Jtam-A4.dvi


JOURNAL OF THEORETICAL

AND APPLIED MECHANICS

52, 1, pp. 181-188, Warsaw 2014

DETERMINATION OF MECHANICAL PROPERTIES OF P91 STEEL BY

MEANS OF MAGNETIC BARKHAUSEN EMISSION

Katarzyna Makowska

Motor Transport Institute, Warsaw, Poland, e-mail: katarzyna.makowska@its.waw.pl

Zbigniew L. Kowalewski

Institute of Fundamental Technological Research Polish Academy of Sciences, Warsaw, Poland and

Motor Transport Institute, Warsaw, Poland, email: zkowalew@ippt.pan.pl; zbigniew.kowalewski@its.waw.pl

Bolesław Augustyniak, Leszek Piotrowski

Gdansk University of Technology, Gdańsk, Poland

e-mail: bolek@mif.pg.gda.pl; lesio@mif.pg.gda.pl

In this work, an attempt at determination of mechanical properties by means of a method
based onmagnetic Barkhausen emissionmeasurements was proposed. The specimens made
of P91 steel were subjected to creep or plastic flow which were interrupted after a range of
selected time periods in order to achieve specimenswith an increasing level of strain. Subse-
quently, measurements of magnetic Barkhausen emission were carried out, and then static
tensile tests were performed in order to check variations of basic mechanical parameters. It
is shown that evident relationships between the yield point/ultimate tensile strength and
some parameters of the Barkhausen emission exist.

Key words: mechanical properties, magnetic Barkhausen effect, plastic strain

1. Introduction

Themethod based on measurements of the magnetic Barkhausen emission/noise (MBE) might
be applied to diagnostics of parts made of ferromagnetic materials. The MBE is a result of
the irreversible movement of magnetic domain walls during a magnetisation cycle (Blaow et
al., 2007). The domain walls are pinned by microstructural barriers (like grain boundaries,
precipitation, dislocation tangles) and released abruptly in the changing magnetic field (Jiles,
2000). The Barkhausen jumps of the domain walls are detected as voltage pulses induced in a
magnetic reading head or in a searching coil.

Due to sensitivity of theMBE to the microstructure of the material, this technique is com-
monly applied to provide information about differentmaterial properties, i.e. material structure
(Saquet et al., 1999; Kleber et al., 2008), grain size (Yamuara et al., 2001), texture (Augusty-
niak, 2003), as well as mechanical properties, i.e. hardness (Gorkunov et al., 2000), residual
stress (Vahista and Paul, 2009) or ultimate tensile stress (Kleber et al., 2008). For example,
Kleber et al. (2008) achieved a linear relationship between the volume fraction of ferrite and
amplitude of MBE peak in a lower magnetic field intensity for a steel with carbon content of
343× 10−3%C, whereas Yamuara et al. (2001) found a linear dependency between rms of the
magnetic Barkhausen emission and grain size for pure iron.

Moreover, microstructural degradation during creep (Mohopatra et al., 2008) or plastic de-
formation (O’Sullivan et al., 2004) of the material may be detected by means of the MBE.
Intensity of theMBE is very sensitive tomaterial damage development, particularly at the ear-
ly stage of degradation (Mohapatra et al., 2010). Mohopatra et al. (2008) examined 5Cr-0.5Mo
steel subjected to creep at 600◦C/60MPabymeans of theMBEmethod.The same testwas also



182 K.Makowska et al.

performed on specimens cut out from a tubemade also from 5Cr-0.5Mo steel. The rms voltage
of the magnetic Barkhausen signal for the virgin specimens decreased at the initial stage of the
creep life due to the newly formed carbides. For the specimen cut from the tube exploited by
15 years the growth of carbides had already taken place, and the rms voltage ofMBE increased
even during the initial stage of laboratory creep testing. As fast as the void started to form in
the specimens (both virgin and service exposed), the rate of the increase of the rms voltage of
the Barkhausen signal started to decrease.
TheMBEmight bealso used to estimatemechanical properties.Kleber et al. (2008) achieved

a linear relationship between the ultimate tensile stress of five commercial DP steels (DP450,
DP500,DP600,DP780,DP1000) andMBEamplitudeaswell asMBEpeakposition for example.
Trillon et al. (2012) found linear relationships between the hardness andMBE peak position of
medium-carbon specimens earlier submitted to a standardized Jominy test.

The aim of this work is to evaluate relationships between selected mechanical parameters
(yield point, ultimate tensile strength) of P91 steel and parameters coming from the non-
destructive magnetic method (amplitude and integral of the MBE). The thesis of this work
is that the basic mechanical properties as the yield point or ultimate tensile stress could be
estimated on the basis magnetic parameters only, without conducting of a static tensile test. It
would be possible in the case when functional relationships between parameters coming from
non-destructive and destructivemethodswould be found.The relationships between parameters
coming from considered magnetic and mechanical tests could be varied for different types of
steel in dependence on, among others, carbon content, grain size andmicrostructure of the steel.
Non-destructive measurements may replace conventional mechanical tests in special cases, for
example when the cutting out specimens from an installation or other devices is impossible.
Determination of mechanical properties using magnetic parameters can be very helpful in es-
timation of the stage of material degradation. This issue is especially important for the power
industry applications since it may lead to reduction of maintenance inspections.

2. Experimental procedure

P91 low-carbon steel, typically used in power industry for structural components such as plates
and tubes, was tested. Its chemical composition is shown in Table 1.

Table 1.Chemical composition of P91 steel

C Si Mn P S Cr Mo

0.085 0.27 0.30 0.015 < 0.01 8.2 0.86

Ni Al Cu Ti Nb V Fe

0.16 0.010 0.14 < 0.01 0.098 0.19 rest

Themicrostructure of the material in the as-received state is presented in Fig. 1.

P91 steel consists of a temperedmartensite phase (Fig. 1). According to Panait et al. (2009)
the microstructure of P91 steel is stabilized by M23C6 carbides (where M = Cr,Fe,Mo) and
MX carbonitrides (where M =V,Nb and X =C,N).

A creep test (T = 500◦C, σ = 290MPa) and a tensile test (T = 25◦C, V = 1mm/min)
were performed on plain specimens having a rectangular cross section of 5mm× 7mm and
gauge length of 40mm.The creep process was interrupted to obtain different deformation levels
i.e.: 0.85%, 1.85%, 3.15%, 4.60%, 5.90%, 7.90% and 9.30%. In the case of the tensile test, the
deformation levels, were as follows: 2.00%, 3.00%, 4.50%, 5.50%, 7.50%, 9.00% and 10.50%.

The magnetic Barkhausen emission signal was measured using the measuring set whoose
diagram is presented in Fig. 2.



Determination of mechanical properties of P91 steel... 183

Fig. 1. Microstructure of P91 steel, etched state (2%HNO3 + etlhyl alcohol), light microscope,
magnification 500×

Fig. 2. Themeasuring set for magnetic properties and Barkhausen emission investigations
(1-specimen, 2-solenoid, 3-pick-up coil, 4-core)

Specimen (1) was magnetized by solenoid (2). Core (4) was used to close magnetic flux.
Pick-up coil (3) provided the induced voltage signal U0. This voltage contains two components:
the low frequency component (related to the main change of the magnetic flux) and a high
frequency component due to the Barkhausen effect. It should be stressed that the second one is
some orders smaller than the first one. The voltage signal U0 was integrated in order to obtain
the hysteresis loop B(H). In order to evaluate intensity of the MBE, the second component
of U0 was separated bymeans of a high-pass filter. TheMBE intensity envelopes of this voltage
were calculated as the rms value Ub according to the equation (Augustyniak, 2003)

Ub =

√

√

√

√

√

1

τ

τ
∫

0

U2
tb1
(t) dt (2.1)

where Ub is the root mean square of the coil output voltage, Utb1 – fast-variable component
defining the voltage separated by means of the high-pass filter from induced voltage in the
pick-up coil, τ – integration time.
Next, the parameter Ubpp – evaluated as the difference between the maximum Ubmax and

the lowest value Ubmin of Ub was calculated using the relation

Ubpp =Ubmax−Ubmin (2.2)

The second descriptor ofMBE – the integral over a half-period of the voltage Usb (signal Ub
corrected on the noise level) was also calculated using formula

Int(Ub)=

+Ugmax
∫

−Ugmax

UsbdUg (2.3)

where

Usb =
√

U2
b
−U2

tb
(2.4)



184 K.Makowska et al.

Theparameter Utb is a rootmean square of the backgroundvoltage. Ug is a voltage proportional
to the magnetizing current intensity. The magnetic parameters were normalized in respect to
their values for the non-deformed specimen (Ubppnorm, Int(Ub)norm).

After magnetic measurements, static tensile tests at room temperature were performed in
order to evaluate the yield point and ultimate tensile strength. Determination of relationships
between the parameters obtained frommechanical (destructive) andmagnetic (non-destructive)
tests was the last point of the experimental programme.

3. Results

Figure 3 presents the MBE envelopes obtained for P91 steel at three extreme stages: (a) non-
deformed, (b) after plastic deformation by a tensile test and (c) after a creep test (with the
highest deformation levels for each type of damage). These plots reveal variation ofMBE inten-
sity for one period ofmagnetization with an evident hysteresis. The arrows depict the ‘direction’
of magnetic field change.

Fig. 3. The rms envelopes of the magnetic Barkhausen emission for P91 steel: (a) non-deformed,
(b) after plastic deformation up to 10.5%, (c) after creep up to 9.3%

In the case of the non-deformed specimen, the MBE intensity plot reveals broad maximum
with two peaks which can be attributed to ‘soft’ and ‘hard’magnetic phases. In P91 steel, these
phases can be attributed to micrograins with low and high density of dislocations (Pesicka et
al., 2003). In comparisonwith the non-deformed specimen, theBarkhausen rms envelopes of the
strained specimens (Figs. 3b-c) have different shapes. The two-peak broad maximum observed
in the non-deformed specimen transforms to a single maximum for specimens strained up to
10.5% in the tensile test and up to 9.3% in the creep test. The shape (height andwidth) of that
maximum varies for all the deformed samples.

Figures 4a,b present how two magnetic parameters calculated from these envelopes
(Ubppnorm and Int(Ub)norm) depend on the deformation level.



Determination of mechanical properties of P91 steel... 185

Fig. 4. Amplitude of the magnetic Barkhausen emission versus pre-strain for P91 steel (a);
integral of a half-period voltage signal of the magnetic Barkhausen emission

versus pre-strain for P91 steel (b)

The amplitude (Ubpp) of the Barkhausen emission of specimens after the creep process
seems to be more sensitive to the strain level than of those after plastic deformation achieved
(Fig. 4a). In the case of the creep test, the abruptpeak of the parameter Ubpp is observed for the
deformation level of about 1%, and – subsequently, a decrease down to nearly the initial level
appears. For specimens deformed bymeans of plastic deformation, only a slight andmonotonic
decrease of this parameter is observed. However, the lack of data around the deformation level
of 1% does not allow one to conclude explicitly whether a maximum of Ubpp in this case also
occurs or not.

A decrease of the magnetic Barkhausen emission integrals Int(Ub) for plastic deformation
can be observed in Fig. 4b. However, its values for specimens after creep are systematically
higher in comparison with those after plastic deformation. Additionally, the plot for the creep
test demonstrates also a single maximum at low values of deformation.

For both types of damage tests, the variations of the MBE amplitude (Ubpp) and inte-
gral, (Int(Ub)) depend on changes occurring in the microstructure. We argue that performed
plastic deformation ofP91 steel leads to an increase of dislocation density. A relatively high level
of plastic deformation (above 2%) is probably associated with formation of extensive disloca-
tion tangles which reduces the mean free path of domain wall displacement as well as increases
the domain wall pining force (Baldev et al., 2001). Movement of the domain walls becomes less
effective, and theMBE signal intensity decreases monotonically (O’Sullivan et al., 2004).

In the case of creep experiment, the observed initial increase of MBE intensity is probably
related to theabruptdecrease of dislocationdensity atmartensite subgrains (temperingprocess).
Material recovery by dislocations cross slips also occurs. The applied stress of a high level
(290MPa) leads also, at the same time, to poligonization of thematerial by climb of dislocations.
The second process is the hardening of the steel by introducing high density dislocation tangles
to the material. The significant amount of dislocation tangles impedes much more movement
of the magnetic domain walls, and the amount of Barkhausen jumps decreases as it was in the
case of plastic flow at room temperature. It should be thusmentioned that the performed creep
experiment is amixed process consisting of ‘creep’ and ‘plastic deformation’ due to a high stress
level.

These differences in P91 microstructure after plastic flow and creep experiments should be
revealed by mechanical tests. The results of static tensile tests of pre-strained specimens are
presented inFig. 5a (yield point) and inFig. 5b (ultimate tensile strength). The ultimate tensile
strength of P91 steel subjected to creep decreases very much whereas the yield point of this
material is insensitive to the deformation level. In the case of plastic flow, a prior deformation
leads to the hardening effect: both mechanical parameters increase.



186 K.Makowska et al.

Fig. 5. Variation of yield point (a) and ultimate tensile strength (b) of P91 steel

It should be noticed that, generally, higher values of magnetic parameters are related to
lower R0.2 and Rm (creep process) and conversely – lower values of the magnetic parameters
are related to higher values of R0.2 and Rm (plastic deformation).
In the next step of the experimental programme possible, relations between the mechanical

andmagnetic parameters are evaluated. Since the yield point of P91 steel subjected to creepwas
insensitive to the deformation level induced by this process, it is not considered in the further
analysis.
Figures 6 and 7 show relationships between two magnetic parameters of MBE (Ubpp) and

Int(Ub) and two mechanical parameters (yield point and ultimate stress). The magnetic para-
meters are normalized to their values for the non-deformed specimen. The numbers in figures
denote the level of prior deformation. Figures 6a,b allow one to conclude that both parameters
(amplitude Ubpp and integral Int(Ub)) of the Barkhausen noise may be used to estimate the
level of the yield point of plastically deformed specimens.

Fig. 6. Variation of the yield point of P91 steel versus amplitude of the magnetic Barkhausen
emission Ubppnorm (a) and versus integral over a half-period voltage signal of the magnetic Barkhausen

emission Int(Ub)norm (b)

Also, the ultimate tensile strength of P91 steel subjected to prior plastic flowmay be deter-
mined using the relationship between Rm and Ubppnorm or Int(Ub)norm (Figs. 7a,b) but only
for the flow experiment. It has to be emphasised that a non unique relationship between Rm
and Ubppnorm and Int(Ub)norm was found for the steel pre-strained by creep.
The relations in Figs. 6 and 7 may be explained as follows. Steel after plastic deformation

with higher R0.2 and Rm is characterised by lower values of the magnetic parameters because
it contains more dislocation tangles that impede movement of the domain walls. On the other
hand, higher values of the magnetic parameters are related to lower Rm of steel after creep.
Our results make evident that MBE intensity varies very much due to microstructure mo-

dification but in different ways: its intensity decreases monotonically after plastic flow (for de-



Determination of mechanical properties of P91 steel... 187

Fig. 7. Variation of the ultimate tensile strength of P91 steel versus amplitude of the magnetic
Barkhausen emission Ubppnorm (a) and versus integral over a half-period voltage signal of the magnetic

Barkhausen emission (b)

formation higher than 2%) and increases (several times) and then decreases during creep. Such
a non-monotonic function makes the direct estimation of themechanical state impossible when
only one magnetic parameter is used for such estimation. Addressing the issue of practical ap-
plication of MBE measurement for the assessment of mechanical properties of damaged steel,
we argue that it is possible but analysing at least twomagnetic parameters with respect to their
initial values. One can find fromFigs. 7a,b that a relative decrease of Ubpp and Int(Ub) denotes
plastic flow, while important increase of Int(Ub) is specific for the early stage of creep damage.
The most difficult case – when advanced creep is in question – can be detected knowing that
the level of Ubpp is higher or close to its initial value, while the level of Int(Ub) decreases. It
should be stressed that such analysis can be enhanced by watching the variation of MBE peak
shape.

4. Conclusions

Themagnetic Barkhausen effect can be helpful for the assessment of selected basic mechanical
parameters like the yield point or ultimate tensile strength for example.

It is shown that the yield point of P91 steel subjected to plastic flow might be determined
by means of the amplitude/integral calculated from the rms magnetic Barkhausen emission
envelopes. Since the yield point of P91 steel was insensitive to the creep pre-strain deformations,
such relationships cannot be formulated for such a kind of damage.The yield point aswell as the
ultimate tensile strength of the tested steel can be determined on the basis of the Barkhausen
noise properties using simultaneously twomagnetic parameters Ubpp and Int(Ub).

Acknowledgments

The research was performed within the financial support of the Motor Transport Institute
(No. 6019/CBM) and theMinistry of Science and Higher Education of Poland (grant No. R15004904).

References

1. Augustyniak B., 2003,Magnetoelactic Effects and their Application in the Non-Destructive Te-
sting of Materials (in Polish),Wydawnictwo Politechniki Gdańskiej, Gdańsk, Poland

2. BaldevR., JayakumarT.,MoorthyV.,Vaidyanatham S., 2001,Characterisationofmicro-
structures, deformation and fatigue damage in different steels usingmagnetic Barkhausen emission
technique,Russian Journal of Nondestructive Testing, 37, 11, 789-798



188 K.Makowska et al.

3. Blaow M., Evans J. T., Shaw B. A., 2007, The effect of microstructure and applied stress
on magnetic Barkhausen emission in induction hardened steel, Journal of Materials Science, 42,
4364-4371

4. Gorkunov E.S., Dragoshanskii Yu.N., Mikhovski M., 2000, Barkhausen noise and its uti-
lization in structural analysis of ferromagnetic materials (Review article V), Russian Journal of
Nondestructive Testing, 36, 6, 389-417

5. Jiles D.C., 2000, Dynamics of domain magnetization and the Barkhausen effect, Czechoslovak
Journal of Physics, 50, 893-988

6. Kleber X., Hug-AmalricA.,Merlin J., 2008,Evaluation of proportion of phases andmecha-
nical strength of two-phase steels usingBarkhausennoisemeasurements: application to commercial
dual-phase steel,Metallurgical and Materials Transactions A, 1308-volume 39A

7. Mohapatra J.N., Bandyopadhyay N.R., Gunjan M.K., Mitra A., 2010, Study of high-
temperature ageing and creep on bainitic 5Cr-0.5Mo steel by magnetic NDE techniques, Journal
of Magnetism and Magnetic Materials, 322, 589-595

8. Mohopatra J.N., Ray A.K., Swaminathan J., Mitra A., 2008, Creep behaviour study of
virginandserviceexposed5Cr-0.5MosteelusingmagneticBarkhausenemissions technique,Journal
of Magnetism and Magnetic Materials, 320, 2284-2290

9. O’Sullivan D., Cotterell M., Cassidy S., Tanner D.A., Mészáros O., 2004, Magneto-
acoustic emission for the characterisation of ferritic stainless steel microstructural state, Journal
of Magnetism and Magnetic Materials, 271, 381-389

10. Panait C., Bendick W., Fuchsmann A., Gourgues-Lorenzon A.-F., Besson J., 2009,
Study of the microstructure of the Grade 91 steel after more than 100’000h of creep exposure at
600◦, 2nd ECCC Creep Conference, Zurich, Switzerland

11. Pesicka J., Kuzel R., Dronhofer A., Eggeler G., 2003, The evolution of dislocation den-
sity during heat treatment and creep of tempered martensite ferritic steels, Acta Materialia, 51,
4847-4862

12. Saquet O., Chicois J., Vincent A., 1999, Barkhausen noise from plain carbon steel: analysis
of the influence of microstructure,Materials Science and Engineering, A269, 3-82

13. TrillonA., Deneuville F., Petit S., BisiauxB., 2012,Magnetic Barkhausen noise for hard-
ness checkingonsteel,18thWorldConference onNondestructiveTesting, 16-20April 2012,Durban,
South Africa

14. Vahista M., Paul S., 2009, Correlation between surface integrity of ground with Barkhausen
noise parameters and hysteresis loop characteristics,Materials and Design, 30, 1595-1603

15. Yamuara S., Furuya Y., Watanabe T., 2001, The effect of grain boundarymicrostructure on
Barkhausen noise in ferromagnetic materials,Acta Materialia, 49, 3019-3027

Manuscript received April 25, 2013; accepted for print August 22, 2013