Monitoring and Analysis of Damage Development in Pipelines
due to Exploitation Loadings
Zbigniew L. Kowalewski*, **, Tadeusz Skibiński***, Jacek SzeląŜek*, Sławomir Mackiewicz*
*
Institute of Fundamental Technological Research
ul. Świętokrzyska 21, 00-049 Warsaw, Poland
zkowalew@ippt.gov.pl, jszela@ippt.gov.pl, smackiew@ippt.gov.pl
**
Institute of Motor Transportation, ul. Jagiellońska 80, Warsaw, Poland
zbigniew.kowalewski@its.waw.pl
***
Institute of Power Engineering, ul. Augustówka 5, 02-981 Warsaw, Poland
tadeusz.skibinski@ien.com.pl
ABSTRACT
The paper presents experimental results of damage development due to creep of steels (A336, 40HNMA, P91) under
uniaxial tension at elevated temperatures. A damage of materials is assessed using destructive and non-destructive
methods. In the destructive method applied the specimens after different amounts of creep prestraining were stretched
up to failure and variations of the selected tension parameters were taken into account for damage identification. The
ultrasonic investigations were selected as the non-destructive method of damage development evaluation. In the case
of this method the elasto-acoustic coefficient and acoustic birefringence coefficient were used to identify damage
development in the tested materials. In order to get more thorough understanding of the phenomena associated with an
influence of deformation history on mechanical properties of materials the paper considers two types of deformation
processes: besides of deformation due to creep, a deformation due to plastic flow caused by the monotonically
increasing load is considered.
Introduction
A majority of structural elements and devices are subjected to loadings having long time character and often associated
with elevated temperatures. Such conditions stimulate development of creep phenomenon usually defined as a process
which takes place under constant loads giving stress levels typically lower than the conventional yield point of a
material. In the creep process three periods can be distinguished; primary, secondary and tertiary. From previous
microscopic investigations it is well known that damage of a material already appears at primary creep, and in the next
stages damage changes only its velocity [1]. Therefore, it seems to be interesting to assess how the damage processes
in the subsequent creep stages can change mechanical properties of materials. Moreover, it is also interesting to know
how the same mechanical parameters variations are related to a type of loading process leading to the same amount of
strain. The available results are still not sufficient, since they have been achieved mainly from uniaxial stress state tests,
and moreover, they cover too narrow group of materials [2-6]. Hence, it is important to carry out investigations under
multiaxial stress conditions which give greater opportunity to increase our knowledge dealing with a complex problem of
deformation taking place in the commonly used devices or responsible elements of constructions. Due to extremely high
costs and complexity of such type investigations, their number is limited, and as a consequence, the data available are
still insufficient to create relatively simple, but simultaneously, reasonable and effective constitutive models. In order to
get more thorough understanding of the phenomena associated with an influence of deformation history on mechanical
properties of materials the paper considers two types of deformation processes: deformation due to creep and
deformation due to plastic flow caused by the monotonically increasing load.
Details of the Materials and Experimental Programme
All tests presented in this paper were carried out on three kinds of steel commonly applied at elevated temperatures:
A336 (used mainly at chemical industry), 40HNMA and P91 (both used mainly at power plants). Chemical compositions
of these materials are given in Tables 1, 2 and 3.
Table 1. Chemical composition of A336 GR5
C
Mn
Si
P
S
Cr
Ni
Mo
V
Cu
0.15
0.57
0.10
0.007
0.008
3.2
0.17
1.04
0.25
0.05
Cr
0.60
Ni
1.25
Mo
0.15
V
W
Ti
Cu
As
0.90
1.65
0.25
Max
0.05
Max
0.20
Max
0.05
Max
0.25
Max
0.08
Cr
8.00
Ni
Mo
0.85
V
0.18
Al
Nb
0.06
1.05
0.25
Table 2. Chemical composition of 40HNMA steel
C
0.37
Mn
0.50
Si
0.17
0.44
0.80
0.37
P
S
Max
0.03
Max
0.025
Table 3. Chemical composition of P91 steel
C
0.08
Mn
0.30
Si
0.20
0.12
0.60
0.50
P
S
Max
0.015
Max
0.010
max
0.40
9.00
Max
0.04
0.10
The experimental programme comprised tests for the materials in the as-received state and for the same materials
subjected to prior deformation due to creep at elevated temperatures, Fig.1, and due to plastic flow at room
temperature, Fig.2.
P
P
P=f(t)
P=const
0
t
Fig. 1. Loading of specimens for creep
0
t
Fig. 2. Loading of specimens for plastic deformation
Uniaxial tension creep tests were carried out on A336, 40HNMA and P91 steels using plane specimens, Fig.3. For each
steel all tests were conducted in the same conditions. In the case of A336 the stress level was equal to 425 MPa, and
temperature - 698 K, for 40HNMA they were 250 MPa, and 773K, whereas for P91 – 290 MPa, and 773K, respectively.
In order to assess a damage development during the process of creep the tests for A336 steel were interrupted for a
range of the selected time periods 50h, 75h, 100h, 128h, 135h which correspond to the increasing amounts of creep
strain equal to 1.72%, 1.98%, 2.57%, 5.5%, 7.0%, respectively (Fig. 4). In the case of 40HNMA the tests were
interrupted after 100h (0.34%), 241h (0.8%), 360h (1%), 452h (1.1%), 550h (1.2%), 792h (2,3%), 929h (4.0%) and 988h
(6.5%), Fig. 5, while for P91 after 40h (0.85%), 180h (1.85%), 310h (3.15%), 390h (4.6%), 425h (5.9%), 440h (7.9%)
and 445h (9.3%), Fig. 6.
The same magnitudes of deformation were applied to prestrain specimens by means of plastic flow at room
temperature. After each prestraining test a damage of specimen was assessed using the non-destructive methods. Two
non-destructive methods were applied: magnetic (the results are reported elsewhere) and ultrasonic. In the next step of
experimental procedure, the same specimens were mounted on the hydraulic servo-controlled MTS testing machine and
then stretched until the failure was achieved. The last step of the experimental programme contains microscopic
observation using optical and scanning microscopes.
60°
18
R1
R1
T = 698 [K]
Creep strain [%]
14
Φ 10 H7
12
8
12
7
10
=5
40
8
5
6
4
4
15
30
15
12 8
40
8
3
2
1
2
0
30
0
140
40
80
120
160
Time [h]
Fig. 3. Specimen
Fig. 4. Creep curve of A336 steel with points representing
interrupted creep tests
12
σ = 250 [MPa]
T = 773 [K]
10
8
8
6
7
4
6
2
1
2
5
3 4
0
Creep strain [%]
12
Creep strain [%]
6
σ = 425 [MPa]
16
10
σ = 290 [MPa]
T = 773 [K]
8
7
6
6
4
4
2
2
1
5
3
0
0
200 400 600 800 1000 1200
Time [h]
Fig. 5. Creep curve of 40HNMA steel with points
representing interrupted creep tests
0
100
200
300
Time [h]
400
500
Fig. 6. Creep curve of P91 steel with points representing
interrupted creep tests
Standard Tension Tests as a Tool of Damage Assessments
Variations of the basic mechanical properties of all tested steels, i.e. Young’s modulus, conventional yield point, ultimate
tensile strength and elongation, due to deformation achieved by prior creep or plastic flow were determined. The
selected results are illustrated in Figs. 7, 8 and 9. The entire tension characteristics are presented only for 40HNMA
steel. They were shown as an example to give illustration of typical variations observed. On the basis of tension
characteristics a selected mechanical parameters were determined. For each of the tested materials the Young’s
modulus is almost not sensitive on the magnitude of creep and plastic deformations. Contrary to the Young’s modulus
the other considered tension test parameters, especially the conventional yield point and the ultimate tensile strength,
exhibit clear dependence on the level of prestraining, Figs. 8ab and 9ab. Taking into account the results for A336 steel
we can conclude that assessments of creep damage development only on the basis of mechanical parameters
determined from the standard tension tests for the material subjected to various amounts of creep prestraining do not
allow to evaluate accurately whether some exploitation elements can further work safely or should be exchanged to
protect a construction against the premature failure. However, in the case of 40HNMA and P91 steels such tests well
described damage due to creep. As we can see, a sensitivity of tension test parameters on creep damage depends on a
type of material and conditions of creep investigations. To get more complete knowledge concerning damage it is
necessary in some cases (A336) to carry out additional tests using specimens with similar magnitudes of prestraining as
those during creep achieved, but induced in a different way. An example of the comparative results concerning
variations of the conventional yield point for A336 and 40HNMA steels tested after two different kinds of prestraining is
presented in Figs.8a and 9a.
(b)
1400
1200
As-received
1100
state
1000
6
900
7
4
800
32 1
8
700
5
600
500
400
300
200
100
0
0.00 0.05 0.10 0.15 0.20 0.25
Strain
1200
Stress [MPa]
Stress [MPa]
(a)
1
1000
800
4
3
600
2
As-received
state
400
200
0
0.00
0.05
0.10
Strain
0.15
0.20
Fig. 7. Tension characteristics of 40HNMA: (a) after various creep deformation, (b) after various plastic deformation
(numbers in (a) correspond to those presented in Fig. 5, and numbers 1 to 4 in (b) correspond to plastic deformation
equal to 0.5%, 1.5%, 7.0%, 10.5%, respectively)
(b)
670
700
650
600
550
500
450
400
Rm [MPa]
R0,2 [MPa]
(a)
660
650
640
630
0
2
4
6
Prior deformation [%]
8
0
2
4
6
Prior deformation [%]
8
Fig. 8. Selected results for A336 steel subjected to prior deformation: (a) conventional yield limit;
(b) ultimate tensile stress; (solid lines – after creep, broken lines – after plastic flow)
(b)
1200
1400
1000
1200
800
Rm [MPa]
Yield point [MPa]
(a)
600
400
200
1000
800
600
400
200
0
0
0
2
4
6
8
10
Prior deformation [%]
12
0
2
4
6
8
10
Prior deformation [%]
12
Fig. 9. Selected results for 40HNMA steel subjected to prior deformation:
(a) conventional yield limit; (b) ultimate tensile stress (solid lines – after creep, broken lines – after plastic flow)
Creep Damage Evaluation Using Non-Destructive Testing Methods
The ultrasonic wave velocity and attenuation are acoustic parameters most often used to assess a material damage due
to creep or fatigue. Results of investigations [7, 8] show that the attenuation of ultrasonic waves is in practice stable
until the last creep or fatigue stages. It was also observed that velocity changes due to creep or fatigue are small, and
therefore, an application of velocity measurement for damage evaluation, at industrial conditions, is very difficult.
Difficulties in attenuation and velocity measurements, or their combinations, are result of heterogeneous acoustic
properties of technical materials, like steel for example. The second reason is dependence of both attenuation and
velocity of ultrasonic waves on numerous factors other than material damage. This observation is confirmed by the
results presented in [3] where steel samples subjected to 10% plastic deformation and subjected to 140 000 hour loads
at elevated temperature were measured. The results showed that ultrasonic wave attenuation is not influenced by
deformation or long time, high temperature load exposure.
In this work in order to evaluate a damage progress in samples made of three kinds of steel, instead of velocity and
attenuation measurement, the elastoacoustic coefficients β and the acoustic birefringence B were measured [4, 5].
Specimens were subjected to creep deformations according to the programmes presented in Figs. 4, 5 and 6.
The elastoacoustic coefficient β is the measure of elastoacoustic effect and describes ultrasonic wave velocity changes
due to stress variation. A value of β depends on the material grade, mode of ultrasonic wave and correlation between
ultrasonic wave propagation, polarization and stress directions. The highest value of β is observed for longitudinal wave
propagated along the stress.
For this wave, propagated along the axis of the sample subjected to tensile test, elastoacoustic constant is calculated
as:
β=
(t 0 − t σ )
(1)
tσ σ
where:
to – time of flight of ultrasonic pulse for zero stress level,
tσ – time of flight corresponding to applied stress,
σ – stress level.
A scheme illustrating the method for the elastoacoustic coefficient experimental determination is presented in Fig. 10.
Receiving
probe
Transmitting
probe
Wave paths in the
sample
40
Fig. 10. Distribution of probes during wave propagation measurements for the elastoacoustic coefficient determination
under tension
It was shown that this parameter is sensitive to prior creep deformation in the case of A336 steel, for 40HNMA and P91
steels, however, it was completely not damage sensitive.
The acoustic birefringence B is a measure of material acoustic anisotropy. It is based on the velocity difference of two
shear waves polarized in perpendicular directions. In the specimen subjected to creep the shear waves propagated in
the specimen thickness direction and were polarized along its axis and in the perpendicular direction. The birefringence
B was calculated as:
B=
2 (t I − t p )
tI + tp
= B0 + Bp
(2)
where:
tl – time of flight of ultrasonic shear wave pulse for wave polarization direction parallel to the sample axis,
tp – time of flight of ultrasonic shear wave pulse for wave polarization perpendicular to the sample axis,
B0 – acoustic birefringence for the material in the virgin state (before creep test),
BP – acoustic birefringence for the material after creep.
A scheme of the birefringence measurements is presented in Fig.11. Figure 12 shows mean values of the acoustic
birefringence measured in specimens after creep test. In the deformed part of specimen a value of the birefringence
depends on the deformation amount. The same feature can be observed for all kinds of steel under the question. The
results show that the acoustic birefringence can be a good indicator of material degradation and can help to localize the
regions where material properties are changed due to creep. This is consistent with microscopic observations of tested
steels performed after all creep tests. The microscopic investigations allowed to observe gradual development of
damage due to creep. As an example of macroscopic observations, the results for 40HNMA steel are presented in
Fig.13. It shows microscopic views of creep prestrained specimens without etching. For the primary and secondary
creep periods the number of voids was limited. It started to increase rapidly from the beginning of the tertiary creep.
A
B
Fig. 11. Scheme of birefringence measurement, A – probe at gauge length
B - probe at gripping part of specimen
(a)
(b)
0.003
0.002
0.001
0.000
-0.001
B
B0
0 1 2 3 4 5 6 7
Creep deformation [%]
Acoustic
birefringence [-]
Acoustic
birefringence [-]
0.004
0.0160
0.0140
0.0120
0.0100
0.0080
0.0060
0.0040
0.0020
0.0000
-0.0020
B
Bo
0
2
4
6
8
10
12
Creep deformation [%]
Fig. 12. Acoustic birefringence B variation due to prior creep deformation for: (a) A336 steel; (b) 40HNMA steel
(a)
(b)
(c)
Fig. 13. Microscopic view of specimens after creep carried out up to: (a) 360 h; (b) 550 h; (c) 988 h
In order to check sensitivity of the birefringence coefficient into damage evaluation, it was also measured in the part of
specimen fixtures, where a texture of material was assumed to be unchanged during creep test, Fig. 12. Values of
birefringence measured in the fixture show some scatter around zero value. This scatter is a picture of birefringence
evaluation accuracy and an initial acoustic homogeneity of specimens material. In the deformed part of specimen a
value of the birefringence depends on deformation amount. It can be noticed that birefringence variations due to creep
are significantly higher than birefringence scatter in a virgin material. The same feature is also observed for 40HNMA
steel.
In the case of 40HNMA steel the birefringence coefficient was also determined for the material subjected to prior
deformation induced by means of plastic flow due to monotonic loading at room temperature. The results are illustrated
in Fig. 14.
Acoustic
birefringence [-]
0.0020
Bo
0.0000
-0.0020
B
-0.0040
-0.0060
-0.0080
-0.0100
0
2
4 6 8 10 12 14 16 18 20
Prior deformation [%]
Fig. 14. Acoustic birefringence B variation due to prior plastic deformation of 40HNMA steel (Bo – acoustic
birefringence in the fixture part of specimen)
As it is clearly seen from comparison of the birefringence values presented in Figs. 12b and 14, this parameter is
sensitive not only on a damage development, but also on a type of deformation process, and therefore, can be treated
as a very promising tool in many engineering applications related to different kinds of diagnostics.
Having determined the parameters of destructive and non-destructive methods for damage development evaluation it is
worth to analyze courses of their variation in order to find possible correlation. This is because of the fact that typical
macroscopic creep investigations give the macroscopic parameters characterizing the lifetime, strain rate, ductility,
without any information concerning microstructural damage development and material microstructure variation. On the
other hand, the non-destructive methods enable to provide a knowledge concerning damage development in a particular
time of the entire working period of an element, however, without sufficient information dealing with the microstructure
and time of its further secure exploitation. Therefore, it seems to be reasonable to plane future damage development
investigations in the form of interdisciplinary tests connecting results achieved using destructive and non-destructive
methods with microscopic observations in order to find mutual correlation between their parameters. In our case a good
correlation of the selected mechanical and ultrasonic parameters identifying creep damage is observed for 40HNMA
steel. It is easy to see such correlation looking, for example, at Figs. 9a, 12b and 14. The yield point of the steel
subjected to creep decreases. It corresponds to the increase of the acoustic birefrigerence. In the case of the same
material prestrained due to monotonic loading the yield point increases with the increase of prior deformation
magnitude. It corresponds to the decrease of the acoustic birefrigerence. Such mutual relationships enables to predict
the yield point variation on the basis of the ultrasonic damage measurements, and as a consequence, allows to find
corresponding point on the creep curve. This information enables determination of the remain time to rupture.
Conclusions and Remarks
The effects associated with different types of prior deformation are studied for three kinds of steel: A336, 40HNMA and
P91.
The results show that an influence of prior deformation on the subsequent material behaviour depends not only on the
magnitude of prestrain, but also on the way in which such deformation is achieved.
The prior deformation induced during the monotonic loading of A336 steel at room temperature causes hardening. In the
case of deformation induced by means of creep at elevated temperature also hardening effect was observed, however,
its magnitude was significantly lower.
The results presented in this paper exhibits that future damage development investigations should be planned in the
form of interdisciplinary tests connecting results achieved using destructive and non-destructive methods with
microscopic observations in order to find mutual correlation between their parameters. Such approach can be treated as
a very promising tool in planning of necessary repairs and predictions of a safety margin not only in pipelines working at
power plants, but also in many responsible elements working in other industrial plants.
Acknowledgement
The support of the State Committee for Scientific Research (KBN) under grant 4 T07A 018 26 is greatly acknowledged.
References
[1] Boettner R.C., Robertson W.D., Study of growth of voids in copper during creep process by measurement of
accompanying change in density, Trans. Metall. Soc. A.I.M.E., 221, 613, 1961.
[2] Dyson B.F., Rodgers M.J., Prestrain, cavitation and creep ductility, Metal Sci., 8, 261, 1974.
[3] Dyson B.F., Loveday M.S. and Rodgers M.J., Grain boundary cavitation under various states of applied stress
Proc. R. Soc. Lond. A, 349, 245, 1976.
[4] Marlin R.T., Cosandey F. and Tien J.K., The effect of predeformation on the creep and stress rupture of an oxide
dispersion strengthened mechanical alloy, Metallurgical Transactions A, 11A, 1771, 1980.
[5] Pandey M.C., Mukherjee A.K., Taplin D.M.R., Prior deformation effects on creep and fracture in inconel alloy X750, Metallurgical Transactions A, 15A, 1437-1441, 1984.
[6] Trąmpczyński W.A., The influence of cold work on the creep of copper under biaxial states of stress, Acta Metall.,
30, 1035-1041, 1982.
[7] Narayan R., Green R. E., Jr., Ultrasonic attenuation monitoring of fatique damage in nuclear pressure vessel steel
at high temperature, Materials Evaluation, 25- 36, February 1975.
[8] Fel D., Hsu D.K., Warchol M., Simultaneous velocity, thickness and profile imaging by ultrasonic scan, Journal of
Nondestructive Evaluation, 8, 95-112, 2001.