Research Article • DOI: 10.

2478/amst-2019-0010 AMST • 44(2) • 2020 • 57–61

Advances in Manufacturing Science and Technology

The application of Chaboche model in uniaxial

ratcheting simulations
Marta Wójcik* and Andrzej Skrzat

Department of Materials Forming and Processing, Faculty of Mechanical Engineering and Aeronautics,
Rzeszow University of Technology, 8 Powstańców Warszawy Ave., 35-959 Rzeszów, Poland

Abstract
This article presents the application of Chaboche nonlinear kinematic hardening model in simulations of uniaxial
ratcheting. First, the symmetrical strain-controlled cyclic tension/compression tests for PA6 aluminum samples
were done. Using the experimental stress–strain curve, initial material hardening parameters were determined by the
ABAQUS software. The experimental curve was compared with the numerical one. For better fitting of both curves, the
optimization procedure based on the least-square method was applied. Using the determined hardening parameters,
numerical simulations of the ratcheting were done by the finite element analysis software. Numerical results were then
compared with the experimental data obtained in the stress-controlled cyclic loading test.
Keywords
cyclic plasticity, ratcheting, Chaboche model, nonlinear kinematic hardening, isotropic hardening

Nomenclature

b – material constant Bauschinger effect, cyclic hardening, relaxation, and ratcheting

B – error norm are the examples of phenomena related to the cyclic plasticity
c – material constant [2].
dεp – the inelastic (plastic) strain increment Among the cyclic plasticity phenomena, ratcheting can result
dp – equivalent plastic strain increment in the additional damage of materials and the shortening of
dx – backstress increment their fatigue life. Ratcheting is defined as a progressive strain
dxi – the kinematic hardening rule accumulation in a material under the stress-controlled cycling
E – Young’s modulus loading with nonzero mean stress [3]. Correct prediction of the
ε0p – translation of the yield surface ratcheting can prevent the catastrophic failure of structures.
σ – stress Ratcheting has been extensively studied in both experimental
σappr– approximated value of stress and numerical researches in the recent years [4–8].
σSAT – saturation stress The appropriate constitutive models that can accurately describe
σγ – yield stress the material behavior under cyclic loading are developed.
Q – material constant Frederick-Armstrong, Chaboche model, which is the extension
φ – true (logarithmic) strain of Frederick–Armstrong model, is commonly used in modeling
p – equivalent plastic strain of uniaxial and multiaxial ratcheting. Even though a lot of
r(p) – isotropic hardening function research has been done in modeling of cyclic plasticity including
γ – material constant ratcheting, none of the existing models is versatile and robust to
υ – Poisson’s ratio simulate it accurately [9].
x – backstress The precise identification of material hardening parameters
is essential for the prediction of ratcheting. It was noted
that kinematic hardening is considered as a main reason
1. Introduction
for ratcheting. On the other hand, this phenomenon is very
sensitive to the isotropic hardening and depends on the isotropic
The cyclic plasticity concerns an elastoplastic stress–strain hardening parameters values; thus, the gradual deceleration or
response of materials in closed and repeated loading paths [1]. even the blocking of ratcheting might occur [10].

*
Corresponding author: Marta Wójcik
E-mail: [email protected]

57
Advances in Manufacturing Science and Technology

In this research, the Chaboche nonlinear kinematic model,

with three nonlinear terms as well as the Voce isotropic one,
is used to simulate the ratcheting for the PA6 aluminum.
Hardening parameters for the Chaboche model are
determined based on the experimental stress–strain curves
obtained in strain-controlled cyclic tension/compression
tests by the ABAQUS software. The numerical stress–strain
curve was generated using initially determined hardening
parameters. For better fitting of numerical and experimental
curves, the optimization procedure with the use of the least-
square method was applied. The computed set of hardening
parameters is later on used in simulations of a uniaxial
ratcheting. The numerical results are compared with the
Figure 1. Kinematic hardening model with three backstresses.
experimental curve for a stress-controlled nonsymmetrical
cyclic loading test. For the evolution of the yield surface size in the stress space, the
Voce isotropic hardening model was considered in this study (Eq. 3).

2. Constitutive model
dr (=
p ) b ( Q − r ) dp (3)

Many constitutive models have been developed for the

prediction of the material behavior subjected to the cyclic
3. Experimental investigations
loading. In this research, the Chaboche nonlinear kinematic
model, including the translation of the yield surface during
plastic deformation in the stress plane, is applied. The For the initial selection of hardening parameters for the
kinematic hardening rule for Chaboche model is described in PA6 aluminum, the strain-controlled symmetrical tension/
the following Eq (1). compression test was carried out using the ZWICK/ROELL
Z100 testing machine. The geometry of a specimen used in
(1) this study is shown in Figure 2. The deformation ±2% of the
measuring base was applied. The experiment was carried out
in which the superposition of some hardening rules is included in triplicate, and the results were calculated as an average
(Eq. 2). value from theses series. The experimental true stress–true
N
strain curve obtained is shown in Figure 3.
dx = ∑dxi (2)
i =1

Equation 1 consists of the linear purely kinematic term 2 cdå p

3
derived from Prager–Ziegler model and relaxation (recall)
Figure 2. The geometry of specimens used in this research.
term γxdp, which introduces the nonlinearity and includes the
fading memory effect of the strain path [11].
Three decomposed hardening rules are recommended in
simulations of stable hysteresis loop (Figure 1) [12]. The first
rule ( x1 ) should start hardening with a very large modulus and
stabilize quickly. The second rule ( x2 ) should simulate the
transient nonlinear portion of the stable hysteresis curve. The
third rule ( x3 ) should almost linear hardening rule ( γ 3 = 0))
to represent the subsequent linear part of a hysteresis curve
at the high strain rate [13]. According to Bari and Hassan [9],
the introduction of small nonlinearity ( γ 3 ≠ 0)) might improve
the simulation of ratcheting.

Figure 3. The experimental stress–strain curve for PA6 aluminum.

58
 Wójcik and Skrzat: The application of Chaboche model in uniaxial ratcheting simulations

In this research, the unknown parameters were c1, c2, c3, γ1, 4. Approximation of hardening parameters
γ2, and γ3 (Chaboche model) as well as Q and b (Voce model).
In Chaboche model, each c parameter is the initial kinematic For better fitting of experimental and numerical curves, the
hardening moduli, and each γ parameter defines the rate optimization approach using the least-square method was
at which the kinematic hardening moduli deceases with the applied. This procedure assumes that stress deviations from
increase of the plastic deformation. For Voce model, parameter their approximate values (the error norm) are being minimized
Q determines the saturation stress, and b parameter defines (Eq. 5). More information about this procedure is included in
the rate in which the material saturates. Wójcik and Skrzat [14].
The Chaboche parameters were identified using automated
x
calibration in ABAQUS program. Parameters Ci and γ i
∫ε (σ − σ appr ) d ε
2
were determined from the stabilized cycle test data. The
=B exp
N
c
constrained optimization problem was solved with ∑ γ i = σ SAT (5)
x i
equality and γ3 < 9 inequality constraints [12]. Based on the
experimental stress–strain data converted into true stress– The minimum of Eq. (5) error norm (12%) was obtained for the
plastic strain relation (Figure 4), ABAQUS automatically following material’s hardening parameters: Q = 140 MPa, c1 =
calibrates the parameters for Chaboche model. The yield 1544 MPa, γ1 = 611, c2 = 1028 MPa, γ2 = 612, c3 = 358 MPa,
stress was determined using the 0.2% offset method. The γ3 = 20, and b = 18. The comparison of experimental and
plastic strain was calculated using the additive decomposition numerical curves after the optimization procedure is shown
of generalized strain (Eq. 4). in Figure 7.
σi
ε ip= ε i − (4)
E
The material hardening parameters for Chaboche selected
by means of ABAQUS procedure are Q = 150 MPa,
c1 = 4735 MPa, γ1 = 299, c2 = 1511 MPa, γ2 = 291,
c3 = 1554 MPa, γ3 = 22, and b = 11. The numerical curve for
such obtained parameters is shown in Figure 5.
It is clearly seen in Figure 5 that a relatively good correlation
was obtained between the experimental and numerical
curves. However, the stabilized cycle test data ensure the
better selection of parameters for the last stabilized hysteresis
loop than for the first one.
As a main part of this research, the nonsymmetrical stress-
controlled cyclic tension-compression test for PA6 aluminum
was carried out using the testing machine mentioned before.
The registered hysteresis loops are shown in Figure 6. Figure 5. The comparison of experimental and numerical stress–
strain curves for the Chaboche model obtained by ABAQUS
program.

Figure 6. The stress–strain data for stress-controlled cyclic tension–

Figure 4. The calibration stress–plastic strain curve. compression test.

59
Advances in Manufacturing Science and Technology

The kinematic and isotropic hardening parameters were

then applied to create a model in ABAQUS for simulating of 6. Summary and conclusions
ratcheting in PA6 aluminum.
In this article, Chaboche model was used in numerical
simulations of uniaxial ratcheting for the PA6 aluminum.
5. The modeling of the uniaxial ratcheting for PA6
The parameters for Chaboche model were determined from
aluminum the stabilized cycle test data in ABAQUS program based
on the experimental stress–strain curve. For better fitting of
The ABAQUS software was applied in numerical simulations of experimental and numerical curves, hardening parameters
nonsymmetrical stress-controlled cyclic tension/compression were enhanced using the least square method. The relatively
tests. The model is shown in Figure 8. Due to the axial good agreement between numerical and experimental curves
symmetry of the specimen, the 2D axisymmetric analysis is was obtained for the Chaboche nonlinear kinematic hardening
made. The model was meshed by two-dimensional quadratic model with three decomposed rules.
quadrilateral elements. The numerical simulation of the nonsymmetrical stress-
The following material data are used in analyses: E = 70 controlled cyclic tension-compression test for the PA6
× 103 MPa, υ = 0.33, and σ y = 412 MPa. The hardening aluminum was done by the finite element method analysis.
parameters identified in the cyclic tension/compression test A relatively good agreement between experimental and
and optimized by means of the least-square method were numerical curves was noted for Chaboche nonlinear kinematic
applied in numerical calculations. hardening model with three kinematic rules.
The specimen is constrained at the left end in the longitudinal Further research will focus on the application of improved
direction, and a load is applied to the right end of the models, for example, Chaboche model with threshold on
specimen. The load history applied is shown in Figure 9. backstress or Chaboche model with fourth rule, in modeling
Figure 10 presents the results of the ratcheting simulation for of ratcheting. These models include the linear segment on the
stress-controlled hysteresis loop with the use of Chaboche hysteresis curve within the threshold level, and therefore, the
model with three kinematic hardening rules. It is observed that rate of ratcheting is reduced improving the simulations of the
the simulated hysteresis loop traces the experimental curve ratcheting.
relatively close. The good agreement between both curves
confirms the proper selection of the hardening parameters.

Figure 9. The load history used in numerical simulations.

Figure 7. The comparison of experimental and numerical stress–

strain curves for Chaboche model after the optimization by the least-
square method.

a)
Figure 10. Numerical and experimental stress–strain curves for a
Figure 8. The sketch of the model.
nonsymmetrical stress-controlled loading test.

60
 Wójcik and Skrzat: The application of Chaboche model in uniaxial ratcheting simulations

References [8] H. WANG, et al.: Uniaxial ratcheting behaviour of 304L stainless

steel and ER308L weld joints. Mater. Sci. Eng. A, 708(2017),
[1] S. LI: Investigation of elastoplastic ratchetting behavior of Stain- 21-42.
less Steel 316 under cyclic uniaxial asymmetric loading at room [9] S. BARI, T. HASSAN: Anatomy of coupled constitutive models
temperature. Lehigh University – Lehigh Preserve, Bethlehem for ratcheting simulation. Int. J. Plast., 16(2000)3-4, 381-409.
2018. [10] X. JIANG, et al.: Constitutive model for time-dependent ratchet-
[2] S.K. PAUL: A critical review of experimental aspects in ratchet- ting of SS304 stainless steel: simulation and its finite element
ing fatigue: microstructure to specimen to component. J. Mater. analysis. J. Theor. App. Mech.-Pol., 51(2013)1, 63-73.
Res. Technol., 8(2019)5, 4894-4914. [11] V. BUDAHAZY, L. DUNAI: Parameter-refreshed Chaboche mod-
[3] C.B. LIM, K.S. KIM, J.B. SEONG: Ratcheting and fatigue behav- el for mild steel cyclic plasticity behaviour. Period. Polytech. Civil
ior of a copper alloy under uniaxial cycling loading with mean Eng., 57(2013)2, 139-155.
stress. Int. J. Fatigue., 31(2009)3, 501-507. [12] S. BARI, T. HASSAN: An advancement in cyclic plasticity model-
[4] S. GOYAL, et al.: Low cycle fatigue and cyclic plasticity behavior ing for multiaxial ratcheting simulation. Int. J. Plast., 18(2002)7,
of Indian PHWR/AHWR primary piping material. Procedia Eng., 873-894.
55(2013), 136-143. [13] J.L. CHABOCHE: Modeling of ratcheting: evolution of various
[5] K. DUTTA, K.K. RAY: Ratcheting phenomenon and post-ratch- approaches. Eur. J. Mech. A-Solid., 13(1994), 501-518.
eting tensile behaviour of an aluminum alloy. Mater. Sci. Eng. [14] M. WÓJCIK, A. SKRZAT: Fuzzy logic enhancement of mate-
A-Struct., 540(2012), 30-37. rial hardening parameters obtained from tension-compression
[6] S.K. MISHRA, K. DUTTA, K.K. RAY: Fatigue life estimation in test. Continuum Mech. Thermodyn., 2019, 1-8. DOI:10.1007/
presence of ratcheting phenomenon for AISI 304LN stainless s00161-019-00805-y.
steel tested under uniaxial cyclic loading. Int. J. Damage Mech.,
25(2016)3, 431-444.
[7] C.L. PUN, et al.: Ratcheting behaviour of high strength rail steels
under bi-axial compression-torsion loadings: experiment and
simulation. Int. J. Fatigue, 66(2014), 138-154.

61