Available online at www.sciencedirect.

com

ScienceDirect
Procedia Engineering 86 (2014) 150 157

1st International Conference on Structural Integrity, ICONS-2014

Creep Life Prediction of Modified 9Cr-1Mo Steel under

Multiaxial State of Stress
Sunil Goyal*, K. Laha and M. D. Mathew
Mechanical Metallurgy Division, Indira Gandhi Centre for Atomic Research, Kalpakkam- 603 102, India
*
Email id: goyal@igcar.gov.in

Abstract

Creep life prediction of modified 9Cr-1Mo steel under multiaxial state of stress has been carried out in the present investigation.
Creep tests were carried out on smooth and notched specimens of modified 9Cr-1Mo steel in the stress range 150 - 230 MPa at
873 K. The creep rupture life was found to be higher in the presence of notch than that of smooth specimens indicating notch
strengthening behaviour. The extent of strengthening tends to saturate for relatively sharper notches. Finite element analysis was
used to estimate the variation of stresses across the notch throat plane. As the stresses varied across the notch throat plane, the
stresses at the skeletal point were used for estimating the representative stress. The models proposed by Cane, Hayhurst and Nix
were examined for creep life prediction under multiaxial state of stress. The von-Mises stress was found to predominantly govern
the creep rupture life of the steel. The creep rupture life has been predicted based on Canes model using finite element analysis
and continuum damage mechanics.

2014
2014 The
The Authors.
Authors. Published
Published by
by Elsevier
Elsevier Ltd.
Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/3.0/).
Peer-review under responsibility of the Indira Gandhi Centre for Atomic Research.
Peer-review under responsibility of the Indira Gandhi Centre for Atomic Research
Keywords: multiaxial creep, life prediction, representative stress, finite element analysis, continuum damage mechanics

1. Introduction

Notched specimens are widely used for evaluating the effect of multiaxial state of stress under creep
conditions [1-4]. The multiaxial state of stress in the notched specimens can be varied by changing the notch root
radius. The components are generally designed based on uniaxial creep data. However, the components experience
multiaxial state of stress as a result of change in geometry, heterogeneous microstructure as in weld joint and also
due to the mode of loading during service. As the design for creep is generally based uniaxial creep data, it is
important to establish the relation between net applied stress and rupture life to account for multiaxial state of stress.
The representative stress considering the relative contribution of different components of stresses viz., von-Mises,
maximum principal and hydrostatic stresses, can be used for estimation of rupture life under multiaxial state of

1877-7058 2014 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/3.0/).
Peer-review under responsibility of the Indira Gandhi Centre for Atomic Research
doi:10.1016/j.proeng.2014.11.023
 Sunil Goyal et al. / Procedia Engineering 86 (2014) 150 157 151

stresses. Many relationships are available in litterature for multiaxial stress rupture criteria (MSRC) [5]. Hayhurst et
al. [1] proposed the representative stress as an algebraic sum of von-Mises and maximum principal streess, whereas,
the model proposed by Cane assumes that the t cavity growth is constrained by the matrix deform mation under
multiaxial state of stress.

In the present investigation, an attemppt is made to establish the relation between creep rupturre life under
multiaxial state of stress with uniaxial creep coonditions. The approaches proposed by Cane [2], Hayhu urst et al. [1]
and Nix et al. [6] have been used and comparedd for their suitability for estimation of representative stresss for ferritic
steels. The material chosen is modified 9Cr-1M Mo steel which is extensively used in power generating plants
p due to
its excellent high temperature mechanical propeerties.

2. Experimental

Effect of multiaxial state of stress on creep behaviour of the steel has been studied by carryin ng out creep
tests at net stresses in the range 150- 230 MPaa and 873 K on circumferentially U-notched specimens having h notch
root radius of 1.25 mm with notch throat diameeter of 5 mm. Further, in order to study the effect of notcch sharpness
on creep behaviour, creep tests were also carrried out on circumferentially U-notched specimens hav ving various
notch root radii in the range 0.25 - 5 mm by keeping the notch throat diameter 5 mm at net stress of 210 0 MPa and at
873 K. These notch geometries resulted in notch acuity ratio (d/r) of 1 to 20. The geometry of the notcheed specimens
is illustrated in Fig. 1. Creep tests were also carrried out on smooth specimens of the steel to evaluate thee coefficients
required for continuum damage mechanics.

Fig. 1 Notched creep tesst specimen details (all dimensions are in mm).
152 Sunil Goyal et al. / Procedia Engineering 86 (2014) 150 157

2.1 FE Analysis of Stress Distribution

Finite element analysis of stress distribution across the notch throat plane during creep exposure has been
carried out to understand the creep behaviour of the material under constrained condition. The 2D axisymmetric
analysis was carried out using 4 noded quadrilateral elements with ABAQUS 6.10 finite element solver. Nortons
creep law (  s = A n ) relating the steady state creep rate ( s ) with applied stress () was used as a constitutive
equation for the analysis. In the analysis, the elastic modulus was assumed to be 160 GPa and constant A in the
Nortons equation was selected to give a steady state creep rate of 10-5 h-1 at 210 MPa for a given value of stress
exponent n. The calculations were carried out for the n values ranging from 1 to 10. A similar approach was
adopted earlier to estimate stress distribution in the notched specimens [7]. Initially, the material was assumed to
deform elastically followed by creep deformation. The element size was reduced closer to the notch root and elastic
analysis was used to assure that the mesh configuration was sufficiently refined near the notch root for the
theoretical stress concentration factor. The analysis was continued till the stress redistribution across the notch throat
plane attained a stationary state condition. The stress redistribution was generally considered to attain stationary
state condition when the total strain at the notch throat plane reached the elastic strain in the material as stated by
Calladine [8].

3. Results and Discussion

3.1 Creep Rupture Life Variation

The differences in creep rupture life of the steel in the presence of notch of root radius 1.25 mm with those
of plain specimen for different applied stresses are shown in Fig. 2. The creep rupture lives of notched specimens
were found to be higher than those of the plain specimens, exhibiting strengthening in the presence of notch for the
steel. The extent of strengthening increased with notch acuity ratio and tends to saturate for higher notch acuity
ratios, Fig. 3. Creep rupture ductility of the steel decreased in presence of notch especially for higher notch
sharpness.

Fig. 2 Fig. 3

Fig. 2 Variation of creep rupture life as a function of net applied stress for notch of root radius of 1.25 mm at 873 K.
Fig. 3 Notch strengthening as a function of notch acuity ratio for modified 9Cr-1Mo steel.

3.2 Creep Life Prediction Based on Representative Stress Concept

The creep rupture life of material under multiaxial state of stress depends on creep deformation and
cavitation. The mechanisms associated with the creep deformation and cavitation have been discussed extensively
 Sunil Goyal et al. / Procedia Engineering 86 (2014) 150 157 153

under uniaxial and triaxial state of stresses by Nix et al. [6], Cane et al. [2] and Goyal et al. [9]. The creep rupture
life under uniaxial loading is expressed in terms of the applied stress as
t r = M m (1)
where, m is the slope of the uniaxial creep rupture plot. Creep rupture life under multiaxial state of stress can also be
described by an equation similar to uniaxial loading on incorporating a representative stress as
m
t r = M rep (2)
The representative stress, rep, is defined as the stress applied to the uniaxial plain specimen, which would result in
the same creep rupture life as that of notched specimen. If rep > net (where, net is the net applied stress in the
notched specimen), the presence of notch decreases the rupture life of material causes notch weakening; whereas, if
rep < net, the presence of notch increases the rupture life of material resulting in notch strengthening.

Creep rupture under multiaxial state of stress depends on the components of stresses viz., maximum
principal stress, hydrostatic stress and von-Mises stress in governing the creep deformation and cavitation. It is
widely reported that the (i) von-Mises stress controls the deformation and creep cavity nucleation processes [6], (ii)
the maximum principal and hydrostatic stress control the continuum cavity growth whereas (iii) the maximum
principal stress controls the stress directed diffusion controlled intergranular cavity growth [6]. The relative
contribution of each of the stresses to representative stress in governing the creep rupture life under multiaxial state
of stress depends on the materials deformation and fracture response to the stresses.

Many relationships have been proposed and are available in open literature for stress-rupture criteria under
multiaxial state of stress. In the present investigation, the approaches proposed by Cane [2], Hayhurst et al. [1] and
Nix et al. [6] has been used and compared for their suitability for prediction of representative stress for the steel.

D. R. Hayhurst rep = 1 + (1 ) vm (3)

m ( m ) m
B. J. Cane rep = 1 vm (4)
W.D. Nix rep = 2.24 1 0.62( 2 + 3 ) (5)

where m, , and are material constants and 1, 2 and 3 are maximum, intermediate and minimum principal
stresses and vm is von-Mises stress. The representative stress reduces to the maximum principal stress for = m (
= 1) and to the von-Mises stress for = 0 ( = 0).

Since the stresses vary across the notch throat plane with creep exposure and attain the stationary states, it
is difficult to identify the location in notch throat plane at which the stress should be consideration in defining the
representative stress. Hayhurst et al. [10], based on FE analysis, have introduced the skeletal point concept in
determining the representative stress.

3.2.1 Skeletal Point Concept

The skeletal point is the location in notch throat plane for a given notch geometry, where the variation of
stress across the notch throat plane for different stress exponent n in Nortons law, intersects. The stresses
estimated at this point have been implemented to characterize the creep deformation, damage and failure behaviour
of materials under multiaxial state of stress many investigators [5-7].

In order to obtain the skeletal point stresses, FE analysis of stress distribution across the notch throat plane
was carried out for various values of stress exponent n in Nortons law (  s = A n ) ranging from 1 to 10 for net
applied stress of 210 MPa. The value of coefficient A was obtained based on the creep strain rate of 10-5 h-1. The
variation of maximum principal stress across the notch throat plane for different values of stress exponent n for
notch root radius of 0.5 mm is shown in Fig. 4. Similar results were also obtained for other notch root radii. It was
154 Sunil Goyal et al. / Procedia Engineering 86 (2014) 150 157

observed that there exists a skeletal point at which stress variation for different stress exponent intersects. Similar
observations have been reported by Webster et al. [7]. For a given notch geometry, the radial position of skeletal
point remained in a narrow band of distance for all the stresses. The radial position of skeletal point shifted from
close to centre towards the notch root and tends to saturate for higher notch acuity ratio.

Variation of skeletal point stresses with notch acuity ratio is shown in Fig. 5. The maximum principal and
hydrostatic stresses at skeletal point increased with notch acuity ratio and tend to saturate at higher notch acuity
ratio; whereas von-Mises stress decreased with increase in notch acuity ratio. The value of maximum principal stress
at the skeletal point was more than the net applied stress. However, the von-Mises and hydrostatic stresses at the
skeletal point were lower than the net applied stress for all notch acuity ratios. The stresses at the skeletal point for a
given notch acuity ratio and net applied stress can be calculated by multiplying the normalized stresses, Fig. 5, with
the net applied stress.

3.3 Creep Life Prediction Based on Representative Stress

3.3.1 Hayhurst Model

The model proposed by Hayhurst et al. [1] relates the representative stress as algebraic sum of maximum
principal stress and von-Mises stress incorporating a material dependent parameter influencing the contribution of
each stress on multiaxial stress rupture criteria, Eq. 3. In order to determine the parameter () for these steel under
tested conditions, regression analysis was carried out for the representative stress (rep) based on skeletal point
stresses for each notch geometry. The best fit value of was found out to be 0.07 for the steel. The creep rupture
life of the steel under multiaxial state of stress is considered to be governed predominantly by von-Mises stress with
only 7 % maximum principal stress for the steel. Representation of multiaxial creep data in terms of representative
stress as a function of rupture life for the steel with optimized value of is shown in Fig. 6. A high value of
correlation coefficient 0.96 clearly indicates that multiaxial creep rupture behaviour of the material is well
represented by Hayhurst model.

Fig. 4 Fig. 5

Fig. 4 Variation of normalized maximum principal stress across the notch throat plane for different values of stress
exponent (Notch root radius of 0.5 mm).
Fig. 5 Variation of skeletal point stresses as a function of notch acuity ratio.

3.3.2 Cane Model

The model proposed by Cane [2] relates the representative stress as the multiplication of the maximum
principal stress and von-Mises stress incorporating a material dependent parameter influencing the contribution of
 Sunil Goyal et al. / Procedia Engineering 86 (2014) 150 157 155

each stress on multiaxial stress rupture criteria, Eq. 4. In order to determine the parameter () for the steel under
tested conditions, regression analysis was carried out for the representative stress (rep) based on skeletal point
stresses for each notch geometry. The optimized value of for modified 9Cr-1Mo steel was found to be 1.1. The
variation of representative stress as a function of creep life for the steel with optimized value of is shown in Fig. 7.
A high value of correlation coefficient 0.96 indicates that multiaxial creep rupture behaviour of the steel is well
represented by Cane model.

3.3.3 Nix Model

The representative stress considered as principal facet stress determined based on Eq. 5 has been plotted
against the rupture life in Fig. 8. The multiaxial creep rupture lives of the steel were not found to be represented well
based on representative stress calculated base on the Nix model. The multiaxial creep data was found not to
superimpose with the uniaxial creep data. Nix model is particularly applicable for creep cavity prone materials.
However, the steel used in the present investigation are not very prone intergranular creep cavitation.

Both the models proposed by Hayhurst and Cane represented the creep behaviour of the investigated steel
under multiaxial state of stress well. However, the model proposed by Cane [2] would be preferable in terms of the
physical processes occurring in the materials during creep exposure. Aplin et al. [5] commented that the creep
deformation and damage incurred by different components of stresses in multiaxial state of stress under creep
condition are not independent to each other and hence can not be represented as an algebraic sum of their
contribution. Based on this understanding, further analysis of creep damage and rupture life prediction have been
carried out adopting the model proposed by Cane [2].

Fig. 6 Fig. 7

Fig. 6 Presentation of multiaxial creep rupture data based on representative stress calculated based on Hayhurst
model for modified 9Cr-1Mo steel.
Fig. 7 Presentation of multiaxial creep rupture data based on representative stress calculated based on Cane model
for modified 9Cr-1Mo steel.

3.4 Creep Life Prediction Based On FE Analysis Coupled With CDM

FE analysis coupled with continuum damage mechanics (FE-CDM) has been extensively used for creep
damage and rupture life prediction under multiaxial state of stress [11,12]. In this investigation, the prediction of
creep rupture life under multiaxial creep condition has been carried out based on FE analysis coupled with CDM. FE
analysis of creep damage was carried out using ABAQUS finite element solver incorporating the creep rate equation
(Eq. 6) and creep damage law (Eq. 7) proposed by Kachanov [13] in VUMAT subroutine. The values of material
constants used in the analysis are given in Table 1.
156 Sunil Goyal et al. / Procedia Engineering 86 (2014) 150 157

Fig. 8 Fig. 9

Fig. 8 Presentation of multiaxial creep rupture data in terms of representative stress as principal facet stress for
modified 9Cr-1Mo steel.
Fig. 9 Prediction of creep rupture life of notched specimens at various applied stresses and notch acuity ratios at 873
K for modified 9Cr-1Mo steel.

Table 1 Elastic, creep and damage constants of the steel at 873 K

E
Material
(GPa)
A n B
-33
modified 9Cr-1Mo steel 160 0.3 3.5710 12.92 12.47 11.1 0.6610-31

The governing rate equations used in the analysis are as follows;

n 1
3 vm ij'
c ij = A (6)
2 1 1

B rep
 = (7)
(1 )
The creep rupture life of the steel under multiaxial state of stress has been predicted based on the representative
stress associated with creep damage using FE analysis coupled with creep damage mechanics. The rupture life of the
steel predicted based on the continuum damage mechanics coupled with FE analysis was found to be in good
agreement with the experiments within a factor of 3 as shown in Fig. 9.

Conclusions

1. Presence of notch increased the creep rupture strength of the steel. With increase in notch sharpness, creep
rupture strength of the steel increased and tend to saturate for higher sharpness.
2. The model proposed by Cane represented the experimental multiaxial creep data well for the steel. Skeletal
point concept was adopted for estimating the representative stress in notched specimens.
3. The FE analysis coupled with continuum damage mechanics under multiaxial loading was found to predict
the rupture lives of the steel within a factor of 3.
 Sunil Goyal et al. / Procedia Engineering 86 (2014) 150 157 157

References

1. Hayhurst D R, J Mech Phy Solids 20 (1972) 381.

2. Cane B J, In: Proceedings of the Int. Conf. on Mechanical behavior of Materials, (eds) Miller K J and Smith R
F, Pergamon Press, Oxford (1979) p 173.
3. Sunil Goyal, Laha K, Das C R, Panneerselvi S, Mathew M D, Int J Mech Sci 75 (2013) 233.
4. Ng S E, Webster G A and Dyson B F, in Advances in fracture research, ICF-5, (eds) Francois D, (1980) p 1275.
5. Aplin P F and Eggeler G F, In: Proceedings of the Int. Conf. on Mechanics of Creep brittle materials, (eds)
Cocks A C F and Ponter A R S, Elsevier Applied Science (1988) p 245.
6. Nix W D, Earthman J C, Eggeler G and Ilschner B, Acta Metall 37 (1989) 1067.
7. Webster G A, Holdsworth S R, Loveday M S, Nikbin K, Perrin I J, Purper H, Skelton R P and Spindler M W,
Fat Frac Engg Mater Struct 27 (2004) 319.
8 Calladine C R, Proc Royal Soc Lond, A 309 (1969) 363.
9 Sunil Goyal, Laha K, Chandravathi K S, Parameswaran P and Mathew M D, Phil Mag, 91 (2011) 3128.
10 Hayhurst D R and Webster G A, In: Techniques for multiaxial creep testing, (eds) Gooch D J and How I M,
Elsevier Applied Science, Amsterdam (1986) p 137.
11 Hayhurst D R, Dimmer P R and Morrison C J, Phil Trans Royal Soc London, 311 (1984) 103.
12 Hyde T H, Sun W and Tang A, Strain, 34 (1998) 83.
13 Kachanov L M, The Theory of Creep, (eds) Kennedy A J, National Lending Library, Boston Spa, UK, 1967.