Fatigue behaviour of high frequency hammer peened ultra high strength steels

Accepted Manuscript Fatigue behaviour of high frequency hammer peened ultra high strength steels Jörn Berg, Natalie Stranghöner PII: DOI: Reference:

S0142-1123(15)00264-9 http://dx.doi.org/10.1016/j.ijfatigue.2015.08.012 JIJF 3688

To appear in:

International Journal of Fatigue

Received Date: Revised Date: Accepted Date:

28 December 2014 3 July 2015 9 August 2015

Please cite this article as: Berg, J., Stranghöner, N., Fatigue behaviour of high frequency hammer peened ultra high strength steels, International Journal of Fatigue (2015), doi: http://dx.doi.org/10.1016/j.ijfatigue.2015.08.012

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

International Journal of Fatigue Fatigue behaviour of high frequency hammer peened ultra high strength steels Jörn Berg a,*, Natalie Stranghöner a a

University of Duisburg-Essen, Institute for Metal and Lightweight Structures, Universitätsstr. 15, D-45141 Essen, Germany

* Corresponding author. Tel.: +49 201 183 2765; fax: +49 201 183 2710. E-mail address: [email protected] (J. Berg).

Keywords: High frequency hammer peening (HFHP) Post weld treatment Fatigue life improvement Ultra high strength fine grained structural steel (UHSS) Low cycle fatigue (LCF)

Abstract Existing design recommendations for the consideration of high frequency hammer peening (HFHP) are limited to steel grades of S960 and plate thicknesses of 5 mm and higher. The influence of HFHP treatment on the fatigue behaviour of welded ultra high strength steels with yield strengths of 960 MPa and higher - loaded in the upper finite and low cycle fatigue life region - has not been investigated sufficiently so far. For this reason, fatigue tests have been performed on four typical welded notch details of mobile crane structures made of S960, S1100 and S1300 to determine the influence of HFHP on the fatigue strength. The fatigue strength of HFHP treated specimens was at least twice the fatigue strength of the as welded toe condition. A fatigue life improvement due to HFHP treatment can be observed at load cycles of 10,000 and higher. In accordance with existing investigations, the slope of the S-N-line increases to approximately m ~ 5 due to HFHP treatment if the fatigue cracks start from the treated weld toes. The classification of the test results for the HFHP treated toe condition shows, that fatigue classes (FAT) of existing design proposals are conservative. Further improvements of the proposed FAT classes are possible which shows the potential use of UHSS with steel grades higher than S960 in combination with HFHP treatment.

1. Introduction The fatigue behaviour of welded steel joints is predominated by crack propagation from crack like imperfections at the weld toe and weld root regions. Therefore, the fatigue design of welded steel joints is independent from the yield strength according to different design rules [1,2]. Existing FAT classes for welded joints base predominantly on fatigue tests covering the mid and lower part of the finite fati gue life region. Currently, modern ultra high strength fine grained structural steels (UHSS) with yield strengths up to 1300 MPa are offered by steel producers. The use of UHSS in welded, fatigue loaded structures is only reasonable for applications with high dead loads or high stress ranges like mobile crane structures, see Fig. 1 a. Due to applied fatigue loads during operation the lifetime of these welded structures, see Fig. 1 b, is limited and can be classified into the upper finite and low cycle fatigue (LCF) life region. One possibility to improve the fatigue behaviour is the application of post weld treatment methods like high frequency hammer peening (HFHP). HFHP mainly induces compressive residual stresses in the treated weld toe region resulting in a translation and a rotation of the S-N-curve. 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 on steels with yield strengths less than or equal to 960 MPa within research activities. Consequently, existing design recommendations for the influence of HFHP on the fatigue strength are limited to steel grades of S960 and plate thicknesses of 5 mm and higher. For this reason, further investigations on four mobile crane typical notch details have been performed to transfer the results and applicability of HFHP to welded UHSS with steel grades up to S1300 and plate thicknesses higher than or equal 4 mm in the LCF and upper finite fatigue life region. a)

b)

Fig. 1. Application of ultra high strength fine grained structural steels in mobile crane structures (a) and detail of a telescopic boom with welded stiffeners on a box-girder (b)

International Journal of Fatigue

2

Nomenclature   C m AW EC 3 fy FAT HFHP HiFIT IIW LCF m n Nf Nf,AW,exp. Nf,HFHP,exp. PIT R t UHSS

nominal stress nominal stress range reference value of fatigue strength at 2 x 106 load cycles corresponding to 95 % survival probability mean value of fatigue strength at 2 x 106 load cycles corresponding to 50 % survival probability as welded Eurocode 3 – EN 1993-1-9 yield strength fatigue class referred to S-N-line of fatigue design standards, which is the stress range in MPa at 2 x 106 load cycles corresponding to 95 % survival probability high frequency hammer peening High Frequency Impact Treatment International Institute of Welding low cycle fatigue slope of S-N-line in finite fatigue life region Number of fatigue tests load cycles until failure load cycles until failure of the test specimens with as welded toe condition load cycles until failure of the test specimens with high frequency hammer peened toe condition Pneumatic Impact Treatment stress ratio defined by min / max plate thickness ultra high strength steels

2. State of the Art 2.1. High frequency hammer peening High frequency hammer peening represents an advancement of common hammer peening methods [3,4] concerning its effectiveness, Fig. 2 (a & b). In comparison to established post weld treatment methods like grinding, TIG dressing or shot peening, HFHP yields in a combin ed effect of weld toe rounding, modification of the residual stress distribution and edge layer hardening. The influence of HFHP on the fatigue behaviour of welded joints has been increasingly investigated scientifically in the past years. By the application of HFHP the local, critical weld toe surfaces are plastically deformed, Fig. 3. The depth e of the treatment line is approximately 0.1 to 0.3 mm [5-8] and depends on the geometric shape of the indenter´s tip and predominantly on the strength of the treated material and its hardness, respectivel y. The plastic deformation of the weld toe region results in cold hardening of the near surface microstructure. Micro hardness measurements show that the hardened zone can reach depths up to 1.5 mm [7,9]. The application of HFHP yields a modified residual stress state by inducing compressive residual stresses influencing the fatigue behaviour of the treated weld toes significantly. Compressive residual stresses transverse to the welding direction are decisive for the effectiveness of the treatment method as these stresses superimpose with the stresses resulting from operational fatigue loads. The magnitude of compressive residual stresses depends on the yield strength f y of the treated material and increases with increasing yield strength. Residual stress measurements [5-8, 10-14] show that compressive residual stresses at the treated surfaces transverse to the welding direction can reach values of approximately 75 % of f y and values that are higher than f y because the local strength can be increased due to cold hardening. Induced compressive residual stresses can appear up to a depth of 1 to 2 mm from the treated surfaces. The plastic deformation of the weld toe geometry can also result in rounding of the weld toe decreasing the local stress concentration. Rounding of the weld toe grinding TIG dressing

Modification of the residual stress distribution stress relieving shot peening hammer / needle peening high frequency hammer peening

b)

Stress range [MPa]

a)

200

R=0 100 10.000

100.000 1.000.000 Load cycles Nf

Fig. 2. Differentiation of post weld treatment methods with regard to their effectiveness (a) and their influence on the fatigue behaviour of welded joints acc. to [4] (b)

The S-N-lines of HFHP treated notch details have a shallower slope with m ~ 5 in comparison to FAT classes of as welded notch details with m ~ 3, Fig. 4. This rotation of the S-N-line can be attributed particularly to the modified residual stress state resulting in an extension of the crack initiation period. In highly stressed structures, e. g. with high stress peaks or high mean stresses, residual stresses can relax resulting in a decrease of the beneficial effect of HFHP. In addition to the rotation of the S-N-line, the reference fatigue strength C translates to higher fatigue strengths, Fig. 4. The favourable effect on the fatigue strength due to HFHP treatment increases with increasing material´s yield strength [15] as this is related to the higher induced compressive residual stresses for higher steel grades.

International Journal of Fatigue direc t appli ion of catio n

a)

3

b)

angle of application  ~90°

d)

c)

of nter inde ol P to HFH



depth e

plastic deformation of weld toe

weld metal radius r

base material Fig. 3. Local treatment of the weld toe region due to high frequency hammer peening: a) angles of application, b) application at the weld toe region of a longitudinal stiffener, c) resulting weld toe geometry parameters and d) resulting plastic deformation of weld toe region for a butt weld

1000 Intersection

Stress range  [MPa]

mHFH > 3

100

LCF Low Cycle Fatigue

10 104

Fatigue life increase

Increase of fatigue strength

mAW = 3

untreated high frequency hammer peening 4104

106 2106

105

107

Load cycles Nf Fig. 4. Schematic comparison of S-N-lines of untreated and HFHP treated notch details

2.2. Fatigue design concepts considering HFHP treatment The positive effect of a HFHP treatment on the fatigue strength of a welded notch detail is actually not incorporated into the fatigue design rules, yet. However, within the literature different proposals exist to consider the influence of a HFHP treatment. Here, the FAT class (detail category) of the as welded notch detail is increased by an improvement factor k depending on the steel grade, the geometry of the notch detail, the stress ratio R, the applied stress level and, if applicable, the slope of the S-N-line is increased to m = 3.5 [16] or m = 5 [3,6,17,18]. In order to be in conjunction with the uni-spaced FAT classes of fatigue design guides [1,2] the fatigue strength improvement can be also expressed by an improvement of the reference FAT class of the as welded toe condition to the next higher FAT classes. The yield strength dependent, improved FAT classes according to different design proposals are visualized in Fig. 5 (a) exemplary for the notch detail of HFHP treat ed transversal stiffener. According to the approach of Yildirim [17] the fatigue life improvement can be expressed by a four fatigue class increase for low strength steels and a further one fatigue class increase for every increase of yield strength of 200 MPa in comparison to the reference yield strength of 355 MPa. The maximum possible FAT class for a steel grade of S960 is FAT 180 [17] corresponding to a fatigue strength increase at 6 N = 2 x 10 load cycles by a factor of 2.3 in comparison to FAT 80 for the as welded toe condition [1,2]. As the slopes of the S-N-lines for the HFHP treated (m = 5) and as welded (m = 3) toe condition are different, see Fig. 5 (b), this increase of fatigue strength corresponds to a fatigue life increase by a factor of 1.7 at a stress range leading to 10,000 cycles for the as welded toe condition and by a factor of 7.8 at a stress range leading to 100,000 load cycles for the as welded toe condition. These numbers are only valid for a stress ratio of R = 0,1. For higher mean stresses, a decrease of the beneficial effect of HFHP treatment has to be considered which is not discussed in detail within this contribution . Guidance on this topic can be found among others in [3,6,16-18]. The main part of available literature results of fatigue tests on HFHP treated notch details covers steel grades S355 and S690. Existing design proposals for the consideration of the positive effect due t o HFHP treatment are limited to maximum steel strengths of S690 or S960 and plate thicknesses of 5 mm and higher. These design proposals do not cover the LCF region and steel grades S960 and higher sufficiently. Further investigations are necessary to transfer these design proposals to UHSS with steel grades S960 and higher. The analysis of existing fatigue tests on as welded and HFHP treated notch details of UHSS shows the necessity for further research activities. Within this contribution the following questions will be clarified: - Can the FAT classes of existing fatigue design rules [1,2], which have been determined experimentally at specimens of steel grades with lower yield strengths, be applied within the upper finite fatigue life and LCF region, Fig. 4, with load cycles to failure of 10,000 to 40,000? - How much is the influence of HFHP on the fatigue strength improvement at UHSS with yield strengths of 960 MPa and higher? - Does the application of HFHP also result in a fatigue life increase in the upper finite fatigue life and LCF region with comparatively high maximum stresses? - At which stress ranges or number of load cycles, respectively, do the S-N-lines of untreated and HFHP treated notch details intersect?

International Journal of Fatigue

200

Yild irim [17] m = 5 (HFHP )

180 160

REFRE SH [6] m = 5 (HFHP )

140

Dürr [1 8] m = 5 (HFHP )

120

960

1300

100

IIW [3] m = 3 (HFHP )

EC 3/IIW [1,2 ] m = 3 (A W)

80

b)

Stress range  [MPa]

FAT class due to HFHP [MPa]

a)

4

400 m=5 1 200

100

LCF low cycle fatigue

160

1 m=3

400

600 800 1000 yield strength f y [MPa]

50 10.000

1200

R = 0.1

125 112 80

As Welded

60 200

180

1.000.000 40.000 100.000 Load cycles Nf

Fig. 5. Yield strength dependent FAT classes (a) and maximum possible S-N-lines (b) for a HFHP-treated weld toe of notch detail of welded transversal stiffener according to different design proposals

3. Experimental investigations 3.1. Test programme The Institute for Metal and Lightweight Structures of University of Duisburg-Essen has performed fatigue tests on UHSS S960, S1100 and S1300 to determine the influence of HFHP on the fatigue behaviour of the different welded notch details, namely, a) longitudinal stiffener, b) transversal stiffener, c) cover plates and d) butt weld with and without transition in thickness, Table 1 [19]. For the notch detail of the double-side welded transversal stiffener made of S1100, one test series with continuous fillet welds along the complete width of the base plate (width of the stiffener equals the width of the base plate, series 11-6-QSd) and one test series with a reduced width of the stiffener in comparison to the width of the base plate have been performed to investigate the effect of start and stop positions along the weld line on each specimen.. The weld toe condition varied in as welded and HFHP treated specimens. In total 119 fatigue tests have been performed. Table 1 Test programme Notch detail

5/6 5/5

Steel grade S1100 S1300

t max / fy [mm] 6 25-55 % 4 30-50%

11-6-QSd2) 11-6-QS3) S13-4-QS3)

5/6 5/5 5/5

S1100 S1100 S1300

6 6 4

45-70 % 35-75 % 30-50 %

Cover plate

96-7-La S11-6-La S13-4-La

4/5 6/6 5/5

S960 S1100 S1300

7.5 6 4

40-65 % 30-55 % 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 %

Test series

n1)

Longitudinal stiffener

S11-6-LS S13-4-LS

Transversal stiffener

1)

2)

3)

Number of tests in as welded and high frequency hammer peened toe condition, respectively. The stiffener has been welded continuously along the complete width of the base plate (width of the stiffener equals the width of the base plate). A reduced width of the stiffener in comparison to the width of the base plate has been used to include start and stop positions along the weld line on each specimen.

3.2. Test specimens The test specimens have been produced from ultra high strength, water-quenched and tempered fine grained heavy plates of steel grades S960, S1100 and S1300 and with plate thicknesses of 4 mm to 8 mm. The plates have been cut by water jet cutting to avoid increased hardness values and grooves due to thermal cutting. All test specimens have been welded manually by MAG process (MAGM, 135) with filler material Union X90. Herewith, the nominal value of the yield strength of the filler material is below the yield strength of the used base materials. However, it can be assumed that the weld material is alloyed by the base material during the welding process resulting in higher strengths in the weld toe regions in comparison to the yield strength of the weld material. Actually no adequate filler materials with yield strengths higher than 960 MPa are available for welding steel grades S1100 and above so that the use of weld material with lower strengths is common during the fabrication of mobile crane structures (undermatching) [20]. Tensile tests according to EN ISO 6892-1 [21] and EN ISO 4136 [22] have been performed to determine the mechanical properties of the used materials, Table 2. The resulting mechanical properties of the tensile tests fulfill the minimum requirements concerning to yield strength, tensile strength and elongation after fracture.

International Journal of Fatigue

5

Table 2 Mechanical technological properties resulting from the performed tensile tests according to EN ISO 6892-1 [21] (base material) and EN ISO 4136 [22] (weld material) Steel grade Product name Plate thickness Yield strength Tensile strength Ductility Elongation after fracture t Rp0,2 Rm Rm / Rp0,2 A (mm) (MPa) (MPa) (-) (%) S960 XABO 7.5 968 1031 1.07 15 S1100 XABO 6 1185 1386 1.17 11 S1100 XABO 8 1189 1400 1.18 10 S1100 WELDOX 6 1300 1474 1.13 12 S1100 WELDOX 8 1263 1438 1.14 10 S1300 WELDOX 4 1424 1567 1.10 10 1) G 89 6 M21 Mn4Ni2CrMo Union X90 4 899 982 1.09 Not determined 1) Tensile specimens transverse to the welding direction have been taken from a butt welded plate.

All attachments have been welded single layered. 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 and remaining root weld. The specimens of test series S11-6-S have been welded both sided with transition in thickness (alignment on the root side, removal of the weld root and welded sealing run), Fig. 6. Due to the deliberate misalignment of the neutral axis of 1 mm, an additional bending moment is induced when the specimens are axially loaded resulting in an increased stress state at the fatigue critical weld toe (top layer, 6 mm). This stress increase can be considered by multiplying the part of membrane stresses resulting from the axial load by a stress magnification factor km, Eq. (1) [1,2]. In conjunction with the detail categories of EC 3 [1], this stress magnification can be also considered on the resistance side by a decreased FAT class of the S-N-line, Eq. (2).

km 

t 1.5 1 6e  1  1.5 1 1.5  1,394 ks t1 t1  t2

(1)

with: e = 1 mm, t1 = 6 mm and t2 = 8 mm.

FAT  71 ks  a)

[mm]

8

25°

71  51 MPa 1.394

(2)

b)

25°

6

Fig. 6. Execution of the weld (a) and macrograph (b) for the notch detail of butt weld with transition in thickness

In addition to visual inspection of the welds by the fabricator who has done all the welding operations, all specimens have b een classified into quality levels concerning selected geometrical surface imperfections and imperfections of the weld geometry acc ording to EN ISO 5817 [23], Tables 3 and 4. With exception of the angular misalignment and the test series of the one side welded butt weld (96-7-S) all specimens can be classified into the quality level B according to EN ISO 5817. This classification correlates to the common and increased requirements onto the weld fabrication of mobile crane structures. The linear misalignment of all butt welds is considerably lower than the maximum allowed linear misalignment of 10 % of plate thickness for the quality level B. However, the results of strain gauge measurements as well as numerical investigations [24,25] confirm that the angular misalignment is the main influencing imperfection on the performance and results of the fatigue tests - in particular due to the rather low plate thicknesses and small specimen dimensions. Table 3 Classification of the fillet welds for selected imperfections (weld geometry) in quality levels (B to D) according to EN ISO 5817 [23] Notch detail Test series 1.10 1.12 3.3 Excessive Incorrect Angular convexity weld toe misalignment Longitudinal stiffener S11-6-LS B B B S13-4-LS B B C Transversal stiffener 11-6-QSd B B B 11-6-QS B B C S13-4-QS B B C Cover plate 96-7-La B B B S11-6-La B B B S13-4-La B B B Table 4 Classification of the butt welds for selected imperfections (weld geometry) in quality levels (B to D) according to EN ISO 5817 [23] Notch detail Test series 1.9 1.11 1.12 Excessive weld overfill Excessive root overfill Incorrect weld toe Butt weld 96-7-S C B C Butt weld with S11-6-S B B B transition in thickness

3.1 Linear misalignment B B

3.3 Angular misalignment B B

3.3. Post weld treatment by high frequency hammer peening After welding and cutting, approximately half of all test specimens have been treated manually by HFHP. The HFHP treatment has been mainly performed by Pneumatic Impact Treatment (PIT) with a frequency of 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 performed by personal of the manufacturer of PIT and HiFIT within the laboratory of University of Duisburg-Essen. The local treatment has been applied at the weld toes to the base material, Fig. 3. At the notch details of welded cover plate and longitudinal stiffener the HFHP treatment has been focused near the attachment ends only, Fig. 7, as these weld toes are the fatigue critical positions and HFHP treatment along the complete weld lines is not necessary for these notch details. For the notch details of butt weld and transversal stiffener, the HFHP treatment has been applied continuously along all weld toes to the base material as fatigue cracks can initiate from all positions of the weld toes along the complete length of the welds. The cross section of one specimen of the

International Journal of Fatigue

6

notch detail of butt weld with transition in thickness before and after HFHP treatment is exemplary visualized in Fig. 8. It can be clearly seen that the HFHP treated weld toe regions show plastic deformations resulting in a rounding of the weld toe surfaces. During post weld treatment the resulting treatment lines have been visually checked concerning the treatment quality, possible micro cracks and the treatment depth e. In exceptional cases the weld metal slightly overlapped at the treated weld toe so that the overlapped weld toe regions have been treated extensively to compress these defects, Fig. 9.

International Journal of Fatigue a)

7

b)

~40-50 mm

HFHP treated weld toe

HFHP-treatment only at attachment ends

HFHP-treatment along complete weld toe HFHP treated weld toe Fig. 7. Regions of HFHP treatment for the different notch details (a) and HFHP treated weld toes with resulting plastic deformations for the notch detail of longitudinal stiffener (b)

a) untreated

b) HFHP treated 3

1 2

4

Fig. 8. Comparison of weld toes of untreated (a) and HFHP-treated (b) condition

a)

b)

Overlapped weld metal

Extensive HFHP treatment

Fig. 9. Overlapped weld metal (a) and extensive HFHP treatment (b)

Residual stresses have been analyzed at one as welded and at one HFHP treated specimen of the notch detail butt weld with transition in thickness made of S1100 (S11-6-S) by X-ray diffraction method with a mobile X-ray equipment at BAM, Berlin, Table 5. The near surface residual stresses have been measured transversal and longitudinal to the welding direction at all four weld toes, Fig. 8. Due to the HFHP treatment the residual stresses have been modified significantly. The results of the measured residual stresses transversal to the welding direction, which superimpose with membrane stresses from external fatigue loads, show that tensile residual stresses exist for the as welded toe condition at 0.15 mm depth and shift to compressive residual stresses for the HFHP treated toe condition with average values of approxim ately 75 % of the yield strength confirming existing results of residual stress measurements at lower steel grades from the literature. Table 5 Results of the residual stress measurements by X-ray diffraction method Measuring position1) AW surface AW (0.15 mm depth) HFHPsurface res,long.2) res,trans.3) res,long.2) res,trans.3) res,long.2) res,trans.3) 140 -768 1: top layer (6 mm) 149 42 -7 -760 136 -774 2: bottom layer (6 mm) -205 -14 -231 -600 150 -687 3: top layer (8 mm) -47 -4 81 -505 58 -1008 4: bottom layer (8 mm) -225 129 -214 -630 Standard deviation of results:  10-42 MPa. 1) Measuring positions according to Figure 8. 2) Residual stresses longitudinal to welding direction in MPa. 3) Residual stresses transversal to welding direction in MPa.

3.4. Experimental testing and analysis 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 reach load cycles until failure of 10,000 to 40,000 for the as welded specimens to cover the upper finite fatigue life and LCF region as this fatigue life region is an important operational region of the investigated steel grades, especially regarding fatigue stressed components of mobile crane structures. These fatigue loads resulted in maximum nominal stresses of 45 % to 75 % of the nominal value of the yield strength depending on the notch detail, Table 1. The HFHP treated specimens have been tested at the same fatigue loads to compare the load cycles until failure Nf to that ones of the as welded toe condition. For the evaluation of S-N-lines the results of all fatigue tests have been analyzed by linear regression. The reference fatigue strength C at 6 NC = 2 x 10 load cycles has been calculated for a 75 % confidence level of 95 % probability of survival for log N according to EC 3 [1] and the corresponding background document [26]. The mean value and the reference value of the fatigue strength  m,50% and C,95% have been

International Journal of Fatigue

8

evaluated with a fixed slope of m = 3 for the as welded and m = 5 and additionally with variable slope for the HFHP treated t oe condition for each test series. For the determination of C the S-N-lines have been extrapolated to the lower finite fatigue life region as the fatigue tests have been focused on the LCF and upper finite fatigue life region. 4. Experimental results and discussion 4.1. 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 (transvers al stiffener) or the weld root region (longitudinal stiffener and cover plate), Table 6 and Fig. 10. This crack initiation change mainly occurs at relatively low stress ranges as the induced compressive residual stresses do not relax the same way as at higher loads. For the notch detail of welded transversal stiffener – with the lowest local stress concentration of all investigated notch details - made of S1100 only 2 of 11 HFHP treated specimens with the highest stress ranges - 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. During post weld treatment a low weld toe angle made it difficult to achieve a straight treatment line along the treated weld toe. Consequently, the decreased stability of the HFHP tool caused a few plastic deformations (“misguided treatments”) on the surf aces of adjacent base material in some exceptional cases, Fig. 11. These “misguided treatments” are often surveyed skeptically by experts concerning its notch effect and are part of many discussions although no literature test results with crack initiation from these notches exist. During the performed fatigue tests no crack initiation starting from these “misguides treatments” could be observed so that existent skeptic could not be confirmed by experimental testing. It can be assumed that compressive residual stresses result at the material surface due to the plastic deformation affecting the fatigue strength in a positive way. Therefore, the fatigue behaviour of HFHP treated notch details is not influenced by these “misguided treatments” negatively. Table 6 Failure modes of HFHP treated specimens during fatigue tests Notch detail Test series Steel grade Longitudinal stiffener Transversal stiffener

Cover plate

S11-6-LS S13-4-LS 11-6-QSd 11-6-QS S13-4-QS 96-7-La S11-6-La S13-4-La 96-7-S S11-6-S

Butt weld Butt weld with transition in thickness  1) Edge or surface notches in the base material.

S1100 S1300 S1100 S1100 S1300 S960 S1100 S1300 S960 S1100

Crack initiation from Weld toe 5 2 1 1 4 2 4 5 4 12

Weld root 1 1 2 1 -

Base material1) 2 2 2 1 1 1 -

Clamping 3 2 -

40

5

9

5

HFHP treatment line

Fig. 10. Failure modes of HFHP treated specimens during fatigue tests

misguided treatments

misguided treatments

Fig. 11. Plastic deformations in adjacent base material (“misguided treatments”)

4.2. Fatigue life and fatigue strength The load cycles until failure (Nf) for the as welded specimens range from 5,000 to 110,000 and cover the LCF and upper finite fatigue life region, Fig. 12. As the yield strength does not influence the fatigue life of the as welded specimens, the different steel grades are not explicitly visualized in Fig. 12. Furthermore, the results of the as welded specimens justify to base the fatigue design of as welded notch details by one

International Journal of Fatigue

9

single FAT class for each notch detail according to EC 3 [1] and IIW [2]. The test results of all as welded specimens are above the S-N-lines of relative FAT classes according to EC 3 [1] and IIW [2]. Furthermore, the reference fatigue strength C of test results of all as welded specimens are above the relative FAT classes according to EC 3 [1] and IIW [2], Fig. 13. This conservative classification of the test results into existing FAT classes of fatigue design guides confirms that the results of the as welded specimens are representative and can be used as a reference in order to investigate the influence of a post weld treatment by HFHP. For the notch detail of transversal stiffener of S1100, weld start and stop positions do not influence the fatigue behaviour significantly. For this reason, the test results for both test seri es with (11-6-QS) and without (11-6-QSd) weld start and stop positions were evaluated as results from the same sample. The results for the notch detail of butt weld with transition in thickness treated by two different HFHP devices PIT and HiFIT do not show any significant variation which would require a separate evaluation of the results for each HFHP device. Consequently, the results of PIT and HiFIT treated specimens were evaluated as results from the same sample. Existing investigations confirm that different available HFHP devices lead to comparable results and that it is not required to distinguish between the results of different available HFHP devices [10]. By the application of HFHP, the number of load cycles N f increases rapidly in comparison to the as welded toe condition, Fig. 12. The slopes of the S-N-lines are much shallower than for the as welded toe condition due to induced compressive residual stresses, Table 5, extending the crack initiation period and due to residual stress relaxation at higher stress ranges. Considering the results with crack initiation starting from the HFHP treated weld toes, the slopes of the S-N-lines increase to m ~ 5 which is consistent to existing investigations and design proposals for steel strengths S960 and lower from literature [6,17,18]. Due to the relatively low number of fatigue tests per test series all results of each test series have been used for the statistical evaluation of the fatigue strength  which include results with crack initiation from different notches, Table 6. However, this procedure of evaluating the test results shows satisfactory results as the results of HFHP treated specimens with crack initiation from the treated weld toes (full symbols) and from differing positions (empty symbols) do not scatter significantly, Fig. 14. 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. Overall the scattering of test results for the HFHP treated toe condition is low, although it is higher than for the as welded toe condition due to variable crack initiation positions at lower stress ranges. For this reason, the results reflect a high post weld treatmen t quality which can be achieved by the application of HFHP if provisions for quality assurance are met [6,27,28]. 6 The reference value of the fatigue strength C at N = 2 x 10 cycles increases due to HFHP treatment by factors of 2.0 to 2.8 in comparison to the as welded toe condition, when a fixed slope of m = 5 is used, Table 7. These factors are only valid for the special case of constant amplitude fatigue loading with a constant stress ratio of R= 0,1 without the consideration of acciden tal overloads. In order to check a potential influence of the base material´s yield strength on the fatigue strength of HFHP treated specimens, the fatigue strength  m,HFHP of the HFHP treated specimens has been normalized to the mean value  m,AW of the as welded toe condition for each test series. An influence of the yield strength on the fatigue strength improvement can be observed for the notch details longitudinal stiffener (LS) and butt weld (S) where the fatigue strength improvement increases by approximately 15 % and 10 % with an increase of the yield strength from 1100 to 1300 MPa (LS) and 960 to 1100 MPa (S), respectively. 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 N f of as welded and HFHP treated specimens intersect theoretically in the LCF region with load cycles ranging from 3,300 to 4,700 depending on the notch detail, as indicated by the shaded regions in Fig. 14, if the relationship between log Nf,AW,exp. and log Nf,HFHP,exp. is extrapolated based on a linear regression. If the maximum stresses of the applied stress ranges reach values of yield strength magnitude, the most beneficial effect of HFHP treatment - the induced compressive residual stresses - is gone. For these high stresses a fatigue life improvement can only be referred to the effect of weld toe rounding. For this reason, the maximum applied stresses and stress ranges within spectrum loading have to be limited if the beneficial effect of compressive residual stresses has to be taken into account during fatigue design, see chapter 4.3. However, within the upper finite fatigue life region, which is important for the application of these types of steel grade in fatigue loaded structures, with load cycles Nf of 10,000 to 40,000 for the as welded toe condition, the fatigue life can be improved by factors of 2 to 10 due to HFHP treatment depending on the notch detail and yield strength. With increasing stress range  the fatigue life improvement effect of the HFHP treatment decreases which can be related to a faster relaxation of the induced compressive res idual stresses. The influence factor of the yield strength cannot be analyzed explicitly due to different plate thicknesses, partially different local weld geometries and angular distortion. 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 [24,25] are not covered on the action part within the nominal stress based design concept which influences the comparability of the test results. Therefore, the evaluation of test results applying structural hot spot stress and notch stress design concepts is content of further publications [29,30]. Furthermore, an evaluation of the yield strength effect cannot be reduced to the base material´s yield strength only, as all specimens have been welded with the same filler material X90. For this reason, a mismatch connection exists within the heat affected zone so that the strength at the local weld toe which influences the compressive residual stresses due to HFHP treatment significantly is limited by the material properties of the filler material.

Normalised stress range /m,

AW

10

cover plate long. stiffener transv. stiffener butt weld

HFHP

as welded

1

1

3 m

 C /  m,AW

 m /  m,AW

T

n

3.0

0.90

1.00

1:1.21

50

HFHP 5.0

2.07

2.47

1:1.39

Symbols Series AW

1 1.000

5

10.000

100.000 Load cycles Nf

59

1.000.000

Fig. 12. Results of the fatigue tests of all specimens within a normalised S-N diagram

International Journal of Fatigue

350

m [MPa]

█ HFHP - m = var.

█ AW - m = 3

300

FAT class acc. to EC 3 (AW) [1]

250

10

█ HFHP - m = 5

FAT class acc. to IIW (AW) [2]

200 150

m=3 m=3

100 50 0

1)

96-7-La

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

S960 1)

S1100

S1300

S1100

S1300

S1100

S1300

96-7-S

S11-6-S

S960

S1100

The results f or both test series with (11-6-QS) and without (11-6-QSd) weld start and stop positions were ev aluated as results f rom the same sample as the specimens incl. weld start and stop positions do not inf luence the f atigue behav iour signif icantly .

Fig. 13. Comparison of the mean value of the fatigue strengths m with the FAT classes according to EN 1993-1-9 [1] and IIW [2]

a) longitudinal stiffener

b) transversal stiffener 1.000.000

1.000.000

400,200 100.000 Nf,HFHP,exp.

Nf,HFHP,exp.

172,600 100.000

19,200 10.000 S1100 (cracks weld toe) S1100 (diff. crack init.) S1300 (cracks weld toe) S1300 (diff. crack init.)

3,300 1.000 1.000

10.000 100.000 Nf, AW, exp.

1.000 1.000

1.000.000

100.000

Nf,HFHP,exp.

100.000

Nf,HFHP,exp.

10.000 100.000 Nf, AW, exp.

1.000.000 560,000

422,400

26,400

10.000

1.000 1.000

S1100 (cracks weld toe) S1100 (diff. crack init.) S1300 (cracks weld toe) S1300 (diff. crack init.)

d) butt weld with and without transition in thickness

1.000.000

3,800

10.000 3,800

1.000.000

c) cover plate

25,700

S960 (cracks weld toe) S960 (diff. crack init.) S1100 (cracks weld toe) S1100 (diff. crack init.) S1300 (cracks weld toe)

10.000 100.000 Nf, AW, exp.

1.000.000

25,200 10.000 4,700 1.000 1.000

960 (cracks root layer) S1100 (cracks top layer) S1100 (cracks bottom layer)

10.000

100.000 Nf, AW, exp.

1.000.000

Fig. 14. Comparison of load cycles until failure Nf for the as welded and HFHP treated toe condition for the notch details of a) longitudinal stiffener, b) transversal stiffener, c) cover plate and d) butt weld with and without transition in thickness. Each point refers to the results of two specimens tested at the same load level with untreated and HFHP-treated weld toe condition Table 7 Characteristic fatigue strengths C of all test series Notch detail Steel grade Longitudinal stiffener Transversal stiffener

Cover plate

n

S1100 S1300 S1100

C1) [MPa] As Welded 82 82 92

S1300 S960 S1100 S1300 S960 S1100

88 85 80 66 71 75

5 4 6 5 4 6

Butt weld Butt weld with transition in thickness 1) Evaluation of the fatigue strength with m = 2) Evaluation of the fatigue strength with m = 3) Evaluation of the fatigue strength with m = 4) Evaluation of the fatigue strength with m =

4.3. Design proposals

5 5 10

C2) [MPa] HFHP 167 204 2253) 1564) 178 205 171 188 193 195

3 = constant. 5 = constant. 5 considering all test results with different crack initiations. 3 considering only test results with crack initiation differing from the weld toe.

n

C,HFHP /C,AW

6 5 11 9 5 5 6 5 4 12

2.0 2.5 2.5 1.7 2.0 2.4 2.2 2.8 2.7 2.6

International Journal of Fatigue

11

For the consideration of the influence of HFHP on the fatigue strength, different design proposals exist which are only valid for steel grades up to maximum S960. At first, the results of the in-house fatigue tests for the HFHP treated toe condition have been classified into proposed FAT classes by computing each FAT class for the maximum valid yield strength of each design proposal, Table 8. The reference values C of the test results for the HFHP treated toe condition are all above the FAT classes of different design proposals with exception of the notch detail transversal stiffener made of S1100 (11-6-QS) results of which have been evaluated with a differing slope of m = 3 due to predominant crack initiation from edge or surface notches in the base material or the clamping area, Fig. 15. Furthermore, the results of all specimens with HFHP treated toe condition are visualized within S-N-diagrams for each notch detail in comparison to the S-N-lines of the relative FAT classes according to the different design proposals, Fig. 16. Results with both fatigue crack initiation from the treated weld toes (full symbols) and fatigue crack initiation from adjacent notches (empty symbols) are displayed within the diagrams. With exception of the notch detail of transversal stiffener made of S1100 with predominant crack initiation from notches in the base material or the clamping area, the results do not scatter significantly. The results, especially for the notch details of longitudinal stiffener and transversal stiffener, extend the results of existing fatigue data bases on HFHP treated toe conditions [15] into the upper finite fatigue life region. The classification of results for the notch details longitudinal stiffener, transversal stiffener and butt weld shows that proposed FAT classes of existing design recommendations are conservative. The design proposal of Yildirim [17] shows the best agreement with the results because this approach is based on fatigue tests covering yield strengths up to 960 MPa. In a second step, the maximum possible FAT classes satisfying the results of the fatigue tests of all steel grades with a conservative approach have been evaluated (last column of Table 8) [29]. These values are only based on the fatigue tests and have not been analyzed by further statistical evaluation, yet. However, the numbers illustrate the potential use of UHSS with steel grades higher than S960 in combination with HFHP treatment. In particular, for the notch detail of butt weld with transition in thickness a further improvement to FAT class 180 seems possible. For the notch detail of welded cover plate, no specific design proposal exists until now. However, the general IIW design recommendations according to Haagensen and Maddox [3] can be applied as long as crack initiation starting from the weld root can be excluded. Especially within the finite fatigue life region the IIW approach is too conservative making a furth er improvement of the FAT class due to HFHP treatment as well as an adaption of the slope of the S-N-line to m = 5 possible, Fig. 16. The results of the HFHP treated specimens with all crack initiation positions are all above the plotted FAT class 180. Concerning the evaluation of the test results, it has to considered that the fatigue tests for the HFHP treated toe condition have been performed predominantly within the upper finite fatigue life region. For this reason, the S-N-lines have been extrapolated to the lower finite fatigue life region for the determination of the 6 reference value  C related to 2  10 load cycles. Within the lower finite fatigue life region adjacent notch details like cut edges or weld roots can be relevant for fatigue design which limit the lifetime of cyclic loaded structures treated by HFHP. In the context of th e present contribution no extension of existing design proposals to steel grades up to S1300 is given. The evaluation of a fatigue design concept for the consideration of HFHP at welded ultra high strength steels up to S1300 will be content of Berg [29]. Table 8 Maximum FAT classes of different design proposals for the consideration of HFHP Design proposal EC 3 (AW) [1] IIW [2,3] Adapted design standard EC 3 IIW Slope m 3 3 Max. steel grade S690 S900 FAT class C Longitudinal stiffener 56 901)

Dürr [18] EC 3 5 S690

REFRESH [6] EC 3 5 S690

Yildirim [17] IIW 5 S960

Berg [29]4) EC 3 5 S1300

-

1123)

160

160

Transversal stiffener

80

1121)

1252)

1603)

180

180

Cover plate

56

71

-

-

-

180

Butt weld

71

1001)

-

1403)

180

200

Butt weld with transition in thickness

515)

80

-

1003)

140

180

1) 2) 3) 4) 5)

The intersection with FAT 160 with m = 5 has to be considered. The intersection with FAT 80 with m = 3 has to be considered. The S-N-lines are only valid for load cycles N  105. Maximum possible FAT classes satisfying the results of the fatigue tests of all steel grades with a conservative approach. FAT 51 = FAT 71 x 0.72, see Eq. (2).

International Journal of Fatigue

12

350 Yildirim[17]

m [MPa]

300

REFRESH [6]

█ HFHP - m = var.

Dürr [18]

IIW [2-3]

█ HFHP - m = 5

250 200 150

m=3

100

50 0 96-7-La S960

1)

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

1)

The results f or both test series with (11-6-QS) and without (11-6-QSd) weld start and stop positions were ev aluated as results f rom the same sample with a constant slope of m = 3 as most of the specimens did not f ailed f rom the HFHP treated weld toes. Fig. 15. Comparison of the mean value of the fatigue strengths m,50% with FAT classes of different design proposals for the HFHP treated toe condition

International Journal of Fatigue

13

a) longitudinal stiffener

b) transversal stiffener

800

400 200

[17] - m = 5 [6] - m = 5 [3] - m = 3 [1] - m = 3 (AW) S11 00 S13 00

100

160 112

1

m

10.000

[17] - m = 5 [6] - m = 5 [18] - m = 5 [3] - m = 3 [1] - m = 3 (AW) S11 00 S13 00

200 100 50 1.000

1.000.000

c) butt weld

200

200

180

[29] - m = 5 [17] - m = 5 [6] - m = 5 [3] - m = 3 [1] - m = 3 (AW) S96 0

10.000

1

140

m

Stress range  [MPa]

Stress range  [MPa]

125

m

112

10.000

80

100.000 Load cycles Nf

1.000.000

800

400

50 1.000

160

1

d) butt weld with transition in thickness

800

100

180

Open symbols indicate cracks with initi. diff. from weld toes.

56

100.000 Load cycles Nf

400

90

Open symbols indicate cracks with initi. diff. from weld toes.

50 1.000

Stress range  [MPa]

Stress range  [MPa]

800

400 200

[29] - m = 5 [17] - m = 5 [6] - m = 5 [3] - m = 3 [1] - m = 3 (AW) S11 00

100

100

Open symbols indicate cracks with initi. from the bottom layer.

71

50 1.000

100.000 1.000.000 Load cycles Nf [-]

10.000

180 140

100

1 m

80 51

100.000 1.000.000 Load cycles Nf

e) cover plate

Stress range  [MPa]

800 400

200

[29] - m = 5 [3] - m = 3 [1] - m = 3 (AW) S96 0 S11 00 S13 00

100

180

1 71

m

Open symbols indicate cracks with initi. diff. from weld toes.

50 1.000

10.000

100.000 Load cycles Nf

56 1.000.000

Fig. 16. Classification of test results in existing design proposals for the HFHP treated toe condition for the notch details a) longitudinal stiffener, b) transversal stiffener, c) butt weld, d) butt weld with transition in thickness and e) cover plate

4.4. 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 material´s yield strength, Table 1. Accidental overloads which can appear during operation of mobile cranes decrease the improvement effect of post weld treatment methods like HFHP due to residual stress relaxation [31] if the local von Mises stress at the weld toe exceeds the local material´s yield strength [32]. Therefore, existing design proposals for the consideration of HFHP limit the maximum stresses  max and stress ranges  of load spectrums to 80 % and 90 % of f y ,respectively [3,16,17]. However, investigations in [6] show that in particular at high strength steels a stable behaviour of induced compressive residual stresses under cyclic loading can be observed and no significant influence on the fatigue behaviour of HFHP treated weld toes by separate overloads or quasi static preloads occur. This phenomenon has to be related especially to edge layer hardening due to HFHP treatment [33,34]. Furthermore, the benefit of HFHP treatment decreases with an increase of mean stresses as this is related to residual stress relaxation. Consequently, for increased stress ratios a drop of the beneficial effect of HFHP treatment has to be considered within fatigue design. The application of HFHP treatment increases the fatigue strength locally. During design of HFHP treated weld toes, adjacent notch details like surface defects, cut edges or weld roots can be relevant for design which limit the lifetime of cyclic ally 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 post weld treatment. Furthermore, full penetration welds or fillet welds with larger throats should be used for welds at the ends of cover plates and longitudinal stiffeners to decrease the risk of weld root cracking. Special attention has to be paid to quality assurance of HFHP treatment. In [6,27,28] quality assurance guidelines are given for HFHP treatment considering aspects of operator training, preparation of weld surface, safety, calibration, application, quality control and documentation. 5. Conclusions Within fatigue tests the influence of HFHP treatment on the fatigue behaviour of welded, ultra high strength fine grained str utctural steels S960, S1100 and S1300 has been investigated. Due to HFHP treatment significant fatigue life improvements by approximately 10,000 cycles could be achieved in comparison to the as welded toe condition. Considering the results with crack initiation from the HFHP treated weld toes

International Journal of Fatigue

14

only, the slopes of the S-N-lines of HFHP treated specimens are much shallower due to induced compressive residual stresses within the weld toe regions and fit quite well with a proposed value of m = 5 according to different design recommendations. The fatigue strength of HFHP treated specimens was at least twice the fatigue strength in comparison to the as welded toe condition. For the notch details longitudinal stiffener (LS) and butt weld (S) the fatigue strength improvement increases by approximately 15 % and 10 % with an increase of yield strength from 1100 to 1300 MPa (LS) and 960 to 1100 MPa (S), respectively. Due to different slopes of the S-N-lines of as welded and HFHP treated toe conditions, the number of load cycles until failure N f for untreated and HFHP treated toe condition intersect theoretically at approximately 4,000 load cycles, if the relationship between logarithm of the load cycles for the untreated and HFHP treated toe condition are extrapolated based on a linear regression. Within the LCF region with load cycles Nf from 10,000 to 40,000 for the as welded toe condition, the fatigue life due to HFHP treatment can be increased by the factors 2 and 10 depending on the notch detail and yield strength. 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. Further improvements of the proposed FAT classes are possible which shows the potential use of UHSS with steel grades higher than S960 in combination with HFHP treatment. Due to the local fatigue life improvement adjacent notch details like cut edges or weld roots can be re levant for design. The achieved results on the beneficial effect of HFHP treatment are only valid for the special case of constant amplitude fatigue loading with a constant stress ratio of R= 0,1 without the consideration of accidental overloads. 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 concept for the consideration of HFHP treatment at welded UHSS up to S1300 will be developed [29]. Acknowledgements The presented investigations have been performed at the Institute for Metal and Lightweight Structures of the University of Duisb urg-Essen, Germany and are part of the research project “Influence of high frequency hammer peening on the fatigue strength of welded notch details of ultra high strength steels with steel grades S960, S1100 and S1300” [19] 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, Duisburg, SSAB EMEA AB, Oxelösund, DEPA Gesellschaft für Kranauslegerteile mbH, Leverkusen, PITec GmbH, Heudorf, DYNATEC Gesellschaft für CAE und Dynamik mbH, Braunschweig and Federal Institute for Materials Research and Testing (BAM), Division Weld Mechanics, Berlin. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10]

[11] [12]

[13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27]

[28] [29] [30] [31] [32] [33] [34]

EN 1993-1-9:2010-12, Eurocode 3: Design and construction of steel components – Part 1-9: Fatigue; 2010. Hobbacher, A. Recommendations for fatigue design of welded joints and components. International Institute of Welding, Doc. IIW-1823-07 ex XIII-2151r4-07/XV-1254r4-07; 2008. Haagensen, P. J., Maddox, S. J. IIW recommendations on methods for improving the fatigue strength of welded joints. International Institute of Welding, Woodhead Publishing, 1st Edition, Cambridge; 2013. Maddox, S. J. Fatigue strength of welded structures. Abington Publishing, 2nd Edition, Cambridge, 1991. Kuhlmann, U., Ummenhofer, T., Kudla, K., Weidner, P. Untersuchungen zur Anwendung höherfrequenter Hämmerverfahren im Stahlwasserbau (In German). Final report, Bundesanstalt für Wasserbau (BAW); 2013. Ummenhofer, T., Herion, S., Hrabowski, J., Rack, S., Weich, I., Telljohann, G. et al. REFRESH – Extension of the fatigue life of existing and new welded steel structures (In German). Final report project P 702, Forschungsvereinigung Stahlanwendung e. V. FOSTA, Düsseldorf; 2011. Kuhlmann, U., Günther, H.-P. Experimentelle Untersuchungen zur ermüdungssteigernden Wirkung des PIT-Verfahrens (In German). Test report, University of Stuttgart; 2009. Kuhlmann, U., Dürr, A., Bergmann, J., Thumser, R. Fatigue strength improvement for welded high strength steel connections due to the application of post-weld treatment methods (In German). Final report project P 620, Forschungsvereinigung Stahlanwendung e. V. FOSTA, Düsseldorf; 2006. Statnikov, E. S., Korostel, V. Y., Manelik, A. D. On identify in UIT preparation for comparative testing and field application. International Institute of Welding, Doc. XIII-218007; 2007. Yildirim, H. C., Marquis, G. B., Barsoum, Z. A round robin study of high-frequency mechanical impact (HFMI)-treated welded joints subjected to variable amplitude loading. Weld World 2013;57:437-447. doi:10.1007/s40194-013-0045-3. Puthli, R., Herion, S., Hrabowski, J., Ummenhofer, T., Weich, I., Faber, T. et al. Detail solutions regarding fatigue and the use of high-strength steels for towers of offshore wind energy converters (In German). Final report project P 633, Forschungsvereinigung Stahlanwendung e. V. FOSTA, Düsseldorf; 2008. Cheng, X., Fisher, J. W., Prask, H. J., Gnäupel-Herold, T., Yen, B. T., Roy, S. Residual stress modification by post-weld treatment and its beneficial effect on fatigue strength of welded structures. Int J Fatigue 2003;25:1259-1269. doi:10.1016/j.ijfatigue.2003.08.020 Janosch, J. J., Koneczny, H., Debiez, S., Statnikov, E. C., Troufiakov, V. J., Mikhee, P. P. Improvement of fatigue strength in welded joints (in HSS and in aluminium alloys) by ultrasonic hammer peening. Weld World 1996;37;72-83. Suominen, L., Khurshid, M., Parantainen, J. Residual stresses in welded components following post-weld treatment methods, Proceedings of the 5th Fatigue Design Conference; 2013 Nov 27-28, Senlis, France, in Procedia Eng 2013:66;181-191. Yildirim, H. C., Marquis, G. B. Overview of fatigue data for high frequency mechanical impact treated welded joints. Weld World 2012;56:82-96. doi: 10.1007/BF03321368. BS 7608:2014-03, Guide to the fatigue design and assessment of steel products, British Standards Institution, UK; 2014. Yildirim, H. C. Design aspects of high strength steel welded structures improved by high frequency mechanical impact (HFMI) treatment. PhD-Thesis, Aalto University; 2013. Dürr, A. Zur Ermüdungsfestigkeit von Schweißkonstruktionen aus höherfesten Baustählen bei Anwendung von UIT-Nachbehandlung (In German), PhD-Thesis, University of Stuttgart; 2007. Stranghöner, N., Berg, J. Influence of high frequency hammer peening on the fatigue strength of welded notch details of ultra high strength steels with steel grades S960, S1100 and S1300 (In German). Final report project P 938, Forschungsvereinigung Stahlanwendung e. V. FOSTA, Düsseldorf; 2015 (in press). Gundel, W., Hamme, U., Herion, S. Ermüdungsfestigkeit geschweißter Konstruktionen aus hoch- und höchstfesten Feinkornbaustählen in der Praxis (In German). DVS reports, Volume 267, 2010, p. 220-224. EN ISO 6892-1:2009-12, Metallic materials – Tensile testing – Part 1: method of test at room temperature, 2009. EN ISO 4136:2013-02, Destructive tests on welds in metallic materials – Transverse tensile test, 2013. EN ISO 5817:2014-06, Welding – Fusion-welded joints in steel, nickel, titanium and their alloys (beam welding excluded) – Quality levels for imperfections, 2014. Stranghöner, N., Berg, J., Schiborr, M., Neuhoff, R. Einfluss des höherfrequenten Hämmerns auf die Ermüdungsfestigkeit geschweißter ultrahochfester Stähle (In German). Proceedings 16. Duisburger Schweißtage; 2011 Jun 30; Duisburg, Germany. 2011. p. 75-85. Berg, J., Stranghöner, N. Höherfrequentes Hämmern am Beispiel geschweißter ultrahochfester Stähle (In German). Proceedings 18. DASt-Kolloquium; 2012 Oct 16-17; Aachen, Germany. Deutscher Ausschuß für Stahlbau; 2012. p. 125-130. Brozetti, J., Hirt, M. A., Ryan, I., Sedlacek, G., Smith, I. F. C. Background information on fatigue design rules – Statistical evaluation – Chapter 9 – Document 9.01, Eurocode 3 Editorial Group, 1989. Marquis, G., Barsoum, Z. Fatigue strength improvement of steel structures by high-frequency mechanical impact: proposed procedures and quality assurance guidelines. Weld World 2014;58:19-28. doi: 10.1007/s40194-013-0077-8. Statnikov, E. S., Trufiakov, V. I., Mikheev, P. P., Kudryavtsev, Y. Specification for weld toe improvement by ultrasonic impact treatment. Weld World 2000;44:5-7. Berg, J. Zum Einfluss des Höherfrequenten Hämmerns auf die Ermüdungsfestigkeit geschweißter ultrahochfester Feinkornbaustähle (In German). PhD-Thesis, University of Duisburg-Essen, 2015 (in preparation). Berg, J., Stranghöner, N. Application of design concepts at high frequency hammer peened welded ultra high strength steels. Proceedings of the 37th IABSE Symposium; 2014 Sep 3-5; Madrid, Spain. 2014. p. 1579-1586. Sonsino, C. M. Effect of residual stresses on the fatigue behaviour of welded joints depending on loading conditions and weld geometry, Int J Fatigue 2009;31:88-101. doi:10.1016/j.ijfatigue.2008.02.015. Farajian, M. Welding residual stress behavior under mechanical loading. Weld World 2013;57:157-169. doi: 10.1007/s40194-013-0024-8. Weich, I. Edge layer condition and fatigue strength of welds improved by mechanical post-weld treatment. Weld World 2011;55:3-12. doi:10.1007/BF03263510. Weich, I., Ummenhofer, T., Nitschke-Pagel, Th., Dilger, K., Eslami, H. Fatigue behaviour of welded high strength steels after high frequency mechanical post-weld treatments. Weld World 2009;53:R322-R332.

International Journal of Fatigue doi:10.1007/BF03263475.

Highlights (maximum of 85 characters, including spaces, per highlight)     

119 fatigue tests on S960, S1100 and S1300 with as welded and HFHP condition. The negative inverse slope of the S-N-line increases to m ~ 5 due to HFHP treatment. An improvement due to HFHP can be observed at load cycles of 10,000 and higher. Increase of FAT classes of design proposals is possible if HFHP is applied at UHSS. By HFHP treatment, other untreated notch details can be relevant for design.

15