EUROSTEEL 2014, September 10-12, 2014, Naples, Italy

HIGH FREQUENCY HAMMER PEENING AT WELDED ULTRA HIGH

STRENGTH STEELS
Jrn Berg, Natalie Stranghner
University of Duisburg-Essen, Institute for Metal and Lightweight Structures, Essen, Germany
joern.berg@uni-due.de, natalie.stranghoener@uni-due.de

INTRODUCTION
Ultra high strength fine grained structural steels (UHSS) with yield strengths up to 1100 N/mm2 are
often used in mobile crane structures to minimize dead loads. Due to applied fatigue loads during
operation the lifetime of these welded structures is limited and can be classified into the upper finite
respectively low cycle fatigue life region. The fatigue life of welded steel joints is dominated by
crack propagation from crack like imperfections at the weld toe and weld root areas. Therefore, the
fatigue design of welded steel joints according to different design rules is independent from the
yield strength. By the application of post weld treatment methods like high frequency hammer
peening (HFHP) the fatigue behaviour can be improved. HFHP mainly induces compressive
residual stresses in the treated weld toe area resulting in a translation and a rotation of the S-Ncurve. Therefore, the S-N-curves of as welded (m = 3) and HFHP treated (m =5) notch details
intersect theoretically in the upper finite fatigue life region. Up to now, the influence of HFHP on
the fatigue behaviour has mainly been investigated at steels with yield strengths smaller than or
equal to 960 N/mm2 within research activities. Consequently, existing design recommendations for
the influence of HFHP on the fatigue strength are limited to maximum steel grades of S960 and
plate thicknesses of 5 mm and higher. For this reason, further investigations have been performed to
transfer the results and applicability of HFHP to welded UHSS with steel grades up to S1300 in the
LCF and upper finite fatigue life region.
1

STATE OF THE ART

The influence of HFHP on the fatigue behaviour of welded joints has been increasingly investigated
scientifically in the past years. The application of HFHP yields in a local, plastic deformation of the
weld toe surface resulting in a modified residual stress state by inducing compressive residual
stresses. Furthermore, HFHP also results in cold hardening of the near surface region and rounding
of the weld toe. The size of compressive residual stresses depends on the yield strength fy of the
treated material and increases with increasing yield strength. Residual stress measurements [1-8]
show that compressive residual stresses at the treated surface transverse to the welding direction can
reach values of approximately 75 % of fy and values that are higher than fy because local strength
can be increased due to cold hardening. Due to the modified residual stress state the slope of the
S-N-line of a HFHP treated weld toe increases to approximately m ~ 5.
[3]
[10]
[11]
[12]
[13], [14]

180
160
140
120
100

960

80

1300

As Welded

60

R = 0.1

40
200

400

600

800

1000

yield strength fy [N/mm2]

1200

1000
Stress range [N/mm2]

FAT class due to HFH [N/mm2]

200

m=5
1

100
LCF
low cycle
fatigue
10
10,000

180
160
125

m=3
80

112

40,000 100,000
1,000,000
Load cycles Nf [-]

Fig. 1. Design FAT classes (left) and maximum possible design S-N-lines (right) for HFHP treated notch detail of
welded transversal stiffener according to different design proposals

Nomenclature
C
m
AW
fy
HFHP
LCF
m
Nf
PIT
R
t
UHSS

characteristic value of fatigue strength at 2 106 load cycles corresponding to 95 % survival probability
mean value of fatigue strength at 2 106 load cycles corresponding to 50 % survival probability
as welded
yield strength
high frequency hammer peening
low cycle fatigue
slope of S-N-line in finite fatigue life region
load cycles until failure
Pneumatic Impact Treatment
stress ratio defined by min / max
plate thickness
ultra high strength steels

In high stressed structures residual stresses can relax resulting in a decrease of the improvement
effect of HFHP. The improvement effect of the fatigue strength due to HFHP treatment increases
with increasing materials yield strength [9] as this is related to the higher induced compressive
residual stresses at higher steel grades. Therefore, fatigue class improvements due to HFHP
treatment are proposed in different design recommendations depending on the steel grade [3,
10-12]. However, these design proposals are limited to maximum steel strengths of S960 and plate
thicknesses of 5 mm and higher, see Fig. 1.
The analysis of existing fatigue tests at as welded and HFHP treated notch details of ultra high
strength fine grained structural steels showed demand for further research activities. Therefore,
within this contribution the following questions will be clarified:
How much is the influence of HFHP on the fatigue strength improvement at UHSS with
yield strengths of 960 N/mm2 and higher?
At which stress ranges respectively number of load cycles do the S-N-lines of untreated and
HFHP treated notch details intersect?
Does the application of HFHP also result in a fatigue life increase in the LCF respectively
upper finite fatigue life region?
2

EXPERIMENTAL INVESTIGATIONS

2.1 Test programme

The Institute for Metal and Lightweight Structures of the University of Duisburg-Essen has
performed fatigue tests on the four different welded notch details longitudinal stiffener, transversal
stiffener, cover plates and butt weld of UHSS S960, S1100 and S1300 to determine the influence of
HFHP on the fatigue behaviour, see Table 1. The weld toe condition of the test specimens varied in
as welded and HFHP treated. In total 119 fatigue tests have been performed.
2.2 Test specimens
The test specimens have been produced from ultra high strength, waterquenched and tempered fine
grained heavy plates of steel grades S960, S1100 and S1300 and with plate thicknesses of 4 mm to
8 mm. All test specimens have been welded manually by MAG process with filler material X90.
Herewith, the nominal value of the yield strength of the filler material is below the yield strength of
the base materials. However, during welding process the weld metal is alloyed by the base material
(undermatching). Actually no adequate filler materials with yield strengths higher than
960 N/mm2 are available for welding steel grades S1100 and above so that the use of
undermatching weld metal is common during the fabrication of mobile crane structures [15]. For
the production of the notch details cover plate and longitudinal stiffener no weld starts and stops
have been positioned at the attachment ends as these points are the fatigue critical regions of as
welded specimens. For the butt welded test specimens two different welding executions have been
investigated. The specimens of test series 96-7-S have been welded one sided with full penetration.
The specimens of test series S11-6-S have been welded both sided with transition in thickness and
alignment on the root side.

Table 1.

Programme of the fatigue tests

Notch detail

Test series

No. of tests1)

Steel grade

t [mm]

Longitudinal
stiffener

S11-6-LS

5/6

S1100

max / fy
25-55 %

S13-4-LS

5/5

S1300

30-50%

11-6-QSd

5/6

S1100

45-70 %

11-6-QS

5/5

S1100

35-75 %

S13-4-QS

5/5

S1300

30-50 %

96-7-La

4/5

S960

7.5

40-65 %

S11-6-La

6/6

S1100

30-55 %

S13-4-La

5/5

S1300

30-45 %

Butt weld

96-7-S

4/4

S960

7.5

35-65 %

Butt weld with

transition in thickness

S11-6-S

6 / 12

S1100

6/8

35-60 %

Transversal stiffener

Cover plate

1)

Number of tests in as welded and high frequency hammer peened toe condition, respectively.

2.3 Post weld treatment by high frequency hammer peening

After welding, approximately half of all test specimens have been treated by HFHP at the weld toes.
The HFHP treatment has been mainly performed by Pneumatic Impact Treatment (PIT) with 90 Hz
and with radii of the indenters of 2 mm for the steel grades S960 and S1100 and 1.5 mm for S1300.
Six specimens of test series S11-6-S have been treated by high frequency impact treatment (HiFIT)
with a radius of indenters of 1.5 mm. The HFHP treatment has been applied at all weld toes to the
base material. At the notch details of welded cover plate and longitudinal stiffener HFHP treatment
has been focused near the attachment ends as HFHP treatment along the complete weld lines is not
necessary for these notch details. The HFHP process and the resulting plastic deformations at the
weld toes can be seen in Fig. 2.
direc
t
appli ion of
catio
n

angle of
application
~90

of
nter
inde
ol
P to
HFH

plastic deformation
of weld toe

depth e
radius r

weld metal

base material

Fig. 2. Local application of high frequency hammer peening at weld toe region

2.4 Experimental testing

The fatigue tests have been performed load controlled with a sinus shaped load-time-function and
constant load amplitudes resulting only in tensile stresses. The specimens have been axially loaded
with a stress ratio of R = 0.1. The fatigue tests were carried out until failure of the specimens by
fracture. The fatigue loads were iteratively determined in order to achieve load cycles until failure

of 10,000 to 40,000 for the as welded specimens to cover the LCF. This fatigue life region is an
important operational region of the investigated steel grades, especially regarding fatigue stressed
components of mobile crane structures. The HFHP treated specimens have been tested at the same
fatigue loads.
3

EXPERIMENTAL RESULTS AND DISCUSSION

3.1 Failure modes

As expected all as welded test specimens failed due to crack initiation from the weld toes to the
base material. Due to HFHP treatment the crack initiation changes in 32 % of all cases to edge or
surface notches in the base material or the clamping area (transversal stiffener) or the weld root area
(longitudinal stiffener and cover plate), see Table 2 and Fig. 3. The crack initiation mainly changes
at relatively low stress ranges as the induced compressive residual stresses do not relax the same
way as at higher loads.
Table 2.
Notch detail

Longitudinal stiffener

Transversal stiffener

Cover plate

Failure modes of HFHP treated specimens during fatigue tests

Test series
S11-6-LS

S1100

Crack initiation from

Weld toe
Weld root
5
1

S13-4-LS

S1300

11-6-QSd

S1100

11-6-QS

S1100

S13-4-QS

S1300

96-7-La

S960

S11-6-La

S1100

S13-4-La

S1300

S960

S1100

8 / 42)

40

Butt weld
96-7-S
Butt weld with transition in
S11-6-S
thickness

Steel grade

1)
2)

a)

Base material1) Clamping

-

Edge or surface notches in the base material.

Eight specimens failed due to crack initiation from weld toe at top layer. Four specimens with angular
misalignment oriented to the top side failed due to crack initiation from weld toe at sealing run.
crack initiation from weld toe: max = 0.55 fy

b)

crack initiation from weld root: max = 0.30 fy

HFHP treatment line

Fig. 3. Fracture type and fracture surface of two test specimens of the notch detail welded cover plate treated by HFHP

For the notch detail of welded transversal stiffener of S1100 only 2 of 11 HFHP treated specimens
failed due to crack initiation from the weld toes. Consequently, especially at lower stress ranges the
fatigue life of HFHP treated weld toes can be limited due to adjacent notches in the base material or
in the weld root.
3.2 Fatigue life and fatigue strength
The load cycles until failure Nf for as welded specimens range from 5,000 to 110,000 and cover the
LCF and upper finite fatigue life region. The test results of all as welded specimens are above the SN-lines of relative FAT classes according to EC 3 [14] and IIW [15] and their statistical evaluation
shows good agreement with the given FAT classes, see Fig. 4. As expected, the yield strength does
not influence the fatigue behaviour of as welded specimens.
By the application of HFHP, the number of load cycles Nf increases rapidly in comparison to the as
welded toe condition. The slopes of the S-N-lines are much shallower than for the as welded toe
condition and increase to m ~ 5 due to residual stress relaxation, see Fig. 5. One exception are the
results of the notch detail of welded transversal stiffener (test series 11-6-QSd, 11-6-QS) with
mainly crack initiation from notches in the base material or clamping area, where the slope of the
S-N-line remains at m ~ 3. The mean value of the fatigue strength m has been evaluated with a
fixed slope of m = 3 for the as welded and m = 5 and additionally with variable slope for the HFHP
treated toe condition for each test series, see Fig. 4. Because of the relatively low number of fatigue
tests per test series all results of each test series have been used for the statistical evaluation which
includes results with crack initiation from different notches.
The mean value of the fatigue strength m increases due to HFHP treatment by factors of 2.16 to
2.65 in comparison to the as welded toe condition, when a fixed slope of m = 5 is used. For the
notch details longitudinal stiffener and butt weld the fatigue strength improvement increases by
approximately 15 % and 10 % with an increase of yield strength from 1100 to 1300 N/mm2
(longitudinal stiffener) respectively 960 to 1100 N/mm2 (butt weld). The results of the notch details
transversal stiffener and cover plate do not show any increasing effect of HFHP treatment with
increasing yield strength.
The number of load cycles until failure Nf of the test specimens with as welded respectively HFHP
treated toe condition intersect theoretically in the LCF region with load cycles of approximately
4,000, see Fig. 5. The fatigue life can be improved by factors of 2 to 10 due to HFHP treatment in
the upper finite fatigue life region with load cycles Nf of 10,000 to 40,000 for the as welded toe
condition.
The influence factor of the yield strength cannot be analysed explicitly due to different plate
thicknesses and partially different local weld geometries. It is assumed that the residual stress state
due to HFHP treatment at low plate thicknesses cannot be realized in the same way as at larger plate
thicknesses. The scatter of local weld geometries at different test series and partially observed
angular distortion due to the low plate thicknesses are not covered on the action part within the
nominal stress design concept which influences the comparability of the test results.
350

AW - m = 3.0

HFH - m = var.

HFH - m = 5.0

[14] EC 3 (AW)

[15] IIW (AW)

250
200
150
100
m = 3.0
m = 3.0

m, 50% [N/mm2]

300

50
0

96-7-La
S960

S11-6-La S13-4-La S11-6-LS S13-4-LS 11-6-QS S13-4-QS

S1100

S1300

S1100

S1300

S1100

S1300

96-7-S

S11-6-S

S960

S1100

Fig. 4. Comparison of the mean value of the fatigue strength m, 50 % with the FAT classes according to [14] and [15]

Therefore, the test results will be evaluated considering structural hot spot stress and notch stress
design concepts which will be content of further publications. Furthermore, the strength at local
weld toe, which influences the compressive residual stresses due to HFHP treatment, is limited by
the material properties of the filler material.
1,000,000

[-]

10

related stres range /m,

AW

376,400

HFHP

100,000

cover plate

long. stiffener

5.0

transv. stiffener

1
1,000

Nf, HFH, exp. [-]

As Welded

3.0

100,000
Load cycles Nf [-]

10,000

cover plate
long. stiff.
transv. stiff.
butt weld

4,100

butt weld
10,000

24,100

1,000
1,000

1,000,000

10,000
100,000
Nf, AW, exp. [-]

1,000,000

Fig. 5. Related fatigue tests results (left) and comparison of load cycles until failure Nf (right) of the test specimens
with as welded and high frequency hammer peened toe condition

10
1,000

Stress range [N/mm2]

10,000

100

1000

56

180
m

140

[3] - m = 5
[10] - m = 5
[11] - m = 3
[14] - m = 3 (AW)
96-7-S-PIT

71

10,000
100,000
Load cycles Nf [-]

1,000,000

[11] - m = 3
[14] - m = 3 (AW)
96-7-La-PIT
S11-6-La-PIT
S13-4-La-PIT
diff. crack initiation
diff. crack initiation

10,000

180
160
100

10
1,000

1
[3] - m = 5
[10] - m = 5
[11] - m = 3
[12] - m = 5
[14] - m = 3 (AW)
11-6-QS-PIT
S13-4-QS-PIT
diff. crack initiation

10,000

1000

100

100

10
1,000

90

100,000
1,000,000
Load cycles Nf [-]

10
1,000

112

[3] - m = 5
[10] - m = 5
[11] - m = 3
[14] - m = 3 (AW)
S11-6-LS-PIT
S13-4-LS-PIT
diff. crack initiation
diff. crack initiation

Stress range [N/mm2]

Stress range [N/mm2]

100

1000

Stress range [N/mm2]

160

Stress range [N/mm2]

1000

1000

125

112
80

100,000
1,000,000
Load cycles Nf [-]

140
100

10
1,000

[3] - m = 5
[10] - m = 5
[11] - m = 3
[14] - m = 3 (AW)
S11-6-PIT
S11-6-HiFIT
diff. crack initiation
diff. crack initiation

10,000

100
m

80
51

100,000
1,000,000
Load cycles Nf [-]

71
56

100,000
1,000,000
Load cycles Nf [-]

Fig. 6. Classification of test results in existing design proposals for high frequency hammer peened toe condition

3.3 Design proposals

Proposed FAT classes of design recommendations are conservative in comparison to the test results
of HFHP treated specimens, see Fig. 6. The design proposal of Yildirim [10] shows the best
agreement with the test results because this approach covers yield strengths up to 960 N/mm2. The
results at HFHP treated cover plates show that the proposal of Haagensen and Maddox [11] is too
conservative especially in the finite fatigue life region. A significant improvement of the fatigue
class due to HFHP treatment with a change of the slope to m = 5 is possible if crack initiation from
the weld root side can be excluded. For the notch detail butt weld with transition in thickness a
further improvement to FAT class 160 is possible.
3.4 Transferability of results and quality assurance
Due to HFHP treatment the fatigue life can be improved in the upper finite fatigue life region with
maximum stresses max up to 75 % of the materials yield strength. Separate overloads which
appear during operation of mobile cranes can decrease the improvement effect of HFHP due to
residual stress relaxation. Therefore, existing design proposals limit the maximum stresses max of
load spectrums to 80 % of fy [11]. The application of HFHP treatment increases the fatigue strength
locally. During design of HFHP treated weld toes, adjacent notch details like cut edges or weld
roots can be relevant for design which limit the lifetime of cyclic loaded structures. If HFHP is used
during fabrication of new structures, the welded notch details should have high welding quality and
existing flaws have to be removed before treatment. Special attention has to be paid to quality
assurance of HFHP treatment. In [16] quality assurance guidelines are given for HFHP treatment
considering aspects of operator training, preparation of weld surface, safety, calibration, application,
quality control and documentation.
4

SUMMARY

Within this contribution the results of fatigue tests at as welded and HFHP treated specimens of
UHSS S960, S1100 and S1300 have been discussed. The slopes of the S-N-lines of HFHP treated
specimens increase to m ~ 5 when crack initiation starts from weld toes. The fatigue strength of
HFHP treated specimens was at least twice the fatigue strength in comparison to the as welded toe
condition. An influence of the yield strength on the fatigue strength improvement can be observerd
for the notch details longitudinal stiffener and butt weld. The S-N-lines of as welded and HFHP
treated toe conditions intersect theoretically at approximately 4,000 load cycles. In the LCF region
with load cycles Nf of 10,000 and 40,000 for the as welded toe condition, the fatigue life due to
HFHP treatment can be increased by the factors 2 and 10. Proposed FAT classes of existing design
recommendations for the consideration of HFHP treatment are conservative in comparison to the
test results of HFHP treated specimens. Due to the local fatigue life improvement adjacent notch
details like cut edges or weld roots can be relevant for design. Further investigations are necessary
to prove the influence of HFHP treatment at variable amplitude loading considering overloads and
preloads. Based on the presented experimental results, a design proposal for HFHP treatment of
welded UHSS will be developed.
ACKNOWLEDGEMENT
The research activities at the IML are funded by the German Research Association for Steel
Application (FOSTA research project No. P938). The authors want to thank FOSTA and the
following companies for their support during the project: ThyssenKrupp Steel Europe AG, SSAB
EMEA AB, DEPA Gesellschaft fr Kranauslegerteile mbH, PITec GmbH and SLV Duisburg.

REFERENCES
[1] Yildirim, H. C., Marquis, G. B., Barsoum, Z., 2013. A round robin study of high-frequency
mechanical impact (HFMI)-treated welded joints subjected to variable amplitude loading, Welding in
the World, Vol. 57, No. 3, pp. 437-447.
[2] Kuhlmann, U., Ummenhofer, T., Kudla, K., Weidner, P., 2013. Untersuchungen zur Anwendung
hherfrequenter Hmmerverfahren im Stahlwasserbau (in German), Final report, Bundesanstalt fr
Wasserbau (BAW).
[3] Ummenhofer, T. et al., 2011. REFRESH - Lebensdauerverlngerung bestehender und neuer
geschweiter Stahlkonstruktionen, Final report project P 702, Forschungsvereinigung Stahlanwendung
e. V. FOSTA, Dsseldorf.
[4] Kuhlmann, U., Gnther, H.-P., 2009. Experimentelle Untersuchungen zur ermdungssteigernden
Wirkung des PIT-Verfahrens (in German), Test report, University of Stuttgart.
[5] Puthli, R., et al., 2008. Detaillsungen bei Ermdungsfragen und dem Einsatz hochfester Sthle bei
Trmen von Offshore-Windenergieanlagen (in German), Final report project P 633, Forschungsvereinigung Stahlanwendung e. V. FOSTA, Dsseldorf.
[6] Kuhlmann, U., Drr, A., Bergmann, J., Thumser, R., 2006. Effizienter Stahlbau aus hherfesten
Sthlen unter Ermdungsbeanspruchung (in German), Final report project P 620,
Forschungsvereinigung Stahlanwendung e. V. FOSTA, Dsseldorf.
[7] Cheng, X., Fisher, J. W., Prask, H. J., Gnupel-Herold, T., Yen, B. T., Roy, S., 2003. Residual stress
modification by post-weld treatment and its beneficial effect on fatigue strength of welded structures,
International Journal of Fatigue, Vol. 25, No. 9-11, pp. 1259-1269.
[8] Janosch, J. J., Koneczny, H., Debiez, S., Statnikov, E. C., Troufiakov, V. J., Mikhee, P. P., 1996.
Improvement of fatigue strength in welded joints (in HSS and in aluminium alloys) by ultrasonic
hammer peening, Welding in the World, Vol. 37, No. 2, pp. 72-83.
[9] Yildirim, H. C., Marquis, G. B., 2012. Overview of fatigue data for high frequency mechanical impact
treated welded joints, Welding in the World, Vol. 56, No. 7-8, pp. 82-96.
[10] Yildirim, H. C., 2013. Design aspects of high strength steel welded structures improved by high
frequency mechanical impact (HFMI) treatment, PhD-Thesis, Aalto University, Finland.
[11] Haagensen, P. J., Maddox, S. J., 2010. IIW recommendations on post weld fatigue life improvement of
steel and aluminium structures, International Institute of Welding, Doc. XIII-2200r7-07.
[12] Drr, A., 2007. Zur Ermdungsfestigkeit von Schweikonstruktionen aus hherfesten Bausthlen bei
Anwendung von UIT-Nachbehandlung (in German), PhD-Thesis, University of Stuttgart.
[13] EN 1993-1-9:2010-12, Eurocode 3: Design and construction of steel components Part 1-9: Fatigue.
[14] Hobbacher, A., 2008. Recommendations for fatigue design of welded joints and components,
International Institute of Welding, Doc. IIW-1823-07 ex. XIII-2151r4-07/XV-1254r4-07.
[15] Gundel, W., Hamme, U., Herion, S., 2010. Ermdungsfestigkeit geschweiter Konstruktionen aus
hoch- und hchstfesten Feinkornbausthlen in der Praxis (in German), DVS Berichte, Vol. 267, pp.
220-224.
[16] Marquis, G., Barsoum, Z., 2014. Fatigue strength improvement of steel structures by high-frequency
mechanical impact: proposed procedures and quality assurance guidelines, Welding in the World, Vol.
58, No. 1, pp. 19-28.