Journal of Materials Processing Tech.

258 (2018) 310–318

Contents lists available at ScienceDirect

Journal of Materials Processing Tech.

journal homepage: www.elsevier.com/locate/jmatprotec

A quantitative description of machining effects to mechanical behavior of T

sintered powder metals
⁎
Long Zhanga, Huang Yuanb,
a
School of Mechanical Engineering, Beijing Institute of Technology, Beijing, China
b
School of Aerospace Engineering, Tsinghua University, Beijing, China

A R T I C LE I N FO A B S T R A C T

Keywords: Sintered powder metals play increasingly important role in industry. The most remarkable mechanical char-
Sintered porous iron acteristics of the sintered material is the high porosity, which characterizes deformations and failure of the
Machining effects material. It is confirmed additionally that manufacturing process affects mechanical behavior and fatigue per-
Micro-cracks formance of the material. In the present work a continuum damage mechanics model is used to describe the
Multi-axial loading
damage evolution in machined sintered iron. It is confirmed that machining effects are localized in the sub-
Continuum damage modeling
surface layer of the mechanical part, and the damage can be quantitatively described by the damage model. The
experimental results from the fabricated specimen have to be separated into mechanical behavior of the sub-
surface layer and the specimen core. The damage model provides an effective way to describe mechanical
performance of a machined part of the sintered iron.

1. Introduction by the porosity. Actually various mechanisms contribute to the damage

Reihanian et al. (2011) claimed sintered powder metallurgy is a
special technology with advantages in zero waste, accurate dimension
and artificial alloy components for near-net shape manufacturing of
complex structures. With improvement of powder metallurgy more
components made of sintered alloys are applied under complex and
high loading conditions. Mower (2014) and Salak et al. (2005) pre-
sented that for certain applications the sintered parts may need further
machining, such as drilling, milling etc. Due to the highly porous mi-
crostructure of the sintered powder metals the macroscopic mechanical
property depends on local void morphology and distribution. Ma-
chining can affect material property significantly.
Rausch et al. (2010) has mentioned that the microstructure of the
sintered metal is characterized by connecting voids and connecting
alloy particles, which is the most obvious feature of the sintered metal.
Schneider and Yuan (2012) showed that all voids are inter-connected
and form a 3D void mesh. The material can be seen as a mixture of the
3D void mesh and a 3D particle mesh. Fig. 1 shows the typical cross
section of the microstructure and morphology of the sintered metal. The
porosity of the material depending on machining parameters is varying
between 7% and 15% in weight. A higher porosity will make the ma-
terial useless. In comparison with casting metals with ca. 0.5% porosity,
the sintered metals contain generally high porosity. Gilmas et al. (2016)
believed that mechanical behavior of sintered metals is characterized
of sintered metals. Chen et al. (2013) revealed that plastic deformation
of the sintered metal consist of matrix particle distortion and separation
among particles. The latter takes 30–40% of the total plastic deforma-
tion.
Recent detailed investigation on sintered metals has shown sig-
nificant difference in deformation and damage mechanisms from cast or
forged metals. Computational and experimental analysis by Schneider
and Yuan (2012) confirmed that the damage process of the sintered
metals cannot be described by the known continuum damage models,
such as GTN model. Due to the complexity of micro-structure and high
initial porosity the material degradation cannot be represented by the
porosity. For this reason Ma and Yuan (2013) proposed a continuum
damage model that both the elastic damage driven by stress and the
plastic damage driven by plastic strain constitute the total damage.
Arunachalam et al. (2004) has reported in cast or forged metals the
influences are mainly limited in residual stresses as well as surface
roughness, respectively. Saleem et al. (2013) argued that the damage
arising from different machining processes for the same material shows
an obvious distinction in cyclic loading condition. Moreover, machining
process is able to affect material microstructure, such as void mor-
phology, grain orientations etc., so that the material property is
changed by machining. For instance in titanium alloys Zhang et al.
(2014) found that machining can induce phase transformation, so that
the part surface shows significantly different mechanical property from

⁎
Corresponding author.
E-mail address: yuan.huang@tsinghua.edu.cn (H. Yuan).

https://doi.org/10.1016/j.jmatprotec.2018.04.003
Received 1 March 2018; Accepted 3 April 2018
Available online 11 April 2018
0924-0136/ © 2018 Published by Elsevier B.V.
L. Zhang, H. Yuan Journal of Materials Processing Tech. 258 (2018) 310–318

Nomenclature σ͠ eq effective stress

σH hydrostatic stress
D damage variable σy0 initial yield stress
De elastic damage variable σu tensile stress
Dp plastic damage variable σij stress tensor
Desa saturation of elastic damage εij strain tensor
Dpcr critical plastic damage εcr tensile fracture strain
0
DMAZ initial damage εth threshold strain
DMAZ damage of machining affected zone p plastic strain
DCORE damage of core area pcr plastic fracture strain
De0 initial elastic damage p0 initial plastic strain
Dp0 initial plastic damage K coefficient of Ramberg–Osgood model
E0 initial elasticity modulus n strain-hardening exponent of Ramberg–Osgood model
∼
E equivalent elasticity modulus b parameter of elastic damage model
H1 area weight of core area α exponent of plastic damage model
H2 area weight of machining affected zone η stress triaxiality
FeD potential function for elastic damage η0 initial stress triaxiality
FpD potential function for plastic damage Y energy density release rate
R(r) plastic strain hardening Y0 threshold energy density release rate
ψp plastic energy dissipation Z material resistance with damage
σeq Misses stress

the core material. Generally, machining procedure is a significant factor
to the behavior of the mechanical part.
Since the sintered metal is softer than the corresponding cast or
forged material, the conventional turning and drilling can affect the
void morphology as well as the microstructure of particles. Beiss (2013)
has reported that the mechanical properties such as elasticity modulus,
yield stress and fatigue strength can be changed by machining process.
The empirical correlation of the material behavior for different porous
sintered metals can be expressed in a power-law of the mass density,
which is related with the porosity directly. It is essential for engineering
to represent and to quantify the machining effects to the damage evo-
lution and mechanical property of the sintered metals. Desbiens et al.
(2012) and Robert et al. (2013) investigated the influence of machining
and sintering on tensile properties and fatigue strength of sintered alloy
FLC-4608 and confirmed that the sequence of machining and sintering
influenced the mechanical behavior of sintered alloy. Robert et al.
(2007) proposed drilling parameters for green PM parts and argued that
machining process of green parts had no effect on the final dimension.
Veldhuis et al. (2010) found that the micro-hardness in machined sur-
faces of the sintered nickel base superalloy ME16 decreased with
turning speed. Kwong et al. (2009) studied drilled and milled surfaces
of the sintered alloy RR1000 and observed that the residual stress and
the machining affected zone grew with material removal rate. Alao and
Fig. 1. Microstructure of the sintered iron with ca. 8% porosity. The powder
particle size are 50–100 μm.
Yin (2016, 2015) investigated machining induced cracks in sintered
zirconia and characterized elastic-plastic deformations of the material.
The sintered metals are influenced by machining significantly. Quan-
tifying machining effects is of importance for designing and manu-
facturing mechanical parts of sintered metals.
In the present work, the damage evolution of the machined sintered
metal is investigated under uniaxial and multi-axial loading conditions.
The aim of the work is to find a quantitative method to describe me-
chanical behavior of machined specimens. Based on experimental and
computational works the initial damage from machining is considered
explicitly under multi-axial loading condition. A method is developed
based on the continuum damage model to quantify the damage evo-
lution in machined sintered material. Machining effects depend on
loading configuration and are integrated into the continuum damage
framework.
2. Materials and experiments
In the present work iron powders of Hoeganaes Corporation are
adopted for sintered disks manufacturing by the GKN Sinter Metals
GmbH, Germany. Schneider and Yuan (2012) confirmed that the fun-
damental mechanical property of sintered iron has no influence with
powder size. The lubricant has been developed by adding 0.6% HDL-
wax uniformly into the raw powder mixture during compacting process.
Then the green material has been sintered at 1120 °C for 60 min around
by 95% N2–5% H2. During sintering the wax is eliminated by the high
temperature. The density of the sintered iron is 7.2 g/cm3, which is
lower than that of the regular iron because of the porosity. According to
engineering experience 7.2 g/cm3 is the highest reachable density for
the sintered iron, which means a porosity of ca. 8% in weight.
Quasi-static experiments (the strain rate is 10−4/s) are controlled
by an extensometer with a gauge length of 25 mm at room temperature.
Small loading-unloading cycles are introduced to quantify change of the
elastic stiffness of the specimen and to determine the variation of
Young's modulus. Because the tubular is thin-walled specimen, the
stress gradients are negligible.
To study effects of machining, two kinds of specimens are fabricated
in the present work (Fig. 2). The rectangular specimen represents the
originally sintered metal without machining and is used to identify
basic mechanical property. Effects of complex loading paths are tested
in the tubular specimens, as shown in Fig. 2(b). The tubular specimen
cannot be sintered directly and have to be fabricated from sintered

311
L. Zhang, H. Yuan Journal of Materials Processing Tech. 258 (2018) 310–318

Fig. 2. Specimens tested in the present work. (a) The tensile specimen without machining. (b) The tubular specimen with machining.

disks. To study effects of manufacturing, two various processes are
adopted:
• Conventional Machining in the following sections denotes conven-
tional drilling and turning. The specimen are mechanically cut from
sintered disks and drilled at first. Then the accurate dimension of the
specimen is turned and drilled with a cutting speed of 20–50 m/min
and a feed rate of 0.6–1.0 mm/r by coated carbide tool. During the
machining the specimen is sufficiently cooled so that the material
should not strongly be affected by heating. From microscopic ob-
servations the local heat influence seems not obvious. Due to high
porosity in the sintered material, the turning and drilling processes
severely distort the surface material and change its mechanical be-
havior. The present work should develop a method based on con-
tinuum damage mechanics to quantify machining effects.
• WEDM process stands for wire electrical discharge method, in which
the specimen is cut into the form of a tube. The final shape of the
specimen is approached by fine turning and grinding. Effects of the
recasting from the wire electrical discharge are removed in the final
machining step. Ma and Yuan (2013) reported that the wire elec-
trical discharge has almost no influence to test results. All the ten-
sion-torsion stress-strain curves as well as fatigue performance
reached practically the same results as the sintered material.
However, conventional turning and drilling show significant ma-
chining effects to the sintered metals. To quantify machining effects,
additional tubular specimens were fabricated by drilling the hole and
then turning external surfaces. The fabrication of the tubular specimen
induces material damage into the specimen surfaces and affects test
results. Micro-cracks from machining internal and external surfaces are

Fig. 3. Micro-cracks and distortions induced by conventional machining process (turning and drilling) in the cross-section.

312
L. Zhang, H. Yuan Journal of Materials Processing Tech. 258 (2018) 310–318

summarized in Fig. 3. From Fig. 3(a) and (b) it is clear that cracks were plastic damage,
nucleated in nearly 45° direction which implies shear stress damage of
D˙ = D˙ e + D˙ p , (2)
the machining tools. The maximum length of the cracks is 100–300 μm.
The surface cracks change mechanical property of the surface material where De and Dp stands for elastic damage and plastic damage respec-
layer and reduce the loading capacity of the specimen. tively. The elastic damage driven by the stress occurs in elastic state,
Micro-cracks from the conventional machined internal and external whereas the plastic damage is dominated by the plastic strain.
surfaces are summarized in Fig. 3. From the figure it is clear that cracks
are nucleated in nearly 45° direction which implies shear stress damage 3.2. Elastic damage evolution
of the machining tools. The maximum length of the cracks is
100–300 μm. The surface cracks change mechanical property of the Within the framework of thermodynamics Voyiadjis and Kattan
surface material layer and reduce the loading capacity of the specimen. (2008) defined D was related to the strain energy density release rate Y,
The test results from such specimens contain the surface damage from
2
machining, which is representative for machined mechanical part. 1∼ e e σ͠ eq
Y= Eijkl ϵ ijϵ kl = f (η),
Besides microcracks on the surfaces, Fig. 3(c) and (d) have respec- 2 2E0 (3)
tively shown that turning and drilling processes introduce strong dis-
where σ͠ eq represents the effective stress, and it is defined by,
tortions near the surfaces. In the figures it is obvious that the direction
of distortion is uniform depending on the machining procedure. This σij
σ͠ ij = .
kind of distortions changes the microstructure including shapes of (1 − D) (4)
particles and pores. The porosity near the surfaces differs from the Besides, in Eq. (3) f(η) is a function representing the influence of the
original one. The microscopy of the axial cut surface is shown in Fig. 4 stress triaxiality, η = σH/σeq, with σH as the hydrostatic stress and σeq as
and demonstrates that the machining process has induce grooves on the the Misses stress of the nominal stress state.
surface. The porosity around the grooves seems obviously lower than The elastic damage of sintered iron is related to a potential function
that near the core. Few micro-cracks are in the perpendicular direction defined by Lemaitre and Desmorat (2005) and determined as,
to the surface. Damage in the longitudinal plane is not so significant as
that in the cross-section plane. FeD = Y − Z (De ), (5)
with Z represents the material resistance against the damage. The po-
3. Continuum damage model (CDM) for sintered iron tential function predicts the initiation and growth of elastic damage at a
low stress state. Cicekli et al. (2007) have defined Z as,
3.1. Damage variables
1 ⎛ Desa ⎞
Z = Y0 + ln ⎜ ⎟,

For the sintered metal Ma and Yuan (2013, 2015, 2017) suggested a b ⎝ Desa − De ⎠ (6)
continuum damage model (CDM) within the frame of thermodynamics.
where Y0 is the initial energy density release rate as a threshold value, b
In analogy to the definition of Kachanov, the damage variable is
is a material constant and Desa represents the elastic damage saturation.
identified as degradation of the elastic modulus, i.e.
And by the rule of the maximum dissipation, the elastic damage evo-
∼ lutes as following,
E
D=1− ,
E0 (1) ∂F D
D˙ e = λ˙ b e = λ˙ b .
∼ ∂Y (7)
where E denotes to the equivalent elastic modulus in the maximum
principal strain direction under current loading conditions and E0 re- And the consistency relation has defined the multiplier λ̇b as,
presents the initial elastic modulus. In experiments the current elastic
D ∂F D ∂Z ˙
modulus is determined in the stage of unloading. F˙ e = e Y˙ − λb = 0.
∂Y ∂D (8)
It is confirmed that the sintered iron is damaged in a low stress level,
so that the damage has to be decomposed into elastic damage and According to the equations above, the elastic damage can be written as,

Fig. 4. Micro-cracks induced by conventional machining process in the longitudinal cut section. (a) The damaged external surface turned. (b) The damaged surface of
the drilled hole.

313
L. Zhang, H. Yuan Journal of Materials Processing Tech. 258 (2018) 310–318

1 K2 p
(Dpcr )1/ α = f (η)ln ⎛ cr ⎞
⎜ ⎟

α 2E0 S0 ⎝ εth ⎠ (17)

where εth, pcr represents a threshold value of strain for the initiation of
plastic damage and a critical plastic strain matching to the critical value
of plastic damage respectively.
Considering the situation of uniaxial tensile loading f(η) = 1 and the
fracture strain pcr=εcr, the equation above can be rewritten into,

1 K2 ε
(Dpcr )1/ α = ln ⎛ cr ⎞.
⎜ ⎟

α 2E0 S0 ⎝ εth ⎠ (18)

Integrating Eq. (16) over [Dp , Dpcr ] for D and [p, pcr] for p results in,

1 K2 p
(Dpcr − Dp)1/ α = f (η)ln ⎜⎛ cr ⎟⎞.
α 2E0 S0 ⎝ p⎠ (19)

Combining equations above follows an explicit expression for plastic

Fig. 5. Comparison of uniaxial tensile results of the three differently machined damage as,
specimens. The stress is nominally defined as F/A0 and the nominal strain is Δl/
ln(p / εth )
l0. Dp = Dpcr ⎧1 − [1 − f (η )]α ⎫ .
⎨
⎩ ln(εcr / εth ) ⎬
⎭ (20)

D˙ e = Desa b exp[−b Y − Y0 ] Y˙ , (9) The expressions above are only valid for the sintered material without
initial damage.
where 〈 · 〉 are Macaulay brackets. Integrating Eq. (9) over [Y0, Y] for Y,
the explicit elastic damage expression can be obtained as,
4. Damage evolution in machined specimens
De = Desa {1 − exp[−b Y − Y0 ]}. (10)
4.1. Mechanical behavior of machined specimens

3.3. Plastic damage evolution It is known that machining can induce damage into specimens.
Fig. 5 shows tensile test results of three differently fabricated speci-
It is assumed that the plastic damage in the sintered metal is mainly mens: Convent. Machined denotes the tubular specimen conventionally
induced by the plastic deformation, which is common for ductile me- drilled and turned, as shown in Fig. 2(b). Base Material stands for the
tals. The plastic damage dissipation F is defined by two items as, specimen without machining (Fig. 2(a)). Ma and Yuan (2013) observed
F = FpD + ψp , WEDM specimen does not show machining effects, while the conven-
(11)
tional machining obviously induced plastic deformations and dramati-
where and ψp represent the plastic energy with damage and the
FpD cally reduced fracture strain of the specimen (Fig. 5).
plastic energy dissipation, respectively. Abu Al-Rub and Kim (2010) Further comparison among the three specimens is summarized in
introduced that the yield function can be defined as, Table 1. In the table all values are calculated based on the nominal
ψp = σ͠ eq − σ͠ y0 − R (r ) dimension of the specimens. Local damage and local changes in the
(12)
specimen from machining are not considered. From the table the initial
and σeq denotes the Misses stress. Bonora (1997) has defined the da- Young's modulus, E, and the initial yield stress, σy0, show a tiny dis-
mage dissipation potential as, tinction after various machining processes, which means machining
2 cr (α − 1)/ α
affects specimen performance only locally. The initial properties of the
1 ⎛ Y ⎞ S0 ⎤ (Dp − Dp)
FpD = ⎡ ⎜ ⎟
⎢ 2 S0 1 − D ⎥ (2 + n )/ n
, specimen are almost independent of machining. Nevertheless, the ul-
⎣ ⎝ ⎠ ⎦ p (13) timate tensile stress, σu, and the fracture strain, ϵf, differs from each
other. It illustrates that the specimen performance related with damage
where S0 is a parameter for damaged material, α is the damage ex-
and fracture is influenced by machining. It can be concluded that the
ponent describing the plastic damage curve, n is the hardening para-
fracture strain reduction is caused by the machining damage, as shown
meter and p is the accumulated plastic strain.
in Fig. 3. Microcracks on the machined surfaces become initial crack
Bonora et al. (2006) claimed that the plastic damage accumulates
sites for the specimen. The final failure of the specimen is directly re-
when the plastic strain p reaches the threshold ϵth in the condition of the
sulted from crack growth, without crack nucleation.
positive hydrostatic stress, and it is identified by,
More detailed tests are performed on the machined specimens to
cr ∼
Y (Dp − Dp)
(α − 1)/ α quantify damage evolution. In Fig. 6(a) the actual elastic modulus E is
D˙ p = (2 + n )/ n
p˙ . determined for both conventionally as well as WEDM machined speci-
S0 p (14)
mens and plotted as a function of the nominal strain ε. The WEDM
Combining the Ramberg–Osgood law and Eq. (4) the definition of the specimen provides practically the same results as from the base mate-
effective stress can be described as, rial. However, the machining effect to the evolution of Young's modular
σeq
σ͠ eq = = Kp1/ n . Table 1
(1 − D) (15)
Comparison of nominal mechanical properties of the three variously machined
Eq. (14) can be induced as with the above equation, specimens.
cr (α − 1)/ α Specimen Types E (GPa) σy0 (MPa) σu (MPa) ϵf (%)
K 2 (Dp − Dp)
D˙ p = p˙ .
2E0 S0 p (16) Base material 162 135 262 12.7
WEDM 162 135 265 13.3
The critical value of plastic damage Dpcr
in the sintered iron can be Convent. machined 160 134 248 4.5
obtained by integrating Eq. (16) over [0, Dpcr ] for D and [εth, pcr] for p,

314
L. Zhang, H. Yuan Journal of Materials Processing Tech. 258 (2018) 310–318

Fig. 6. Comparison of Young's modulus and damage variable under uniaxial tensile loading with various machining states.

is significant for conventional machining. Using the definition of the for both core and MAZ, Dcore and DMAZ can be defined correspondingly.
damage indicator D in Eq. (1), the damage evolution for both specimens The damage evolution of both Dcore and DMAZ can be described by the
is shown in Fig. 6(b). Even though both damage evolution curves share damage model discussed in the previous section. However, DMAZ starts
the similar shape, damage in the conventionally machined specimens is from an initial damage induced by machining. The nominal damage for
obviously higher than that of the WEDM. Although the initial damage a machined specimen can be expressed as,
value for the conventionally machined specimen cannot be accurately
D = H1 D MAZ + H2 Dcore . (21)
measured in the test, both figures generate a clear impression about
machining damage to the specimen. Above H1 and H2 are the area weight for MAZ as well as the core, re-
spectively, and have to be identified from tests and simulations. DMAZ
4.2. CDM modeling of machined specimens represents for the damage of the machined material, while Dcore denotes
the damage of the base material without machining effects. During
From microscope microcracks can be clearly found near machined turning and drilling, micro-cracks occur on both internal and outside
surfaces, as shown in Fig. 3, and the mechanical behavior of the near surface of the specimen, which results in initial damage and initial
surface zone is changed by machining. The machining affected zone plastic deformation in MAZ. As in Eq. (2), the initial damage from
(MAZ) is always located directly under the machined specimen sur- machining can be divided into elastic damage De0 and plastic damage
faces, the microcracks and the MAZ thickness must be much smaller Dp0 ,
than the specimen thickness and are related to machining parameters. 0
DMAZ = De0 + Dp0 . (22)
Based on discussions in the previous sections, the damage evolution in
specimens is affected by machining and even by the specimen size. For Since the initial plastic strain in MAZ has been much higher than the
a large specimen, the effects of a given MAZ can be smaller than those elastic limitation, the initial elastic damage De0 must have reached the
in a small specimen by the identical machining procedure. To describe saturation with the maximum Desa .
influence of machining, a general approach based on continuum da- The initial plastic damage in MAZ, Dp0 , is accompanied by the initial
mage mechanics is necessary. plastic strain, p0. Integrating Eq. (16) over [Dp0 , Dpcr ] for D and [p0,
A machined specimen contains generally a MAZ near the surface
and the core, in which the material behavior is not affected by the
machining. Schematically the specimen cross section can be re-
presented by two different zones: The machining affected zone (MAZ)
and the core zone, as shown in Fig. 7. The mechanical property of the
core material can be described by the constitutive model with CDM,
whereas MAZ contains microcracks and is damaged during machining.
Since MAZ was the same material before machining, it can be described
as core area by the constitutive model. To describe the machining effect
on the specimen, the initial damage can be found in MAZ. The concept
to give a reliable description of the machined specimen is based on the
quantification of mechanical behavior of the MAZ.
To establish a constitutive description of the MAZ, the machining
affected zone has to be taken as homogeneous so that the microcracks
will not be considered deterministically. For the present tubular spe-
cimen the MAZ consists of two rings near the machined surfaces. The
area fraction of the MAZ is defined as H1 = AMAZ/A, where AMAZ de-
notes the cross section area of the machining affected zone and A is the
total cross section area. The area fraction of the core is H2 = 1 − H1.
Both H1 and H2 are less than 1 and represent the area weights for MAZ Fig. 7. Schematic representation of the machining affected zone (MAZ) near
and the core, respectively. machined specimen surfaces. Obviously, the MAZ size depends on machining
By taking the damage definition introduced in Eq. (1), the damage parameters and must be much less than the specimen thickness.

315
L. Zhang, H. Yuan
pcr + p0] for p follows,
1 K2 p + p0 ⎞
(Dpcr − Dp0)1/ α = f (η0)ln ⎜⎛ cr ⎟.
α 2E0 S0 ⎝ p0 ⎠
The equation above states the relation between initial plastic damage
Dp0 and initial plastic strain p0 as well as the initial stress triaxiality η0. In
the integration variations of η are neglected. For uniaxial tensile tests η
is constant and equals to 1/3.
Assuming the initial plastic damage can be represented as uniaxial
tensile loading, it follows f (η0)=1. Eq. (23) can be simplified into,
1 K2 εcr + p0 ⎞
(Dpcr − Dp0)1/ α = ln ⎜⎛ ⎟,
α 2E0 S0 ⎝ p0 ⎠
with εcr as measured fracture strain under tensile loading condition for
specimens with machining effects. Integrating Eq. (16) over [Dp , Dpcr ]
for D and [p + p0, pcr + p0] for p gives,
1 K2 p + p0 ⎞
(Dpcr − Dp)1/ α = f (η)ln ⎛⎜ cr ⎟.
α 2E0 S0
⎝ p + p0 ⎠
Combining Eq. (25) with Eq. (24) follows an explicit expression for the
plastic damage with machining effects as,
α Machining of the tubular specimen is uniform and the initial damage in
ln[(p + p0 )/ p0 ]
Dp = Dpcr − (Dpcr − Dp0) ⎡1 − f (η) ⎤ .
⎢ ln[(εcr + p0 )/ p0 ] ⎥
⎣ ⎦
In the equations above parameters such as Dpcr , K, S0 and α are for the
base material and can be taken from experiments of the base material.
In MAZ the machining effect acts on the initial damage through the
initial plastic strain. Therefore, the parameters with respect to the in-
itial plastic damage Dp0 and the initial plastic strain p0 have to be
identified from experiments of the machined specimens.
4.3. Parameter identification from uniaxial tensile tests
The fundamental assumption of the present paper is the sintered
iron can be described by the unique constitutive model. The machining
effects can be represented by introducing initial plastic deformations
and initial damage in the model. Based on such consideration model
parameters for the sintered iron without machining effects can be taken
from Ma and Yuan (2013) and are summarized in Table 2. All the
parameters are identified from tension and compression tests by WEDM
specimens. The model can predict material behavior under most gen-
eral multi-axial loading conditions.
The model parameters for machining effects in the present work
include initial damage DMAZ, initial plastic strain p0, and fraction of the
machining affected zone H1. The sintered iron is damaged by tensile
stress, since the damage generally occurs in the maximum principal
strain plane. The machining damage can be taken as tensile damage
with η = 1/3 (f(η0) = 1), and the initial plastic damage and initial
plastic strain are related via Eq. (23).
From tensile tests of the machined and base specimens, the differ-
ence of the fracture strain is approximately 0.082, which can be taken
as the initial plastic strain p0. From Eq. (23) the initial plastic damage
Dp0 can be calculated. The initial elastic damage reaches the saturation
De0 = Desa , which is identified by the uniaxial test of the base material
and listed in Table 2. It follows the total initial damage from machining.
Finally H1 has to be determined from the experimental curve and the
fraction of MAZ in the cross section. Parameters are summarized in
Table 3.
Based on the parameters summarized in Tables 2 and 3, the damage
evolution in uniaxial tensile tests with and without machining effects is
illustrated in Fig. 8. In the figure symbols denote experimental records,
whereas the curves represent predictions from the CDM model. Com-
parison between the experiments and predictions shows that small
deviations exist in the beginning phase, while the elastic damage
Journal of Materials Processing Tech. 258 (2018) 310–318
dominates. Generally, for the small damage region measuring damage
value may contain large relative errors. Generally, the CDM description
gives an acceptable agreement to experiments.
(23)
4.4. Prediction of damage evolution under multi-axial loading conditions
To verify the CDM, more tests on the tubular specimen under multi-
axial loading conditions are performed. Tensile-torsional tests are
controlled by a given loading biaxiality, ε/γ, as shown in Fig. 9, where ε
and γ denotes the axial strain and shear strain respectively. ε/γ =∞ is
the uniaxial tension and ε/γ = 0 denotes the pure torsion. As known,
loading biaxiality is directly related to the stress triaxiality η which
(24) influences the damage evolution, as shown by f(η).
Damage evolution from both experiments and CDM model with
machining effects is plotted in Fig. 10 for different loading paths. Note
that the model parameters for the multi-axial tests are identified from
the uniaxial tests, so the multi-axial influence to damage evolution is
purely predicted by the CDM. The result in Fig. 10 shows a reasonable
agreement for all tested loading paths and stress triaxilities.
(25)
Computational results match experiments for all loading paths ex-
cept the pure torsional test. It can be seen that the deviation between
the pure torsional test and the CDM prediction is more than others.
all specimens can be assumed as constant. If the machining can be re-
(26) garded as the first loading step, the test is the second loading step for
specimens. Different loading path means a different direction of da-
mage accumulating in comparison with the first loading step.
Representation of the damage evolution is still a topic expecting further
detailed investigation. In the present work, the damage is treated as a
scalar so that the damage direction can only be represented in an
averaged form. Sudden changes in loading direction cannot be con-
sidered accurately. In drilling and turning the material is mainly cut
due to the high localized tensile strain, in this sense the material is
mainly damaged by the tensile stress from machining. It follows that
the damage from torsion in the tests cannot accurately be represented
by the CDM model at the beginning. With progress of the test, the in-
fluence of initial damage and initial deformation has disappeared gra-
dually and the CDM prediction is controlled by the damage evolution
from the test loading.
4.5. Machining effects on fracture strain
Due to initial damage in the machined specimen the final failure
occurs generally with a less tensile strain. More results can be found in
Fig. 10. In comparison with the results from WEDM specimens, con-
ventionally machined specimens reveal significantly less fracture strain.
The stress triaxiality is not related to the machining process, so the
reduction of the fracture strain is duo to the initial plastic strain and
effectively independent of the loading configuration.
Ma and Yuan (2013) developed an estimate based on maximum
damage value measured from tests for the sintered iron. Since the
maximum damage for the sintered iron is hardly affected by the loading
path, the fracture strain can be determined from the CDM directly. The
CDM result agrees with experiments excellently, as shown in Fig. 11.
Table 2
CDM parameters for the sintered iron without machining effects.
Elastic/plastic properties
E0 = 162,000 MPa; ν = 0.27; σy0 = 135 MPa
n = 4.714; K = 549 MPa
Elastic damage evolution law
Desa = 0.091; b = 19.961; Y0 = 0.04 MPa
Plastic damage evolution law
ϵth = 0.00038; α = 0.47
ϵcr = 0.12522; Dpcr = 0.205; Dcr = 0.296
316
L. Zhang, H. Yuan
Table 3
Model parameters for the sintered iron with the machining effect.
Machining effected zone fraction
H1 = 0.22; H2 = 0.78
Initial damage under tension assumption (η0 = 1/3)
p0 = 0.082; Dp0 = 0.1396
Initial damage under shear assumption(η0 = 0)
p0 = 0.082; Dp0 = 0.075
Fig. 8. Experimental and computational results of sintered iron with/without
machining effects. Symbols denote experimental results and curve denotes
computational results.
Fig. 9. Loading paths of multi-axial tests for verifying the CDM model.
For the machined specimen the damage distributed non-uniformly.
Furthermore, the damage evolution laws in the core and MAZ are dif-
ferent so that the fracture strain cannot be directly estimated from the
damage.
However experimental results show that the fracture strain of the
machined specimens can be effectively shifted from the base material
curve, that is,
pcrM = pcrB − p0 .
Above PcrB denotes the fracture strain of the base material without ma-
chining effects, whereas PcrM represents the fracture strain of the ma-
chined specimen. p0 is the plastic strain from machining introduced in
the CDM modeling. The results in Fig. 11 shows an acceptable agree-
ment between experiments and predictions for the whole loading
biaxiality.
Journal of Materials Processing Tech. 258 (2018) 310–318
Fig. 10. Experiments vs CDM prediction of damage evolution on machined
tubular specimens under multi-axial loading conditions. Symbols denote ex-
perimental results and curves are CDM predictions.
Fig. 11. Comparison of relationship between critical plastic strain and stress
triaxiality with and without machining effect.
4.6. Effects of the initial damage assumption
In the previous section it is assumed that the initial damage from
machining is tensile. However, the tensile stress from a turned and
drilled specimen is local, the overall loading configuration for the
specimen seems to be torsional. It becomes interesting to study effects
of the initial damage assumption to the damage evolution in machined
specimens.
With η0 = 0 for the pure torsional initial damage the model para-
meters can be determined, as summarized in Table 3. Using the CDM
one finds evolution of damage in different loading paths, as shown in
Fig. 12. Since the pure torsional curve is served for the parameter
identification, the torsion computation agrees with the experiment well.
Deviations for tension-dominate curves seem more significant, in
comparison with Fig. 10. Even though the prediction from the torsional
initial damage shows a more conservative result than tensile initial
damage assumption under multi-axial loading paths, the different as-
(27) sumptions of the initial damage form seem not to change the prediction
substantially.
5. Conclusions
In the present work the damage evolution of sintered iron specimens
with machined effects is studied experimentally and computationally.
The machining effects to the sintered iron specimen are represented in
317
L. Zhang, H. Yuan
Fig. 12. Damage prediction by the CDM model based on the shear damage
assumption for machining effects. Symbols denote experimental results and
solids curves are CDM predictions for the multi-axial loading.
the CDM model by introducing a machining affected zone and initial
damage into the specimen. This method can directly be applied to more
general machined mechanical parts. The following conclusions can be
drawn:
• Machining affects mechanical behavior of the material in the surface
layer and can be characterized by the machining affected zone. The
MAZ size should be determined mainly by the base material beha-
vior and machining process. The core material under the MAZ is not
changed by machining.
• The mechanical behavior of the MAZ material is characterized by
the initial damage and initial plastic strain in the continuum damage
model. The overall mechanical behavior of the machined specimen
is a superposition of the MAZ and core material. The superposition
model provides a reasonable prediction of both damage evolution
and material failure for all multi-axial loading conditions.
• The machining damage can be described by the tensile damage
model, while the shearing model gives a slightly different results,
especially for the tension loading. The damage model can give a
reasonable prediction of the machining effects in the machined
parts.
• The fracture strain of the machined specimen is determined by the
MAZ failure and described by the plastic strain reduction.
Acknowledgements
This work was supported by the National Natural Science
Foundation of China (Grant No. 11572169 and 51775294).
318
Journal of Materials Processing Tech. 258 (2018) 310–318
References
Abu Al-Rub, R.K., Kim, S.-M., 2010. Computational applications of a coupled plasticity-
damage constitutive model for simulating plain concrete fracture. Eng. Fract. Mech.
77 (10), 1577–1603.
Alao, A.R., Yin, L., 2015. Nanoindentation characterization of the elasticity, plasticity and
machinability of zirconia. Mater. Sci. Eng. A 628, 181–187.
Alao, A.R., Yin, L., 2016. Assessment of elasticity, plasticity and resistance to machining-
induced damage of porous pre-sintered zirconia using nanoindentation techniques. J.
Mater. Sci. Technol. 32 (5), 402–410.
Arunachalam, R.M., Mannan, M.A., Spowage, A.C., 2004. Residual stress and surface
roughness when facing age hardened inconel 718 with CBN and ceramic cutting
tools. Int. J. Mach. Tools Manuf. 44 (9), 879–887.
Beiss, P., 2013. Pulvermetallurgische Fertigungstechnik. Springer Berlin Heidelberg.
Bonora, N., 1997. A nonlinear CDM model for ductile failure. Eng. Fract. Mech. 58 (1),
11–28.
Bonora, N., Ruggiero, A., Esposito, L., Gentile, D., 2006. CDM modeling of ductile failure
in ferritic steels: assessment of the geometry transferability of model parameters. Int.
J. Plast. 22 (11), 2015–2047.
Chen, J., Yuan, H., Schneider, M., 2013. Investigation of micromechanical deformation
mechanisms in sinter powder metals. Adv. Mater. Res. 668, 351–355.
Cicekli, U., Voyiadjis, G.Z., Abu Al-Rub, R.K., 2007. A plasticity and anisotropic damage
model for plain concrete. Int. J. Plast. 23 (10-11), 1874–1900.
Desbiens, J., Robert-Perron, E., Blais, C., Chagnon, F., 2012. Effect of green machining on
the tensile properties and fatigue strength of powder metallurgy sinter-hardenable
steel components. Mater. Sci. Eng. A 546, 218–222.
Gilmas, M., Chottin, J., Dougan, M.J., Hug, E., 2016. Evolution of damage and fracture in
two families of niccucmo sinter-hardened steels with various initial porosities. Mater.
Sci. Eng. A 654, 85–93.
Kwong, J., Axinte, D.A., Withers, P.J., 2009. The sensitivity of Ni-based superalloy to hole
making operations: influence of process parameters on subsurface damage and re-
sidual stress. J. Mater. Process. Technol. 209 (8), 3968–3977.
Lemaitre, J., Desmorat, R., 2005. Background on Continuum Damage Mechanics.
Ma, S., Yuan, H., 2013. Damage evolution and modeling of sintered metals under multi-
axial loading conditions. Comput. Mater. Sci. 80, 123–133.
Ma, S., Yuan, H., 2015. Computational investigation of multi-axial damage modeling for
porous sintered metals with experimental verification. Eng. Fract. Mech. 149,
89–110.
Ma, S., Yuan, H., 2017. A continuum damage model for multi-axial low cycle fatigue of
porous sintered metals based on the critical plane concept. Mech. Mater. 104, 13–25.
Mower, T.M., 2014. Degradation of titanium 6Al-4V fatigue strength due to electrical
discharge machining. Int. J. Fatig. 64, 84–96.
Rausch, T., Beiss, P., Broeckmann, C., Lindlohr, S., Weber, R., 2010. Application of
quantitative image analysis of graphite structures for the fatigue strength estimation
of cast iron materials. Proc. Eng. 2 (1), 1283–1290.
Reihanian, M., Asadullahpour, S.R., Hajarpour, S., Gheisari, K., 2011. Application of
neural network and genetic algorithm to powder metallurgy of pure iron. Mater. Des.
32 (6), 3183–3188.
Robert, P.E., Blais, C., Pelletier, S., 2013. Tensile properties of sinter hardened powder
metallurgy components machined in their green state. Powder Metall. 52 (1), 80–83.
Robert, P.E., Blais, C., Pelletier, S., Thomas, Y., 2007. Drilling of high quality features in
green powder metallurgy components. Mater. Sci. Eng. A 458 (1-2), 195–201.
Salak, A., Selecka, M., Danninger, H., 2005. Machinability of Powder Metallurgy Steels.
Cambridge International Science Publishing.
Saleem, M., Toubal, L., Zitoune, R., Bougherara, H., 2013. Compos. Part A: Appl. Sci.
Manuf. 55, 169–177.
Schneider, M., Yuan, H., 2012. Experimental and computational investigation of cyclic
mechanical behavior of sintered iron. Comput. Mater. Sci. 57 (0), 48–58.
Veldhuis, S., Dosbaeva, G., Elfizy, A., Fox-Rabinovich, G., Wagg, T., 2010. Investigations
of white layer formation during machining of powder metallurgical Ni-based ME 16
superalloy. J. Mater. Eng. Perform. 19 (7), 1031–1036.
Voyiadjis, G.Z., Kattan, P.I., 2008 May. A comparative study of damage variables in
continuum damage mechanics. Int. J. Damage Mech. 18, 315–340.
Zhang, X.P., Shivpuri, R., Srivastava, A.K., 2014. Role of phase transformation in chip
segmentation during high speed machining of dual phase titanium alloys. J. Mater.
Process. Technol. 214 (12), 3048–3066.