
Gen3 980T steel sheet. .......................................................................................................................... 61
Figure 3.34 – Thickness profile of the Gen3 980T steel sheet after Erichsen cupping test carried out
until the fracture. ................................................................................................................................... 62
Figure 3.35 – Punch displacement versus punch force reaction during the ECT. ................................. 62
Figure 3.36 – An element of sheet: (a) the undeformed state with the circle and square grids and (b)
the deformed state with the grid circles deformed to ellipses of major diameter 𝑑1and minor diameter
𝑑2 (Marciniak et al, 2002). ................................................................................................................... 63
Figure 3.37 – Deformation modes defined by the strainratio in the principal strains space. Adapted
from Marciniak et al. (2002). ................................................................................................................ 65
Figure 3.38 – A typical forming limit diagram for the isotropic case (Kumar et al., 2016). ................ 66
Figure 3.39 – DIC results obtained from the uniaxial tension specimen: (a) major and (b) minor
strains. ................................................................................................................................................... 67
Figure 3.40 – Major and minor truestrains distributions obtained along section 1. ............................. 68
Figure 3.41 – Plane strain specimen geometry and dimensions [mm], Schwindt et al. (2015). ........... 69
Figure 3.42 – Horizontal and vertical sections used to measure the major and minor strains in the
planestrain specimen. ........................................................................................................................... 69
Figure 3.43 – Selected strain fields to define the: (a) major and (b) minor strains in the doublenotched
planestrain specimens. ......................................................................................................................... 70
Figure 3.44 – Major strain and minor strain distributions determined from the (a) vertical and (b)
horizontal sections for a specimen in the rolling direction. .................................................................. 70
Figure 3.45 – Schematic of the tooling used in the Nakajima tests. Dimensions in mm. ..................... 71
Figure 3.46 – Blank specimen geometries used to reproduce: (a) the left side of FLC and the (b) right
side of FLC from the Nakajima testing procedure. ............................................................................... 72
Figure 3.47 – Comparison of the Gen3 980T steel limit strains obtained from flat specimens and
Nakajima tests. ...................................................................................................................................... 73
Figure 3.48 – (a) Principle of the hydraulic bulge test (Lazarescu et al., 2013), and (b) variables used
for evaluation of stress and strain (Campos et al., 2014). ..................................................................... 73
Figure 3.49 – Biaxial true stressstrain curve obtained from the bulge test. ......................................... 76
Figure 3.50 – Cut slices prepared from the gauge length region for the VF analysis. .......................... 76
Figure 3.51 – Gen3 980T steel sheet in the asreceived condition: (a) SEM micrograph (1,500x) and
(b) corresponding binarized image. ....................................................................................................... 78
Figure 3.52 – Microvoid analysis results: (a) void density, (b) mean void size, and (c) void area
fraction versus the true strain. ............................................................................................................... 78
Figure 4.1 Representative volume element (RVE): definition of effective stress. .............................. 82
Figure 4.2 Experimental void area fraction as a function of the true longitudinal plasticstrain for the
Gen3 980T steel sheet under uniaxial tension and with respect to the rolling direction. ...................... 83
Figure 4.3 Comparison between the experimental true stress and true plasticstrain data and the
predicted workhardening curves with respect to the rolling direction. ................................................ 83
Figure 4.4 Schematic approach adopted to obtain the initial void volume fraction from the measured
void area fraction................................................................................................................................... 84
Figure 4.5 Yield loci parameters (𝑞1, 𝑞2) identified from the micromechanical predictions of
Faleskog et al. (1998) using the experimental data of Gen3 980T steel sheet. ..................................... 85
Figure 4.6 Softening effects in uniaxial tension as a function of the fN parameter in the GTN model
(Abaqus, 2009). ..................................................................................................................................... 86
Figure 4.7 Boundary conditions for the uniaxial tensile specimen with 1/4 symmetry. ...................... 87
Figure 4.8 Experimental uniaxial tensile test loadelongation and predicted results obtained (a) with
the parameters of the fully dense matrix (Table 4.1), and (b)with the parameters of the fully dense
matrix (Table 4.1) and the GTN parameters (Table 4.2)....................................................................... 88
Figure 4.9 Major strain predictions of the uniaxial tensile test without the failure parameters: (a)
regions at the necking and far from the necking, and (b) corresponding major strain history. ............. 89
Figure 4.10 Proposed identification procedure for the critical void volume fraction 𝑓Cvalue: contours
plots of the (a) major strain and (b) void volume fraction and (c) void volume fraction measures as a
function of the major strain in the necking region. ............................................................................... 90
Figure 4.11 Images of the uniaxial tensile specimen during the testing at (a) the unloaded state, (b)
immediately before the fracture, and (c) after the fractured. ................................................................. 91
Figure 4.12 (a) Finite element simulations of the uniaxial tensile test of the Gen3 980T steel sheet: (a)
experimental digital image and forecasted deformed specimen and (b) void volume fraction. ............ 92
Figure 4.13 Experimental uniaxial tensile test loadelongation and predicted results obtained with the
full set parameters of the GTN damage model for the Gen3 980T steel. .............................................. 92
Figure 4.14 Different mesh sizes in the gauge length region of the uniaxial tensile specimen: (a) 1.6
mm, (b) 0.8 mm, (c) 0.4 mm, and (d) 0.2 mm....................................................................................... 93
Figure 4.15 Influence of the mesh size on the loadelongation curve. ............................................... 94
Figure 4.16 Contour plot distributions of the (a) major strain and (b) minor strain determined by the
numerical simulation of the uniaxial tension test with a flat specimen. ................................................ 96
Figure 4.17 Experimental and predicted major and minor strains at the centerline section of the
uniaxial tensile specimen. ..................................................................................................................... 96
Figure 4.18 Limit strains from the uniaxial tensile test simulations using the GTN model. ............... 97
Figure 4.19 Mesh refinement adopted in the flat doublenotched specimen ....................................... 97
Figure 4.20 Distribution of the (a) major strain and (b) minor strain obtained by the numerical
simulation of the plane strain test with a flat doublenotched specimen. .............................................. 98
Figure 4.21 Experimental and predicted major and minor strains determined from the doublenotched
planestrain specimen: (a) vertical and (b) horizontal sections. ............................................................ 98
Figure 4.22 Limit strains from the planestrain tension test simulations using the GTN model. ........ 99
Figure 4.23 Experimental true stressstrain curve of the plane strain test and predicted results obtained
with the full set of parameters of the GTN model. .............................................................................. 100
Figure 4.24 Void volume fraction (VFF), void volume fraction growth (VVFG), and void volume
fraction nucleated (VVFN): (a) before and (b) after the fracture. ....................................................... 100
Figure 4.25 Details of the 3D finite element model of the Nakajima test: (a) mesh refinement of the
specimen W25 and (b) the rigid tooling assembly. ............................................................................. 102
Figure 4.26 Distribution of the (a) major and (b) minor strain obtained by numerical simulation of the
Nakajima test for the W25 specimen. ................................................................................................. 102
Figure 4.27 Limit strains defined from the Nakajima test numerical simulation with the W25
specimen using the GTN model. ......................................................................................................... 103
Figure 4.28 Comparison between the experimental lefthand side of the forming limit curve and the
numerical predictions determined for the Gen3 980T steel sheet using the GTN model. .................. 104
Figure 4.29 Influence of the friction coefficient on the result of the numerical simulation of the
Nakajima test. ...................................................................................................................................... 105
Figure 4.30 – (a) VVF distribution for W25 specimen and (b) VVF of the critical element as a function
of the equivalent plastic strain for all specimen geometry numerically simulated. ............................ 105
Figure 4.31 Comparison between the result of experimental and numerical simulation of hole
expansion test in terms of punch forcedisplacement. ........................................................................ 106
Figure 4.32 Predicted stress triaxiality values determined from the uniaxial tensile and hole
expansion tests. ................................................................................................................................... 107
Figure 4.33 Distribution of the void volume fraction in the hole expansion test.............................. 108
Figure 4.34 Comparison between the experimental and predicted punch forcedisplacement curves
determined from the Erichsen cupping test of the Gen3 980 steel sheet............................................. 108
Figure 4.35 (a) Distribution of void volume fraction for the numerical simulation of ECT, and (b)
comparison between experimental and numerical simulation result of deformed ECT specimen profile.
............................................................................................................................................................. 109
Figure 4.36 Unit cell models submitted to (a) uniaxial tension, (b) planestrain tension, and (c)
equibiaxial stretching deformation modes. ......................................................................................... 110
Figure 4.37 Stress triaxiality factor as a function of the equivalent plastic strain for the calibrated
GTN model parameters for different deformation modes. .................................................................. 110
Figure 4.38 Void volume fraction as a function of the equivalent plastic strain for the calibrated GTN
model parameters for different deformation modes. ........................................................................... 111
Figure 4.39 Void volume fraction due to the (a) growth and (b) nucleation as a function of the
equivalent plastic strain for the calibrated GTN model parameters for uniaxial tension, plane strain,
and equibiaxial tension. ....................................................................................................................... 112
Figure 4.40 Comparison of the numerical predictions of the GTN model for a single element and
critical element of different deformation modes ................................................................................. 113
LIST OF TABLE
Table 2.1 – Chemical composition and tensile properties of the 3rd generation of AHSS. .................. 14
Table 2.2 – GTN model parameters available in the literature. Adapted from Gholipour et al. (2019),
and Yildiz and Yilmaz (2020) ............................................................................................................... 22
Table 3.1 – Calculated theoretical 𝑅𝛼 and 𝑅𝛾 using Cu radiation (Jatczak, 1980). ............................. 33
Table 3.2Integrated intensity per angular diffraction peak from Lorentz fitting peak (𝐼𝛼ℎ𝑘𝑙, 𝐼𝛾ℎ𝑘𝑙). 34
Table 3.3 – The volume fraction of RA of Gen3 980T steel................................................................. 34
Table 3.4– Mechanical properties of Gen3 980T steel sheet in the rolling direction............................ 40
Table 3.5– Mechanical properties of Gen3 980T steel as a function of the angular orientation. .......... 43
Table 3.6– Planar and normal anisotropy of Gen3 980T steel. ............................................................. 43
Table 3.7– Parameters of Hollomon’s workhardening equation. ........................................................ 45
Table 3.8– Parameters of Ludwik’s workhardening equation. ............................................................ 46
Table 3.9 – Parameters of Swift’s workhardening equation. ............................................................... 47
Table 3.10 – Parameters of Voce’s workhardening equation. ............................................................. 48
Table 3.11 – Parameters of the modified workhardening equations.................................................... 49
Table 3.12 – Vickers hardness of the Gen3 980T steel sheet. ............................................................... 51
Table 3.13 – Disc compression test results for the Gen3 980T steel sheet. .......................................... 53
Table 3.14 – Biaxial anisotropy coefficient (𝑟𝑏) value of the Gen3 980T steel sheet. ......................... 54
Table 3.15 – Hole expansion test results determined for the Gen3 980T steel sheet. ........................... 59
Table 3.16 – Erichsen cupping test results determined for the Gen3 980T steel. ................................. 63
Table 3.17 – Minor and major true strains in uniaxial tension for the Gen3 980T steel. ...................... 68
Table 3.18 – Minor and major limit strains inplane strain tension for Gen3 980T steel. .................... 71
Table 3.19 – Limit strains obtained from the Nakajima tests* of the Gen3 980T steel sheet. .............. 72
Table 4.1 –Elastic properties and parameters of the isotropic effective workhardening. .................... 88
Table 4.2 – Initial porosity, material, and void nucleation parameters. ................................................ 88
Table 4.3 – Full set parameters of the GTN damage model for the Gen3 980T steel. .......................... 91
Table 4.4 – Mesh sensitivity – minimum standard error of loadelongation curve, CPU time and
number of elements. .............................................................................................................................. 95
Table 4.5 – Types and number of elements for each part used in the Nakajima test simulation. ....... 101
Table 4.6 – Major and minor limit strains obtained from the numerical simulations of the Nakajima
test for the Gen3 980T steel sheet. ...................................................................................................... 103
LIST OF ABBREVIATIONS
A3 Annealing temperature, where the material is composed of austenite
ACEA European Automobile Manufacturers Association
AHSS Adavanced High Strength Steel
AUST.SS Austenitic Stainless Steel
BCC BodyCentered Cubic
BIW BodyInWhite
CNC Computer Numerical Control
CP ComplexPhase steel
DIC Digital Imagem Correlation
DP DualPhase steel
ECT Erichsen Cupping Test
EDM Electrical Discharge Mechining
EI Erichsen Index
FCC FaceCentered Cubic
FE Finite Element
FLC Forming Limit Curve
FLD Forming Limit Diagram
FZ Fusion Zone
GTN GursonTvergaardNeedleman model
HAZ HeatAffected Zone
HER Hole Expansion Ratio
HET Hole Expansion Test
HRC Rockwell Hardness Cscale
HV Vicker Hardness
LIP Aladded lightweight steel with Induced Plasticity
LW Laser Welding
MAG Metal Active Gas welding
MART Martensitic steel
MMnS MediumMn Steel
MS Martensite start temperature
MVS Mean Void Size
NR&S Nanophase Refinement and Strenthening
OM Light Optical Microscopy
PI Performance Index
Q&P Quenching and Partitioning
RA Retained Austenite
RD Rolling Direction
RSW Resistence Spot Welding
RVE Representative Volume Element
SE Minimum Standard Error
SEM Scanning Electon Microscope
SIMT StrainInduced Martensite Transformation
TBF TRIPaided Bainitic Ferrite steel
TD Transverse direction
TRIP Transformation Induced Plasticity steel
TWIP Twinning Induced Plasticity steel
VAF Void Area Fraction
VD Void Density
VF Void Formation
VG Void Growth
VN Void Nucleation
VVF Void Volume Fraction
VVFG Void Volume Fraction due to Growth
VVFN Void Volume Fraciton due to Nucleation
XRD XRay Diffraction
δTRIP TRIP with densityreduced steel
LIST OF NOTATIONS
𝛼 Slope of line equation
𝐴 Area
𝐴0 Initial Area
𝐴𝑓 Final Area
𝐴𝑉 Area of voids
𝛽 yintercept of line equation
𝑑 Average length of diagonals of the Vickers hardness indentation
𝑑0 Initial spacing of the grid to measure the strains during sheet metal forming
𝑑1 Measure of deformed grid in the major axis
𝑑2 Measure of deformed grid in the minor axis
𝑖
𝐷0° Initial longitudinal diameter
𝑖
𝐷90° Initial transverse diameter
𝑓
𝐷0° Final longitudinal diameter
𝑓
𝐷90° Final transverse diameter
𝐷ℎ Average hole diameter after the specimen rupture
𝐷0 Initial hole diameter
𝐷𝑐𝑣 Spherometer diameter
𝐷𝑠𝑡 Current diameter of a measured circle in the hydraulic bulge test
𝐷𝑠𝑡0 Intial diameter of a measured cicle in the hydraulic bulge test
∆𝐿 Elongation of the specimen gauge length
∆𝑅 Planar anisotropy coefficient
∆𝑊 Elongation of the specimen width
𝜹 Secondorder identity tensor
𝐸 Young’s Modulus
𝑒𝑡 Total elongation
𝑒 Engineering strain
𝑒𝑊 Engineering width strain
𝑒𝐿 Engineering length strain
𝜀 True strain
𝜀𝑊 True width strain
𝜀𝑡 True thickness strain
𝜀𝐿 True length strain
𝜀0 Prestrain
𝜀̇ 𝑝 Plastic strain rate
𝜀̇0 Strain rate of reference
𝑝
𝜀𝑇𝐷 True plasticstrain in the transverse direction
𝑝
𝜀𝑅𝐷 True plasticstrain parallel to rolling direction
𝜀𝑁 Mean value of characteristic nucleation plasticstrain
𝜀 𝑝 True plastic strain
𝑝
𝜀𝑡 True plastic thickness strain
𝜺̇ 𝒑 Plastic strainrate secondorder
𝜀̅̇𝑝 Equivalent plastic strainrate
𝜀1 True major strain
𝜀2 True minor strain
𝐹 Force
𝑓𝑁 Nucleated void volume fraction
𝑓𝐶 Critical void volume fraction
𝑓𝐹 Void volume fraction in the fracture
𝑓0 Initial void volume fraction
𝑓 ∗ Void volume fraction parameter
𝑓𝐴 Void area fraction
ℎ Dome height in the hydraulic bulge test
𝐼𝛼 , 𝐼𝛾 Integrated intensity per angular diffraction peak for αiron and γiron, respectively
𝐾 Strength coefficient
𝐿0 Initial gauge length
𝑚 Strainrate sensitivity index
𝑛 Strainhardening exponent
𝜈 Poisson’s ratio
𝑃 Load
𝑝 Hydraulic pressure
𝑞1 , 𝑞2 , 𝑞3 Material parameters of GTN model
𝑟0 Initial porosity
𝑟𝑏 Biaxial anisotropy coefficient
𝑅 Rvalue, Lankford coefficient
𝑅𝛼 , 𝑅𝛾 Tabulated parameters for αiron and γiron, respectively
𝑅0 , 𝑅45 , 𝑅90 Rvalue with three angular orientation 0º, 45º, and 90º with respect the rolling direction
𝑅̅ Normal anisotropy coefficient
𝜌 Curvature radius for hydraulic bulge test and strainratio for limit strains
𝜌1 Radius of the curved surface at midthickness for the major axis
𝜌2 Radius of the curved surface at midthickness for the minor axis
𝜌𝑏 Curvature radius for membrane thery in the hydraulic bulge test
𝑆 Engineering stress
𝑺 Deviactoric stress tensor
𝑆𝑢 Ultimate tensile strength
𝑆𝑁 Standard deviation of the mean characteristic nucleation plasticstrain
𝑆𝑦 Yield strength
𝛴 Stress triaxiality factor
𝛴𝜀𝛼 Sum of the true plasticstrain
𝜎 True stress
𝜎𝑦 True yeld strength
𝜎1 True major stress
𝜎2 True minor stress
𝜎̅ Von Mises equivalent stress
𝜎̃ Effective workhardening
𝝈 Cauchy secondorder stress tensor
𝜎𝑏 Biaxial true stress
𝜎ℎ𝑎𝑟𝑑_𝑒𝑞𝑛 Workhardening equation
𝜎ℎ Hydrostatic stress
𝑡 Current thickness
𝑡𝑖 Initial thickness in the disk compression test
𝑡0 Initial thickness
𝑡𝑓 Final thickness
𝑉 Current volume
𝑉0 Initial volume
𝑉𝛾 Retained austenite volume fraction
𝑊 Width
𝑊0 Initial width
𝜃 Bragg angle
𝜗 Angle between opposite faces of the pyramid in the Vickers hardness test
TABLE OF CONTENTS
Chapter 1 .................................................................................................................................... 1
1. INTRODUCTION .................................................................................................................. 2
1.1. Objectives ........................................................................................................................ 3
1.2. Thesis organization .......................................................................................................... 4
1.3. List of publications .......................................................................................................... 5
Chapter 2 .................................................................................................................................... 6
2. LITERATURE REVIEW ....................................................................................................... 7
2.1. Definition of thirdgeneration of AHSS .......................................................................... 7
2.1.1. Formability of Q&P 980 steel ................................................................................. 15
2.1.2. Microstructural characterization of Q&P 980 steel ................................................. 17
2.2. Damage behavior with the GTN model ......................................................................... 20
Chapter 3 .................................................................................................................................. 31
3. EXPERIMENTAL PROCEDURES .................................................................................... 32
3.1. Microstructural characterization of Gen3 980T steel .................................................... 32
3.2. Uniaxial tensile test ........................................................................................................ 35
3.2.1. Preparation of the uniaxial tensile test .................................................................... 38
3.2.2. Mechanical properties in the rolling direction ........................................................ 39
3.2.3. Parameters of anisotropy ......................................................................................... 40
3.2.4. Workhardening behavior ....................................................................................... 45
3.2.5. Strainrate sensitivity .............................................................................................. 48
3.3. Vickers hardness test ..................................................................................................... 50
3.4. Disc compression test .................................................................................................... 51
3.5. Hole expansion test ........................................................................................................ 54
3.5.1. Preparation of HET specimens ................................................................................ 55
3.5.2. Development of a HET device ................................................................................ 56
3.5.3. Hole sheared edge ................................................................................................... 57
3.5.4. Hole Expansion Ratio (HER) .................................................................................. 58
3.6. Erichsen cupping test ..................................................................................................... 60
3.7. Limit strains ................................................................................................................... 63
3.7.1. Uniaxial tension....................................................................................................... 66
3.7.2. Plane strain .............................................................................................................. 68
3.7.3. Nakajima test ........................................................................................................... 71
3.8. Hydraulic bulge test ....................................................................................................... 73
3.9. Void formation analysis ................................................................................................. 76
Chapter 4 .................................................................................................................................. 79
4. MODELING OF PLASTIC BEHAVIOR ............................................................................ 80
4.1. Calibration of the GTN parameters ............................................................................... 80
4.1.1. Effective workhardening ........................................................................................ 81
4.1.2. Initial porosity ......................................................................................................... 84
4.1.3. Material parameters (𝑞1, 𝑞2, 𝑞3) ........................................................................... 85
4.1.4. Void nucleation parameters (𝜀𝑁, 𝑆𝑁, 𝑓𝑁) ............................................................. 86
4.1.5. Failure parameters (𝑓𝐶, 𝑓𝐹) ................................................................................... 89
4.1.6. Mesh sensitivity....................................................................................................... 93
4.2. Numerical prediction of the FLC ................................................................................... 95
4.2.1. Uniaxial tension mode ............................................................................................. 95
4.2.2. Planestrain mode .................................................................................................... 97
4.2.3. Nakajima test ......................................................................................................... 101
4.2.4. Holeexpansion and Erichsen cupping tests .......................................................... 106
Chapter 5 ................................................................................................................................ 114
5. CONCLUSIONS AND FUTURE WORKS ...................................................................... 115
5.1. Final conclusions ......................................................................................................... 115
5.2. Future works ................................................................................................................ 117
References ............................................................................................................................... 118
1
Chapter 1
Introduction
2
1. INTRODUCTION
Climate changes are becoming ever more a central topic of global discussions.
According to the European Union and concerns about CO2 emissions, high fossil fuel prices
harm the global economy. Thus, reducing the dependence on oil by increasing the efficiency
and sustainability of new passenger cars and light commercial vehicles is a priority (Council of
the European Union, 2014). The European Union agreed to set a target of 37.5% CO2 reduction
for cars by 2030 with the 2021 baseline (Mock, 2019). In response to the European
Commission’s proposal, the European Automobile Manufacturers Association (ACEA)
advocates a realistic target to reduce 20% CO2 emission (ACEA, 2018). In this context, the
Advanced High Strength Steels (AHSS) have gained considerable importance in the automotive
industry. This steels class has been developed, aiming to satisfy the requirements of vehicles’
performance and weight optimization. These materials have improved tensile strength and
elongation, which allow the manufacturing of components with reduced thickness and, thus,
lighter automobiles with fuel economy and fewer emissions of greenhouse gas. Different AHSS
steel grades are applied in the vehicle design, aiming to increase the crashworthiness level,
which is achieved by improving the material toughness.
Generations commonly classify the AHSS sheet materials. The first generation
comprises the DualPhase steel (DP), the Transformation Induced Plasticity (TRIP), Complex
Phase (CP), and martensitic (MART). The second generation of AHSS includes the Twinning
Induced Plasticity (TWIP), Aladded lightweight steel with Induced Plasticity (LIP), and the
Austenitic Stainless Steel (AUST.SS). According to Matlock and Speer (2008), the first
generation of AHSS, which has ferritebased microstructures, is already well established.
Concerning the second generation, the automotive sector considers these materials attractive
due to their excellent formability. However, the 2nd AHSS generation is more expensive due to
the high alloy addition required to stabilize the austenite at room temperature. In particular, Mn
contents between 20 and 30 wt.%. The 3rd generation of AHSS refers to a recent class with
mechanical properties between the 1st and 2nd generations. The 3rd generation steels have a
multiphase microstructure with a significant amount of retained austenite (Grajcar et al., 2012;
Aydin et al., 2013).
Wang and Speer (2013) reported that in 2009 the Baosteel group was the first company
to process the 3rd generation of AHSS sheets industrially. Other companies are also developing
their products to meet market demands. However, there is still little information on marketable
3
products. In 2015, General Motors announced the 3rd generation of AHSS in the SIACGM’s
Chevrolet LOVA RV, reducing selected body components' weight by approximately 20%. This
outcome from a joint venture between GM Pan Asia Automotive Technical Center (PATAC),
Baosteel, and the University of Tongji to develop solutions for introducing the 3rd generation
of AHSS in the GM vehicles (Shen, 2015). According to Macek and Lutz (2020), the U.S. Steel
developed a thirdgeneration AHSS with a 980 MPa minimum tensile strength, commercially
known as 980 XG3 steel. The 980 XG3 steel exhibits an excellent combination of strength and
ductility, with ultimate tensile strength around 1000 MPa and total elongation around 25%.
Stamping tests were carried out with the 980 XG3 steel to manufacture components of the body
of the Jeep Cherokee (Macek and Lutz, 2020).
The introduction of AHSS steel grades has been driven by market requirements and
customers’ demands (Kempken, 2018). With the insertion of the 3rd generation AHSS steel in
the production of BodyInWhite (BIW), it is expected to improve the performance of vehicles
in terms of safety, fuel consumption, and drivability. However, the 3rd generation of AHSS is
still under development and, thus, some challenges related to its application need to be better
investigated. It is necessary to study the limits of the application of the material and the
influence of the forming and joining processes currently available in the automotive industry to
produce the components.
The contribution of this Doctoral Thesis work aimed to perform an experimental
characterization and modeling of the plastic behavior of a recently 3rd generation AHSS steel.
The material evaluated in this study is commercially available, and to attend the confidentiality
requirements, the identity of the steel producer was kept blind. Likewise, a generic terminology
referred to as Gen3 980T steel was attributed to the studied 3rd generation AHSS.
1.1. Objectives
1. Evaluate the mechanical properties of the thirdgeneration Gen3 980T AHSS steel
based on a series of laboratory tests aiming to model its mechanical and damage
behavior.
2. Propose a simple methodology for calibrating the GTN model based on the
literature’s works, uniaxial tensile testing, microvoid analysis, and numerical
simulations.
4
3. Verify the pertinence of the proposed methodology by comparing numerical
predictions of the GTN model and the experimental data from uniaxial tension and
other deformation modes.
1.2. Thesis organization
The thesis manuscript is organized into five chapters. The description of each chapter is
as follows:
1. Introduction: A brief motivation about the importance of studying the third generation
of advanced high strength steels and the objectives outlined for this work is presented.
2. Literature review: A literature review including the definition of the third generation
of AHSS steels, the main manufacturing strategies, the established formability targets,
the current state of development of the third generation of AHSS steel, and its main
characteristics are detailed. The aspects related to the modeling of mechanical behavior
using the GTN model are also addressed, such as calibration of the model parameters
and microvoids experimental analysis.
3. Experimental characterization: The techniques adopted in the laboratory testing are
detailed, aiming to define the mechanical and anisotropic properties, the limit strains,
and the hole expansion ratio of the GEN3 980T steel. The proposed microvoids and
retained austenite volume fraction analyses are also detailed.
4. Modeling of plastic behavior: The damage behavior of the GEN3 980T steel sheet is
described using the classical GTN damage model. Firstly, a calibration procedure for
the GTN material parameters is proposed, and a mesh sensitivity analysis is also done.
Then, numerical simulations of mechanical tests were performed to compare with the
results obtained experimentally. Lastly, the limitations of the proposed methodology are
discussed.
5. Conclusions: The concluding remarks are outlined, and the perspectives of future
research are proposed.
5
1.3. List of publications
Santos, R. O., Pereira, A. B., Butuc, M. C., Vincze, G., Festas, A. J., Moreira, L. P.
Development of a device compatible with universal testing machine to perform hole expansion
and Erichsen cupping tests. Machines, vol. 8 (2020), pp. 18.
https://doi.org/10.3390/machines8010002
Santos, R. O., Silveira, L. B., Moreira, L. P., Cardoso, M. C., Silva, F. R. F., Paula, A. S.,
Albertacci, D. A. Damage identification parameters of dualphase 600800 steels based on
experimental void analysis and finite element simulations. Journal of Materials Research and
Technology, vol. 8 (2019), pp. 644459. https://doi.org/10.1016/j.jmrt.2018.04.017
Pereira, A. B., Santos, R. O., Carvalho, B. S., Butuc, M. C., Vincze, G., Moreira, L. P. The
evaluation of laser weldability of the thirdgeneration advanced high strength steel. Metals, vol.
9 (2019), pp. 113. https://doi.org/10.3390/met9101051
Castanheira, B. C., Moreira, L. P., Nascimento, G. O., Santos, R. O., Santiago, G. A. A.,
Oliveira, I. S., Silva, F. R. F. Void formation and straininduced martensitic transformation in
TRIP 780 steel sheet submitted to uniaxial tensile loading. Material Research, vol. 22 (2019),
pp. 111. https://doi.org/10.1590/19805373mr20190245
Silveira, L. B., Moreira, L. P., Silva, L., Santos, R. O., Silva, F. R., Cardoso, M. C., Freitas,
M. C. S. Limit strains analysis of advanced high strength steels sheets based on surface
roughness measurements. Materials Science Forum (online), vol. 930 (2018), pp. 349355.
https://doi.org/10.4028/www.scientific.net/MSF.930.349
6
Chapter 2
Literature review
7
2. LITERATURE REVIEW
2.1. Definition of thirdgeneration of AHSS
According to the literature, different steel types and strategies have been adopted to
develop this new generation of AHSS. Among them, the Quenching and Partitioning (Q&P),
MediumMn (MMnS), TRIPaided Bainitic Ferrite (TBF), densityreduced steel (δTRIP), and
NanoSteel (Yi et al., 2018; Sugimoto et al., 2017 and Branagan et al. 2018). In 2003, Speer et
al. (2003) first proposed an approach designated as the Q&P process to exploit novel martensitic
steels containing retained austenite. Wang and Speer (2013) explained that since 2003 the Q&P
steel has gained interest due to its improved properties of strength and ductility with a
composition comparable to TRIP steel and has been proposed as thirdgeneration automotive
steel. In principle, any steel with both strength and elongation between the first and second
generations (Paykani et al., 2016), and lower production costs than the second generation could
also be candidates for the third generation. In terms of the ultimate tensile strength (𝑆𝑢 ) and
total elongation (𝑒𝑡 ) values obtained from uniaxial tensile testing, there is still no consensus on
the definition to be used for the 3rd generation of AHSS. For Matlock et al. (2008) and Sugimoto
et al. (2017), the product of 𝑆𝑢 × 𝑒𝑡 required for the thirdgeneration AHSS must be higher
than 30 GPa%, being desirable at least 𝑆𝑢 = 1000 MPa and 𝑒𝑡 = 30%. The US Department
of Energy set two targets, namely, one steel grade with 𝑆𝑢 = 1200 MPa and 𝑒𝑡 = 30%
(36 GPa%), and another steel grade with 𝑆𝑢 = 1500 MPa and 𝑒𝑡 = 25% (37.5 GPa%)
(Yi et al., 2018). Hance (2018) stated that the 3rd generation AHSS should have, by definition,
the performance of 𝑆𝑢 × 𝑒𝑡 greater than 20 GPa%.
8
Hance (2018) also proposed an alternative performance classification system for AHSS,
known as Performance Index (PI=𝑆𝑢 ∙√𝑒𝑢 ×𝑙𝑛(𝐴0 ⁄𝐴f )). The PI reflects not only strength (𝑆𝑢 )
and global formability (𝑒𝑢 ), but also the intrinsic local formability (𝑙𝑛(𝐴0 ⁄𝐴f )) and, accordingly,
the 3rd generation of AHSS must provide the PI > 30 GPa%. Except for the medium Mn steels,
most developed AHSS have the PI between 20 GPa% and 30 GPa%, and are commonly referred
to as “current 3rd generation”. Figure 2.1 shows the global formability diagram, in which a
region between the first and second generation is highlighted as an opportunity region for the
development of the thirdgeneration AHSS.
80
5 10 15 20 25 30 35 40 45 50 55 60 65 70 (GPa%) 3rd Gen AHSS
Future Opportunity
70
LIP 2nd GEN
60
Total Elongation (%)
AUST.SS TWIP
50 IF
IFHS 3rd GEN
40 Mild ISO
70
Matlock (2008)
BH Sugimoto (2017) Yi (2018)
30
CMn Yi (2018)
20
TRIP Current
HSLA
DP, CP
10 MART
MART
1st GEN 5
0
0 200 400 600 800 1000 1200 1400 1600 1800
Ultimate Tensile Strength (MPa)
Figure 2.1 – Global formability diagram for various automotive steels: the future opportunity
and definitions of the third generation AHSS. Adapted from Matlock et al. (2008).
Wang and Speer (2013) announced the start of industrial implementation of the third
generation of AHSS through the Q&P 980 and Q&P 1180 steels. In terms of metal formability,
it was observed that these steels are comparable to the DP 780 steel grade, but with the strength
of the DP 980 steel. Regarding the performance in welded joints, Wang and Speer (2013) tested
the resistance spot welding (RSW), laser welding (LW), and metal active gas (MAG) welding.
They concluded that by using the correct welding parameters, the Q&P 980 steel could be
successfully welded. Chen et al. (2017) also compared the formability of Q&P 980 with the DP
980 by evaluating the forming limit diagram (FLD) based on the Marciniak testing procedure.
9
exhibited poor weldability owing to the shrinkage of porosities and the segregation of Mn and
C in the fusion zone, which, in turn, resulted in total uniaxial tensile elongation less than 3%
and intergranular fracture.
Zhou et al. (2019) investigated the quality of the edge and its effect on the uniaxial
tensile properties and fracture patterns of a Q&P 980 steel sheet with 1.4 mm nominal thickness.
Three different cutting processes were tested: blanking, wire electrical discharging machining
(EDM), and laser cutting. A smaller die clearance can improve the formability of Q&P 980 to
some extent. According to these authors, the recommended die clearance is between 7% to 11%
of the thickness. The total uniaxial elongation and fracture strain considerably decrease for die
clearances higher than 14%. Also, Zhou et al. (2019) observed that the wireEDM provides a
higher level of total elongation 18.00±0.55%, than 17.25±0.65% obtained with the laser cutting.
The heat produced during laser cutting increases the fraction of martensite in the microstructure
near the edge of the specimen. This fact was verified through hardness measurements. The
hardness value for the material as received was approximately 305 HV. In the lasercut samples,
close to the cut (0.3 mm), hardness measurements of the order of 500 HV were obtained.
Since the 1970s, different grades of carbon transformationinduced plasticity (TRIP)
aided bainitic steels containing high silicon content have been developed (Sugimoto et al.
2017). The mediumcarbon contents are known as TRIPaided bainitic ferrite (TBF). Generally,
the Q&P steels display 𝑆𝑢 × 𝑒𝑡 equivalent to 20 GPa%, whereas the TBF steels present
25 GPa% (Sugimoto et al. 2017). Ebner et al. (2018) used coldrolled sheet steel with 1.1mm
thickness and nominal composition of Fe0.2C(0.5/1.0)Si(2.2/2.7)Mn(0/0.5)Cr wt.% to
investigate the TBF and Q&P heat treatments. Figure 2.2, after Ebner et al. (2018), shows the
difference between the two thermal cycles used to change the microstructure and mechanical
properties of TBF and Q&P steels, respectively. The static recrystallization of the
microstructure and carbide dissolution takes place in the heating stage. The material is kept at
temperatures higher than A3 to reach the full austenitization (Krizan et al., 2018). After the
soaking step, the material is abruptly cooled to a specific temperature. The isothermal bainitic
transformation occurs with temperatures near to 400ºC. For the Q&P steel, the ratio of the
martensite and the retained austenite is adjusted by quenching above the MS temperature. Then,
the steel is heated up to promote the carbon partitioning from the martensite to the austenite
followed by the martensite tempering and some minor bainite formation.
By comparing the values obtained from the uniaxial tensile testing, namely, the uniform
truestrain and area reduction, the steels subjected to Q&P heat treatments presented local
11
formability. Conversely, the steels subjected to the TBF heat treatments are superior regarding
global formability. In this manner, it would be possible to adapt the mechanical properties of
the same steel composition to the requirements of a given application by selecting either TBF
or Q&P processing, as suggested by Ebner et al. (2018).
B
Full austenitization Full austenitization
A 3 A 3
Temperature
Temperature
Isothermal bainitic
transformation Partitioning
M S M S
Quenching
Time Time
(a) (b)
Figure 2.2 – Schematic thermal cycles to produce the (a) TBF and (b) Q&P steels. Adapted
from Ebner et al. (2018).
Li et al. (2016) proposed a Quenching and Dynamic Partitioning (QDP) process for a
Fe0.3C1.4Si1.8Mn1.3Cr0.3Mo wt.% bainitic steel. The resulting microstructure and
mechanical properties that are comparable with the Q&P processing route. Q&P steel presents
a 𝑆𝑢 = 1450 MPa, and a 𝑒𝑡 = 20.1% with a 4.9% volume fraction of retained austenite, while the
QDP process provided 𝑆𝑢 =1550 MPa, 𝑒𝑡 =20.3%, and 6.9% of retained austenite volume
fraction. The impact toughness of the QDP treated steel exhibited a maximum value of 108 J,
which is nearly 20% higher than that of the specimen treated by the Q&P process (90 J).
A TRIP 800 steel sheet with chemical composition F30.23C1.23Si1.5Mn wt.% was
evaluated by Ariza et al. (2018a, 2018b) using the Q&P process and the hotstamping followed
by Q&P (HSQ&P). The HSQ&P process results in a final shape formed product with a
reasonable amount of available retained austenite between 15 and 20%. This phase has enough
stability to favor the TRIP effect, which may give an additional advantage in producing high
strength parts with good ductility.
Intuitively, when it is necessary to reduce the mass of a component, the first strategy
adopted is to replace the material by a lighter (lower density) one. However, Jeswiet et al. (2008)
explained that the manufacturing of lightweight components in metal forming is submitted to
12
the analysis of specific stiffness (tensile modulus/density). The tensile modulus is metal
dependent and cannot be changed by alloying. Even if a given metal has a lowdensity value, a
thicker crosssection will be needed to meet the imposed design requirements. Both tensile
modulus and density must be improved, aiming to reduce the mass. In this sense, δTRIP steel
became a promising candidate for the 3rd generation of AHSS for automotive applications based
on the tensile mechanical properties, the weldability, and the lowdensity characteristic and no
decrease of Young’s modulus (Yi, 2014). One of the advantages of δTRIP is the reduced
density achieved by the addition of aluminum. The addition of light elements into steel
decreases the specific mass of the alloy owing to the combination of atomic mass reduction and
lattice parameter expansion. Adding approximately 3.5 wt.% of aluminum to the δTRIP steel
will reduce about 5% of the density.
Yi et al. (2011) carried out tensile tests at room temperature and a nominal strainrate of
3.3 × 103 s1 for an δTRIP sheet steel (Fe0.41C0.26Si1.53Mn2.3Al0.49Cu wt.%) with 1.8
mm of thickness. As result, the material showed 𝑆𝑢 =885 MPa and 𝑒𝑡 =37%, that is, of
𝑆𝑢 × 𝑒𝑡 = 32.7 GPa%, which represent an excellent formability.
On the other hand, high Al content in the δTRIP steel may bring a series of problems,
such as many δferrite dendrites and microsegregation. Since Al is the ferrite promoting
element, high Al content makes the δphase more stable. It inhibits the phase transformation of
δferrite to austenite at high temperature, resulting in a high amount of δferrite, which usually
exists in the form of coarse dendrites as in casting steels (Liu et al., 2019).
Among the 3rd generation AHSS steels, mediumMn steels (MMnS) stand out from the
others for exceeding 35 GPa% in 𝑆𝑢 × 𝑒𝑡 , as shown in Figure 2.3. Initially, MMnS steels
showed a location on the global formability diagram like the TRIP steels, as reported by Aydin
et al. (2013). Later, Hu et al. (2017) introduced mediumMn steels containing between 5 and
10 Mn wt.%. They stated that this steel brought much attention due to its improved mechanical
properties in the uniaxial tensile testing, which exhibit values similar to the highMn TWIP
austenitic steel, but with the lower production costs.
Choi et al. (2017) characterized a medium Mn steel with Fe0.3C6,0Mn3,0Al1,5Si
wt.% through tensile testing for different intercritical annealed temperatures, namely, 760 °C,
800 °C, 840 °C, and 880 °C. According to the authors, the intercritical annealing at 760 °C
displayed the most significant yield strength and elongation values but the lowest tensile
strength because of the low strain hardening behavior. Conversely, the intercritical annealing
at 880°C presented the highest tensile strength with the lowest yield strength and elongation.
13
Hu et al. (2018) also analyzed the effect of intercritical rolling temperature on the mechanical
properties of an MMnS with 6.4 Mn wt.%. Three temperature of intercritical rolling was chosen,
750 °C, 850 °C, and 950 °C, which resulted from XRD analysis an austenite volume fraction
of 33.2%, 34.6%, and 33.9%, respectively. After the intercritical annealing, the tensile strength
was enhanced, and the elongation was slightly reduced. The intercritical rolling annealed
sample at 750 °C presented the best combination of yield strength (705 MPa), tensile strength
(966 MPa), and total elongation (42.6%).
60
[19] [21] [22]
35
30
[20]
50
[19]
[20] 70 GPa%
[12] [19]
40 [13]
Elongation (%)
[20]
[16]
[17]
30 [14]
[21]
[15] [7]
[5] [18]
[14] [6] [8]
[1] [3]
[4] [9]
20
[2] [9]
[1] [11] [20]
MMnS [1] [10]
10 Q&P
TBF
TRIP
NanoSteel 5 GPa%
0
600 800 1000 1200 1400 1600
Utimate Tensile Strength (MPa)
Figure 2.3 – Global formability diagram of the 3rd generation of AHSS from the literature.
Reference number according to Table 2.1.
The combination of ultrahighstrength and ductility is generally obtained through the
intercritical annealing with a large fraction of metastable austenite. It has been reported that
strength over 1000 MPa, together with elongation above 40% have been achieved for MMnS.
However, the long yield elongation, high yield ratio, and serrated plastic flow deteriorate the
homogeneous and stability of deformation. The Quenching and Tempering (Q&T) process was
applied by Li et al. (2019) into experimental coldrolled medium Mn steel. According to the
authors, the complex microstructure (martensite, nanocarbide, and retained austenite) produces
approximately 1.5 GPa of strength with 30% ductility. For Yi et al. (2018), the mediumMn
steel challenge is associated with low work hardening. Also, the Lüders band in the continuous
annealing line MnTRIP would lead to the localized thinning and, therefore, deteriorate the
14
formability and crash performance of stamped components. In resume, it is challenging to
ensure the robustness of commercial products with complex phase, including retained austenite
in all the current concepts of new ductility high strength steels.
Table 2.1 – Chemical composition and tensile properties of the 3rd generation of AHSS.
Chemical composition wt. (%) RA 𝑆𝑢 𝑒𝑡
GPa% Reference Fig. 2.3
C Mn Al Si Cu Cr P Mo (%) (MPa) (%)
Q&P
0.23 1.50 1.23 8.1 836 18.0 10.03 Ariza et al. (2018) [1]
0.23 1.50 1.23 10.4 916 22.7 20.79 Ariza et al. (2018) [1]
0.23 1.50 1.23 13.3 1031 15.0 15.46 Ariza et al. (2018) [1]
977 18.8 18.37 Zhou et al. (2019) [2]
11.6 1005 20.2 20.30 Wang & Speer. (2013) [3]
11.5 1054 23.4 24.66 Chen et al. (2017) [4]
0.25 2.35 0.04 1.33 0.56 12.6 1067 26.6 28.38 Li et al. (2018) [5]
0.20 2.41 0.06 1.31 0.01 11.0 1081 26.0 28.11 Guo et al. (2018) [6]
0.18 1.95 1.46 0.53 0.08 950 30.1 28.60 Tan et al. (2019) [7]
0.30 2.50 1.50 0.80 17.9 1500 21.5 32.25 Wang et al. (2017) [8]
TBF
0.30 1.80 1.40 1.30 0.3 5.8 1550 17.6 27.28 Li et al. (2016) [9]
0.30 1.80 1.40 1.30 0.3 4.9 1450 20.1 29.15 Li et al. (2016) [9]
0.20 2.70 1.00 0.50 10.0 1580 10.9 17.22 Ebner et al. (2018) [10]
0.23 1.41 0.31 1.18 0.22 4.0 1095 18.0 19.71 Benedetti et al. (2017) [11]
δTRIP
0.36 1.96 2.22 0.73 0.52 0.02 12.0 710 42.4 30.10 Yi (2011) [12]
0.41 1.53 2.30 0.26 0.49 19.6 882 38.0 33.25 Yi et al. (2011) [13]
MMnS
0.12 7.00 0.30 29.8 760 22.3 16.95 Wang et al. (2019) [14]
0.12 7.00 0.30 29.8 1035 30.2 31.26 Wang et al. (2019) [14]
0.06 11.7 2.90 0.20 0.01 891 30.1 26.82 Shen et al. (2019) [15]
0.05 12.0 3.00 824 32.6 26.92 Lu et al. (2019) [16]
0.14 7.16 0.23 30.0 1015 30.7 31.16 He et al. (2019) [17]
0.18 10.2 2.78 1.00 20.5 1270 24.2 30.73 Sun et al. (2019) [18]
0.18 7.50 2.80 0.50 0.08 33.2 970 44.2 42.87 Hu et al. (2018) [19]
0.18 8.10 2.80 0.50 0.08 34.6 815 46.7 38.06 Hu et al. (2018) [19]
0.18 8.60 2.80 0.50 0.08 33.9 720 56.7 40.82 Hu et al. (2018) [19]
0.30 6.00 3.00 1.50 930 49.2 45.76 Choi et al. (2017) [20]
0.30 6.00 3.00 1.50 995 47.0 46.77 Choi et al. (2017) [20]
0.30 6.00 3.00 1.50 1130 35.0 39.55 Choi et al. (2017) [20]
0.30 6.00 3.00 1.50 1360 19.2 26.11 Choi et al. (2017) [20]
0.52 7.75 2.78 1520 30.1 45.70 Li et al. (2019) [21]
0.52 7.75 2.78 1090 56.3 61.40 Li et al. (2019) [21]
NanoSteel
1211 57.0 69.03 Branagan et al. (2018) [22]
Data not informed in the reference.
A new class of the 3rd generation of AHSS is termed as NanoSteel. Branagan et al.
(2018) presented the results of the tensile test of commercially produced NXG 1200 sheets with
a thickness of 1.4 mm. During the uniaxial tensile testing up to failure, the initial structure
undergoes Nanophase Refinement and Strengthening (NR&S), leading to the formation of the
final structure. The NR&S during cold working involves a complex interaction of dislocation
dominated deformation mechanisms with phase transformation, nanoscale phase formation,
nanoprecipitation, and dynamic strain aging effects. The NR&S steel achieves approximately
1200 MPa of tensile strength and 56% of total elongation, which gives 𝑆𝑢 × 𝑒𝑡 = 67.2 GPa%.
15
The current scenario of the 3rd generation of AHSS from the literature resumed hereabove, and
related research is summarized in Table 2.1 and Figure 2.3.
2.1.1. Formability of Q&P 980 steel
In terms of production, the Q&P 980 steel is the thirdgeneration AHSS with the most
advanced development among the other candidates. Wang and Speer (2013) report that in 2009,
the Baosteel company produced the first industrially processed coldrolled Q&P steel sheet,
with tensile strength above 980 MPa and ductility greater than 15%. The works of Li et al.
(2018) and Guo et al. (2018) were conducted using the Q&P 980 steel with a thickness of 1.36
mm provided by the Baoshan Iron & Steel Co., Ltd (Baosteel Group). This Q&P 980 steel has
a formability performance of 28 GPa%, with an elongation superior to 24%. Figure 2.4 shows
a literature compilation of the engineering uniaxial stressstrain curves of the Q&P 980 steel.
1200
1100
1000
900
Engineering stress (MPa)
800
700
600
500
400 Wang and Speer (2013)
300 Chen et al. (2017)
Guo et al. (2018)
200 Li et al. (2018)
Finfrock et al. (2020)
100
0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28
Engineering strain (%)
Figure 2.4 – Engineering stressstrain curves of Q&P 980 steel obtained by several authors.
In the work of Li et al. (2018), the hole expansion (HET) and Erichsen cupping (ECT)
tests were performed in a Q&P 980 steel sheet with 1.36 mm thick. Figure 2.5 shows the results
obtained from the HET. The cracks are detected with partial penetration along the whole edge,
and the macro fracture took place in the rolling direction. The hole expansion ratio that
represents the increase of hole diameter was equal to 28.03%. The punch forcedisplacement
reached about 11.5 kN and 9.5 mm, respectively.
16
12
11
10
9
8
Punch force (kN)
7
6
5
4
3
2 Li et al. (2018)
1
0
0 1 2 3 4 5 6 7 8 9 10
Punch displacement (mm)
(a) (b)
Figure 2.5 – Results of hole expansion test for Q&P 980 steel: (a) punch loaddisplacement
curve and (b) appearance of cracks on the rim of an expanded hole (Li et al. 2018).
The Erichsen cupping test results obtained for the Q&P 980 sheet, shown in Figure 2.6,
the fracture also occurred in the rolling direction, for which the Erichsen index (EI) is the
corresponding punch displacement of ~ 10 mm.
60
50
40
Punch force (kN)
30
20
10 Li et al. (2018)
0
0 1 2 3 4 5 6 7 8 9 10 11
Punch displacement (mm)
(a) (b)
Figure 2.6 – Results of Erichsen cupping test of Q&P 980 steel: (a) punch loaddisplacement
curve and (b) failure location (Li et al. 2018).
Chen et al. (2017) determined the forming limit curve (FLC) for a Q&P 980 steel with
a thickness of 1.2 mm, supplied by the Baosteel, as shown in Figure 2.7. In comparison to a
similar steel grade, DP 980 sheet, Chen et al. (2017) observed that the main gain in formability
occurs in the uniform deformation range. However, in terms of localized deformation, both
17
sheets of steel have similar behavior. They also reported that during the forming limit testing,
the Q&P 980 specimens could flow provided adequate lubrication between the sample, and the
punch is achieved, allowing a uniform deformation. If the lubrication is not sufficient enough,
local deformation will occur instead. In this condition, Q&P 980 steel would perform poorly,
and the high formability level of the Q&P 980 steel might not be reflected.
Major strain, 1
0.4
0.3
0.2
0.1
Chen et al. (2017)
usually applied with a concentration of 4%. The lightest regions represent the martensite phase,
whereas the darkest is the ferrite. However, the SEM images present some issues to reveal the
retained austenite with reasonable accuracy. It is a nontrivial microstructural characterization
that identifies all phases of the Q&P steel, which requires some special preparation techniques,
such as different etching and powerful microscopes. The retained austenite phase and carbide
precipitates could not be distinguished by optical microscopy due to its limited resolution, as
reported by Wu et al. (2013). To reveal the retained austenite, Radwanski et al. (2015) employed
the Klemm and Le Pera etching methods of color metallography. According to the authors,
Klemm’s etchant suggests the presence of small particles of retained austenite that was
confirmed by XRD analysis. On the other hand, Le Pera highlights martensite plus retained
austenite. However, there is no visible distinction between the martensite and retained austenite.
(a) (b)
Figure 2.8 – (a) OM and (b) SEM images of a Q&P 980 steel (Guo et al., 2018).
Finfrock et al. (2020) applied optical microscopy to reveal the Q&P microstructure, in
which lightetching referred to ferrite and darketching martensite and retained austenite,
depicted in Figure 2.9(a). Polarized light shows ferrite as blue and martensite and austenite as
brown, shown in Figure 2.9(b). The authors also utilized the posttreatment of images by
applying color thresholding to account for the percentage of ferrite (Figure 2.9(c)). In this case,
an average of ten images revealed an area of 23.4% ferrite. In the XRD analysis, two types of
peaks are presented: the peaks referring to the BCC crystal structure inherent to the ferrite and
martensite phases and the peaks associated with the FCC crystal structure that corresponds to
the retained austenite phase.
19
(a) (b) (c)
Figure 2.9 – Optical micrographs of a Q&P 980 steel (a) in black and white, (b) photographed
with polarized light, and (c) color thresholding overlay (red), Finfrock et al. (2020).
From the XRD measurements, it is possible to quantify the volume fraction of the
retained austenite. In this way, the three phases constituents of the Q&P 980 steel can be
determined. The study by Guo et al. (2018) evaluated the volume fraction of austenite retained
of a Q&P 980 steel sheet from XRD measurements performed in the asreceived condition and
after applying different strainrates in uniaxial tensile test specimens. The XRD results in
Figure 2.10 indicate that the asreceived Q&P 980 steel has about 11% of retained austenite.
After testing, the retained austenite volume fraction is reduced to values below 4%. This
reduction is related to the retained austenite to martensite transformation phase. It is worth
observing in the study of Guo et al. (2018) that there is no complete transformation of retained
austenite in martensite. Also, the authors do not explain the relationship between the strainrate
and the amount of transformed martensite.
(a) (b)
Figure 2.10 – (a) XRD patterns and (b) correlation between the strain rate and the volume
fraction of retained austenite (Guo et al. 2018).
20
Chen et al. (2017) evaluated the volume fraction of retained austenite of a Q&P 980
steel via XRD, with the longitudinal strain in the uniaxial tensile test, as depicted in Figure 2.11.
The asreceived volume fraction of retained austenite is ~ 11% and decreased very quickly at
the beginning of the uniaxial deformation reaching about 3% close to the uniform elongation.
After the uniform elongation, the Q&P 980 steel microstructural composition became similar
to the DP 980 steel due to the transformation of the retained austenite into martensite. As a
result, QP980 steel has a practically dualphase microstructure composed of martensite and
ferrite, which explains the similar formability of Q&P 980 and DP 980 steels after necking.
12
Chen et al. (2017)
Retained austenite (RA), vol. %
10
0
0.00 0.05 0.10 0.15 0.20 0.25
Longitudinal strain
Figure 2.11 – The volume fraction of retained austenite of a Q&P 980 steel as a function of the
longitudinal tensile strain (Chen et al. 2017).
2.2. Damage behavior with the GTN model
According to Sun et al. (2020), the constitutive model proposed by Gurson (1977) is the
most established damage theory that describes the ductile fracture of metallic materials caused
by the growth of microvoids. Afterward, the Gurson model has been improved by several
researchers. The extension developed by Tvergaard and Needleman (1984), commonly referred
to as the GTN damage model, has achieved wide acceptance by the scientific community.
According to Li and Cui (2020), who wrote a review of the theoretical and practical aspects of
plastic mesodamage mechanics, the main advances in the GTN model are the considerations
of the void shape, strainhardening, and the modification for shear.
21
The adequate use of the GTN model in the simulation of engineering problems depends
on the calibration of the nine constitutive parameters associated with the material (𝑓0 , 𝑞1 , 𝑞2 ,
𝑞3 , 𝜀𝑁 , 𝑆𝑁 , 𝑓𝑁 , 𝑓𝐶 , 𝑓𝐹 ). The complete determination of material parameters is very complicated.
For this reason, some numerical approaches to identify these parameters are generally adopted,
such as response surface methodology or artificial neural networks. Such techniques aim to
quantitatively describe the stressstrain response by minimizing the difference between the
experimental results and the numerical predictions determined with the GTN model. The details
of these calibration techniques are given in the works of Abbasi et al. (2011), Kami et al. (2015),
Gholipour et al. (2019), and Sun et al. (2020).
Some works were devoted to applying the GTN model in the 1st generation of AHSS,
namely, dualphase steels. Cha and Kim (2014) quantitatively determined the microcrack
amount for rollformed products in DP780 steel using the GTN model. Zhao et al. (2016)
applied a modified GTN model to describe shear damage for DP600 steel. Santos et al. (2019)
proposed an experimental identification of the GTN damage model based on void analysis and
finite element simulations of DP600 and DP800 grades. In literature, the GTN model has been
applied to a wide range of metals. Slimane et al. (2015) carried out a parametric study of the
GTN model and the effect of the finite element mesh size. According to the authors, the
refinement of the mesh has very little influence on the material’s elastoplastic domain.
Conversely, it significantly affects the void initiation point. The variation of the parameter 𝑓𝑁
(void volume fraction due to the nucleation) does not have a significant influence on the
material elastoplastic behavior, but it has an essential role from the necking to the fracture. The
authors also tested different combinations of the parameters introduced by Tvergaard (𝑞1 , 𝑞2 ,
𝑞3 ), which exhibited a significant impact on void growth. As noted in Table 2.2, after Tvergaard
and Needleman (1984) established the material parameters (𝑞1 , 𝑞2 , 𝑞3 ), several researchers have
used the value 𝑞1 = 1.5, 𝑞2 = 1, and 𝑞3 = 2.25, for different type of metals. However,
Faleskog et al. (1998) proposed a procedure to identify the values of these parameters as a
function of the mechanical properties, namely, the strainhardening exponent and the ratio
between the yield strength and Young’s modulus. Kossakowski (2012), Wcislik et al. (2016),
and Santos et al. (2019) adopted the procedure proposed by Faleskog et al. (1998).
22
Table 2.2 – GTN model parameters available in the literature. Adapted from Gholipour et al.
(2019), and Yildiz and Yilmaz (2020)
Ref. Material 𝑓0 𝑞1 𝑞2 𝑞3 𝜀𝑁 𝑆𝑁 𝑓𝑁 𝑓𝐶 𝑓𝐹
Tvergaard (1984) 1.50 1.00 2.25 0.300 0.100 0.04000 0.1500 0.2500
Schmitt et al. (1997) 20MnMoNi55 1.50 1.00 2.25 0.300 0.100 0.00200 0.0600 0.2120
Hambli (2001) Carbon steel 1.50 1.00 2.25 0.300 0.100 0.04000
Rachik et al. (2002) Steel DD13 1.50 1.00 2.25 0.300 0.100 0.04000 0.1000 0.1010
Springmann et al. (2005) Steel 0.00100 1.50 1.00 2.25 0.300 0.100 0.01000 0.0100 0.1500
Lemiale et al. (2009) Mild steel XES 1.50 1.00 2.25 0.200 0.100 0.04000 0.1500 0.2500
Marouani et al. (2009) FeSi (3wt.%) steel 1.50 1.00 2.25 0.300 0.100 0.04000 0.1100 0.1200
Chen and Dong (2009) AA6111T4 1.50 1.00 2.25 0.300 0.100 0.04000 0.1500
Kossakowski et al. (2012) S235JR steel 0.00170 1.91 0.79 3.65 0.300 0.050 0.04000 0.0600 0.6000
Abbasi et al. (2012) IFsteel 0.00020 1.50 1.00 2.25 0.100 0.100 0.01060 0.0134 0.0216
Abbasi et al. (2013) Mild steel 1.50 1.00 2.25 0.300 0.220 0.04000 0.0800 0.1200
Yan et al. (2013) Silicon steel 0.00250 1.55 0.90 0.240 0.010 0.04000 0.1010 0.1550
Kiran et al. (2014) ASTM A992 1.50 1.00 2.25 0.450 0.050 0.02000 0.0300 0.5000
Achouri et al. (2014) HSLA 0.00150 1.20 0.80 1.44 0.200 0.100 0.02000 0.0800 0.1300
Zhou et al. (2014) Zircaloy4 1.50 1.00 2.25 0.100 0.020 0.01200 0.0300 0.0800
Mansouri et al. (2014) Mild Steel 0.00100 1.50 1.00 2.15 0.210 0.100 0.03900 0.0601
Cha and Kim (2014) DP780 0.00300 1.50 1.00 2.25 0.102 0.206 0.01800 0.0130 0.2070
Kami et al. (2015) AA6016T4 0.00035 1.50 1.00 2.25 0.300 0.100 0.05000 0.0500 0.1500
Slimane et al. (2015) Csteel A48AP 0.00001 1.50 1.00 2.25 0.300 0.100 0.00400
Gatea et al. (2015) Pure Ti 0.00138 1.50 1.00 2.25 0.300 0.100 0.01700 0.2593 0.3025
Li et al. (2015) DC06 steel 0.00100 1.50 1.00 2.25 0.300 0.100 0.00400 0.0400 0.0600
Sun et al. (2015) AA2219T6 0.00328 1.50 1.25 2.25 0.120 0.300 0.01000 0.0110 0.0150
Balan et al. (2015) DP steel 0.00002 1.50 1.00 2.25 0.0002 0.2300
Zhao et al. (2016) DP600 0.00080 1.50 1.00 2.25 0.200 0.100 0.02000 0.0280 0.0900
Jiang et al. (2016) Al 2024 T3 0.00400 1.50 1.00 2.25 0.200 0.100 0.02000 0.0250 0.1500
Zhong et al. (2016) 3Cr1MoV 0.61 0.56 0.37 0.124 0.083 0.17900 0.2400 0.2610
Chalal (2017) AA6016T4 0.00035 1.50 1.00 2.25 0.300 0.100 0.05000 0.0500 0.1500
Teng et al. (2017) 5A06O Al alloy 0.00120 1.50 1.00 2.25 0.250 0.080 0.03200 0.0340 0.0420
Liu et al. (2018) 22MnB5 steel 0.00200 1.50 1.00 2.25 0.300 0.100 0.01550 0.0500 0.1300
Nguyen et al. (2018) AA6061T6 0.00160 1.32 2.58 0.115 0.054 0.05200 0.0870 0.1400
Gholipour et al. (2019) SAE 1010 0.00107 1.50 1.00 2.25 0.300 0.100 0.00716 0.0100 0.1500
Santos et al. (2019) DP600 0.00500 1.86 0.80 3.46 0.090 0.030 0.01040 0.0168 0.0180
Santos et al. (2019) DP800 0.00700 1.76 0.83 3.09 0.085 0.200 0.04450 0.0239 0.0260
Sun et al. (2020) Silicon steel 0.00200 1.50 1.00 2.25 0.344 0.100 0.21700 0.2340 0.3510
Yildiz and Yilmaz (2020) 6061 Al alloy 0.00025 0.224 0.070 0.01285 0.0349 0.0629
Maire et al. (2008) applied the insitu 3D Xray diffraction microtomography technique
in DP600 steel to observe the influence of nucleation and the growth of voids in the damage
process during the uniaxial tensile test. Figure 2.12(a) shows the 2D and 3D microtomography
just before the necking and the fracture, and the DP600 steel microvoids population. Later,
Requena et al. (2014) also used the Xray diffraction microtomography to analyze the
contributions of the nucleation mechanism and the void growth on the damage of DP980 steel.
Figure 2.12(b) shows the results from the separation procedure of the existing voids before
deformation, growth, and nucleation. Cao et al. (2014) also used Xray microtomography
combined with a macro mechanical test to identify the GTN model parameters and characterize
the ductile damage of high carbon steel. Figure 2.12(c) exhibits the 3D views of the damage
evolution in the uniaxial tensile specimen center at different strain levels (Cao et al., 2014).
Balan et al. (2015) proposed a damage evolution law for dualphase steels based on
Gursontype models. For this purpose, Balan et al. (2015) performed an insitu void volume
fraction measurement using Xray diffraction microtomography during the uniaxial tensile test.
23
In the GTN model, the void volume fraction due to the nucleation follows a standard normal
distribution, and the parameter 𝑓𝑁 (nucleated void volume fraction) is defined by the maximum
value of this cumulative normal distribution. Balan et al. (2015) observed from experiments
that the nucleation phenomenon has its monotonically increased intensity for dualphase steel.
Thus, it does not reach a specific level described by the GTN model parameter 𝑓𝑁 .
(a)
(c)
(d)
Figure 2.12 – 3D Xray microtomography applied to void volume fraction determination for
different references in the literature (a) Maire et al. (2008), Requena et al. (2014), and (c) Cao
et al. (2014).
In work carried out by Samei et al. (2016), the influence of deformation modes on
nucleation and void growth in DP600 steel 1.5 mm thick was analyzed. Specimens were tested
using the Marciniak methodology for uniaxial tension, plane strain, and biaxial stretching
deformation modes. The samples were deformed at different strain levels to observe the voids
in the material’s microstructure. To avoid deformation inherent in the sample cutting process,
the authors adopted a wire EDM cutting procedure. The samples were polished with a diamond
paste of 3 µm and with alumina of 1 µm and 0.05 µm and, then, etched with Nital 2% for 30 s.
The micrographs were analyzed with a scanning electron microscope (SEM) model FEI Quanta
24
200 FEG. In Figure 2.13, the phases of ferrite (F), martensite (M), microvoids (V), and
inclusions (I) are shown. Figure 2.13(a) presents the voids related to the microcracks at the
interfaces of the martensite islands for the uniaxial tension loading. Figure 2.13(b) exhibits the
nucleated voids generated in the inclusion interface resulting from the planestrain deformation
mode.
(a) (b)
Figure 2.13 – SEM images highlighting voids, inclusions, ferrite, and martensite phases for
the (a) uniaxial tension, and (b) plane strain deformation modes, Samei et al. (2016).
Regarding inclusions, Santos et al. (2019) observed for dual phase steels, namely DP600
and DP800, that with the increase in plastic strain, the inclusions serve as sites for nucleation
and formation of microvoids. The authors identified the inclusions employing energy
dispersive Xray spectroscopy (EDS) analysis (Figure 2.14(a)). They verified through SEM
images the microvoids produced by interface decohesion between the ferritic matrix and
inclusion for DP800 steel at 18% of plastic strain.
(a) (b)
Figure 2.14 – (a) EDS analysis of the inclusion in DP800 steel and (b) SEM image with
elongated void areas formed around the inclusions in uniaxial tension at a strain level of 18%.
25
For the quantitative analysis of the microvoids, Samei et al. (2016) polished the sample
with a diamond paste of 3 µm and with alumina of 1 µm and 0.05 µm. Then, they applied Nital
10% for 5 s to remove the polishing layer without revealing the microstructure. The porosities
were observed using an optical microscope model Leiz Labourlux 12 ME. Figure 2.15 shows
the binarized images of different strain levels, where the black spots represent the voids.
Figure 2.15 – Voids in the DP600 steel microstructure resulting from deformation modes under
(ab) uniaxial tension, (cd) plane strain, and (ef) biaxial tension specimens. After Samei et al.
(2016).
From the binarized images, Samei et al. (2016) analyzed the deformed samples in terms
of void density, void area fraction, void aspect ratio, and mean void size. Figure 2.16(a) and
Figure 2.16(b) show that the void density and the void area fraction under planestrain is greater
than the uniaxial tension condition. For the same strain level, the difference reaches about 15%.
Figure 2.16(b) indicates that the void growth rate increases significantly with the increase in
the stress triaxiality factor η, which is defined by the ratio between the equivalent stress and the
hydrostatic stress. Assuming the isotropic von Mises yield criterion, the void area fraction for
the biaxial stretching mode (Σ = 0.67) is about five times greater than the uniaxial tension mode
(Σ = 0.34). In Figure 2.16(c), the biaxial stretching showed less variation between the length
and width ratio of the voids. Samei et al. (2016) explained that spherical voids have a lower
level of stress concentration when compared to the ellipsoidal voids. The damage evolves more
slowly under biaxial stretching conditions, delaying the onset of necking as observed by the
mean void size plotted in Figure 2.16(d).
26
200 0.012
180 Uniaxial Tension Uniaxial Tension
Plane Strain 0.010 Plane Strain
Void density [number/mm²]
Void Area (%)
120
100 0.006
80
0.004
60
40 0.002
20
0 0.000
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35
Major strain Major strain
(a) (b)
2.0 2.4
1.9 Uniaxial Tension Uniaxial Tension
Plane Strain 2.0 Plane Strain
1.8 Biaxial Tension Biaxial Tension
1.7 Mean Void Size (m)
Void Aspect Ratio
1.6
1.6
1.5 1.2
1.4
0.8
1.3
1.2
0.4
1.1
1.0 0.0
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35
Major strain Major strain
(c) (d)
Figure 2.16 – Microvoid quantities determined for the DP600 steel sheet deformed under
uniaxial tension, planestrain, and biaxial stretching modes: (a) void density, (b) void area
fraction, (c) void aspect ratio, and (d) mean void size. Adapted from Samei et al. (2016).
Saeidi et al. (2014) investigated the void nucleation mechanism of the dualphase steel
DP780. Saeidi et al. (2014) adopted three different specimen geometries to perform the uniaxial
tensile testing, one without a notch and two with notches using different radii (1.5 and 7.5 mm).
The tests were conducted at room temperature with a nominal crosshead speed of 0.03 mm/s.
The fractured samples were cut using a precision cutting machine. The SEM images were
examined to estimate the number of voids per unit area as a function of the imposed plastic
strain, shown in Figure 2.17. Figure 2.18(a) shows the void area fraction as a function of the
true thickness strain for DP780, and Figure 2.18(b) the corresponding void density for the same
strain levels considering the three specimens geometries.
27
Figure 2.17 – Schematic representation of the specimens cutting method used by
Saeidi et al. (2014).
Firstly, the stress triaxiality factor increases in specimens with a decreasing notch radius.
Then, the sample without a notch can be understood as having a notch of an infinite radius.
Specimens with low triaxiality have the voids evolution faster than the specimens with high
triaxiality (Saeidi et al., 2014). Moreover, the detailed SEM observation of the microstructure
showed that regardless of the stress state, the void nucleation mechanism is the same in smooth
and notched DP780 steel sheet specimens. The nucleation primarily occurs by fragmentation
of the interface decohesion of martensite particles, mainly at the narrow ferrite regions between
two adjacent martensite islands.
2.0 50
Void Area Density per m x 1000
1.4
1.2 30
1.0
0.8 20
0.6
0.4 10
0.2
0.0 0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.0 0.1 0.2 0.3 0.4 0.5 0.6
True thickness strain True Thickness strain
(a) (b)
Figure 2.18 – (a) Void area fraction and (b) void density as a function of the true thickness
strain for the DP780 steel, adapted from Saeidi et al. (2014).
Castanheira et al. (2019) analyzed the damage behavior of coldrolled TRIP780 steel,
prone to the straininduced martensite transformation (SIMT), through void formation in the
uniaxial tensile test. The authors adopted a similar procedure for quantifying microvoids used
28
by Samei et al. (2016). However, the analysis was performed both in longitudinal and transverse
directions, as shown in Figure 2.19(a). Unlike dualphase steels, TRIP steels present phase
transformation during deformation with an increase in the martensite fraction, as shown in
Figure 2.19(b). The results of void density and void area fraction for the parallel and transverse
crosssections are plotted in Figure 2.19(c) and (d), respectively.
100
Volume fraction of transformed martensite (%)
90
80
70
60
50
40
30
20
10
0
0.00 0.05 0.10 0.15 0.20
Truestrain
(a) (b)
Void Area Fraction
Void Area Fraction
8 8
(%) x 100
(%) x 100
6 6
4 4
2 2
Transverse crosssection Parallel crosssection
(Void number/m ) x 1000
0 60
(Void number/m ) x 1000
0 60
Void Density
45 45
2
Void Density
2
30 30
15 15
0 0
0.00 0.05 0.10 0.15 0.20
0.00 0.05 0.10 0.15 0.20
True strain
True strain
(c) (d)
Figure 2.19 – (a) Specimen cutting sections for microvoid analysis, (b) volume fraction of
straininduced martensite as a function of the longitudinal truestrain in uniaxial tension, (c)
void density and void area fraction in the parallel, and (d) transverse crosssections. From
Castanheira et al. (2019).
Concerning the parameter of void volume fraction due the nucleation (𝑓𝑁 ), Cingara et
al. (2009) proposed quantify the nucleated void area fraction consider the voids size. First, the
authors quantified the void area fraction of DP600 steel as a function of the thickness strain,
Figure 2.20(a). Then, the voids were classified into two groups, recently nucleated voids (those
< 1µm) and grown voids (>1 µm). For the lower size range curve, the intercept along the
thickness strain axis occurs at a thickness strain value of 0.15 (or an equivalent strain of 0.3),
as shown in Figure 2.20(b). Voids were found to initiate at a low level of straining (ɛ ≤ 0.15).
The void nucleation rate is expected to be a function of many parameters such as the volume
29
fraction, size and distribution of martensite, and hydrostatic stress level. Cingara et al. (2009)
found from Figure 2.20(b) an average area fraction of nucleated voids of 8 × 10−4 /unit
thickness strain, and the void area fraction due to exponential void growth. Cingara et al. (2009)
conclude that the significant void nucleation mechanism is decohesion at the ferrite/martensite
interface. Void nucleation generally occurs at the interfaces perpendicular to the tensile axis,
and those voids grow along the ferrite grain boundaries. A small quantity of voids is nucleated
due to the fracture of the martensite phase, with nominal growth. A small number of large size
voids is nucleated at the inclusions.
1.8 0.016
1.2
0.010
1.0
0.008
0.8
0.006
0.6
As-received DP600 0.004
0.4
0.002
0.2
0.000
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Thickness Strain Thickness Strain
(a) (b)
Figure 2.20 – DP600 steel (a) evolution of the void area fraction with thickness strain for tensile
specimen, and (b) variation of the void area fraction with the thickness strain for two categories
of void size. After Cingara et al. (2009).
To determine the parameter 𝑓𝐶 (critical void volume fraction) Zhao et al. (2016) carried
out a tensile test until the necking to appear. The morphology of microvoids was observed
through SEM first and then the value of 𝑓𝐶 was obtained from the statistics of the fraction of
microvoids. The morphology of microvoids near the fracture surface is shown in Figure 2.21.
The authors used the same method to obtain the parameter 𝑓𝐹 (void volume fraction at fracture)
from the rupture specimen. The values of 𝑓𝐶 and 𝑓𝐹 of DP600 steel were estimated as 0.028 and
0.09, respectively. In the research developed by Wcislik (2016), the value of the void volume
fraction in the fracture, 𝑓𝐹 , was estimated from the tensile test and SEM analysis, for steel
identified as S355J2G3. The author adopted cylindrical specimens with a diameter of 12 mm,
with a central notch with a radius of 10 mm. The thinnest diameter of the sample was 8 mm,
see Figure 2.22(a). According to the sample geometry, the calculated stress triaxiality factor
was equal to 0.516, which was considered low by the author. The tensile tests were carried out
30
until the fracture (Figure 2.22(b)), then the fractured surface was examined in the SEM model
QUANTA FEG 250 with a magnification of 1,000x.
Figure 2.21 –Morphology of voids at the crack initiation zone for DP600, Zhao et al. (2016).
Figure 2.22(c) shows the image of the fractured surface of the specimen, while
Figure 2.22(d) presents the same image of the fractured surface; however, after applying the
image binarization. A total of 46 microscopic images were examined. However, the author
explains that it is worth to note that the estimated 𝑓𝐹 value refers to a surface fraction of voids
and not a volumetric fraction of voids.
(a) (b)
(c) (d)
Figure 2.22 –(a) Specimen with the ring notch during a tensile test, (b) fracture surface of
specimen (c) microscopic photograph of the fracture surface, and (d) surface fracture
photograph after binarization, Wcislik (2016).
31
Chapter 3
Experimental Procedure
32
3. EXPERIMENTAL PROCEDURES
To meet the confidentiality requirements, the sheet of the 3rd generation of AHSS
analyzed in this work is referred to as Gen3 980T steel. The material was received as an
uncoated coldrolled sheet with a nominal thickness of 1.58 mm.
3.1. Microstructural characterization of Gen3 980T steel
Figure 3.1 presents the micrographs of the Gen3 980T steel obtained with the SEM
Hitachi model SU70. The microstructural characterization was carried out after standard
metallographic sample procedures. It was applied a mechanical grinding, conventional
polishing, and electropolishing using a Struers A2 electrolyte. The process was completed by
electrolytic polishing and etching using the Struers LectroPol5 automatic equipment operated
at 35 V.
RA
F
MA
(a) (b)
Figure 3.1 – SEM micrographs of the Gen3 980T steel sheet with (a) 1,000x, and (b) 10,000x
highlighting the ferrite (F), martensiteaustenite (MA), and retained austenite (RA).
33
From the comparison of the microstructural analysis conducted by Finfrock et al. (2020)
for the Q&P steels and the martensitic structure characterization performed by NavarroLopez
et al. (2017), the dark regions are ferrite (F), the regions exhibiting fine internal laths surrounded
by white blocks are martensiteaustenite islands (MA), and the elongated white regions are
retained austenite (RA), as shown in Figure 3.1(b).
In order to determine the volume fraction of retained austenite, the methodology
described in the ASTM E97513 standard was adopted. The ASTM standard establishes a
procedure to estimate the amount of retained austenite from XRay diffractograms. The method
is applied to carbon and alloy steels with a nearrandom crystallographic orientation of both
phases, and the retained austenite contents must be from 1% by volume and above. The retained
austenite volume fraction is calculated as:
𝐼𝛾
⁄𝑅
𝛾
RA% = 𝑉𝛾 = (3.1)
𝐼𝛼 𝐼𝛾
⁄𝑅 + ⁄𝑅
𝛼 𝛾
where the index γ and α represent the austenite (γiron: facecentered cubic) and ferrite plus
martensite (αiron: bodycentered cubic), respectively. The parameter 𝐼 is the integrated
intensity per angular diffraction peak, and 𝑅 is a parameter that depends upon interplanar
spacing (hkl), the Bragg angle θ, the crystal structure, and composition of the phase being
measured. 𝑅𝛼 and 𝑅𝛾 can be calculated from basic principles (ASTM E97513, 2013), and
Table 3.1 presents its calculated values for Cu radiation.
Table 3.1 – Calculated theoretical 𝑅𝛼 and 𝑅𝛾 using Cu radiation (Jatczak, 1980).
hkl 2θ R
α 110 44.6 233.8
α 200 65.0 31.9
α 211 82.2 60.9
α 220 98.9 20.6
γ 111 43.6 182.8
γ 200 50.8 81.6
γ 220 74.6 44.4
γ 311 90.6 51.3
34
The XRay diffraction (XRD) analysis was conducted in the Rigaku SmartLab XRay
Diffractometer with a Cu tube, and Kalpha radiation operated at 40 kV and 30 mA using a
scanning range from 40º to 110º with a scan speed of 2º/min. The Xray diffractogram of
Gen3 980T steel for the asreceived condition is displayed in Figure 3.2, in which the peaks are
identified as αiron or γiron.
20000
()
Gen3 980T steel
Intensity (au)
15000
10000
()
()
()
()
5000
()
()
()
0
40 45 50 55 60 65 70 75 80 85 90 95 100 105 110
2 (deg)
Figure 3.2 – XRay diffractogram patterns of Gen3 980T steel sheet determined from the as
received condition.
Table 3.2 exhibits the integrated intensity per angular diffraction peak, that is, the peak
area above the background. Each peak was fitted with the Lorentz fit peak tool available on
OringLab. From the Equation (3.1) and Table 3.1 and Table 3.2 contents, the estimated volume
fraction of the retained austenite for the Gen3 980T steel sheet was equal to 12.2% (Table 3.3),
which is very close to the RA presented for Q&P 980 steel, see Table 2.1.
Table 3.2Integrated intensity per angular diffraction peak from Lorentz fitting peak (𝐼𝛼ℎ𝑘𝑙 , 𝐼𝛾ℎ𝑘𝑙 ).
2θ 43.6 44.6 50.8 65.0 74.6 82.2 90.6 98.9
Peak γ(111) α(110) γ(200) α(200) γ(220) α(211) γ(311) α(220)
Integrated
2079.47 6100.57 105.83 1615.14 214.08 2643.91 159.17 573.00
intensity
Table 3.3 – The volume fraction of RA of Gen3 980T steel.
True strain RA%
as received
12.2
condition
35
3.2. Uniaxial tensile test
The uniaxial tensile tests and the stressstrain curves are widely used to provide design
information on the strength of metallic materials. The engineering stressstrain curve is defined
from the loadelongation measurements usually obtained from standard specimens. Figure 3.3
exemplifies a uniaxial tensile specimen for flat products, where 𝐿0 represents the original gauge
length, 𝑊0 the original width, and where 𝐿0 represents the initial gauge length, 𝑊0 the initial
width, and the original sheet thickness is 𝑡0 . The engineering stress is obtained by dividing the
measured load, 𝐹, by the initial crosssection area of the uniaxial tensile specimen, 𝐴0 , that is,
𝑆 = 𝐹 ⁄𝐴0 . The engineering strain is the ratio between the elongation of the specimen gauge
length, ∆𝐿, and its original length, 𝐿0 .
Figure 3.3 – Nomenclature of the main dimensions of the uniaxial tensile test specimen.
On the other hand, the true stressstrain curve considers the instantaneous variation of
the specimen crosssection area. The truestress and the truestrain measures can be defined as
a function of the engineering stress and strain values, as:
𝜎 = 𝑆(𝑒 + 1) (3.2)
𝜀 = ln (𝑒 + 1) (3.3)
Equations (3.2) and (3.3) are valid only in the uniform strain domain, with the
assumption of volume constancy and a homogeneous distribution of strain along the gage length
of the specimen.
36
From the uniaxial tensile test, it is possible to determine important mechanical properties
for understanding the material behavior, such as Young’s modulus, Poisson’s ratio, yield stress,
ultimate tensile strength, uniform elongation, and total elongation. The Young’s modulus and
the Poisson’s ratio are associated with the material elastic domain. Young’s modulus, also
known as the modulus of elasticity, is a property that measures the material stiffness. The higher
Young’s modulus, there will be a smaller amount of elastic deformation resulting from the
application of a given load.
Figure 3.4 – Young’s modulus determination from a uniaxial tensile test.
For most metals, the stressstrain relationship in the elastic regime can be considered as
linear, with the elasticity modulus being interpreted mathematically as the slope of the line,
schematically shown in Figure 3.4, and defined as:
𝑆2 − 𝑆1 𝑆
𝐸= | = (3.4)
𝑒2 − 𝑒1 𝑒𝑙𝑎𝑠𝑡𝑖𝑐 𝑒
Figure 3.5 shows the volume elements of the specimen gage region in the original
unloaded configuration (1) and after and elastic strain resulted from a given external load (2).
The Poisson’s ratio is defined as the ratio between the width and length strains as:
|𝑒𝑊 | 𝑒𝑊
𝜈= =− (3.5)
𝑒𝐿 𝑒𝐿
where,
37
∆𝑊 ∆𝐿
𝑒𝑊 = ; 𝑒𝐿 = (3.6)
𝑊0 𝐿0
Figure 3.5 – The Poisson’s effect during a uniaxial tensile loading.
For metal with a discontinuous yielding, the yield stress is defined using the offset
method. The yield stress value is determined by the intersection of the engineering stressstrain
curve with a straight line parallel to the elastic portion offset with a predefined small strain,
usually 0.2% (𝑒 = 0.002), as schematically shown in Figure 3.6, and defined by:
𝐹(𝑒=0.002)
𝑆𝑦 = (3.7)
𝐴0
Figure 3.6 – Engineering stressstrain curve showing the definitions of yield strength, ultimate
tensile strength, uniform elongation, and total elongation.
38
Figure 3.6 also exhibits the engineering ultimate tensile strength, which is defined as the
maximum measured force during the test divided by the specimen original crosssection area:
𝐹𝑚𝑎𝑥
𝑆𝑢 = (3.8)
𝐴0
The uniform strain, 𝑒𝑢 , is defined as the engineering strain corresponding to the
maximum force, while the total elongation, 𝑒𝑡 , is the measure of engineering strain up to the
specimen fracture.
3.2.1. Preparation of the uniaxial tensile test
In this work, the uniaxial tensile tests were performed by employing the universal testing
machine Shimadzu AGX of 100 kN. The specimen elongation was measured using the system
ARAMIS with the GOM digital image correlation (DIC). Figure 3.7 presents the adopted dog
bone specimen dimensions. The reduced section of the specimen has 75 mm. However, the
gauge length analyzed was the central region with a length of 50 mm.
Figure 3.7 Geometry and dimensions [mm] of the uniaxial tensile test specimen.
From sheet strips cut with the aid of a guillotine, the specimens were manufactured by
CNC milling machining. Before testing, the width of the gauge length and sheet thickness were
measured using a digital caliper with 0.01 mm precision. The burrs from the machining were
removed with fine sandpaper, and then the samples were cleaned with acetone.
According to the adopted DIC system, it is required to paint a high contrast stochastic
pattern over the surface of the gauge length to calculate the strain fields. Firstly, matte white
aerosol ink was used to prepare the background. After approximately 5 to 10 minutes, a matte
black ink was carefully sprayed over the white background. The drying time of the black ink
was around 2 minutes. The tests were carried out right after the painting process, not exceeding
39
30 minutes of curing. After complete cure, the ink loses elasticity and detaches from the surface
during the deformation. The painted speckle pattern was verified utilizing the spray pattern
reference supplied by the DIC program. Figure 3.8 shows the adopted spray pattern reference
for the uniaxial tensile testing.
Figure 3.8 Spray pattern reference with 100 mm × 80 mm (ARAMIS, 2009).
3.2.2. Mechanical properties in the rolling direction
The mechanical properties of the Gen3 980T steel in the sheet rolling direction (RD)
were obtained through the uniaxial tensile tests using a constant crosshead speed of 4.5 mm/min.
Five specimens were tested and Figure 3.9(a) presents the average of the engineering stress
strain curve, while Figure 3.9(b) shows the corresponding true stressstrain curve.
1000 1200
Engineering stress (MPa)
1000
800
True stress (MPa)
800
600
600
400
400
200
200
0
0 2 4 6 8 10 12 14 16 18 20 22 24 0
0 2 4 6 8 10 12 14 16 18
Engineering strain (%)
True strain (%)
(a) (b)
Figure 3.9 – (a) Engineering stressstrain curve and (b) true stressstrain curve of Gen3 980T
steel sheet in the RD.
40
Table 3.4 lists the determined values of the Young’s modulus (𝐸), Poisson’s ratio (𝜈),
yield strength (𝑆𝑦 ), ultimate tensile strength (𝑆𝑢 ), uniform elongation (𝑒𝑢 ), total elongation (𝑒𝑡 ),
and the corresponding standard deviations.
Table 3.4– Mechanical properties of Gen3 980T steel sheet in the rolling direction.
𝐸 [MPa] 𝜈 𝑆𝑦 [MPa] 𝑆𝑢 [MPa] 𝑒𝑢 [%] 𝑒𝑡 [%]
192 ± 5 0.289 ± 0.003 604 ± 7 1040 ± 9 18.0 ± 0.5 23.4 ± 0.2
The excellent mechanical properties of Gen3 980T steel, mainly in terms of ultimate
tensile strength and total elongation, allow its proper positioning in the global formability
diagram within the current thirdgeneration AHSS, as shown in Figure 3.10.
𝑆𝑢 × 𝑒𝑡 = 24.3 GPa% (3.9)
80
5 10 15 20 25 30 35 40 45 50 55 60 65 70 (GPa%)
70
LIP 2nd GEN
60
Total Elongation (%)
AUST.SS TWIP
50 IF
IFHS
40 Mild ISO
70
BH Gen3 980T steel
30
CMn
Current
TRIP
20 3rd GEN
HSLA
DP, CP
10 MART
MART
1st GEN 5
0
0 200 400 600 800 1000 1200 1400 1600 1800
Ultimate Tensile Strength (MPa)
Figure 3.10 – Position of Gen3 980T steel in the global formability diagram.
3.2.3. Parameters of anisotropy
For sheet metals, the plastic anisotropy is evaluated by the Rvalue, which is also known
as the Lankford coefficient. The Rvalue is defined as the ratio between the plasticstrain rates
41
in length and thickness directions of the sheet. Assuming a proportional loading, as in the
uniaxial tensile test, the corresponding plasticstrains are obtained by integrating the plastic
strain rates. The thickness strain can be measured in the specimen, but it is subjected to errors
owing to the surface roughness changes. Also, the correct thickness measurements would
require the specimen unloading to consider only the plasticstraining. The sheet thickness strain
is usually obtained by firstly assuming that the plastic yielding is independent of the hydrostatic
pressure and, therefore, the material volume is kept constant during the plastic straining. In this
case, both undeformed and deformed gauge regions of the specimen have the same volume,
that is:
𝑉 = 𝑉0 ∴ 𝑊 ∙ 𝐿 ∙ 𝑡 = 𝑊0 ∙ 𝐿0 ∙ 𝑡0 (3.10)
or else
𝑡 𝑊0 𝐿0
= (3.11)
𝑡0 𝑊𝐿
and, therefore, the Rvalue is defined by:
𝑊 𝑊0 𝐿0 𝜀𝑊 𝜀𝑊
𝑅 = 𝑙𝑛 ( )⁄𝑙𝑛 ( )= =− (3.12)
𝑊0 𝑊𝐿 𝜀𝑡 (𝜀𝑊 + 𝜀𝐿 )
The suffix 𝛼 indicates the orientation with respect to the rolling direction in which the
Rvalue is defined. Usually, three angular orientations are adopted to evaluate Rvalue, namely,
the rolling, transverse and diagonal directions providing the 𝑅0 , 𝑅45 , and 𝑅90 values,
respectively. Equation (3.13) and (3.14) define the parameters known as planar anisotropy and
normal anisotropy, respectively. Both parameters depend on the values of 𝑅0 , 𝑅45 , and 𝑅90 .
𝑅0 + 𝑅90 − 2𝑅45
∆𝑅 = (3.13)
2
𝑅0 + 2𝑅45 + 𝑅90
𝑅̅ = (3.14)
4
42
If the Rvalues are different, the sheet is said to have a planar anisotropy (∆𝑅). The ∆𝑅
can be positive or negative, although, for steels, it is usually positive. The phenomenon known
as earring in the cup drawing is related to the planar anisotropy. If ∆𝑅 > 0, there are ears at 0º
and 90º, and if ∆𝑅 < 0, ears form at 45º. The closer the values of 𝑅0 , 𝑅45 , and 𝑅90 are, the
smaller the earring amplitude will be. Earing is undesirable because the walls must be trimmed,
creating scrap. If the normal anisotropy ratio 𝑅̅ is greater than unity it indicates that in the tensile
test the width strain is greater than the thickness strain. This is related to a greater strength in
the throughthickness direction and, generally, a resistance to thinning. Also, in drawing deep
parts, a high value of 𝑅̅ allows deeper parts to be drawn and reduce the chance of wrinkling.
The Gen3 980T steel anisotropic behavior was evaluated from the uniaxial tensile tests
performed at seven different angular orientations with respect to the rolling direction (RD),
namely, 0°, 15°, 30°, 45°, 60°,75°, and 90° as shown in Figure 3.11. Five specimens were tested
for each orientation. The uniaxial tensile tests were performed up to specimen fracture, at room
temperature (20°C), under a constant crosshead speed of 4.5 mm/min corresponding to a
nominal strain rate of 103 s1. The strain measurements were recorded with an image acquisition
rate of 1 frame/s.
Figure 3.11 – Angular orientations of the specimens with respect to the rolling direction.
Table 3.5 presents for each selected specimen direction, the respective mechanical
properties: yield strength, ultimate tensile strength, uniform elongation, total elongation, as well
the Lankford’s parameter. The Rvalue was determined by continuously measuring the change
in the width of the specimens during the uniaxial tensile tests, using digital image correlation.
The Rvalue was calculated as the slope of the trend line of the relationship between the width
strain (𝜀𝑊 ) and the thickness strain (𝜀𝑡 ), see Equation (3.12).
43
Table 3.5– Mechanical properties of Gen3 980T steel as a function of the angular orientation.
Orientation
𝑆𝑦 [MPa] 𝑆𝑢 [MPa] 𝑒𝑢 [%] 𝑒𝑡 [%] Rvalue
(º)
0 604 ± 7 1040 ± 9 18.0 ± 0.5 23.4 ± 0.2 0.861 ± 0.003
15 619 ± 6 1006 ± 6 18.4 ± 0.9 24.5 ± 0.9 0.880 ± 0.002
30 623 ± 2 995 ± 5 18.8 ± 0.6 24.5 ± 1.0 0.926 ± 0.002
45 643 ± 4 1015 ± 8 17.9 ± 0.2 23.0 ± 0.6 0.957 ± 0.004
60 654 ± 3 1013 ± 2 18.1 ± 0.5 23.5 ± 0.9 0.945 ± 0.002
75 660 ± 3 1026 ± 2 17.8 ± 0.5 23.1 ± 1.0 0.912 ± 0.002
90 668 ± 7 1023 ± 8 16.9 ± 0.6 21.9 ± 1.3 0.895 ± 0.005
Table 3.6 display the values of planar and normal anisotropy of Gen3 980T steel sheet,
which were calculated by applying the Equations (3.13) and (3.14), respectively. As the value
found for ∆𝑅 is close to zero and 𝑅̅ is also close to the unity, it is reasonable to consider the
plastic behavior of the Gen3 980T steel as isotropic for some applications. This assumption can
be verified through Figure 3.12(a) that shows a comparison between the engineering stress
strain curve for 0º, 45º, and 90º according to the rolling direction. The three stressstrain curves
present relatively close plastic behavior. Also, the listed Rvalues in Table 3.5 were plotted as
a function of orientation angle with respect to the sheet rolling direction, as shown in
Figure 3.12(b). This graph shows that the Gen3 980T steel tends to present small amplitudes of
earing, which is desirable characteristic in forming processes.
Table 3.6– Planar and normal anisotropy of Gen3 980T steel.
∆𝑅 𝑅̅
Gen3 980T steel 0.079 0.917
1.00
1000 0.98
Lankford parameter (Rvalue)
0.96
Engineering stress (MPa)
800
0.94
0.92
600
R
0.90
0.88
400
0.86
0º RD
200 45º RD 0.84
90º RD 0.82 Gen3 980T steel
0 0.80
0 2 4 6 8 10 12 14 16 18 20 22 24 0 15 30 45 60 75 90
Engineering strain (%) (degrees)
(a) (b)
Figure 3.12 – (a) Comparison between engineering stressstrain curves for 0º, 45º, and 90º RD,
and (b) Lankford coefficient of the Gen3 980T steel as a function of orientation angle with
respect to the rolling direction.
44
Figure 3.13 displays the yield stress and ultimate tensile strength values normalized by
the corresponding data obtained in the rolling direction. The yield stress behavior presents a
monotonic increase between 0º and 90º of about 11%. On the other hand, the ultimate tensile
strength varies less than 5%.
1.15
Normalized yield strength
Normalized ultimate tensile strength
1.10
Normalized stress
1.05
1.00
0.95
0.90
0 15 30 45 60 75 90
Tensile direction (º)
Figure 3.13 – Normalized yield stress and ultimate tensile strength as a function of the
orientation angle with respect to the rolling direction.
The average uniform and total elongation values and corresponding standard deviations
are depicted in Figure 3.14 as a function of the specimen angular orientation with respect to the
sheet rolling direction (RD). A small reduction between the rolling and transverse orientations
are observed for both uniform and total elongation values. However, the amount of the average
engineering strain during the necking shows little variation, as indicated by the shaded area.
Therefore, the total elongation reduction is associated with a decrease in the uniform elongation.
28
Uniform elongation
Total elongation
Engineering strain (%)
24
Necking
20
16
0 15 30 45 60 75 90
Tensile direction (º)
Figure 3.14 – Uniform and total elongation as a function of the orientation angle with respect
to the rolling direction.
45
3.2.4. Workhardening behavior
In this work, the plastic behavior of the Gen3 980T steel sheet is described by the
Hollomon, Ludwik, Swift, and Voce workhardening equations with respect the sheet rolling
direction. The Hollomon workhardening equation is defined as:
𝜎 = 𝐾(𝜀 𝑝 ) 𝑛 (3.15)
where 𝐾 is the strength coefficient [MPa], and 𝑛 is the strainhardening exponent.
Figure 3.15 presents the comparison between the experimental truestress and true plasticstrain
curve and the predictions obtained with Equation (3.15). It is possible to observe that the
Hollomon equation does not accurately describe the beginning of the plastic yielding. Table 3.7
lists the workhardening parameters determined for the Hollomon’s equation from using the
nonlinear fitting in the Origin (OriginLab, 2017) software and the resulting Rsquare, which the
closer to the unity the better the fit accuracy.
1300
1200
1100
1000
900
True stress (MPa)
800
700
600
500
400
300
200 Experimental
100 Hollomon
0
0 2 4 6 8 10 12 14 16 18
True plastic strain (%)
Figure 3.15 – Experimental true stress and true plasticstrain data of the Gen3 980T steel sheet
at the rolling direction and Hollomon’s predicted workhardening curve.
Table 3.7– Parameters of Hollomon’s workhardening equation.
The Ludwik workhardening equation parameters are defined by the yield stress, 𝜎𝑦 ,
which describe the initial plastic yielding, the strength coefficient, 𝐾, and the strainhardening
exponent, 𝑛, as:
𝜎 = 𝜎𝑦 + 𝐾(𝜀 𝑝 ) 𝑛 (3.16)
For the Gen3 980T steel, the Ludwik equation showed a better fit than the Hollomon
workhardening equation. However, as displayed in Figure 3.16, the fitted curve still presents
points slightly distant from the experimental values. Table 3.8 lists the parameters obtained for
the Ludwik equation.
1300
1200
1100
1000
900
True stress (MPa)
800
700
600
500
400
300
200 Experimental
100 Ludwik
0
0 2 4 6 8 10 12 14 16 18
True plastic strain (%)
Figure 3.16 – Experimental true stress vs. true plastic strain and the curve fitted by the
Ludwik hardening equation.
Table 3.8– Parameters of Ludwik’s workhardening equation.
𝜎 = 𝐾(𝜀0 + 𝜀 𝑝 )𝑛 (3.17)
The fitted curve with Swift’s equation showed better accuracy than that provided by
Ludwik. Also, Ludwik and Swift equations presented similar Rsquared values for the data
range analyzed. Nevertheless, both workhardening equations overpredicted the experimental
stress curve in strain range between ~ 12 and 16%. Thus, these workhardening equations may
produce significant divergences in extrapolation to larger plasticstrains values. The parameters
obtained for Swift’s workhardening equation are given in Table 3.9.
Table 3.9 – Parameters of Swift’s workhardening equation.
1300
1200
1100
1000
900
True stress (MPa)
800
700
600
500
400
300
200 Experimental
Swift
100
0
0 2 4 6 8 10 12 14 16 18
True plastic strain (%)
Figure 3.17 – Experimental true stress and true plasticstrain data of the Gen3 980T steel sheet
at the rolling direction and Swift’s predicted workhardening curve.
The hardening equation that best suited the experimental values analyzed was the Voce
equation, which is described below:
𝑝
𝜎 = 𝜎𝑦+ 𝐾(1 − e−𝑛𝜀 ) (3.18)
48
Figure 3.18 displays the curve generated from Equation (3.18), which was able to fit
with good precision, both at the beginning and at the end of the plastic strain level. This
precision can be confirmed by the Rsquare value very close to 1 (0.9996) presented in
Table 3.10, together with the material constants of the Voce equation.
1300
1200
1100
1000
900
True stress (MPa)
800
700
600
500
400
300
200 Experimental
Voce
100
0
0 2 4 6 8 10 12 14 16 18
True plastic strain (%)
Figure 3.18 – Experimental true stress and true plasticstrain data of the Gen3 980T steel sheet
at the rolling direction and Voce’s predicted workhardening curve.
Table 3.10 – Parameters of Voce’s workhardening equation.
Equation 𝜎𝑦 [MPa] 𝐾 [MPa] 𝑛 Rsquare
𝑝
𝜎 = [𝜎𝑦 + 𝐾(1 − e−𝑛𝜀 )](𝜀̇ 𝑝 ⁄𝜀̇0 )𝑚 Modified Voce
{ (3.19)
𝜎 = [𝐾(𝜀0 + 𝜀 𝑝 )𝑛 ](𝜀̇𝑝 ⁄𝜀̇0 )𝑚 Modified Swift
where 𝑚 is the strainrate sensitivity index and 𝜀̇0 is the reference strainrate.
The strainrate sensitivity is calculated from the uniaxial tensile tests performed at two
strainrates (𝜀̇1 > 𝜀̇2 ) using the corresponding truestress data, namely:
ln (𝜎2 /𝜎1 )
𝑚= (3.20)
ln (𝜀̇1 /𝜀̇2 )
The engineering stressstrain curves obtained from the nominal strainrates of 103 and
102 s1 are plotted in Figure 3.19.
1200
1000
True stress (MPa)
800
600
400
3 1
200 10 s (4.5 mm/min)
2 1
10 s ( 45 mm/min)
0
0 2 4 6 8 10 12 14 16 18 20
Strain strain (%)
Figure 3.19 – The true stressstrain curves obtained with strain rates of 103 and 102s1.
Figure 3.20 presents the 𝑚value as a function of the true plasticstrain and its average
value, which corresponds to 0.0098. Table 3.11 lists the parameter values of the modified Voce
and Swift workhardening equations. The predicted curves computed with the fitted parameters
presented a good agreement with the experimental data, as shown in Figure 3.21.
Table 3.11 – Parameters of the modified workhardening equations.
Equation 𝜎𝑦 [MPa] 𝐾 [MPa] 𝜀0 𝑛 𝑚 𝜀̇0 (s1)
0.016
0.014
Strainrate sensitivity
0.012
0.010
Average mvalue = 0.0098
0.008
0.006
0.004
0.002
0 2 4 6 8 10 12 14
True plastic strain (%)
Figure 3.20 – Calculated and average strainrate sensitivity parameter 𝑚 of the Gen3 980T steel
sheet plotted as a function of the uniaxial tensile true plasticstrain.
1300
1200
1100
True stress (MPa)
1000
900
3 1
Exp. 10 s (4.5 mm/min)
2 1
800 Exp. 10 s ( 45 mm/min)
3 1
Voce 10 s
700 2 1
Voce 10 s
3 1
600 Swift 10 s
2 1
Swift 10 s
500
0 2 4 6 8 10 12 14
True plastic strain (%)
Figure 3.21 – Experimental and fitted truestress and true plasticstrain along the rolling
direction as a function of the strainrate.
3.3. Vickers hardness test
The Vickers hardness uses a squarebase diamond pyramid as the indenter. The angle
between the opposite faces of the pyramid is equal to 136º. The Vickers hardness (HV) is
defined as the ratios between the measured load and the indentation surface area. In practice,
51
this area is calculated from measurements of the lengths of the diagonals of the indentation.
The HV may be determined as:
2𝑃 sin(𝜗⁄2) 1.854 ∙ 𝑃
HV = = (3.21)
𝑑2 𝑑2
where 𝑃 is the applied load, 𝑑 corresponds to the average length of diagonals, and 𝜗 is the angle
between opposite faces of the pyramid. The sample preparation consisted of cutting a piece of
the sheet with ~ 15 mm x 15 mm using a precision cutter. The measurements were performed
on the sheet plane by employing the Shimadzu HMV 2000 machine with a load of 0.2 kgf and
dwell time of 15 s. Five measurements were performed in the asreceived material. Table 3.12
presents the Vickers hardness results (HV) determined for the Gen3 980T steel sheet. The
measured hardness value for Gen3 980T steel was 304 HV, which is compatible with the
hardness values reported in the literature for 980 grade thirdgeneration steels, namely between
293 HV (Liu et al., 2019) and 305 HV (Zhou et al., 2019).
Table 3.12 – Vickers hardness of the Gen3 980T steel sheet.
𝑃: 0.2 kgf; Dwell time: 15 s
Nº of indentation 𝑑1 (mm) 𝑑2 (mm) 𝑑 (mm) HV
1 0.0351 0.0347 0.0349 304
(a) (b)
Figure 3.22 – (a) Sheet strip cut by wire electrical discharge machining and (b) disc compression
sample dimensions.
The compression tests were conducted at room temperature with a constant crosshead
speed in compression of 0.09 mm/min. The disc compression test was controlled according to
the load level being stopped at the maximum value prescribed, ranging from 25 kN to 50 kN.
Three specimens were tested for each compression load. Both top and bottom disc surfaces in
contact with the cylindrical flathead punches were well lubricated, applying graphite grease,
as shown in Figure 3.23.
Figure 3.23 – Universal testing machine in compression mode and sample lubrication with
graphite grease (in detail).
According to Barlat et al. (2003), the biaxial anisotropy coefficient 𝑟𝑏 refers to the ratio
between the true plasticstrain in the transverse direction, that is, at 90o to the rolling direction,
𝑝 𝑝
𝜀𝑇𝐷 , and true plasticstrain parallel to rolling direction, 𝜀𝑅𝐷 , namely:
𝑝
𝜀𝑇𝐷
𝑟𝑏 = 𝑝 (3.22)
𝜀𝑅𝐷
53
To obtain the true plasticstrains, each disc specimen was photographed before and after
the compression, see Figure 3.24, using a scale with 0.1 mm accuracy. The measurements were
done with the opensource software ImageJ.
(a) (b)
Figure 3.24 – Measurements of the disc compression specimens of the Gen3 980T steel sheet:
(a) undeformed and (b) tested with a maximum load of 50 kN.
In Table 3.13 are listed for each tested disc: the maximum compression load, the initial
𝑖 𝑖
longitudinal diameter (𝐷0° ), the initial transverse diameter (𝐷90° ), the initial disc thickness (𝑡𝑖 ),
𝑓 𝑓
the final longitudinal diameter (𝐷0° ), the final transverse diameter (𝐷90° ), the final thickness
𝑝 𝑝
(𝑡𝑓 ), the longitudinal true plasticstrain (𝜀𝑅𝐷 ), the transverse true plasticstrain (𝜀𝑇𝐷 ) and the
thickness true plasticstrain (𝜀𝑡𝑝 ). The sum of the true plasticstrains ∑ 𝜀𝛼 = 𝜀𝑅𝐷
𝑝 𝑝
+ 𝜀𝑇𝐷 + 𝜀𝑡𝑝
presented a residual value, close to the assumption of volume.
Table 3.13 – Disc compression test results for the Gen3 980T steel sheet.
Force Undeformed values [mm] Deformed values [mm] Calculated true plasticstrains
Test 𝑖 𝑖 𝑡𝑖 𝑓 𝑓 𝑡𝑓 𝑝 𝑝 𝑝
[kN] 𝐷0° 𝐷90° 𝐷0° 𝐷90° 𝜀𝑅𝐷 𝜀𝑇𝐷 𝜀𝑡 𝛴𝜀𝛼
1 6.01 6.01 1.58 6.09 6.07 1.55 0.013 0.010 0.019 0.004
25 2 6.01 6.01 1.58 6.08 6.04 1.55 0.012 0.005 0.019 0.003
3 6.01 6.01 1.58 6.06 6.04 1.55 0.008 0.005 0.019 0.006
1 6.01 6.01 1.58 6.17 6.13 1.51 0.026 0.020 0.045 0.001
30 2 6.01 6.01 1.58 6.17 6.14 1.51 0.026 0.021 0.045 0.002
3 6.01 6.01 1.58 6.17 6.14 1.51 0.026 0.021 0.045 0.002
1 6.01 6.01 1.58 6.30 6.26 1.45 0.047 0.041 0.086 0.002
35 2 6.01 6.01 1.58 6.28 6.26 1.45 0.044 0.041 0.086 0.001
3 6.01 6.01 1.58 6.29 6.24 1.45 0.046 0.038 0.086 0.003
1 6.00 6.00 1.58 6.43 6.39 1.39 0.069 0.063 0.128 0.004
40 2 6.01 6.01 1.58 6.45 6.39 1.39 0.071 0.061 0.128 0.004
3 5.99 5.99 1.58 6.41 6.38 1.39 0.068 0.063 0.128 0.003
1 6.01 6.01 1.58 6.60 6.55 1.33 0.094 0.086 0.172 0.007
45 2 6.01 6.01 1.58 6.61 6.55 1.33 0.095 0.086 0.172 0.009
3 6.01 6.01 1.58 6.61 6.55 1.33 0.095 0.086 0.172 0.009
1 6.01 6.01 1.58 6.81 6.76 1.25 0.125 0.118 0.234 0.008
50 2 6.01 6.01 1.58 6.78 6.74 1.26 0.121 0.115 0.226 0.009
3 6.01 6.01 1.58 6.80 6.75 1.26 0.123 0.116 0.226 0.013
54
The 𝑟𝑏 value is defined as the slope of the linear fit performed on the plot of the true
plasticstrain data determined at the transverse and parallel orientations to the rolling direction,
see Figure 3.25. Table 3.14 list the linear fit results, which provide 𝑟𝑏 = 0.918 ± 0.010. The more
pronounced is the anisotropy, the farther is the biaxial anisotropy coefficient from unity. As in
this case, the Gen3 980T steel sheet provide a 𝑟𝑏 value close to one, the assumption of isotropic
plastic behavior is also reasonable in the biaxial deformation conditions.
Table 3.14 – Biaxial anisotropy coefficient (𝑟𝑏 ) value of the Gen3 980T steel sheet.
Equation 𝑟𝑏 𝑅square
𝑝 𝑝
𝜀𝑇𝐷 = 𝑟𝑏 𝜀𝑅𝐷 0.918 ± 0.010 0.998
0.14
p p
TD = rb RD 50 kN
0.12
0.10
Transverse strain (TD )
45 kN
p
0.08
40 kN
0.06
35 kN
0.04
30 kN
0.02
25 kN
0.00
0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14
p
Longitudinal strain (RD)
Figure 3.25 – Transverse strain vs. longitudinal strain in the disc compression test with graphite
grease lubricant.
3.5. Hole expansion test
The hole expansion test (HET) consists of expanding a hole by a punch with a conical
tip through a die, aiming to form a flange. The ISO 16630 standard establishes a method of
determining the hole expansion ratio (HER) in metallic sheets and strips with a thickness range
of 1.2 to 6 mm and a width of at least 90 mm. The stretchflangeability of sheet metal can be
evaluated by the HER and is calculated as per ISO 16630 (2009):
55
𝐷ℎ − 𝐷𝑂
𝐻𝐸𝑅(%) = × 100 (3.23)
𝐷𝑂
where 𝐷ℎ is the average hole diameter after the specimen rupture and 𝐷𝑂 is the initial hole
diameter. The HET stop criterion is defined when a crack is visible across the entire thickness.
Then, the ruptured hole must be measured with slide calipers in two directions perpendicular to
each other, located away from the cracks.
3.5.1. Preparation of HET specimens
According to ISO 16630 standard, the hole expanding test consists of two steps:
punching a hole and forcing a conical expanding tool into a prepunched hole until any one
crack extends through the test piece thickness of the sheet. For steel sheets with a thickness of
1.58 mm, the diameter of the punch for cutting must be 10 mm and the die 10.4 mm. The holes
were produced by punching in the central region of a 95 mm × 95 mm square strip.
Figure 3.26(a) shows the specimen dimensions.
ollover
Burnish
Test piece one
Fracture
one
Burr
(a) (b)
Figure 3.26 – (a) Specimen dimensions in [mm] for HET and (b) hole sheared edge produced
by the punching process.
The punching process creates four characteristics regions into the hole sheared edge,
namely, the rollover, the burnish zone, the fractured zone, and the burr, shown in Figure 3.26(b).
An ideal cutting would produce only the burnish zone, without rollover, fracture zone, and burr.
However, in the actual punching process, all regions are commonly observed due to the elastic
and plastic strains (Wang et al., 2012). In this way, during the HET, the conical punch must
contact the sheet surface in the rollover zone side instead of the burr.
56
3.5.2. Development of a HET device
A lowcost device capable of performing the HET and compatible with the universal
testing machines has been developed in this work (Santos et al., 2020). The main dimensions
of the device are displayed in Figure 3.27. With an approximated volume of 186 mm × 130 mm
× 500 mm and a maximum punch stroke of 50 mm, this device can be easily adapted on the
bench of a universal testing machine. The material used to manufacture the device structure
was the AISI 1045 steel. The punch was made from AISI D2 steel (quenched and tempered) to
satisfy the minimum hardness of 55 HRC required by the ISO 16630 standard.
(a) (b)
Figure 3.27 – (a) Schematic setup of the device for HET and (b) detailed view of HET parts.
During the test, the operator needs to observe the deformed surface of the sheet material,
aiming to stop the test when the crack occurs. For this purpose, a mirror was installed with an
angular orientation of 45º concerning the punch axis to reflect the region of interest
(Bandyopadhyay et al., 2016). Thus, it is possible to observe and record the test in realtime
using a video camera.
A holder was used to prevent any material drawin from the clamping area during the
test. The clamping force was imposed employing eight M8 bolts class 8.8. The crisscross
pattern was used as a bolt tightening sequence, and the torque applied in each bolt was equal to
10 Nm. The sum of the clamping force produced by the bolts should be around 50 kN. The
punch has a permanent magnet enclosed at the top of the tool. This magnet allows the punch to
57
be linked to the test machine. Thus, it is possible to reset the load cell considering the tool
weight in the load cell presets. To reduce the friction effects during the punch displacement, a
cylindrical bush PAP 2030 P20 was mounted at the holder. Additionally, a thin layer of
lubricant, commonly grease, is applied between the tool contact surface and the specimen.
In this work, the universal testing machine Shimadzu model AGX with 100 kN load
capacity was employed to perform the HET on Gen3 980T steel sheet specimens. The digital
camera was used only to visualize the stretching process. The test was performed with a
constant crosshead speed of 1 mm/min, and a commercially available grease (Liqui Moly LM
47 longlife grease with MoS2 content) was used as the lubricant between the sheet and the
punch.
3.5.3. Hole sheared edge
To analyze the sheared hole edges, the punched specimens were cut with a precision
cutter. The sliced samples were firstly mounted into epoxy resin and then mechanically
prepared using SiC abrasive paper of 180, 240, 600, 800, 1200, and 2400 in a Struers TegraPol
2 machine. The sheared edges were analyzed in the Scanning Electron Microscope (SEM)
Hitachi TM400Plus. Figure 3.28 shows the typical sheared edge profile resulting from the
punching process. Figure 3.28(a) displays the rollover, burnish, fracture, and burr dimensions
measured in the rolling direction, while Figure 3.28(b) exhibits these zones in the perpendicular
orientation to the rolling direction.
(a) (b)
Figure 3.28 – Hole sheared edge at (a) 0º and (b) 90º, according to the sheet rolling direction.
58
It is worth observing, see Figure 3.28, a longer burnish zone and a greater burr zone
with a higher angle in the fractured region in the rolling direction. The transverse section to the
rolling direction showed less distortion in the rollover and burr zones. Conversely, the fractured
zone in this direction is more extensive than in the rolling direction.
3.5.4. Hole Expansion Ratio (HER)
Figure 3.29 shows the punch displacement and the corresponding reaction force curves
obtained from three specimens evaluated in the proposed HET device. A small plateau is visible
in about 5 mm of punch displacement and 9 kN reaction force. This small plateau may be related
to the beginning of crack formation at the hole's outer edge. As the sheet has good strength, the
punch force starts to increase again, and the hole continues to expand until a macrocrack
appears across the full thickness, around 6.5 mm and 10.5 kN.
12
11
10
9
8
Punch force (kN)
7
6
5
4
3
Specimen #1
2 Specimen #2
1 Specimen #3
0
0 1 2 3 4 5 6 7
Punch displacement (mm)
Figure 3.29 – Punch reaction force curves determined from the HET of the Gen3 980T steel.
Figure 3.30 shows the picture taken from the deformed Specimen #3. The corresponding
final diameter measurements were made with the help of the Leica EZ4W stereomicroscope.
Red circles highlight the regions with macro fractures, wherein the cracks were detected across
the entire specimen thickness. As reported by Xu et al. (2012) for a highMn TWIP steel sheet,
the experimental results of the HET for the Gen3 980T steel also showed that cracking occurred
along the rolling direction. Xu et al. (2012) explain that both the driving force for shearing for
59
cracking and the decohesion from the matrix of stringers containing high Mn elongated along
the rolling direction contribute to the observed fracture behavior. As Mn is also applied in the
thirdgeneration of AHSS to stabilize retained austenite, even at lower levels, a similar effect is
suggested to occur in Gen3 980T steel.
Fracture through
the full thickness
Figure 3.30 – Final hole diameter with a fracture through the full thickness.
Table 3.15 resumes the maximum punch force and displacement values, and the hole
expansion ratio obtained for the Gen3 980T steel sheet with 1.58 thick. The average HER value
obtained was 10.9%. A large difference was observed by comparing with the HER of 28.03
obtained for 1.36 mm thick Q&P 980 steel tested by Li et al. (2018). The HER value found for
Gen3 980T steel was about 60% less than the value obtained by Li et al. (2018), as shown in
Figure 2.5(b). In addition to the difference in thickness that influences the stiffness of the
sample during the HET test, other factors can contribute to the results so distinct for two third
generation AHSS with 980 MPa grade, for example, the quality of the sheared edge or the stop
criterion of the HET.
Table 3.15 – Hole expansion test results determined for the Gen3 980T steel sheet.
Maximum punch Maximum punch
Specimen 𝐷𝑂 [mm] 𝐷ℎ [mm] HER [%]
force [kN] displacement [mm]
3.6. Erichsen cupping test
The Erichsen Cupping Test (ECT) is aimed to reproduce stretching conditions on thin
sheets by a small hemispherical punch against a clamped specimen placed between a die and a
blank holder. The punch stretches the specimen until a throughthickness failure will be visible.
The Erichsen’s test formability measure is defined by the corresponding punch displacement,
known as the Erichsen cupping index or else IE, as per ISO 20482 standard.
Figure 3.31 displays the specimen dimensions used in the ECT of the Gen3 980T steel.
For the ECT specimen preparation, it is merely necessary to cut squared strips with a guillotine.
The ECT and the HET procedures are qualitative and comparative sheet metal formability tests.
However, IE values are usually associated with mechanical sheet properties (Sangkharat and
Dechjarem, 2019).
Figure 3.31 – Specimen dimensions in [mm] adopted for the Erichsen cupping test.
The Erichsen cupping tests were performed with the same device developed for the HET
but with different tooling geometry. Figure 3.32(a) shows the ECT setup mounted in the
universal testing machine Shimadzu model AGX with 100 kN, equipped with a digital camera
and a lighting system. The detailed view Erichsen test tooling is depicted in Figure 3.32(b),
which agrees with the ISO 20482 standard. The difference in the setup shown in Figure 3.27(b)
is the die and holder needed for the ECT and the replacement of the conical punch by a
hemispherical tip. The ECT specimen was clamped using eight bolts that confer a load force of
50 kN, and the tightening sequence was the crisscross pattern with 10 Nm of torque. The contact
between the punch and the sheet was lubricated with grease. The crosshead speed adopted was
61
also 1 mm/min. The stopping ECT criterion is set by an abrupt drop of the reaction force. Also,
three specimens were tested for obtaining the IE value of the Gen3 980T steel.
(a) (b)
Figure 3.32 – (a) Universal testing machine ECT setup and (b) detailed view of the tooling.
Figure 3.33 shows the specimen fractures are located along the rolling direction (RD).
During the tests, it was also observed that the crack spreads quickly and produces a loud noise.
#1 #2 #3
Figure 3.33 – Deformed specimens and fracture locations obtained from Erichsen cupping tests
of the Gen3 980T steel sheet.
The specimens were cut in half employing a precision cutter to understand better the
instability in the Erichsen cupping test of the Gen3 980T sheets. Then, they were mounted into
epoxy resin and mechanically prepared with SiC abrasive paper. The deformed specimens were
analyzed in the SEM Hitachi TM400Plus. The fracture occurred about 23 mm from the center
of the dome. The measured profiles of the inner and outer surfaces of a deformed specimen are
plotted in Figure 3.34. Close to the fracture site, the specimen thickness is decreased by 53.8%,
62
while in the center of the dome, the reduction was 46.2%, which corresponds to a thickness of
0.85 mm.
16
outer surface
t = 0.85 mm
Dome height (mm) 12 inner surface
4 t = 0.73 mm
0
16 12 8 4 0 4 8 12 16
Distance from the center (mm)
Figure 3.34 – Thickness profile of the Gen3 980T steel sheet after Erichsen cupping test carried
out until the fracture.
Figure 3.35 shows the experimental results of punch reaction load as a function of the
punch displacement, the latter corresponding to the Erichsen cupping index. The ECT
specimens results in Figure 3.33 are also summarized in Table 3.16. For the Gen3 980T steel,
the average ECT maximum punch force is 63 kN with an average IE of 10.6 mm. Comparing
the results obtained for Gen3 980T steel with Q&P 980 steel tested by Li et al. (2018), in terms
of ECT the results were reasonably close. According to Figure 2.6(a), Li et al. (2018) obtained
a punch force of 56 kN and an Erichsen Index of 9.9 mm. In addition, the appearance of the
fracture was also similar, see Figure 2.6(b).
70
60
50
Punch force (kN)
40
30
20
Specimen #1
10 Specimen #2
Specimen #3
0
0 1 2 3 4 5 6 7 8 9 10 11 12
Punch displacement (mm)
Figure 3.35 – Punch displacement versus punch force reaction during the ECT.
63
Table 3.16 – Erichsen cupping test results determined for the Gen3 980T steel.
Specimen Maximum punch force (kN) IE (mm)
#1 62.61 10.36
#2 63.43 10.87
#3 63;37 10.64
(a) (b)
Figure 3.36 – An element of sheet: (a) the undeformed state with the circle and square grids and
(b) the deformed state with the grid circles deformed to ellipses of major diameter 𝑑1 and minor
diameter 𝑑2 (Marciniak et al, 2002).
In this case, the principal true strains (𝜀1 , 𝜀2 ) can be calculated from the measurements
of the initial (𝑑0 ) and final values (𝑑1 , 𝑑2 ) of the printed grids that is:
𝑑1
𝜀1 = 𝑙𝑛 (3.24)
𝑑0
64
𝑑2
𝜀2 = 𝑙𝑛 (3.25)
𝑑0
On the other hand, the thickness strain can be determined from measurements made with
the aid of a micrometer or else by neglecting the elastic strains assuming the plastic
incompressibility condition, as:
𝑡
𝜀3 = 𝑙𝑛 = −(1 + 𝜌)𝜀1 (3.26)
𝑡0
The strainratio introduce in Eq. (27) is defined by
𝜀2
𝜌= (3.27)
𝜀1
and assumed to be linear and constant in the deformation process.
From the inplane true principal strains, defined by Eqs. (3.24) and (3.25), it is possible
to represent a yield strain locus as a function of the strainratio 𝜌 = 𝜀2 ⁄𝜀1 , as shown in
Figure 3.37. In this plot, the major principal truestrains are placed on the vertical axis, whereas
the principal minor truestrains are put on the horizontal axis. The circle grid changes its shape
̅̅̅̅ strainpath in Figure 3.37
for different deformation modes or else strainratio values. The 𝑂𝐴
̅̅̅̅ path describes the plane strain
describes the equibiaxial stretching, for which 𝜌 = 1. The 𝑂𝐵
condition, where the sheet deforms in only one direction, and the circle becomes an ellipse in
̅̅̅̅ path refers to uniaxial tension mode with the
which the minor axis does not change. The 𝑂𝐶
minor principal truestress 𝜎2 = 0, wherein for an isotropic sheet material 𝜌 = −1/2. In the
̅̅̅̅ path, the inplane principal truestrains are equal and opposite and, thus, the sheet deforms
𝑂𝐷
without thickness change. This strain mode is known as drawing, as it is observed when a sheet
is drawn into a converging region. It is also referred to as a pure shear stressstate and occurs
in the flange of a deepdrawing cup. Lastly, the uniaxial compression defined by the strainpath
̅̅̅̅ , 𝜌 = −2 for an isotropic sheet material (Marciniak et al, 2002).
𝑂𝐸
65
Figure 3.37 – Deformation modes defined by the strainratio in the principal strains space.
Adapted from Marciniak et al. (2002).
The concept of the Forming Limit Diagram (FLD) was introduced by Gensamer in 1946
(Banabic et al., 2021). However, the way in which the FLD is currently presented was proposed
by Keeler in 1965, first for positive values of the principal major strain (𝜀2 > 0) and then, in
1968, it was extended by Goodwin for strains located in the drawing range (𝜀2 < 0). This
diagram is defined by the major and minor truestrains obtained in the plane of the sheet. The
curve which is defined by linear strainpaths, namely, for fixed strainratios (𝜌 = 𝜀2 ⁄𝜀1 = 𝐶),
is known as the Forming Limit Curve (FLC). The FLC is widely used to evaluate the formability
behavior of metallic sheets and tooling design to solve manufacturing problems in tryout steps.
The FLC can be defined depending on the necking or fracture condition, where the localized
necking usually characterizes the state preceding the fracture failure (Freitas et al., 2013).
Figure 3.38 shows that a typical FLC is defined as a function of different deformation
modes for the isotropic case. For 𝜀2 > 0 the FLC domain is the biaxial stretching, whereas for
𝜀2 < 0 the region is drawing, and 𝜀2 = 0 is the planestrain intercept, also known as FLC0. The
FLC is a diagram in which the failure occurs above the limiting curve as, for instance, by the
necking process. Below the curve, the sheet metal forming process can be performed in safety
conditions.
According to Butuc (2004), despite the simplicity of the FLD concept, its experimental
determination is somewhat not trivial. A wide range of tests between uniaxial tension and
equibiaxial stretching is required. There are several experimental procedures proposed to
66
determine the FLC. For instance, uniaxial tension tests using samples with varied geometries,
Nakajima, and Marciniak procedures in which the specimens with varying width are deformed
using a hemispherical punch (Nakajima) or a flatbottom cylindrical punch (Marciniak) and
also bulge testing techniques where blanks are submitted to a hydraulic pressure through
circular or elliptical dies (Banabic, 2010).
Figure 3.38 – A typical forming limit diagram for the isotropic case (Kumar et al., 2016).
In this work, the limitstrains of the Gen3 980T steel sheet were evaluated by three
deformation modes composing the main regions of the Forming Limit Curve (FLC), namely,
uniaxial tension, planestrain tension, and equibiaxial stretching. The ISO 120042:2008
standard was adopted to define the major and minor principal limit strains.
3.7.1. Uniaxial tension
In the uniaxial tension deformation tests, the strainfields were obtained using the
ARAMIS digital image correlation system with a stochastic painting on the specimen surface.
Figure 3.39 shows the major and minor true totalstrain fields in the specimen gauge length,
obtained before the fracture. According to the A AMIS user’s manual (2009), the limit strains
are calculated from at least three sections (0, 1, and 2) equallyspaced with 2 mm.
67
(a) (b)
Figure 3.39 – DIC results obtained from the uniaxial tension specimen: (a) major and (b) minor
strains.
The ARAMIS software in the FLC module automatically calculates the respective major
and minor strain values using the ISO 120042:2008 standard. For each intersection line,
depicted in Figure 3.39, the true strains are then plotted as a function of the length position to
discard the points exposed to the localized necking. The procedure is done by first calculating
the local second derivatives from a parabola fitting of each major and minor true strains located
at the fracture's left and right sides. A fitting curve is adopted to determine the maximum and
minimum values, indicated in Figure 3.40 for section 1.
The uniaxial tensile tests were performed up to specimen fracture at room temperature.
Using a constant crosshead speed of 4.5 mm/min in the universal testing machine Shimadzu
AGX (load capacity of 100 kN). The specimens were manufactured with seven different
angular orientations concerning to the rolling direction, namely, 0°, 15°, 30°, 45°, 60°,75°, and
90°. In this manner, determining the limit strains was repeated five times for each specimen
orientation. The average limit strain values (𝜀2 , 𝜀1 ) and corresponding standard deviations for
each analyzed specimen orientation are listed in Table 3.17.
68
Table 3.17 – Minor and major true strains in uniaxial tension for the Gen3 980T steel.
Orientation
Minor strain, 𝜀2 Major strain, 𝜀1
(degrees)
0 0.148 ± 0.007 0.324 ± 0.020
15 0.151 ± 0.011 0.325 ± 0.027
30 0.162 ± 0.012 0.339 ± 0.026
45 0.168 ± 0.007 0.348 ± 0.017
60 0.166 ± 0.010 0.345 ± 0.023
75 0.158 ± 0.014 0.335 ± 0.031
90 0.149 ± 0.011 0.320 ± 0.028
Figure 3.40 – Major and minor truestrains distributions obtained along section 1.
3.7.2. Plane strain
The planestrain tension tests were also performed using the universal testing machine
Shimadzu AGX with 100 kN load capacity. The limit strains were obtained with the ARAMIS
DIC system. Two angular orientations were tested, 0º, and 90º according to the rolling direction.
Three tests for each angular orientation were performed at room temperature using a constant
crosshead speed of 0.1 mm/min up to specimen fracture. Figure 3.41 shows the double notched
specimen geometry adopted in this work to reproduce a planestrain tension state condition
(Schwindt et al., 2015).
69
Figure 3.41 – Plane strain specimen geometry and dimensions [mm], Schwindt et al. (2015).
According to ISO 120042:2008 standard, the limit strains were evaluated from the
vertical sections shown in Figure 3.42. To verify if the planestrain state condition is satisfied,
the major and minor strains were also measured in the horizontal section representing the
specimen's width notched crosssection.
Figure 3.42 – Horizontal and vertical sections used to measure the major and minor strains in
the planestrain specimen.
The strain fields obtained from the central region of the double notched are depicted in
Figure 3.43. The values refer to the instant immediately before the fracture of the specimen.
For the specimens aligned with the rolling direction, the average values are close to 15%, and
1% were obtained for major and minor strains, respectively. For the specimens at 90o to the
rolling direction, the measured average major and minor strains were 13% and 1%, respectively.
The planestrain tension limit strains and their respective standard deviations are listed in
Table 3.18.
70
(a) (b)
Figure 3.43 – Selected strain fields to define the: (a) major and (b) minor strains in the double
notched planestrain specimens.
Through Figure 3.44(a) and (b), it is possible to observe that the geometry of the sample
does provide a condition very close to the plane strain, but in a very small central region of
about 3 mm x 3 mm. In practical terms, another difficulty in obtaining the limit strains of the
plane strain with the notched specimen is that the notch has a small radius where the stress is
concentrated. The fracture begins to propagate from the notches' radii towards the center of the
sample, which is differently observed in the uniaxial tensile test. Therefore, the instant before
the fracture's beginning may indicate a much earlier stage for the limit condition of plane strain
in the specimen's center. On the other hand, when the fracture starts, there is a change in the
deformation mode.
0.18 0.45
Section 0 Major strain
0.16 0.40
Section 1 Minor strain
0.14 Section 2 0.35
Major strain Section 0 0.30
0.12
Section 1 0.25
True strain
True strain
0.10 Section 2
0.20
0.08
0.15
0.06
0.10
0.04
0.05
0.02 Minor strain 0.00
0.00 0.05
0.02 0.10
0 1 2 3 4 5 6 7 8 9 10 11 0 1 2 3 4 5 6 7 8 9
Deformed vertical sections [mm] Horizontal section [mm]
(a) (b)
Figure 3.44 – Major strain and minor strain distributions determined from the (a) vertical and
(b) horizontal sections for a specimen in the rolling direction.
71
Table 3.18 – Minor and major limit strains inplane strain tension for Gen3 980T steel.
Orientation
Minor strain, 𝜀2 Major strain, 𝜀1
(degrees)
0 0.0109±0.0026 0.1525±0.0062
Specimen
Die
Holder
Punch
Figure 3.45 – Schematic of the tooling used in the Nakajima tests. Dimensions in mm.
In the Nakajima testing procedure, blank specimens with different width values are
employed to vary the strainratio between the equibiaxial stretching and uniaxial tension
deformation modes. Figure 3.46(a) shows the set of specimens used to obtain the FLC's left
side. The specimen with W=25 mm corresponds to the uniaxial tension deformation mode,
whereas W=75 mm is for the plane strain. The specimens adopted to obtain the limit strains on
the FLC's right side are depicted in Figure 3.46(b), in particular, the squared specimen with
W=175 mm, which has been chosen to reproduce the equibiaxial stretching condition. Again,
the limit strains were determined as per ISO 120042 standard.
72
(a)
(b)
Figure 3.46 – Blank specimen geometries used to reproduce: (a) the left side of FLC and the
(b) right side of FLC from the Nakajima testing procedure.
Due to the larger contact area of the FLC right side specimens, it was necessary to apply
a higher blankholder load to assure stretching conditions avoiding either the slip or drawin of
the blank. The Gen3 980T steel has an outstanding mechanical strength, which for a thickness
sheet gauge of 1.58 mm reached the load limit of the testing machine operation. Consequently,
the limit strains of the Gen3 980T steel sheet were defined only for the left side of the FLC.
Table 3.19 resumes major and minor strains obtained from the specimens of the lefthand side
of the FLC. These limit strains are plotted in Figure 3.47, together with the results obtained
using flat specimens under uniaxial and planestrain tension deformation modes.
Table 3.19 – Limit strains obtained from the Nakajima tests* of the Gen3 980T steel sheet.
can be ascribed to the superposition of stretching and bending effects across the specimen
thickness (Schwindt et al., 2015). Therefore, a higher straining level is expected in the Nakajima
hemispherical punch testing than flat specimens. This effect is more pronounced in AHSS steels
than in conventional steels as they exhibit a high hardening rate in the first stages of straining,
which would retard the onset of necking.
Major strain, 1
Plane strain
(flat specimen 0º RD)
0.4 Uniaxial tension
(flat specimen 0º RD)
Plane strain
0.3 (flat specimen 90º RD)
Uniaxial tension
(flat specimen 90º RD)
0.2 Nakazima test
(W=25 mm)
Nakazima test
0.1 (W=50 mm)
Nakazima test
(W=75 mm)
(a) (b)
Figure 3.48 – (a) Principle of the hydraulic bulge test (Lazarescu et al., 2013), and (b) variables
used for evaluation of stress and strain (Campos et al., 2014).
74
The membrane theory is commonly used to determine the flow stress curve, and its
validity is related to a small sheet thickness/bulge diameter ratio. The effects of bending can be
neglected when calculating the average membrane stresses. Considering that thin sheets are
used, the throughthickness stress 𝜎3 is zero, and the relation between stress, sheet geometry,
and bulge pressure can be established:
𝜎1 𝜎2 𝑝
+ = (3.28)
𝜌1 𝜌2 𝑡
where 𝜎1 and 𝜎2 are the principal stresses on the sheet surface, 𝜌1 and 𝜌2 are the corresponding
radius of the curved surface at midthickness, 𝑝 is the hydraulic pressure, and 𝑡 is the sheet
thickness.
Both principal stresses can be taken equivalent and equal to the socalled membrane
stress (𝜎1 = 𝜎2 = 𝜎𝑏 ), considering the axisymmetric case of the bulge test. The same reason
can be pointed out to curvature radius, 𝜌1 and 𝜌2 , which also remain equivalent (𝜌1 = 𝜌2 =
𝜌𝑏 ). The Eq. (3.28) can be simplified, and flow stress can be determined by:
𝑝∙𝜌
𝜎𝑏 = (3.29)
2𝑡
The determination of flow stress requires the calculation of curvature radius 𝜌 and the
current sheet thickness 𝑡 at the top of the dome during testing. The curvature radius is calculated
using the following equation:
(𝐷𝑐𝑣 ⁄2)2 + ℎ2
𝜌= (3.30)
2ℎ
Where 𝐷𝑐𝑣 ⁄2 is the spherometer radius and ℎ is the dome height, measured from tripod pins,
as shown in Figure 3.48(b).
The current sheet thickness 𝑡 can be calculated from the initial sheet thickness 𝑡0 and
the current thickness strain 𝜀𝑡 as:
𝑡 = 𝑡0 ∙ 𝑒𝑥𝑝(−𝜀𝑡 ) (3.31)
75
The thickness strain 𝜀𝑡 is calculated by evoking the assumption of material
incompressibility. If volume remains constant during plastic deformation, the following
relations are verified:
𝜀𝑡 = −(𝜀1 + 𝜀2 ) (3.32)
where 𝜀1 and 𝜀2 are the principal plastic strain components. The hypothesis of equal values for
both surface strains near the pole is assumed, and thickness strain is given by:
𝜀𝑡 = −2 𝜀 (3.33)
where 𝜀 is the membrane strain. The determination of this value is performed by measuring the
expansion of a circle with an initial diameter 𝐷𝑠𝑡0 . This measurement is done by a modified
extensometer device, with two probes that rest in the test specimen, initially positioned onto the
initial circle. The biaxial strain (𝜀𝑏 ) is equal to the true thickness strain in the polar region (𝜀𝑡 ).
During testing, this circle expands, without change of volume, to a diameter 𝐷𝑠𝑡 and current
thickness strain can be calculated as (Campos et al., 2014):
𝐷𝑠𝑡
𝜀𝑏 = 𝜀𝑡 = −2 𝑙𝑛 ( ) (3.34)
𝐷𝑠𝑡0
From Eq. (3.29) the biaxial stress is calculated according to the variables presented in
Figure 3.48(b) (Campos et al., 2014):
𝑝∙ℎ 𝐷𝑐𝑣 2 𝐷𝑠𝑡 2
𝜎𝑏 = ∙ [( ) + 4] ∙ ( ) (3.35)
16 𝑡0 ℎ 𝐷𝑠𝑡0
As the Gen3 980T steel sheet evaluated in this work has 1.58 mm of thickness and high
strength, it was necessary to reduce the sheet thickness to determine the equibiaxial stress 𝜎𝑏
using the available test equipment. The sheet was machined by milling with a low depth of cut
and low feedrates applying sufficient cooling. The resulting thickness from the machining was
0.89 ± 0.04 mm. The biaxial stressstrain curve obtained from the hydraulic bulge test of the
Gen3 980T steel sheet is shown in Figure 3.49. The yield stress value in equibiaxial tension is
equal to 𝜎𝑏 = 603.7 MPa.
76
1400
1200
1000
Stress (MPa)
800
600 b=603.7 MPa
400
Hydraulic bulge test
200
Uniaxial tensile test 0º RD
0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
Equivalent plastic strain (%)
Figure 3.49 – Biaxial true stressstrain curve obtained from the bulge test.
3.9. Void formation analysis
Interrupted uniaxial tensile tests were used to analyze the evolution of the void
formation (VF) fraction due to the imposed straining level. Seven different longitudinal true
strain levels were observed: 0% (asreceived), 2%, 8%, 13%, 18%, 20% and the fractured
condition. After the tensile testing, the specimens were cut according to Figure 3.50, where four
slices with 12.5 mm × 5 mm × thickness were prepared from the gauge length region. From
these slices, three samples were used for the VF analysis. The interrupted uniaxial tensile tests
were carried out with a constant crosshead speed of 4.5 mm/min in specimens machined in the
rolling direction.
Figure 3.50 – Cut slices prepared from the gauge length region for the VF analysis.
77
For the microvoid analysis, samples taken from the red crosssections in Figure 3.50
were cut by a precision metallographic cutter by applying a low cutting feed and high cooling.
The sliced pieces were mounted into epoxy resin and mechanically prepared with SiC abrasive
grinding papers 180, 240, 600, 800, 1200, and 2400grit using a Struers TegraPol2 machine.
The polishing was performed in two steps: applying a diamond paste of 6 and 3 microns,
respectively.
The samples were analyzed in the Hitachi TM4000Plus SEM. Three sections were
analyzed for each straining level. For each analyzed section, 20 images were taken and, thus,
60 images for each longitudinal strain level and 420 images for all strain levels. The ImageJ
software with the “Analy e Particles” function was used in the quantitative analysis of micro
voids to determine the measures of void density (VD), mean void size (MVS), and void area
fraction (VAF), defined as (Samei et al., 2016):
Number of voids
Void Density (𝑣𝑜𝑖𝑑/𝜇𝑚2 ) = × 1000 (3.36)
Total area
∑ Void area
Mean Void Size (𝜇𝑚2 ) = (3.37)
Number of voids
∑ Void area
Void Area Fraction (%) = × 100 (3.38)
Total area
The polished surface of the asreceived sample is shown in Figure 3.51(a), and the
corresponding binarized image is depicted in Figure 3.51(b). In the later, the black regions
represent the voids.
Figure 3.52 exhibits the results of microvoids measures obtained from Eqs. (3.363.38).
The void density has an amplitude between 2 and 7 void/µm2 ×1,000 and the mean void size is
less than 2 µm2, with most results between the range of 0.5 and 1 µm2. The void area fraction,
shown in Figure 3.52(c), ranges from 0.1% to 0.5%.
78
(a) (b)
Figure 3.51 – Gen3 980T steel sheet in the asreceived condition: (a) SEM micrograph (1,500x)
and (b) corresponding binarized image.
B B
8 2.5
Void Density (void/m ) x 1000
7
Mean Void Size (m )
2.0
2
6
2
5 1.5
4
3 1.0
2
0.5
1
0 0.0
0.00 0.05 0.10 0.15 0.20 0.25 0.00 0.05 0.10 0.15 0.20 0.25
True strain True strain
(a) B (b)
0.6
0.5
Void Area Fraction (%)
0.4
0.3
0.2
0.1
0.0
0.00 0.05 0.10 0.15 0.20 0.25
True strain
(c)
Figure 3.52 – Microvoid analysis results: (a) void density, (b) mean void size, and (c) void
area fraction versus the true strain.
79
Chapter 4
Modeling of plastic behavior
80
4. MODELING OF PLASTIC BEHAVIOR
4.1. Calibration of the GTN parameters
The model, initially proposed by Gurson (1977) and later extended by Tvergaard and
Needleman (1984), known as the GTN model, has been widely used to describe the
micromechanical damage effects ductile metals. The following yield function defines the GTN
model:
𝜎̅ 2 3
𝜙 = ( ) + 2𝑞1 𝑓 ∗ 𝑐𝑜𝑠ℎ ( 𝑞2 Σ) − [1 + 𝑞3 (𝑓 ∗ )2 ] = 0 (4.1)
𝜎̃ 2
with
𝑓, 𝑓 < 𝑓c
𝑓∗ = { 𝑓u − 𝑓c (4.2)
𝑓c + (𝑓 − 𝑓c ) , 𝑓c < 𝑓 < 𝑓f
𝑓f − 𝑓c
where 𝑓c is a critical value of void volume fraction, 𝑓u = 1⁄𝑞1 and 𝑓f is the void volume fraction
at fracture. In Equation (4.1) 𝑞1 , 𝑞2 , and 𝑞3 are material parameters whereas 𝑓 ∗ is the damage
parameter defined by Eq. (4.2). The GTN model is an extension of the von Mises isotropic yield
function accounting for the effects of the hydrostatic pressure on the plastic yielding of metals.
In Eq. (4.1), 𝜎̅ = √(3⁄2) 𝑺: 𝑺 is the von Mises equivalent stress measure defined from the
components of the deviatoric stress tensor:
𝑺 = 𝝈 − 𝜎ℎ 𝜹 (4.3)
81
1
in which 𝝈 is the Cauchy secondorder stress tensor, 𝜎ℎ = 3 𝑇𝑟(𝝈) is the hydrostatic stress
Figure 4.1 Representative volume element (RVE): definition of effective stress.
Thus, from the truestress definition 𝜎 = 𝐹 ⁄𝑆 the effective stress measure can be
calculated as:
𝐹 𝜎∙𝐴 𝜎 𝜎
𝜎̃ = = = = (4.6)
𝐴 − 𝐴𝑉 𝐴 − 𝐴𝑉 1 − (𝐴𝑉) 1 − 𝑓𝐴
𝐴
In Equation (4.6), 𝑓𝐴 = 𝐴𝑉 ⁄𝐴 denotes the fraction area of voids, and 𝜎 is the stress
defined from the adopted workhardening equation in which the parameters were fitted to the
experimental uniaxial tensile curve. Assuming the hypothesis of isotropy meaning that the voids
are equally distributed in all directions as well as independent of the mechanical loading path,
tension or compression, the voided fraction area 𝑓𝐴 can thus be viewed as a scalar damage
variable.
Secondly, from the average measured values of the voided area fraction 𝑓𝐴 as a function
of the total longitudinal strain in uniaxial tension, results taken from Figure 3.52 and neglecting
the elastic strains, a linear equation is proposed to describe the evolution of 𝑓𝐴 as a function of
the longitudinal plastic strain, namely:
𝑓𝐴 = 𝛼𝜀 𝑝 + 𝛽 (4.7)
Figure 4.2 shows the linear fitting results determined from the experimental average
void area fraction of Gen3 980T steel. The α and β coefficients are indicated in Figure 4.2.
The value of void area fraction related to the strain level of 18% was disregarded to
obtain a better fitting of the experimental data. The predicted stressstrain curves of each
analyzed workhardening equation obtained from Equation (4.8) are compared with the
83
experimental truestress and true plasticstrain data in Figure 4.3. As presented in section 3.2.4,
both workhardening equations have good accordance with the experimental true plastic stress
strain curve. However, they show a significant difference for the extrapolated plastic behavior,
which will be better discussed in section 4.1.4, in terms of the void nucleation parameter
softening effect.
𝜎ℎ𝑎𝑟𝑑_𝑒𝑞𝑛
𝜎̃ = (4.8)
1 − (𝛼𝜀 𝑝 + 𝛽)
0.006
p
fA = +
0.005
= 0.00542
= 0.00264
Void Area Fraction
0.004
0.003
0.002
0.001 Exp.
Linear fit
0.000
0.00 0.05 0.10 0.15 0.20 0.25
True strain
Figure 4.2 Experimental void area fraction as a function of the true longitudinal plasticstrain
for the Gen3 980T steel sheet under uniaxial tension and with respect to the rolling direction.
1400
1200
True stress (MPa)
1000
800 Experimental
Ludwik (effective)
Swift (effective)
600 Voce (effective)
4.1.2. Initial porosity
The initial porosity parameter or initial void volume fraction was obtained from an RVE,
which is assumed to be equivalent to the binarized image determined from the SEM analysis,
as schematically depicted in Figure 4.4. This RVE is composed of a small single spherical void
embedded concentrically in a large spherical metal matrix. The spherical matrix has a diameter
𝐷 whereas the spherical single void has a diameter 𝑑, as depicted in Figure 4.4, which, in turn,
is assumed to be equivalent to the actual binarized SEM areas of the material matrix and voids,
respectively. The initial void volume fraction (𝑓0 ) can thus be estimated from the initial area of
voids (𝑓𝐴 ) represented by the intercept of Equation (4.7), that is:
𝐴𝑉 𝜋 𝑟 2 𝑑 2
𝑓𝐴 = = = (4.9)
𝐴 𝜋 𝑅2 𝐷2
𝑉𝑉 4⁄3 𝜋 𝑟 2 𝑑 3 3
𝑓𝑉 = = = = (𝑓𝐴 ) 2 (4.10)
𝑉 4⁄3 𝜋 𝑅 2 𝐷3
3 3
𝑓0 = (𝑓𝐴 |𝜀𝑝 =0 )2 = (𝛽)2 (4.11)
3
𝑓0 = (0.00264)2 = 1.356 × 10−4 (4.12)
𝑟0 = (1 − 𝑓0 ) = 0.99986 (4.13)
Figure 4.4 Schematic approach adopted to obtain the initial void volume fraction from the
measured void area fraction.
85
4.1.3. Material parameters (𝑞1, 𝑞2 , 𝑞3 )
For steels, the typical values of the GTN yield locus parameters 𝑞𝑖 (𝑖 = 1,2,3) in
Equation (4.1) are 𝑞1 = 1.5, 𝑞2 = 1.0 and 𝑞3 = 𝑞12 = 2.25 (Tvergaard and Needleman, 1984).
The parameter 𝑞1 affects the loadbearing capacity of the material, namely, 𝑞1 > 0 provides a
decrease of the yield strength leading to a material softening due to void growth in detriment of
the matrix material's workhardening. The second GTN yield locus parameter 𝑞2 is associated
with the stress triaxiality factor in the GTN yield function, Σ = 𝜎ℎ ⁄𝜎̃, which depends on the
hydrostatic stress and the material yield strength.
From micromechanical simulations considering unit cell finite element models with a
homogeneous material and a single spherical void, the former described by the GTN model,
Faleskog et al. (1998) proposed a procedure for the yield locus parameters calibration, revealing
that (𝑞1 ,𝑞2 ) parameters exhibited a dependence on the strainhardening exponent (𝑛) and the
ratio between the yield strength (𝜎𝑦 ) and Young’s modulus (𝐸). This procedure is adopted here
to define these GTN yield loci parameters avoiding extensive experimental and or combined
numerical investigations given that nonuniqueness of the GTN damage model parameters.
From the calibration results of the procedure developed by Faleskog et al. (1998), reproduced
in Figure 4.5 by the plots of the (𝑞1 ,𝑞2 ) parameters as a function of the strainhardening
exponent 𝑛 = 0.187 and the ratio and the ratio 𝜎𝑦 ⁄𝐸 ≅ 0.003, the yield locus parameters found
for Gen3 980T steel sheet are 𝑞1 = 1.74, 𝑞2 = 0.83 and 𝑞3 = 𝑞12 = 3.03.
2.0 1.10
1.9 y / E= 0.001
1.05
y / E= 0.002
1.8
1.74 1.00 y / E= 0.003
1.7 y / E= 0.004
1.6 0.95
q1
1.5 0.90
q2
1.4 0.85
1.3 0.83
y / E= 0.001
0.80
1.2 y / E= 0.002
y / E= 0.003 0.75
1.1
y / E= 0.004
1.0 0.70
0.00 0.05 0.10 0.15 0.20 0.25 0.00 0.05 0.10 0.15 0.20 0.25
n 0.187 n 0.187
(a) (b)
Figure 4.5 Yield loci parameters (𝑞1 , 𝑞2 ) identified from the micromechanical predictions of
Faleskog et al. (1998) using the experimental data of Gen3 980T steel sheet.
86
the nucleation parameters for the Gen3 980T steel was performed by finite element simulations
of the uniaxial tensile test. Here the idea is to vary the set of parameters to obtain a predicted
nominal stressstrain curve close to that of the experimental uniaxial tensile test results,
including postnecking behavior.
As shown in Figure 4.6, the volume fraction of nucleated voids produces a softening of
the plastic behavior. Considering the values of 𝜀𝑁 , 𝑆𝑁, 𝑓0 to be fixed, increasing the 𝑓𝑁 −value
produces a greater softening effect.
Figure 4.6 Softening effects in uniaxial tension as a function of the fN parameter in the GTN
model (Abaqus, 2009).
Bearing in mind the softening effect from the 𝑓𝑁 parameter, it was observed in
Figure 4.3 that although Voce’s equation provides the best fit in the evaluated stressstrain
range, it also results in a saturation of the workhardening close to perfect plasticity. Such
plastic behavior will result in a rapid loss of the loadbearing capacity after the necking. In this
way, the Swift equation was selected to describe the behavior of the fully dense matrix, since it
87
provides a better fit to the experimental stressstrain data when considering the 𝑓𝑁 softening
effect. Besides, both Swift and Ludwik workhardening equations do not exhibit a saturation at
larger plastic strains. However, the Swift workhardening provided a better fit to experimental
data than the Ludwik equation.
Figure 4.7 shows the schematic geometrical model proposed to simulate the uniaxial
tensile test. The simulation test was carried out by fixing one grip section and applying a
prescribed constant speed at the other grip section. The quasistatic numerical simulations were
performed with the ABAQUS/Explicit commercial finite element code, wherein the GTN
model is available as porous metal plasticity.
Figure 4.7 Boundary conditions for the uniaxial tensile specimen with 1/4 symmetry.
Table 4.1 presents the parameters used to simulate the elastoplastic behavior of the fully
dense material. The total reaction force was obtained by the summation of the fixed region's
nodal forces, see the orange area in Figure 4.7, multiplied by four to account for the two planes
of symmetry. The specimen elongation was calculated from the difference in displacements of
points 1 and 2, indicated in Figure 4.7. Figure 4.8(a) compares the experimental and forecasted
displacementforce determined using the parameters listed in Table 4.1. The current numerical
results overpredict the experimental data, mainly the postnecking elongation up to fracture.
From the typical values for metals presented in the literature, several combinations of 𝜀𝑁 , 𝑆𝑁, 𝑓0
were tested, aiming to describe the postnecking displacementforce curve. The set of
nucleation parameters that provided the best fit is listed in Table 4.2. The softening resulting
from the nucleation parameters can be observed in Figure 4.8(b).
88
Table 4.1 –Elastic properties and parameters of the isotropic effective workhardening.
Elastic properties Isotropic effective workhardening (Swift)
𝐸 [MPa] ν 𝐾 [MPa] 𝜀0 𝑛 α β
20 (a)
18
16
14
Load (kN)
12
10
4 Experimental
2 Simulated (Parametes: Table 4.1)
0
0 2 4 6 8 10 12 14 16 18 20
Elongation (mm)
22
20 (b)
18
16
14
Load (kN)
12
10
4 Experimental
Simulated (Parametes: Table 4.1)
2 Simulated (Parametes: Table 4.1 + Table 4.2)
0
0 2 4 6 8 10 12 14 16 18 20
Elongation (mm)
Figure 4.8 Experimental uniaxial tensile test loadelongation and predicted results obtained (a)
with the parameters of the fully dense matrix (Table 4.1), and (b)with the parameters of the
fully dense matrix (Table 4.1) and the GTN parameters (Table 4.2).
89
4.1.5. Failure parameters (𝑓𝐶 , 𝑓𝐹 )
To determine the critical value of the void volume fraction (𝑓C), Figure 3.39(a) was used
as a reference. The picture shows the major strain field immediately before the fracture in the
uniaxial tension mode. According to the digital image correlation analysis, the maximum
longitudinal total true strain reached before the fracture onset is 𝜀1 = 0.54. From the finite
element simulation calibrated with the GTN nucleation parameters, see Figure 4.8, two regions
in the specimen were monitored. One region is the critical region in the localized necking, and
the other one is located at a distance sufficiently far from the necking, shown in Figure 4.9(a).
The plot in Figure 4.9(b) shows, as expected, that up to the necking onset, all the elements of
the gauge length deform uniformly. After starting the necking, the strain far from the necking
remains almost constant while the strain at the necking zone increases rapidly. Thus, once the
necking began, only the specimen central region contributes to the increase of elongation, and
the failure parameters must be calibrated from this region.
(a)
1.5
1.4
1.3
1.2
1.1
1 at the necking
1.0
0.9
0.8
0.7
0.6 = 0.54 (reference strain level)
0.5
0.4
0.3 Necking onset
0.2
0.1
0.0
0.00 0.05 0.10 0.15 0.20
1 far from the necking
(b)
Figure 4.9 Major strain predictions of the uniaxial tensile test without the failure parameters:
(a) regions at the necking and far from the necking, and (b) corresponding major strain history.
90
Figure 4.10 resumes the methodology for selecting the 𝑓C parameter from the major
strain value. The major strain field is observed along the specimen gauge length at the time
where the maximum value is equal to 𝜀1 = 0.54 in the necking region, Figure 4.10(a). Figure
4.10(b) shows the value of the void volume fraction for the same time. The void volume fraction
(VVF), the nucleation (VVFN), and void growth (VVFG) contributions are plotted in Figure
4.10(c) as a function of the major strain. From the major strain 𝜀1 = 0.54, the critical void
volume fraction 𝑓C was found equal to 0.05.
(a)
(b)
0.10
0.09 VVF
VVFG
0.08 VVFN
Void volume fraction
0.07
0.06
fC = 0.05
0.05
0.04
0.03
0.02
0.01
0.54
0.00
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
1 Major strain
(c)
Figure 4.10 Proposed identification procedure for the critical void volume fraction 𝑓C value:
contours plots of the (a) major strain and (b) void volume fraction and (c) void volume fraction
measures as a function of the major strain in the necking region.
91
The 𝑓F parameter denotes the void volume fraction after the complete specimen fracture.
From the critical value, 𝑓C , the void fraction increases rapidly through the void coalescence
until the 𝑓F value and, then, the corresponding finite element is deleted from the mesh of the
uniaxial specimen model. The coalescence is governed by Equation (4.2) in the GTN model.
The 𝑓Fvalue must be greater than the 𝑓C parameter and is limited to unity, that is, 100% voids.
To identify the parameter 𝑓F two experimental results were considered. First, the endpoint of
the displacementforce curve and, then, the width reduction of the specimen after the fracture.
Figure 4.11 shows a sequence of three different moments in the uniaxial tensile test, namely,
(a) undeformed, (b) localized necking before the fracture, and (c) specimen after the fracture.
The width of the specimen in the fractured region measures approximately 9.80 mm, and the
value of the failure parameter that best fits the experimental results for the Gen3 980T steel
under the uniaxial tension deformation mode was 𝑓F = 0.095. Thus, the identified parameters
of the GTN damage model are presented in Table 4.3.
(a) (b) (c)
Figure 4.11 Images of the uniaxial tensile specimen during the testing at (a) the unloaded state,
(b) immediately before the fracture, and (c) after the fractured.
Table 4.3 – Full set parameters of the GTN damage model for the Gen3 980T steel.
𝑟𝑜 𝑞1 𝑞2 𝑞3 𝜀𝑁 𝑆𝑁 𝑓𝑁 𝑓𝐶 𝑓𝐹 Element
Deletion
0.99986 1.74 0.83 3.03 0.18 0.07 0.035 0.050 0.095 yes
92
The contour plot predictions of the void volume fraction determined using the complete
set of the GTN model parameters is shown in Figure 4.12. Figure 4.12(a) depicts one of the
experimental samples placed with the final predicted model shape. Also, the loadelongation
curves showed similar results, see Figure 4.13. The predicted curve is higher than the
experimental data in the elongation range close to the yielding point up to ~ 3 mm elongation.
It may be attributed to the workhardening description.
(a) (b)
Figure 4.12 (a) Finite element simulations of the uniaxial tensile test of the Gen3 980T steel
sheet: (a) experimental digital image and forecasted deformed specimen and (b) void volume
fraction.
22
20
18
16
14
Load (kN)
12
10
8
6
4 Experimental
GTN model
2
0
0 1 2 3 4 5 6 7 8 9 10 11 12
Elongation (mm)
Figure 4.13 Experimental uniaxial tensile test loadelongation and predicted results obtained
with the full set parameters of the GTN damage model for the Gen3 980T steel.
93
4.1.6. Mesh sensitivity
The influence of the mesh size in the predictions of the uniaxial tensile test was
evaluated using the GTN model parameters given in Table 4.3. Four different mesh sizes in the
specimen gauge length region were considered, namely, 1.6 mm, 0.8 mm, 0.4 mm, and 0.2 mm,
as shown in Figure 4.14. In all numerical simulations, the specimen is meshed with the C3D8R
element type: 8node linear brick, reduced integration, and hourglass control (Abaqus, 2009).
The functionality Element Deletion was selected in the finite element Abaqus mesh module.
This option removes the finite element when it loses the strength capacity completely.
(a)
(b)
(c)
(d)
Figure 4.14 Different mesh sizes in the gauge length region of the uniaxial tensile specimen:
(a) 1.6 mm, (b) 0.8 mm, (c) 0.4 mm, and (d) 0.2 mm.
94
Figure 4.15 compares the forceelongation predictions determined with different mesh
sizes with the experimental data with respect to the rolling direction. The four mesh sizes
provided the same behavior until the beginning of the necking. For the postnecking curve,
except for the 1.6 mm mesh that exceeded the experimental curve by ~ 0.5 mm in elongation,
the other finite element mesh sizes showed very close values, as shown in Figure 4.13, which
was obtained with a mesh size of 0.4 mm.
22
20
18
16
14
Load (kN)
12
10
8
Experimental
6 Mesh size = 1.6 mm
4 Mesh size = 0.8 mm
Mesh size = 0.4 mm
2 Mesh size = 0.2 mm
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Elongation (mm)
Figure 4.15 Influence of the mesh size on the loadelongation curve.
The minimum standard error (𝑆𝐸) was adopted to calculate the difference between the
experimental and predicted results
𝑛 1/2
1 2
𝑆𝐸 = ( ∑(𝐹𝑖𝑁𝑢𝑚. − 𝐹𝑖𝐸𝑥𝑝. ) ) (4.14)
𝑛
𝑖=1
The results of the minimum standard errors are listed in Table 4.4. According to Slimane
et al. (2015), that conducted a parametric study of the GTN model, the numerical prediction has
not a significant mesh size dependence in the uniform plastic behavior. During the uniform
elongation the whole elements in the specimen gauge length have the same strain level.
However, from the necking point, only a few numbers of elements continue to deform in the
fracture zone. Thus, macro behavior prediction of the uniaxial tensile test becomes dependent
95
to the element mesh size. A fine mesh triggers the void formation faster than a coarse mesh. As
expressed in Equation (4.14) the minimum standard error is calculated for common values of
the numerical prediction and the experimental result of loadelongation of the uniaxial tensile
specimen. As there is no dependence on the mesh size for the curve before the necking, the SE
values were very close. The mesh size of 0.4 mm shows less error compared to the experimental
data. In addition, it presented a good number of elements for its CPU time. Although the 0.8 mm
mesh has been shown to be able to represent the material’s behavior, its coarse mesh can mainly
provide a poor numerical forecast in the thickness direction.
Table 4.4 – Mesh sensitivity – minimum standard error of loadelongation curve, CPU time and
number of elements.
Mesh size: 1.6 mm 0.8 mm 0.4 mm 0.2 mm
the Lankford values are close to unity justifying the isotropic assumption made in this work.
The calibrated GTN parameters for Gen3 980T steel provided a good agreement with the
experimental strains determined at 0º and 90º angular orientations.
(a)
(b)
Figure 4.16 Contour plot distributions of the (a) major strain and (b) minor strain determined
by the numerical simulation of the uniaxial tension test with a flat specimen.
0.6
Major strain 0º RD (Exp.)
0.5 Minor strain 0º RD (Exp.)
Major strain 90º RD (Exp.)
Minor strain 90º RD (Exp.)
0.4
Major strain (GTN)
Major strain (GTN)
0.3
True strain
0.2
0.1
0.0
0.1
0.2
0.3
0 5 10 15 20 25 30 35 40 45 50 55 60
length of deformed sections [mm]
Figure 4.17 Experimental and predicted major and minor strains at the centerline section of the
uniaxial tensile specimen.
97
The limit strains predictions defined using the ISO 120042:2008 standard are shown in
Figure 4.18. The forecasted major limit strain is equal to 0.346, while the minor limit strain is
– 0.164. For comparison purposes, the corresponding average experimental limit strains values,
listed in Table 3.17, obtained for the specimen at the rolling direction are 0.324 and – 0.148,
respectively.
0.6
Major strain (GTN)
0.5 Minor strain (GTN)
nd
2 derivative of 1 Major strain:
0.4 Selected points
0.346
Fitting major strain
0.3 Fitting minor strain
True strain
0.2
0.1
0.0
0.1
Minor strain:
0.2 0.164
0.3
5 0 5 10 15 20 25 30 35 40 45 50 55 60 65
Deformed vertical sections [mm]
Figure 4.18 Limit strains from the uniaxial tensile test simulations using the GTN model.
4.2.2. Planestrain mode
The finite element simulation of the planestrain with a flat doublenotched specimen
was also performed using a 3D mesh with C3D8R elements. The boundary conditions are
similar to those used in the uniaxial tension model. Figure 4.19 shows the mesh used in the flat
doublenotched specimen, in which the zoomed region shows the refined elements with a mesh
size equal to 0.4 µm.
Figure 4.19 Mesh refinement adopted in the flat doublenotched specimen
98
The major and minor principal strains contour plot predictions that were determined
from the planestrain specimen are shown in Figure 4.20. The doublenotch results in strain
(stress) concentration in the central specimen zone and deformation mode very close to the
planestrain condition. The specimen shows higher levels of strain at the bottom notch radius.
Figure 4.21 compares the experimental results and numerically simulated values of major and
minor strain before the specimen fracture for the sections defined according to Figure 3.42.
Figure 4.21(a) presents the results obtained at 0º and 90º to the rolling direction and the
numerical prediction in the vertical section. On the other hand, Figure 4.21(b) compares the
corresponding results in the horizontal section. The major strain values are slightly below the
experimental results in both vertical and horizontal sections, whereas the minor strain is
reproduced more accurately.
(a)
(b)
Figure 4.20 Distribution of the (a) major strain and (b) minor strain obtained by the numerical
simulation of the plane strain test with a flat doublenotched specimen.
0.24 0.7
Major strain 0º RD (Exp.)
0.22 Minor strain 0º RD (Exp.) (a) 0.6
Major strain (Exp.)
Major strain (GTN)
(b)
0.20 Major strain 90º RD (Exp.)
Minor strain 90º RD (Exp.) 0.5 Minor strain (Exp.)
0.18 Minor strain (GTN)
Major strain (GTN)
0.16 0.4
Major strain (GTN)
0.14
True strain
0.3
True strain
0.12
0.2
0.10
0.08 0.1
0.06 0.0
0.04
0.1
0.02
0.00 0.2
0.02 0.3
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 0 1 2 3 4 5 6 7 8 9 10
Deformed vertical sections [mm] Horizontal section [mm]
Figure 4.21 Experimental and predicted major and minor strains determined from the double
notched planestrain specimen: (a) vertical and (b) horizontal sections.
99
As with the uniaxial tensile test, the response of the mechanical behavior using the GTN
model was compared with the experimental results in terms of the stressstrain curve. The graph
shows the experimental values with samples manufactured at 0º and 90º of the rolling direction
and the stressstrain curve obtained through the finite element model. Although the calibration
of the GTN model parameters was made from the uniaxial tensile test, it was observed that the
selected parameters provide a mechanical behavior close to the experimentally determined
values for the plane strain test. In terms of tension, it was observed that the numerical simulation
was above about 2% of the experimental curve. Regarding the true strain, the value found was
also slightly beyond that recorded experimentally, as shown in Figure 4.23.
Regarding the limit strains for the plane strain condition in the FLC, the major strain
determined from the GTN model was equal to 0.123, and the minor strain equal to 0.012. The
values obtained experimentally for the major and minor strains were 0.152 and 0.011 along the
rolling direction, and 0.134 and 0.010 for the case perpendicular to the rolling direction,
respectively.
0.18
0.16 Major strain (GTN)
Minor strain (GTN)
0.14 Major strain: Fitting major strain
0.123 Fitting minor strain
0.12 2 derivative of
nd
True strain
0.10
0.08
0.06
0.04
0.02
Minor strain:
0.00 0.012
0.02
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Deformed vertical sections [mm]
Figure 4.22 Limit strains from the planestrain tension test simulations using the GTN model.
Both in the experimental and numerical simulation, the fracture was started in the
curvature of the notch. This can be observed from the void volume fraction in Figure 4.24,
wherein the notch radius shows a region with more voids. Figure 4.24(a) presents the result of
the GTN model before the fracture starts for the void volume fraction (VVF), which is
composed of the void volume fraction growth (VVFG) and void volume fraction nucleated
(VVFN). Figure 4.24(b) exhibits the specimen fractured and the results of VVF, VVFG, and
100
VVFN. According to the selected set parameters of the GTN model, before the fracture, the
central region of the specimen presents a higher contribution of the nucleation of new voids
than the growth process.
1400
1200
1000
True stress (MPa)
800
600
400
Experimental 0º RD
200 Experimental 90º RD
Simulated
0
0.00 0.02 0.04 0.06 0.08 0.10 0.12
True strain
Figure 4.23 Experimental true stressstrain curve of the plane strain test and predicted results
obtained with the full set of parameters of the GTN model.
(a) (b)
Figure 4.24 Void volume fraction (VFF), void volume fraction growth (VVFG), and void
volume fraction nucleated (VVFN): (a) before and (b) after the fracture.
101
4.2.3. Nakajima test
The numerical simulation of the Nakajima test was modeled considering the tooling as
rigid surfaces described by discretized 4node shells elements, R3D4 in Abaqus terminology,
and the blank specimens by 3D 8node reduced integration solid elements, C3D8R in Abaqus.
The hemispherical punch has a diameter of 100 mm, and the blank specimens have a width
ranging from 25 to 175 mm. Onefourth of the tooling and blank geometries are considered in
the finite element model, in which displacement symmetry boundary conditions are imposed
along with fixed conditions for the die. Figure 4.25(a) shows the mesh refinement and adopted
in the numerical simulation of the 25 mm wide specimen, whereas Figure 4.25(b) exhibits the
assembly of the Nakajima’s tooling. The specimen's central region is meshed with an element
size of 0.4 mm x 0.4 mm x 0.4 mm. In the lockbead zone, it was also necessary to apply a mesh
refinement to avoid overclosure contact problems, as shown in Figure 4.25(a).
Furthermore, the interaction between the tools and specimen was defined using the
general contact option in the Abaqus/Explicit finite element code and by defining a Coulomb’s
friction coefficient equal to 0.10. The numerical simulation of the Nakajima test was performed
in two steps. The first step refers to the blankholder action to prevent the specimen from
slipping into the die during the punch displacement, thus, assuring the stretching conditions.
The second simulation step consisted of the punch displacement until the specimen fracture.
Table 4.5 resumes the adopted element types according to the Abaqus FE code terminology and
the number of elements for each Nakajima test part.
Table 4.5 – Types and number of elements for each part used in the Nakajima test simulation.
Part Element type Number of elements
Punch R3D4 3.150
Die R3D4 3.420
Holder R3D4 2.790
Specimen W25 C3D8R 15.120
Specimen W50 C3D8R 43.152
Specimen W75 C3D8R 39.480
Specimen W100 C3D8R 51.408
Specimen W125 C3D8R 57.408
Specimen W175 C3D8R 36.800
102
(a) (b)
Figure 4.25 Details of the 3D finite element model of the Nakajima test: (a) mesh refinement
of the specimen W25 and (b) the rigid tooling assembly.
Figure 4.26(a) and (b) show the results of the numerical prediction obtained with the
GTN model parameters of the Gen3 980T steel sheet in terms of the major and intermediate
principal total strain distributions determined from the fracture onset, respectively. Both
contour plots results are shown considering the mirroring reflections defined from the XY and
YZ symmetry planes. The die was also removed for complete visualization of the deformed
specimens.
(a) (b)
Figure 4.26 Distribution of the (a) major and (b) minor strain obtained by numerical simulation
of the Nakajima test for the W25 specimen.
From the results obtained in Figure 4.26(a) and (b), the major and the minor strain were
calculated according to the ISO 120042:2008. Figure 4.27 illustrates the procedure performed
to define the limit strains. The same method was repeated for each specimen, and the resulting
numerical limit strains are listed in Table 4.6.
103
0.7
Major strain (GTN) W25
0.6 Minor strain (GTN)
nd
2 derivative of 1 Major strain:
0.5 Selected points
0.379
Fitting major strain
0.4 Fitting minor strain
True strain 0.3
0.2
0.1
0.0
0.1
Minor strain:
0.2
0.134
0.3
20 15 10 5 0 5 10 15 20
Unwound length [mm]
Figure 4.27 Limit strains defined from the Nakajima test numerical simulation with the W25
specimen using the GTN model.
Table 4.6 – Major and minor limit strains obtained from the numerical simulations of the
Nakajima test for the Gen3 980T steel sheet.
Specimen width [mm] Minor strain Major strain
W25 0.134 0.379
W50 0.104 0.380
W75 0.015 0.260
W100 0.024 0.232
W125 0.065 0.240
W175 0.198 0.199
Figure 4.28 compares the numerical predictions with the experimental limit strains
determined from uniaxial and planestrain tension deformation modes (flat specimens) and
Nakajima (hemispherical punch stretching) tests. The FLC's righthand side, forecasted from
the Nakajima’s test finite element simulations, is also plotted in this figure as a first estimate.
For both inplane and outofplane testing procedures ranging from the planestrain intercept
(𝜌 = 0) and the isotropic uniaxial (𝜌 = −1/2) deformation modes, the predicted limit strains
provided a reasonable agreement with the corresponding experimental data. It is also possible
to observe the increase in the limit strains determined by Nakajima simulations compared to
the flat specimen predictions observed in the experimental results.
104
Exp. Nakajima
Major strain, 1
Exp. flat specimen
0.5 GTN Nakajima
GTN flat specimen
0.3
W100
W125
W75 W175
0.2
=1/2
=1
0.1
there is an exchange in the order between the curves referring to the plane strain and equibiaxial
strain conditions. In Figure 2.16, Samei et al. (2016) showed for DP600 steel that the plane
strain is the critical condition. For a given strain level, the plane strain presents a higher fraction
of voids than in biaxial tension. From Figure 2.16(b), it is also observed that unlike the GTN
model that adopts fixed values for the failure parameters (𝑓𝐶 , 𝑓𝐹 ) regardless of the deformation
mode, experimentally the deformation modes present different critical void fraction values,
being the biaxial tension much more tolerant to the accumulation of voids than the case of plane
strain, which suggests that the limit strain value in the equibiaxial strain condition of
Figure 4.28 and 4.29 are underestimated.
GTN Nakajima ( = 0.03)
Major strain, 1 GTN Nakajima ( = 0.10)
0.5
0.4
0.3
0.2
=1/2
=1
0.1
W125
0.06 W175
0.04
0.02
0.00
0.0 0.1 0.2 0.3 0.4 0.5 0.6
Equivalent plastic strain
(a) (b)
Figure 4.30 – (a) VVF distribution for W25 specimen and (b) VVF of the critical element as a
function of the equivalent plastic strain for all specimen geometry numerically simulated.
106
4.2.4. Holeexpansion and Erichsen cupping tests
The hole expansion and Erichsen tests were numerically simulated using the calibrated
parameters of the GTN model. Both tests' specimens were modeled with 3D elements of type
C3D8R, whereas rigid surfaces described the tools with R3D4 elements assuming ¼ symmetry.
It was observed that the present numerical simulation could not reproduce the
experimental results of the hole expansion test. The punch displacement and load observed from
the complete throughthickness specimen fracture are equal to 6.57 mm and 10.38 kN,
respectively, as listed in Table 3.15. Figure 4.31 compares the experimental loaddisplacement
results and the numerical predictions obtained with the GTN model. The GTN model
parameters adopted in this work (Table 4.3) provided less stiffness of the punch load
displacement for the HET. Also, the punch was able to penetrate towards the hole without the
hole edge showing a fracture. The punching process used to manufacture the HET specimen
was not considered in the present simulations, nor the surface quality of the hole sheared edge.
As shown in Figure 3.28, about 50 to 70% of the sheared hole edge corresponds to the fracture
zone. The quality of the sheared hole edge directly influences the hole expansion ratio
(Paul, 2020). This would partially explain an increase of the predicted hole expansion ratio
compared to the experimental result since the model considered an undamaged circular hole
without any preceding prestraining. However, it was found that the hole was expanding far
beyond what could be attributed to the hole’s finish quality.
16
14
12
Punch force (kN)
10
4
Experimental
2
GTN model
0
0 1 2 3 4 5 6 7 8 9 10 11
Punch displacement (mm)
Figure 4.31 Comparison between the result of experimental and numerical simulation of hole
expansion test in terms of punch forcedisplacement.
107
Figure 4.32 presents a comparison of the stress triaxiality as a function of the equivalent
plastic strain obtained from the critical elements at the edge of the hole expansion specimen
and the uniaxial tensile test specimen. The hole edge has a fixed stress triaxiality equivalent to
the uniaxial tension stressstate that for isotropic von Mises yield criterion is equal to 1/3.
Conversely, the stress triaxiality changes abruptly in the uniaxial tensile specimen's critical
element, as shown in Figure 4.32. Paul (2020), in a recent review on the hole expansion ratio,
points that the stress state at the central hole edge is uniaxial tensile during hole expansion test.
Nevertheless, the hole expansion differs from the uniaxial tensile test owing to the deformation
gradient across the thickness and multiple crack initiation sites. The diffuse necking, followed
by a localized plastic instability, usually occurs during uniaxial tensile tests of ductile metals.
In the hole expansion test, the diffuse necking might be suppressed and or delayed and, thus,
only localized necking will occur. In this work, the methodology for calibrating the GTN
model's parameters was established from experimental and numerical results of the uniaxial
tensile test (section 4.1). In this sense, the increase in triaxiality due to diffuse necking was
accounted for by the selected parameters. Consequently, they are no indicated for forecasting
the forming flanged parts.
Figure 4.32 Predicted stress triaxiality values determined from the uniaxial tensile and hole
expansion tests.
108
Figure 4.33 shows the deformed hole expansion specimen corresponding to the
maximum punch displacement shown in Figure 4.31. The critical elements are concentrated in
the hole edge freecontact region. Even though the punch displacement exceeded about 30% of
the experimental result, the critical void volume fraction value (𝑓C = 0.05) was not reached in
this region.
Figure 4.33 Distribution of the void volume fraction in the hole expansion test.
Concerning the Erichsen cupping test's numerical simulation, the material failed before
the result obtained experimentally, as shown in Figure 4.34. The predicted punch reached a
maximum loading of 51.2 kN and a displacement equal to 7.7 mm. The experimental Erichsen
cupping test provided punch displacement until the material fracture of 10.64 mm and a reaction
force acting on the punch equal to 63.4 kN.
70
60
50
Punch force (kN)
40
30
20
10 Experimental
GTN model
0
0 1 2 3 4 5 6 7 8 9 10 11 12
Punch displacement (mm)
Figure 4.34 Comparison between the experimental and predicted punch forcedisplacement
curves determined from the Erichsen cupping test of the Gen3 980 steel sheet.
109
Figure 4.35(a) shows the void volume fraction distribution before rupture.
Figure 4.35(b) displays the numerical prediction of the dome compared to the measured profile.
The experimentally measured thickness before the fracture is equal to 0.73 mm. From the
numerical prediction obtained using the GTN model, the specimen reached a final thickness of
1.15 mm. This difference in thickness is related to the fact that the model considers that the
void fraction increases more rapidly due to the equivalent plastic strain in equibiaxial stretching
mode. Also, as previously discussed for the W175 specimen from the Nakajima test, the values
of 𝑓C and 𝑓F are the adopted as the same regardless of the strain mode, and both parameters were
calibrated from the uniaxial tensile test.
(a)
16
inner surface GTN model
outer surface GTN model
12 Experimental
Length (mm)
tGTN = 1.15 mm
8
tExp = 0.73 mm
4
0
16 12 8 4 0 4 8 12 16
Length (mm)
(b)
Figure 4.35 (a) Distribution of void volume fraction for the numerical simulation of ECT, and
(b) comparison between experimental and numerical simulation result of deformed ECT
specimen profile.
To further investigate the hole expansion and Erichsen cupping numerical simulations,
a unit cell with only one C3D8R element was subjected to three different deformation modes:
uniaxial tension, planestrain tension, and equibiaxial stretching, schematically shown in
Figure 4.36(a), (b), and (c) respectively. The unit cell represents a critical element.
110
(a) (b) (c)
Figure 4.36 Unit cell models submitted to (a) uniaxial tension, (b) planestrain tension, and
(c) equibiaxial stretching deformation modes.
The result in terms of stress triaxiality factor as a function of the equivalent plastic strain
is shown in Figure 4.37. According to the parameters selected for the GTN model, it was
observed that the critical elements subjected to loads with a lower stress triaxiality factor
showed a higher level of equivalent plastic strain until reaching the rupture condition. In these
numerical simulations, the deformation modes were kept constant until the element failure. For
the case of uniaxial tension, the unit cell reached 0.7 of equivalent plastic strain, while for the
cases of the plane strain and equibiaxial stretching, it was 0.5 and 0.45, respectively. Comparing
the results of uniaxial tension for the single element (Figure 4.37) with the critical elements of
HET and uniaxial tensile test (Figure 4.32), it was observed that the critical element of the
tensile test reaches rupture before reach 0.7 of equivalent plastic strain due to the increase of
stress triaxiality caused by the diffuse necking. On the other hand, as the hole expansion test is
not subject to diffuse necking, the critical element continued to deform according to the uniaxial
tension mode. However, the deformation level generated by the punch displacement could not
produce enough equivalent plastic strain to reach the fracture.
1.0
Uniaxial tension
Plane strain
0.8
Equibiaxial stretching
Stress triaxiality
0.6
0.4
0.2
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Equivalent plastic strain
Figure 4.37 Stress triaxiality factor as a function of the equivalent plastic strain for the
calibrated GTN model parameters for different deformation modes.
111
The results in terms of void fraction versus equivalent plastic strain for the three
deformation modes specified in Figure 4.36 were plotted in Figure 4.38. As for the Nakajima
test prediction's critical elements shown in Figure 4.30(b), the same behavior was observed for
the numerical simulation with a unit cell. The biaxial stretching mode showed a rate of increase
in the void fraction greater than the uniaxial tension and plane strain deformation modes. Park
et al. (2016) studied the limit forming criteria for sheet metal forming, showing that the plane
strain provides lower values of equivalent plastic strain in the fracture compared with uniaxial
tension and biaxial stretching. Figure 4.38 indicates that the GTN model, as defined, is directly
influenced by the increase in the stress triaxiality factor. The higher the stress triaxiality, the
faster the void fraction will increase.
0.12
Uniaxial tension
Plane strain
0.10
Equibiaxial stretching
Stress
Void volume fraction
0.08 triaxiality
factor
0.06 (> )
0.04
0.02
0.00
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Equivalent plastic strain
Figure 4.38 Void volume fraction as a function of the equivalent plastic strain for the calibrated
GTN model parameters for different deformation modes.
As shown in Equation (4.4), the void volume fraction can be decomposed into two
portions, one due to the growth of voids and another associated with the nucleation of new
voids. Figure 4.39(a) and (b) exhibits the decomposition of Figure 4.38 into growth and void
nucleation, respectively. It was observed that only the portion referring to the growth of voids
is subject to the influence of stress triaxiality. For the GTN model, the void nucleation does not
depend on the deformation mode to which the element is subjected.
112
0.14
Uniaxial tension
Void volume fraction of growth
0.12 Plane strain
Equibiaxial stretching
0.10
0.08
0.06
0.04
0.02
0.00
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Equivalent plastic strain
(a)
0.08
Uniaxial tension
Void volume fraction of nucleation
Plane strain
Equibiaxial stretching
0.06
0.04
0.02
0.00
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Equivalent plastic strain
(b)
Figure 4.39 Void volume fraction due to the (a) growth and (b) nucleation as a function of the
equivalent plastic strain for the calibrated GTN model parameters for uniaxial tension, plane
strain, and equibiaxial tension.
The values of void volume fraction as a function of the equivalent plastic strain were
plotted in Figure 4.40 for HET's critical elements and ECT's numerical predictions. These
results were compared with the results presented in Figure 4.38 for the cases of biaxial
stretching and uniaxial tension performed with a unit cell. The critical elements of the cases of
equibiaxial stretching obtained through the Nakajima test with a 175 mm wide specimen and
uniaxial tensile test were also displayed in Figure 4.40. This analysis showed that both the ECT
and Nakajima test follow the unit cell's behavior in equibiaxial stretching. In this sense, the fact
that the prediction of the ECT was not able to deform as much as the experimental ECT, see
113
Figure 4.35(b), reinforces the idea that the numerically predicted results for the right side of the
FLC (Figure 4.28) are conservative in terms of limit strains. Figure 4.40 shows that the result
of the uniaxial tensile test's critical element deviates from the unit cell's result, and the HET
remains with similar behavior. However, the equivalent plastic strain in the HET was very far
from the strain where the fracture occurs. In the HET case, the quality of the hole's sheared edge
was not taken into account in the numerical forecast. The damage caused by the punching
process at the edge of the hole can produce a critical element with a certain amount of equivalent
plastic strain before the HET, which makes numerical prediction of the HET very difficult.
0.12
Critical element:
Uniaxial tension CELL
0.10 Equibiaxial stretching CELL
Uniaxial tensile test
Hole expansion test
Void volume fraction
0.08
Erichsen cupping test
Nakajima test W175
0.06
0.04
0.02
0.00
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Equivalent plastic strain
Figure 4.40 Comparison of the numerical predictions of the GTN model for a single element
and critical element of different deformation modes
114
Chapter 5
Conclusions and future works
115
5. CONCLUSIONS AND FUTURE WORKS
5.1. Final conclusions
In this work, the uncoated coldrolled Gen3 980T steel sheet with 1.58 mm thick was
evaluated by means of mechanical tests and finite element modeling. The Gen3 980T steel sheet
presented a global formability 𝑆𝑢 × 𝑒𝑡 = 24.3 GPa% within the region known as the current third
generation of advanced high strength steels (AHSS). This excellent formability results from its
multiphase microstructure, which is composed of martensite, ferrite, and retained austenite.
From the performed testing procedures and proposed modeling, the main conclusions are
summarized as follows:
1. The standard mechanical testing provided the average mechanical properties, namely,
the yield stress (604 MPa), ultimate tensile strength (1040 MPa), total elongation (23.4%).
Gen3 980T steel showed anisotropy properties appreciated in sheet metal forming processes.
In particular, a planar anisotropy close to zero (−0.079) along with the normal anisotropy
coefficient close to unity (0.917). The biaxial anisotropy coefficient was also close to unity
(0.918).
2. The literature commonly compares the mechanical behavior of the thirdgeneration
AHSS with the firstgeneration of the same grade, which, in this case, is the DualPhase 980.
The higher formability of the Gen3 980T steel can be ascribed to the straininduced martensitic
transformation resulting from the retained austenite. In the asreceived state, the XRD analysis
provided a retained austenite volume fraction of 12.2%, which, in turn, is prone to transform
into martensite during the early stages of plastic straining, as revealed by Chen et al. (2017).
3. An original device was specially developed for performing both hole expansion (HET)
and Erichsen cupping tests (ECT) in a universal testing machine. Such procedures are
qualitative and comparative sheet metal formability tests. The Gen3 980T steel sheet presented
a hole expansion ratio (HER) equal to 10.9% and an Erichsen index (EI) equal to 10.62 mm.
This EI value is in good agreement with a similar Q&P 980 steel grade reported in the literature
(Li et al., 2018). However, the HER is 60% lower than the corresponding value reported for a
similar Q&P 980 steel sheet 1.36 mm thick (Li et al., 2018). The present low value of the HER
suggests special attention in the hole preparation aiming to assure a better finish quality at the
hole edge than the punching process, such as milling and wire EDM.
116
4. One major issue of this project was determining the limit strains for the FLC right side.
Due to the combination of the Gen3 980T elevated yield strength and the 1.58 mm thickness,
very robust laboratory equipment would be required. The FLC left side was defined using the
Nakajima procedure and flat specimens with and without notch submitted to uniaxial tension.
As a result, an FLC shiftup of 10% in major strain was observed in the Nakajima specimens.
The hydraulic bulge test was also performed using a circular specimen with reduced thickness,
which provided biaxial yield stress of 603.7 MPa.
5. Thanks to the anisotropic properties of the Gen3 980T steel, isotropic plasticity could
be adopted as a first approximation in the modeling. The identified damage parameters of the
GTN model were able to reproduce the experimental loadelongation obtained from the uniaxial
tensile test. The mesh sensitivity analysis also showed that the mesh size does not influence the
finite element predictions in the uniform elongation domain. However, as the necking appears,
the smaller the mesh size and, hence, the deformation is more localized. A mesh size of 0.4 mm
in the gauge length zone was enough to fit the experimental data.
6. The numerical predictions determined from the proposed method for calibration of the
GTN model parameters are close to the experimental values of the planestrain and the FLC
lefthand side Nakajima tests. The present simulations also captured the formability increase
due to curvature in the Nakajima test and the friction coefficient reduction. However, the FLC
righthand side predictions provided conservative limit strains values. This behavior was shown
to be highlighted for the equibiaxial stretching condition. An analysis based on the most critical
elements showed that the rate of void volume fraction as a function of the equivalent plastic
strain increases more quickly with the increase of the stress triaxiality factor. As the equibiaxial
state has a greater stress triaxiality factor, it became the most critical case instead of the plane
strain condition. This effect is related only to the void volume fraction due to growth so that the
void nucleation forecast is the same regardless of the deformation mode. Another drawback,
related to the unexpected reduction of the biaxial limit strains, is that the GTN classical model
failure parameters are also independent of the deformation mode. Likewise, the numerical
predictions of the Erichsen test provided limit strains lower than the experimental values.
7. The calibration of the GTN model parameters was established based on the experimental
and numerical results of the uniaxial tensile test. A diffuse necking followed by localized plastic
instability usually occurs during uniaxial tensile tests for ductile metals. Regarding the HET, it
was observed through the stress triaxiality factor that the hole edge stress state is under the
uniaxial tension condition. According to Paul (2020), HET differs from the uniaxial tensile test
117
due to the deformation gradient through the thickness and several crack initiation sites. In the
HET, diffuse necking can be suppressed and/or delayed and, therefore, only localized necking
will occur. The hole also shows dimensional variations and a certain amount of damage due to
the previous punching process. However, in the modeling, an idealized circular hole was
adopted. Due to this difference, it was observed that the calibration of the selected parameters
was unable to reproduce the observed experimental HET.
5.2. Future works
The following topics are considered as potential future works:
1. Analyze the reduction of the retained austenite volume fraction as a function of plastic
strain for the uniaxial tension, plane strain, and equal biaxial stretching deformation modes.
2. From the tested samples, develop a study of the fracture morphology of Gen3 980T steel
and perform local fracture strain measurements for different deformation modes.
3. Due to the difficulty found in this work to obtain the FLC's rightside, it was understood
that it would be interesting to propose or improve the miniaturization techniques of test for
determining the FLC for sheets with high strength.
4. Carry out hole expansion tests for Gen3 980T with different preparation conditions
(machined by milling, EDM, laser cutting, and punching) to verify the influence of the hole
preparation on HER result.
5. Perform an analysis of microvoids for different deformation modes, relate the growth
and nucleation of microvoids with the stress triaxiality factor, and propose a modification of
the GTN model.
118
References
119
Abbasi, M., Bagheri, B., Ketabchi, M., Haghshenas, D. F. Application of response surface
methodology to drive GTN model parameters and determine the FLD of tailor welded blank.
Computational Materials Science, vol. 54 (2012), pp. 368376.
Abbasi, M., Ketabchi, M., Izadkhah, H., Fatmehsaria, D. H., Aghbash, A. N. Identification of
GTN model parameters by application of response surface methodology. Procedia
Engineering, vol. 10 (2011), pp. 415420.
Abbassi, F., Mistou, M., Zghal, A. Failure analysis based on microvoid growth for sheet metal
during uniaxial and biaxial tensile tests. Materials & Design, vol. 49 (2013), pp. 638646.
Achouri, M., Germain, G., Santo, P. D., Saidane, D. Experimental characterization and
numerical modeling of micromechanical damage under different stress states. Materials &
Design, vol. 50 (2013), pp. 207222.
Aramis user’s manual, Version v6.1, GOM mbH Optical Measuring Techniques,
Braubschweig, Germany, 2009.
Ariza, E. A., Masoumi, M. P. Improvement of tensile mechanical properties in a TRIPassisted
steel by controlling of crystallographic orientation via HSQ&P processes. Materials Science &
Engineering A, vol. 713 (2018), pp. 223233.
Ariza, E. A., Poplawsky, J., Guo, W., Unocic, K., Ramirez, A. J., Tschiptschin, A. P., Babu, S.
S. Evaluation of Carbon partitioning in new generation of quench and partitioning (Q&P) steels.
Metallurgical and Materials Transactions A, vol. 49A (2018), pp. 48094823.
ASTM, Standard practice for XRay determination of retained austenite in steel with near
random crystallographic orientation, Standard E97513, American Society for Testing and
Materials, Feb. 15, 2013.
Aydin, H., Essadiqu, E., Jung, I. H., Yue, S. Development of 3rd generation AHSS with medium
Mn content alloying compositions. Materials Science & Engineering A, vol. 564 (2013), pp.
501508.
Banabic, D. Sheet metal forming process – Constitutive Modelling and Numerical Simulation.
SpringerVerlag, Berlin, 2010.
Banabic, D., Bunge, H. J., Pohlandt, K., Tekkaya, A. E., Formability of metallic materials:
plastic anisotropy, formability testing, forming limits. Engineering materials, SpringerVerlag,
1st edition, Berlin, 2000.
Bandyopadhyay, k., Panda, S. K., Saha, P., Hernandez, V. H. B., Zhou, Y. N. Microstructure
and failure analyses of DP980 laser welded blanks in formability context. Materials Science
Engineeriang A, vol. 652 (2016), pp. 250263.
Barlat, F., Brem, J. C, Yoon, J. W., Chung, K., Dick, R. E., Lege, D. J., Pourboghrat, F., Choif,
C. H., Chu, E., Plane stress yield function for aluminum alloy sheets – Part 1: Theory.
International Journal of Plasticity, vol. 19 (2003), pp. 12971319.
Branagan, D., Parsons, G., Machrowicz, T., Cischke, J., Frerichs, A., Meacham, B., Cheng, S.,
Justice, G., Sergueeva, A. Effect of deformation during stamping on structure and property
evolution in 3rd generation AHSS. Open Journal of Metal, vol. 8 (2018), pp. 1533.
Bruschi, S., Altan, T., Banabic, D., Bariani, P. F., Brosius, A., Cao, J., Ghiotti, A., Khraisheh,
M., Merklein, M., Tekkaya, A. E. Testing and modelling of material behavior and formability
in sheet metal forming. CIRP Annals – Manufacturing Technology, vol. 63 (2014), pp. 727
749.
Butuc, M. C., Forming Limit Diagrams. Definition of plastic instability criteria, Ph.D.
dissertation, Engineering Faculty of Porto University, 2004.
Campos, H., Santos, A. D., Martins, B., Ito, K., Mori, N., Barlat, F. Hydraulic bulge test for
stressstrain curve determination and damage calibration for ItoGoya model. Conference –
World Congress on Computational Congress XI, Barcelona (2014), pp. 116.
Cao, T. S., Maire, E., Verdu, C., Bobadilla, C., Lasne, P., Montmitonnet, P., Bouchard, P. O.
Characterization of ductile damage for a high carbon steel using 3D Xray microtomography
and mechanical tests – Application to the identification of a shear modified GTN model.
Computational Material Science, vol. 84 (2014), pp. 175187.
Cha, W., Kim, N. Quantification of microcracks on the bending surface of roll formed products
using the GTN model. Metals and Materials International, vol. 20 (2014), pp. 841850.
Chaboche, J. L. Continuum damage mechanics: Present state and future trends. Nuclear
Engineering and Design, vol. 105 (1987), pp. 1933.
Chalal, H., Meraim, F. A. Determination of forming limit diagram based on ductile damage
models and necking criteria. Latin American journal of solids and structures, 14 (2017), pp.
18721892.
Chen, X., Niu, C., Lian, C., Lin, J. The evaluation of formability of the 3rd generation Advanced
High Strength Steels QP980 based on Digital Image Correlation Method. Procedia
Engineering, vol. 207 (2017), pp. 556561.
Chen, Z., Dong, X. The GTN damage model based on Hill’48 anisotropic yield criterion and
its application in sheet metal forming. Computational Materials Science, vol. 44 (2009), pp.
10131021.
Choi, H., Lee, S., Lee, J., Barlat, F., Cooman, B. C., Characterization of fracture in medium Mn
steel. Materials Science & Engineering A, vol. 687 (2017), pp. 200210.
121
Cingara, G. A., Saleh, C. A. R., Jain, M. K., Wilkinson, D. S. Void nucleation and Growth in
dualphase steel 600 during uniaxial tensile testing. Metallurgical and Materials
Transactions A, vol. 40A (2009), pp. 31173127.
Council of the European Union – European Parliament, Regulation (EU) Nº 333/2014 of the
European Parliament of the Council of 11 March 2014 amending Regulation (EC) Nº 443/2009
to define the modalities for reaching the 2020 target to reduce CO2 emissions from new
passenger cars, Official Journal of the European Union, CELEX: 32014R0333, 2014.
Ebner, S., Suppan, C., Schnitzer, R., Hofer, C. Microstructure and mechanical properties of a
low C steel subjected to bainitic or quenching and partitioning heat treatments. Materials
Science & Engineering A, vol. 735 (2018), pp. 19.
Faleskog, J., Gao, X., Shih, C. F. Cell model for nonlinear fracture analysis – I micromechanics
calibration. International Journal of Fracture, vol. 89 (1998), pp. 355373.
Finfrock, C. B., Clarke, A. J., Thomas, G. A., Clarke, K. D. Austenite stability and strain
hardening in CMnSi Quenching and Partitioning Steels. Metallurgical and Materials
Transactions A, May 2020.
Freitas, M. C., Moreira, L. P., Velloso, R. G. Experimental analysis and theoretical predictions
of the limit strains of a hotdeep galvanized Interstitial Free steel sheet, Materials Research,
vol. 16 (2013) 2, pp. 351366.
Gatea, S., Lu, B., Ou, H., McCartney, G. Numerical simulation and experimental investigation
of ductile fracture in SPIF using modified GTN model. MATEC web of conference, vol. 21
(2015), 04013.
Gatea, S., Xu, D., Ou, H., McCartney, G., Evaluation of formability and fracture of pure
titanium in incremental sheet forming. The International Journal of Advanced Manufacturing
Technology, vol. 95 (2018), pp. 625641.
Gholipour, H., Biglari, F. R., Nikbin, K. Experimental and numerical investigation of ductile
fracture using GTN damage model on insitu tensile tests. International Journal of Mechanical
Sciences, vol. 164 (2019), pp. 105170.
Grajcar, A., Kuziak, R., Zalecki, W. Third generation of AHSS with increased fraction of
retained austenite for the automotive industry. Archives of Civil and Mechanical Engineering,
vol. 12 (2012), pp. 334341.
Guo, W., Wan, Z., Jia, Q., Ma, L., Zhang, H., Tan, C., Peng, P. Laser weldability of TWIP980
with DP980/B1500HS/QP980 steels: Microstructure and mechanical properties. Optics and
Laser Technology, vol. 124 (2020), 105961.
Guo, W., Wan, Z., Peng, P., Jia, Q., Zou, G., Peng, Y. Microstructure and mechanical properties
of fiber laser welded QP980 steel. Journal of Materials Processing Tech, vol. 256 (2018), pp.
229238.
Gurson, A. L. Continuum theory of ductile rupture by void nucleation and growth: part I – yield
criteria and flow rules for porous ductile media. Journal of Engineering Materials and
Technology, vol. 99 (1977), pp. 215.
122
Hambli, R. Comparison between Lemaitre and Gurson damage models in crack growth
simulation during blanking process. International Journal of Mechanical Sciences, vol. 43
(2001), pp. 27692790.
Hance, B. Advanced High Strength Steel (AHSS) Performance Levels, SAE Technical Paper,
(2018).
Hosford, W. F. Metal forming – Mechanics and Metallurgy. Cambridge University Press, 3rd
edition, New York, 2007.
Hu, B., Luo, H., Yang, F., Dong, H. Recent progress in mediumMn steels made new designing
strategies, a review. Journal of Materials Science & Technology, vol. 33 (2017), pp. 14571464.
Hu, Z. P., Xu, Y, B., Zou, Y., Misra, R. D. K., Han, D. T., Chen, S. Q., Hou, D. Y. Effect of
intercritical rolling temperature on microstructuremechanical property relationship in a
medium MnT IP steel containing δ ferrite. Materials Science & Engineering A, vol. 720
(2018) pp. 110.
ISO 16630:2009 – Metallic Materials – Sheet and Strip – Hole Expanding Test, International
Organization for Standardization: Geneva, Switzerland, 2009.
ISO 20482:2013 – Metallic Materials – Sheet and Strip – Erichsen Cupping Test, International
Organization for Standardization: Geneva, Switzerland, 2013.
Jatczak, C. F. Retained Austenite and its measurement by RRay diffraction, SAE Technical
paper series 800426, Society of Automotive Engineers, Congress and exposition, Detroit, Feb.
2529, 1980.
Jeswiet, J., Geiger, M., Engel, U., Kleiner, M., Schikorra, M., Duflou, J., Neugebauer, R. Metal
forming process since 2000, CIRP Journal of Manufacturing Science and Technology, vol.
1(2008), pp. 217.
Jiang, W., Li, Y., Su, J. Modified GTN model for a broad range of stress states and application
to ductile fracture. European Journal of Mechanics – A/Solids, vol. 57 (2016), pp. 132148.
Kami, A., Dariani, B. M., Vanini, A. S., Comsa, D. S., Banabic, D. Numerical determination of
the forming limit curves of anisotropic sheet metals using GTN damage model. Journal of
Materials Processing Technology, vol. 216 (2015), pp. 472483.
Kempken, J. Chapter 16 – Technological challenges and solutions for the production of state
oftheart second and thirdgeneration AHSS grades. In: Roy, T. K. Advanced High Strength
Steel, Lecture Notes in Mechanical Engineering, Dusseldorf, Germany: Springer, (2018), pp.
143148.
Kiran, R., Khandelwal, K., Gurson model parameters for ductile fracture simulation in ASTM
A992 steels. Fatigue & Fracture of Engineering Materials & Structures, vol. 37 (2014), pp.
7183.
Kossakowski, P. G. Simulation of ductile fracture of S235JR steel using computational cells
with microstructurallybased length scales. Journal of theoretical and applied mechanics, vol.
50 (2012), pp. 589607.
123
Krizan, D., Steineder, K., Kaar, S., Hebesberger, T. Development of third generation advanced
high strength steel for automotive applications. 19th International Scientific Conference
TRANSFER 2018, November (2018).
Kumar, S. D., Amjith, T. R., Anjaneyulu, C. Forming limit diagram generation of Aluminum
alloy AA2014 using Nakazima test simulation tool. Procedia Technology, vol. 24 (2016), pp.
386393.
Lazarescu, L., Nicodim, I., Ciobanu, I., Comsa, D. S., Banabic, D. Determination of material
parameters of sheet metals using the hydraulic bulge test. Acta Metallurgica Slovaca, vol. 19
(2013), pp. 412.
Lemiale, V., Chambert, J., Picart, P. Description of numerical techniques with the aim of
predicting the sheet metal blanking process by FEM simulation. Journal of Materials
Processing Technology, vol. 209 (2009), pp. 27232734.
Li, G., Cui, S. A review on theory and application of plastic mesodamage mechanics.
Theoretical and Applied Fracture Mechanics, vol.109 (2020), 102686.
Li, J., Li, S., Xie, Z., Wang, W., Numerical simulation of incremental sheet forming based on
GTN damage model. The International Journal of Advanced Manufacturing Technology, vol.
81 (2015), pp. 20532065.
Li, J., Song, R., Li, X., Zhou, N., Wang, Y. Coupling nanocarbide strengthening with
transformation induced plasticity effect to achieve over 1.5 GPa strength with 30% ductility in
coldrolled mediumMn steel. Vacuum, vol. 167 (2019), pp. 223226.
Li, Q., Huang, X., Huang, W. Microstructure and mechanical properties of a mediumcarbon
bainitic steel by a novel quenching and dynamic partitioning (QDP) process. Materials Science
& Engineering A, vol. 662 (2016), pp. 129135.
Li, W., Ma, L., Peng, P., Jia, Q., Wang, Z., Zhu, Y., Guo, W. Microstructural evolution and
deformation behavior of fiber laser welded QP980 steel joint. Materials Science & Engineering
A, vol. 717 (2018), pp. 124133.
Liu, M., Li, Y., Cui, Z., Yang, Q. High ductility of spray formed low density TRIP steel with
the improvement of δferrite matrix. Material Characterization, vol. 156 (2019), pp. 109828.
Liu, W. Q., Ying, L., Dai, M. H., Hu, P., Wang, D. T. Parameter calibration for a shear modified
GTN model and its application to forming limit prediction. Journal of Physics: Conference
Series, vol. 1063 (2018), pp. 012110.
Macek, B., Lutz, J. Virtual and physical testing of thirdgeneration high strength steel –
Evaluating a new highstrength steel’s ability to improve on an existing stampedsteel
production part. SAE International – Automotive Engineering. August 18th, (2020), Available
in: < https://www.sae.org/news/2020/08/stampingahsstechpaper>, accessed on: September
17th, 2020.
Maire, E., Bouaziz, O., Michiel, M. D., Verdu, C. Initiation and growth of damage in a dual
phase steel observed by Xray microtomography. Acta Materialia, vol. 56 (2008), pp. 4954
4964.
124
Mansouri, L. Z., Chalal, H., Meraim, F. A. Ductility limit prediction using a GTN damage
model coupled with localization bifurcation analysis. Mechanics of Materials, vol. 76 (2014),
pp. 6492.
Marciniak, Z., Duncan, J. L., Hu, S. J. Mechanics of Sheet Metal Forming, 2nd edition,
ButterworthHeinemann, 2002.
Marouani, H., Ismail, A. B., Hug, E., Rachik, M. Numerical investigations on sheet metal
blanking with high speed deformation. Materials & Design, vol. 30 (2009), pp. 35663571.
Matlock, D. K., Speer, J. G., Chapter 11 – Third Generation of AHSS: Microstructure Design
Concepts. In: Haldar, A., Suwas, S., Bhattacharjee, D., Proceedings of International Conference
on Microstructure and Texture in Steels and Other Materials, Jamshedpur, India: Springer,
2008, pp. 185205.
Mock, P., CO2 emission standards for passenger cars and lightcommercial vehicles in the
European Union, ICCT – The International Council on Clean Transportation, 2019.
NavarroLópez, A., Hidalgo, J., Sietsma, J., Santofimia, M. J. Characterization of
bainitic/martensitic structures formed in isothermal treatments below the Ms temperature.
Materials Characterization, vol. 128 (2017), pp. 248256.
Needleman, A., A numerical study of necking in circular cylindrical bars. Journal of the
Mechanics and Physics of Solids, vol. 20 (1972), pp. 111127.
Nguyen, H. H., Nguyen, T. N., Vu, H. C. Ductile fracture prediction and forming assessment
of AA6061T6 aluminum alloy sheet. International Journal of Fracture, vol. 209 (2018), pp.
143162.
OriginLab, Origin User Guide OriginLab Corporation, Northampton, USA, 2017.
Paul, S. K. A critical review on role expansion ratio. Materialia, vol. 9 (2020) 100566.
Paykani, M. A., Shahverdi, H. R., Merismaeilli, R. First and third generation of advanced high
strength steel in FeCrNiBSi system. Journal of Materials Processing Technology, vol. 238
(2016), pp. 383394.
Rachik, M., Roelandt, J. M., Maillard, A. Some phenomenological and computational aspects
of sheet metal blanking simulation. Journal of Materials Processing Technology, vol. 128
(2002), pp. 256265.
Radwanski, K., Wrozyna, A., Kuziak, R. Role of advanced microstructures characterization in
modeling of mechanical properties of AHSS steels. Materials Science and Engineering: A, vol.
639 (2015), pp. 567574.
Requena, G., Maire, E., Leguen, C., Thuillier, S. Separation of nucleation and growth of voids
during tensile deformation of a dual phase steel using synchrotron microtomography. Materials
Science & Engineering A, vol. 589 (2014), pp. 242251.
125
Saeidi, N., Ashrafizadeh, F., Niroumand, B., Forouzan, M. R., Barlat, F. Damage mechanism
and modeling of void nucleation process in a ferritemartensite dual phase steel. Engineering
Fracture Mechanics, vol. 127 (2014), pp. 97103.
Samei, J., Green D. E., Cheng, J., Lima, M. S. C. Influence of strain path on nucleation and
growth of voids in dual phase steels sheet. Materials & Design, vol. 92 (2016), pp. 10281037.
Sangkharat, T., Dechjarem, S. Using image processing on Erichsen cup test machine to
calculate anisotropic property of sheet metal. Procedia Manufacturing, vol. 29 (2019), pp. 390
397.
Santos, R. O., Pereira, A. B., Butuc, M., Vincze, G., Festas, A. J., Moreira, L. P. Development
of a device compatible with universal testing machine to perform hole expansion and Erichsen
cupping test. Machines, vol. 8 (2020) 2, pp. 18.
Santos, R. O., Silveira, L. B., Moreira, L. P., Cardoso, M. C., Silva, F. R. F., Paula, A. S.,
Albertacci, D. A. Damage identification parameters of dualphase 600800 steels based on
experimental void analysis and finite element simulations. Journal of Materials Research and
Technology, vol 8(1) (2019), pp. 644659.
Schmitt, W., Sun, D. Z., Blauel, J. G. Damage mechanics analysis (Gurson model) and
experimental verification of the behavior of a crack in a weldcladded component. Nuclear
Engineering and Design, vol. 174 (1997), pp. 237246.
Schwindt, C. D., Stout, M., Iurman, L., Signorelli, J. W. Forming limit curve determination of
a DP780 steel sheet. Procedia Materials Science, vol. 8 (2015), pp. 978985.
Shen, I. General Motors applies thirdgeneration advanced highstrength steel in new vehicle
for china – Open innovation enables fast adoption of lightweight technology. GM corporate
newsroom, Shanghai, January 1st , (2015), Available in:
https://media.gm.com/media/cn/en/gm/news.detail.html/content/Pages/news/cn/en/2015/dec/1
201_advancedsteel.html , accessed on: September 17th, 2020.
Speer, J., Matlock, D. K., Cooman, B. C., Schroth, J. G. Carbon partitioning into austenite after
martensite transformation. Acta Materialia, vol. 51 (2003), pp. 26112622.
Springmann, M., Kuna, M., Identification of material parameters of the GursonTvergaard
Needleman model by combined experimental and numerical techniques. Computational
Materials Science, vol. 32 (2005), pp. 544552.
Sugimoto, K., Hojo, T., Kobayashi, J. Critical assessment of TRIPaided bainitic ferrite steels.
Materials Science and Technology, vol. 33 (2017), pp. 20052009.
Sun, G. Q. Sun, F.Y., Cao, F. L., Chen, S. J., Barkey, M. E. Numerical simulation of tension
properties for AlCu alloy friction stir welded joints with GTN damage model. Journal of
Materials Engineering and Performance, vol. 24 (2015), pp. 43584363.
Sun, Q., Lu, Y., Chen, J. Identification of material parameters of a shear modified GTN damage
model by small punch test. International Journal Fracture, vol. 222 (2020), pp. 2535.
126
Teng, B., Wang, W., Xu, Y. Ductile fracture prediction in aluminium alloy 5A06 sheet forming
on GTN damage model. Engineering Fracture Mechanics, vol. 186 (2017), 242254.
Tevergaard, V., Needleman, A. Analysis of the cupcone fracture in a round tensile bar. Acta
Metallurgica, vol. 32 (1984), pp. 157169.
Wang, J., Kim, N., Lee, H. Ductile fracture in the shearing process of Zircaloy sheet for nuclear
spacer grids. Metals and Materials International, vol. 18 (2012), pp. 303316.
Wang, L., Speer, J. Quenching and Partitioning Steel Heat Treatment. Metallogr. Microstruct.
Anal., vol. 2 (2013), pp. 268281.
Wcislik, W. Experimental determination of critical void volume fraction fF for the Gurson
Tvergaard Needleman (GTN) model. Procedia Structural Integrity, vol. 2 (2016), pp. 1676
1683.
Wu, R. M., Wang, L., Jin, X. J., Thermal stability of austenite and properties of quenching &
partitioning (Q&P) treated AHSS. Physics Procedia, vol. 50 (2013), pp. 812.
Yan, Y., Sun, Q., Chen, J., Pan, H. The initiation and propagation of edge cracks of silicon steel
during tandem cold rolling process based on the GursonTvergaardNeedleman damage model.
Journal of Materials Processing Technology, vol. 213 (2013), pp. 598605.
Yi, H. L. Review on δTransformationInduced Plasticity (TRIP) steels with low density: The
concept and current progress. The Journal of The Minerals, Metals & Materials Society, vol.
66 (2014), pp. 17591769.
Yi, H. L., Lee, K. Y., Bhadeshia, H. K. D. H. Mechanical stabilization of retained austenite in
δTRIP steel. Materials Science and Engineering A, vol. 528 (2011), pp. 59005903.
Yi, L. H., Sun, L., Xiong, X. C. Challenges in the formability of the next generation of
automotive steel sheets. Materials Science and Technology, vol. 34 (2018), pp. 11121117.
Yildiz, R. A., Yilmaz, S. Experimental Investigation of GTN model parameters of 6061 Al
alloy. European Journal of Mechanics / A Solids, vol. 83 (2020), 104040.
Zhao, P.J., Chen, Z. H., Dong, C.F. Failure analysis based on microvoids damage model for
DP600 steel on insitu tensile tests. Engineering Fracture Mechanics, vol. 154 (2016), pp. 152
168.
Zhong, J., Xu, T., Guan, K., Zou, B. Determination of ductile damage parameters using hybrid
particle swarm optimization. Experimental Mechanics, vol. 56 (2016), pp. 945955.
Zhou, J., Gao, X., Sobotka, J. C., Webler, B. A., Cockeram, B. V. On the extension of the
Gursontype pourous plasticity models for prediction of ductile fracture under sheardominated
conditions. International Journal of Solids and Structures, vol. 51 (2014), pp. 32733291.
Zhou, M., Li, Y., Hu, Q., Li, X., Chen, J. Investigations on edge quality and its effect on tensile
property and fracture patterns of QP980. Journal of Manufacturing Processes, vol. 37 (2019),
pp. 509518.