International Journal of Pressure Vessels and Piping 146 (2016) 198e202

Contents lists available at ScienceDirect

International Journal of Pressure Vessels and Piping

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

Defining a negligible creep temperature curve for Gr. 91 steel

€m
Stefan Holmstro
European Commission, Joint Research Centre, Petten, The Netherlands

a r t i c l e i n f o a b s t r a c t

Article history: Time independent rules for design can be applied if the design temperature of a component is restricted
Received 18 December 2015 below the negligible creep temperature (TNEC). The TNEC curve describes the time and temperature limits
Accepted 18 July 2016 below which accumulated creep strain and damage is insignificant at a specified reference or design
Available online 20 August 2016
stress. For the ferritic/martensitic steel Gr. 91 (X10CrMoVNb9-1) there is currently no TNEC curve available
in the nuclear design code RCC-MRx. In this paper the TNEC curve for Gr.91 is defined using a new
Keywords:
methodology developed in the European FP 7 project MATTER. The TNEC curve is based on time to 0.2%
Negligible creep
creep strain data. Supporting test data from low temperature (375e500
C) creep and tensile tests are
Low temperature
Gr. 91
presented and assessed. The TNEC curve can alternatively be constructed by using time factor corrected
P91 creep rupture data. These time factors can be used for evaluating the safety against significant creep at
X10CrMoVNb9-1 both reference and design stress. The presented methodology is will be used for determination of TNEC
Design rules curves for the next revision of the EN-13445.
© 2016 The Author. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND
license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction
To avoid expensive implementation of surveillance programs
and/or frequent inspections of (monitoring creep damage) required
for components operating in the significant creep regime it is rec-
ommended to design for service below the No Creep temperature
(TNC) or the time dependent Negligible Creep temperature (TNEC).
For nuclear components, it is possible to disregard creep as design
criterion during normal service operation if the temperature, stress
and time limits of negligible creep are respected. It is though a
challenge to define these limits reliably in the temperature range
where creep behaviour (time to strain and/or rupture) is tradi-
tionally not tested.
The TNEC curves in the RCC-MRx design code [1] for austenitic
steels are based on very small creep strain (0.03e0.01%) data. This
type of data is typically not available in public domain literature or
material standard creep property tables [2]. To extract these strain-
time values from data generated by standard dead-weight ma-
chines would in most cases be highly unreliable due to un-
certainties in determining the start of the test from the manual
loading of the specimen. The time to 0.2% creep strain limit, as used
in the ASME design code [3] for deciding if creep-fatigue evaluation
is required, is a better base for the TNEC definition. When defining
the negligible creep criteria, such as allowable creep strain or
E-mail address: stefan.holmstrom@ec.europa.eu.
rupture time fraction, it is clear that also the reference stress (sref)
used for calculation of the TNEC curve plays an important role. The
reference stress can be defined to be equal to the actual design or
allowable stress (Sm) or increased by a correction factor as is done in
the R5 code (1.5$Sm) [4].
In RCC-MRx the TNC temperature is 375
C for Gr. 91
(X10CrMoVNb9-1). The “low temperature” creep properties in the
temperature range 375
Ce500
C are then naturally the most
interesting for studying the limits of negligible creep. The lowest
temperatures tested for creep are in most cases limited to tem-
peratures around 500
C since the interest of the industry is mainly
in the high temperature range for creep resistant steels.
In the European FP7 MATTER project [5] it was shown that the
same TNEC curve can be defined from a higher reference stress when
also using higher strain limits. It is important to remember that the
reference stress, when defined as a constant factor of yield or
tensile strength, is dependent on whether minimum values or
average values are used. For creep it is more conservative to use the
higher yield or tensile strength as base for the calculations since a
higher stress naturally results in shorter creep life.
In this paper the methodology of defining the TNEC is presented
together with low temperature creep strain data for steel Gr.91. The
presented work was conducted in the MATTER project and the
resulting TNEC curve is proposed for implementation in the RCC-
MRx design code. The methodology is also used for the revision
of EN-13445 regarding negligible creep [6]. The initial MATTER
work on TNEC prediction is presented in Ref. [7] and earlier work on

http://dx.doi.org/10.1016/j.ijpvp.2016.07.004
0308-0161/© 2016 The Author. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
 €m / International Journal of Pressure Vessels and Piping 146 (2016) 198e202
S. Holmstro 199

To support the TNEC calculations a series of tensile and creep

Nomenclature tests were performed in the target temperature range
(375e500
C). However, the main data source for determining the
ε Strain [%] higher temperature creep properties (time to strain and rupture at
T Temperature [
C] 500e650
C) was the Gr.91 steel NIMS data sheet [11].
TNC Temperature limit for No Creep (time independent
design)
2.2. Definitions and methods
TNEC Temperature-time curve for Negligible Creep at
reference stress
2.2.1. Negligible creep criteria and limits
Tr Temperature limit for creep rupture at reference
In RCC-MRx design code the negligible creep conditions prevail
stress
as long the sum of time fractions (time at specified service tem-
s Stress [MPa]
perature ti divided by the maximum time allowed timax at that
sref Reference stress [MPa], sref ¼ 0.56,Rm
temperature (from the TNEC curve) is lower than unity (Eq. (1)). The
Sm Allowable stress [MPa] defined by specified design
method for determining timax is shown in Fig. 1 together with the
code
different time-temperature regions for creep, no creep and negli-
Rp0.2 Yield stress [MPa] at specified T
gible creep.
Rm Ultimate tensile strength [MPa] at specified T
Ru/t/T Creep Rupture strength [MPa] to time t at Xn  
ti
temperature T 1 (1)
i¼1
timax
t Time [h]
tr Time to rupture [h] at specified s and T The TNEC curve criteria for austenitic steels in ASME, R5 [2,13]
tε Time to specified strain [h] at s and T and RCC-MRx design codes are given in Table 2. In the case of F/
SCF Stress correction factor used on creep rupture M steels these definitions of reference stress are not suitable
strength because of significantly higher yield to tensile strength ratio.
RTF Rupture time factor; adjusting tr to tNEC; In this work the TNEC curve is calculated on virgin material
tNEC (s, T) ¼ tr (s, T)/RTF properties and time to 0.2% creep strain. The strain limit was cho-
STF Strain time factor; adjusting tε to tNEC; sen to maximize the amount of data available for the assessment
tNEC (s, T) ¼ tε (s, T)/STF and to minimize the measurement uncertainties from the sup-
WE Wilshire model for creep rupture or strain porting creep tests.
The reference stress was chosen to comply with the RCC-MRx
material properties and allowable stress definition. The allowable
stress (Sm) for time independent design using Gr.91 steel is 1/2.7 of
Gr.91 TNEC curves can be found in Ref. [8e10].
the tensile strength, i.e. 0.37 Rm. If Sm is multiplied with a factor of
1.5, as done in R5 (see Table 2.), the resulting in a reference stress is
2. Materials and methods 0.56 Rm. This reference stress is somewhat higher than the design
stress of 2/3$Rp0.2 (0.48e0.54$Rm) used in the EN-13445 [6]
2.1. Material standard.

The Gr.91 test material used for low temperature/low stress
creep testing was 60 mm thick plate from ArcelorMittal (plate
20057). The tested steel was in “as received” product state and has
undergone heat treatment: 1060
C e 4 h, water quenching/
760
Ce3.3 h, air cooling. The material heat was especially designed
to be fully compliant with RCC-MRx material requirements. The
chemical composition of plate 20057 is given in Table 1.
2.2.2. New methodology for calculating TNEC
To calculate the TNEC curve a new methodology was developed
in the MATTER project. The method is based on the Wilshire
equation (WE) for creep rupture and strain [12]. In the WE model
(Eq. (2)) the creep stress is normalized with tensile strength at the
specified temperature.

575
Table 1
Chemical composition of the Gr.91 (heat 20057) test material in comparison to EN- RUPTURE
525 SIGNIFICANT
10028:2 (2009) X10CrMoVNb9-1 composition [2].
Temperature (°C)

CREEP
Tr
Elements Unit Min/Max Plate 20057
475
C % 0.08/0.12 0.12
Mn % 0.30/0.60 0.41 NEGLIGIBLE TNEC
425
Si % 0.20/0.50 0.24 CREEP
Ni % 0.40 0.1
Cr % 8.00/9.50 8.32 375
Mo % 0.85/1.05 1.02 ti-max NO CREEP
Cu % 0.30 0.05 325
Al % 0.020 0.006 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05 1.E+06
S % 0.005 0.001
Time (h)
P % 0.020 0.0091
V % 0.180/0.250 0.235
Fig. 1. Time-temperature regions at specified reference stress. The Negligible Creep
Nb % 0.06/0.10 0.084
temperature curve TNEC (long dash) separates the regions where time dependent and
N % 0.03/0.07 0.041
time independent design has to be applied. The creep rupture temperature curve (Tr
Ti % 0.01 0.002
continuous line) and the No Creep limit (TNC small dash) are also given. The timax for Eq.
Zr % 0.01 0.001
(1) is shown for an example of continuous service at 475
C.
200 €m / International Journal of Pressure Vessels and Piping 146 (2016) 198e202
S. Holmstro

Table 2
Negligible creep definitions [7] for austenitic steels. 0.92 JRC-plate RCC-MRx AVE

Based on: Code Reference stress NEC criterion 0.88 RCC-MRx MIN NIMS

Time fraction 1.5$Rp0.2 Test to avoid Creep-Fatigue

0.84
tu/tr  0.1 evaluation

Rp0.2 /Rm
ASME NH
Creep strain 1.25$Rp0.2
0.80
0.2%
Creep strain 0.9$Rp0.2 Sti/Ti  1
RCC-MRx 0.76
0.01% ~1.05$Sma
Relaxationb 1.5$Sma Sti/Ti  1
0.72
1-s0/srel ¼ 20% R5 1.35$Rp0.2
0.75$Rm
0.68
a
Note that allowable stress Sm and the minimum yield strength Rp0.2 vary from 0 100 200 300 400 500
code to code. Temperature [°C]
b
s0 ¼ peak stress, srel ¼ relaxed stress after hold time.
Fig. 2. Yield over ultimate tensile stress for, RCC-MRx values as well as for plate 20057
and the 3 heats tested by NIMS. The rectangle highlights the temperature and stress
The WE for time to rupture tr or time to strain tε is: range of interest.

"  #ur=ε
1
 
1 s Q
t r=ε ¼  ln ,exp (2) 3.2. Creep tests at target temperatures
kr=ε Rm R,T

For the supporting creep test programme the stress levels of the
where k an u are fitting parameters, s the test stress, Rm the ulti-
creep tests were targeted to reach 0.2% creep strain in about 5000 h.
mate tensile strength, Q the activation energy, R the gas constant
The test results from the low temperature tests are given in Table 4.
and T the absolute temperature.
This model is well suited for calculating negligible creep tem-
3.3. Determining the TNEC curve
peratures since the equation can be rearranged to give the sought
temperature TNEC as:
Individual WE models for time to strain (0.2%, 0.5% and 1%) and
rupture using were fitted for the NIMS data as shown in Fig. 3. The
Q
TNEC ¼    1  (3) parameters for the models are given in Table 5. The supporting low
 uε0:2%
temperature test data for time to 0.1 and 0.2% creep strain is shown
R,ln tε0:2% ,  k 1 ln Rsm
ε0:2% in Fig. 4 together with corresponding RTF corrected rupture models
(see discussion). It can be seen that the generated low temperature
where k an u are optimized for the chosen strain limit (here 0.2%) data complies quite well with the WE model based on higher
and s/Rm is the reference stress (here 0.56). temperatures.
The TNEC curve can now be defined by using Eq. (3). The WE
model parameters (Table 5) for 0.2% strain should be used for the
reference stress of 0.56$Rm. To obtain the most conservative curve
the highest Rm values (NIMS average) were used. The calculated
3. Results
TNEC curve is shown in Fig. 5.
It can be seen that the TNEC temperature for 200 000 h service
3.1. Tensile tests in the 374e500
C temperature range
time coincides with the no creep (TNC) limit of 375
C as defined in
RCC-MRx. The corresponding predicted TNEC temperatures at
The tensile properties of the test material (60 mm plate 20057)
200 000 h calculated on RCC-MRx (min) values is 398
C, for RCC-
are given in Table 3. The tensile strength values are used to
MRx (ave) it is 387
C and for the measured values of the test
normalize the applied creep stresses and for comparison to the
material it is 383
C.
properties given in RCC-MRx. The measured ultimate tensile
strength Rm of the test material is about 5% higher than the RCC-
4. Discussion
MRx average and about 5% lower than for the material heats in
the NIMS data sheet (see Fig. 2) [11].
With both the creep strain model and the rupture model it is
possible to calculate rupture time safety factors (RTF) correspond-
Table 3 ing to different strain limits by overlapping the stain and rupture
Tensile strength and yield stress at temperatures relevant for defining the TNEC
curve. The tensile data for Plate 20057 is readily available in the JRC MatDB data base
[14].
Table 4
Temperature [
C] Rp0.2 [MPa] Rm [MPa] Creep strain tests at temperatures relevant for defining the TNEC curve. For the stress
22 510 680 ratio (s/Rm) calculation the average Rm of NIMS has been used. The creep data for
375 436 542 Plate 20057 is available in the JRC MatDB data base [15].
375 436 541
Temp.
C Stress s s/Rm Time to 0.1% Strain rate Time to 0.2% Strain rate
375 433 544
[MPa] strain, [h] at 0.1% [1/h] strain [h] at 0.2% [1/h]
400 430 534
425 425 524 500 171 0.37 1330 2.9$107 >7050 <5.9$108
425 433 533 450 280 0.55 64 3.7$106 >7970 <3.3$108
425 425 526 425 330 0.63 38 7.0$106 2600 1.1$107
450 410 502 400 380 0.71 14 1.3$105 833 5.7$107
500 395 475 375 415 0.76 12 1.2$105 852 4.5$107
 €m / International Journal of Pressure Vessels and Piping 146 (2016) 198e202
S. Holmstro 201

1 1.0
rupture data Rupture
rupture LT tests 0.2% strain
0.9 0.9
1% data LT tests at 0.1% strain
model 1% model 0.2%
0.8 0.8
0.5% data
model 0.5%
0.7 0.2% data
0.7
model 0.2%
0.6 0.6
σ/Rm

σ/Rm
0.5 0.5

0.4 0.4

0.3 0.3

0.2 0.2

0.1 0.1
1E-23 1E-21 1E-19 1E-17 1E-15 1E-13 1E-23 1E-21 1E-19 1E-17 1E-15 1E-13

tε*exp(-Q/RT) tε*exp(-Q/RT)

Fig. 3. NIMS data and WE models for rupture, 1%, 0.5% and 0.2% strain. The horizontal Fig. 4. Normalized stress plot for the low temperature (LT) test results and the NIMS
0.56$Rm line is the proposed RCC-MRx reference stress and the vertical arrow the data rupture curve. The horizontal 0.56$Rm line is the proposed RCC-MRx reference
location of the TNEC calculation. stress and the vertical arrow the location of the TNEC calculation.

models using Eq. (2). For Gr.91 the optimal RTF that satisfies
tr z RTF$tε for time to 0.1%, 0.2%, and 1% are 10 000, 1000, and 10
correspondingly. The RTF corrected time to rupture models are
500
shown in Fig. 6. This means that using RTFs the safety against
rupture for the TNEC curve based on 0.2% strain is a factor of about
475
1000 in time at the specified reference stress. The actual design
safety margin increases this further when it is related to the actual
450
allowable design stress of 0.37$Rm. The lower stress gives an
Temperature (°C)

additional time factor of ~142.

To check if the proposed TNEC curve complies with the strain 425
equations of RCC-MRx (A3.18AS.5) the time to 0.2% strain was Tr
calculated for temperatures 400, 450 and 500
C at 0.56$Rm (RCC- 400 T-0.2%
MRx average tensile strength). The predictions are shown in Fig. 7 No Creep (NC)
together with the proposed TNEC curve. The RCC-MRx predicted 375
time to 0.2% strain at reference stress and specified temperatures
coincides well with the proposed TNEC curve. 350
As a consequence of the supporting testing and TNEC modelling 100 1000 10000 100000 1000000
it is proposed that the calculation of allowable time above the TNC Time (h)
temperature (Eq. (1)) remains the same for Gr.91 as for austenitic
Fig. 5. Negligible creep curve TNEC for time to 0.2% strain criteria at 0.56$Rm. The Rm
steels. However the reference stress is to be fixed at 0.56$Rm and
are NIMS average values. The Tr for time to rupture at the same reference stress is also
the strain limit set at 0.2% creep strain. The times to specified shown.
strains are interchangeable with RTF corrected time to rupture
values (for Gr.91 t0.2%ztr/1000). The modified approach using RTF
can be applied on standard material properties, i.e. it is possible to 1.0
Rupture
construct the TNEC curves for a number of steels based on their NIMS 1% strain data
0.9
standard strength tables for time to 1% creep strain or time to creep RTF=10
NIMS 0.2% data
rupture. This possibility is currently being applied for the next 0.8
LT tests 0.2% strain
revision of the EN-13445 design standard. RTF=1000
0.7 LT tests at 0.1% strain
RTF=10 000
0.6
σ/Rm

0.5
Table 5
Wilshire creep strain and rupture model parameters for Gr.91 steel (NIMS data).
0.4
Note that time (tr/ε) is in hours, and the total strain in %. Note also that there is a
change in parameters at the stress ratio of 0.45. The activation energy Q is 300 000 J/ 0.3
mol in all cases. The parameters used in the TNEC determination are in bold (s/
Rm ¼ 0.56). 0.2

Model type s/Rm > 0.45 s/Rm  0.45

0.1
1E-23 1E-21 1E-19 1E-17 1E-15 1E-13
k u k u
tε*exp(-Q/RT)
rupture 82.769102 0.135627 75.618032 0.133624
1.0% strain 150.827655 0.141074 31.330443 0.101257
Fig. 6. Negligible creep curve TNEC for time to 0.2% strain criteria at 0.56$Rm. The Rm
0.5% strain 246.987398 0.147260 22.206093 0.086804
are NIMS average values. The corresponding rupture curve Tr at the same reference
0.2% strain 382.358129 0.149313 33.303955 0.090673
stress is also shown.
202 €m / International Journal of Pressure Vessels and Piping 146 (2016) 198e202
S. Holmstro

500 Time safety margin against rupture is a factor of ~142000 at the

allowable stress Sm
475
5 The new methodology will be used in standards and design
450 rules
Temperature (°C)

425 TNEC curve can be generated using standard creep strength ta-
bles with pre-defined RTF values.
Tr
400
T-0.2% Acknowledgments
375 RCC-MRx strain model
No Creep (NC) The research leading to these results is partly funded by the
350
European Atomic Energy Community’s (Euratom) Seventh Frame-
100 1000 10000 100000 1000000 work Programme FP7/2007-2013 under grant agreement No.
Time (h) 269706 (MATTER project) and in the framework of the EERA (Eu-
ropean Energy Research Alliance) Joint Programme on Nuclear
Fig. 7. Comparison of 0.2% time to strain predictions calculated using the RCC-MRx Materials.
creep strain model at reference stress and the proposed TNEC curve. The Tr curve is
also plotted.
References

5. Conclusions
1 New supporting creep strain data (time to 0.2%) was generated
at low temperatures for negligible creep in the temperature
range 375e500
C
The measured low temperature time to 0.2% creep strain com-
plies with the WE predictions based on higher temperature (NIMS)
data.
2 A reference stress for Gr.91 steel was defined using the RCC-MRx
time independent allowable stress (Sm ¼ 0.37$Rm) and a stress
correction factor of 1.5, i.e. sref ¼ 0.56$Rm.
3. The TNEC curve for Gr.91 was defined at the reference stress
using a new methodology based on the Wilshire equations.
The TNEC curve is based on time to 0.2% creep strain data (NIMS
and supporting tests).
Time to 0.2% strain calculations using RCC-MRx strain equations
comply with the proposed TNEC curve.
4 The conservativeness in time of the generated TNEC curve can be
described as follows:
Time safety margin against rupture is a factor of ~1000 at sref
[1] RCC-MRx. Design and construction rules for mechanical components of nu-
clear installations, Section III e Tome 1 e Subsection B, RB 3216.1, AFCEN.
2012.
[2] EN 10028. Flat products made of steels for pressure purposes. 2009.
[3] ASME. Section III Div. 1 sub-section NH. New York USA: ASME; 2007.
[4] R5 assessment Procedure for the high temperature Response of structures.
Gloucester, UK: British Energy; 2001.
[5] T. Lebarbe, S. Marie, C. Fazio,; S. Gavrilov, P. Agostini, Presentation of FP7
MATTER project, General overview, 2011 Pressure Vessels & Piping Confer-
ence, PVP2011-57620, pp. 401e408, 2011.
[6] EN 13445-3, Unfired pressure vessels, Part 3-Design: 2009.
[7] Holmstro €m S, Luzginova N. Negligible creep of P91 steel. In: Proc. ECCC con-
ference, creep & fracture in high temperature components, design & life
assessment, May 5e7th; 2014.
[8] Riou B. Improvement of ASME NH for Grade 91 negligible Creep and Creep
Fatigue STP-NU-13. 2008.
[9] Riou B. et. Al, Proc. of ASME Pressure Vessels and Piping Division Conference,
PVP2006-ICPVT-11, paper: 93408. Canada, 2006.
[10] Ren W., Lillo T., Totemeier T., Assessment of Negligible Creep, Off-Normal
Welding and heat treatment of Gr91 Steel for nuclear reactor pressure
vessel application, ORNL/GEN4/LTR-06e032, 2006.
[11] NIMS Atlas of creep deformation property. No. D-1, Creep deformation
properties of 9Cr1MoVNb steel for boiler and heat exchangers, 2007.
[12] Wilshire B, Scharming PJ, Hurst R. A new approach to creep data assessment.
Material Sci Eng A 2009;510e511:3e6.
[13] Design and Construction Rules, DCR report N1, Negligible creep for 316L, R5,
Vol 2-3, 1988.
[14] Stefan Holmstro € m. Uniaxial tensile data for P91 in the temperature range
20
C to 600
C, version 1.0. European Commission JRC Institute for Energy and
Transport; 2016 [Catalog], http://dx.doi.org/10.5290/6.
[15] Stefan Holmstro € m. Uniaxial creep data for P91 in the temperature range
375
C to 600
C, version 1.0. European Commission JRC Institute for Energy
and Transport; 2016 [Catalog], http://dx.doi.org/10.5290/8.