## Abstract

It is well known that the fatigue strength of welded structures is in general independent from the material strength. In case of high-strength steels, however, a significant improvement in the fatigue behaviour can be realised through post-treatment processes. This paper deals with the effect of high-frequency mechanical impact (HFMI) on the fatigue behaviour of a range of steels starting from mild construction steel (S355) to ultra high-strength steel (S960). The experiments involve fatigue tests at a stress ratio of R = 0.1 on butt welds, T-joints, and longitudinal attachments on 5 mm, thin-walled specimens. The fatigue assessment was performed in accordance to the nominal and the notch stress approach taking the HFMI condition into account. Finally, a novel method is outlined to evaluate the notch stress fatigue behaviour of HFMI-treated joints made of high-strength steel. Applicability of this new HFMI notch stress approach is shown through fatigue assessment of about 330 HFMI post-treated specimens taken from both literature and own test results. Further work focuses on the expansion of the introduced HFMI notch stress model for load spectra influence covering overloads and multiaxial fatigue.

Access provided by Technische Universität Graz Universitätsbibliothek

1 Introduction

In the finite lifetime area, the fatigue behaviour of welded high-strength steel joints is mostly beneficial due to the high yield strength of the base material. In regard to the high-cycle fatigue zone, the notch topography, microstructure in the heat-affected zone, and the residual stress state have a significant influence on the fatigue lifetime. According to the International Institute of Welding (IIW) recommendation [1], the fatigue life is independent of the yield strength of welded steel components. However, additional high-frequency mechanical impact (HFMI) treatment offers the possibility to increase the fatigue life, especially for high-strength steels. To assess the local fatigue strength of welded and HFMI post-treated high-strength steel joints experimentally, fatigue tests using three different thin-walled (t eff = 5 mm) joints are investigated. Figure 1 illustrates the investigated specimen types, ranging from butt joint, root surface in grinded condition, over T-joint to longitudinal attachment.

Fig. 1

figure 1

Investigated joint types

Full size image

Two low-alloyed high-strength steels S690 and S960, and for comparative purposes a common construction steel S355, with a sheet thickness of 5 mm are used as different base materials. All plates were sand blasted before welding. For the as-welded condition, the structural detail-dependent nominal stress range is defined in [1] as FAT, characteristic fatigue strength at two million cycles considering a probability of survival of 97.7 %. Figure 2 depicts the nominal stress FAT classes of the three probed joints.

Fig. 2

figure 2

Recommended nominal stress FAT values for the investigated steel joints [1]

Full size image

In the nominal stress concept, the fatigue strength enhancement for higher strength steel welds (f y > 355 MPa) improved by hammer or needle peening is expressed by a bonus factor of 1.5 [1]. This factor is applied to the recommended stress range. An overview of the existing post-weld treatment methods, their application, and the proposed benefit in fatigue is given in [2]. Recent research results [3–5] observed that the fatigue strength of improved HFMI-treated welds increases with material yield strength. In [3], an extensive study including 228 experimental test data points showed that the fatigue benefit can be expressed by an increase of 12.5 % for every 200 MPa increase in the material strength, choosing f y,0 = 355 MPa as base material strength reference.

2 Objectives

In this contribution, an evaluation of the fatigue behaviour of HFMI-treated joints using the nominal and the notch stress concept for the investigated joints is presented. The goal is to assess methodically the benefit of the HFMI treatment for thin-walled high-strength steels. The work packages undertaken in this study are:

Experimental fatigue tests to ascertain the influence of the base material yield strength and the stress concentration factor at the weld toe on the fatigue behaviour of as welded (without post-treatment) and HMFI-treated steel joints.

Application of an existing procedure [3] to consider HFMI in the nominal stress concept and verification of the own experimental data compared to the published results.

Proposal of a novel S/N model to consider the HFMI treatment by the notch stress approach. Validation of this HFMI-enhanced notch stress approach by extensive experimental data.

3 Experimental work

Preliminary investigations [6–8] concerning the fatigue testing of welded joints, showed that a ratio 10:1 between width and thickness of the base plate encourages technical crack initiation in plate centre and a semi-elliptical crack growth through the plate thickness. After welding and cooling down to room temperature, half of each specimen lot is additionally post-treated by HFMI. The post-treatment procedure is shown in Fig. 3 for the three investigated joints.

Fig. 3

figure 3

High-frequency mechanical impact treatment

Full size image

The HFMI treatment is done using the PIT (Pitec) [9] device. For the pneumatic actuation energy, a common industrial air pressure of 6 bar (0.6 MPa) is needed. Comparative tests showed that a sufficient post-treatment quality can be achieved when the treatment velocity is in between v = 20–30 cm/min and at an operating frequency of 90 Hz. The radius of the hardened pin used in this study is R = 2 mm, which fits to the weld seam size of the investigated specimens [10]. Post-treatment is applied on the welded specimens without additional static prestressing or former cyclic damage. Due to the HFMI treatment, the pulsed transaction of the hardened pin rounds out the curvature of the weld toe region and thereby reduces the geometric stress concentration factor at the weld toe. In addition, compressive stresses are induced which counteract the superposition of residual weld and external load applied tensile stresses [11].

A further effect is an increase of the local hardness. For the evaluation of the local hardness distribution in the vicinity of the weld toe, a Vickers macrohardness measurement device with a test load of 3 kg weight is used. A grid of 0.25 mm spacing in horizontal and vertical direction is applied for the measurement. Figure 4 depicts the hardness mappings on a S960 longitudinal attachment in as-welded and HFMI-treated condition. In case of the HFMI-treated specimen, a slightly increased local hardness value of 360 HV3 in the area beneath the treated weld toe was found. In the untreated condition, a hardness value of about 340 HV3 was measured in the heat-affected zone.

Fig. 4

figure 4

Hardness measurements (HV3) for S960 longitudinal attachments

Full size image

In total, over 300 specimens were tested with a constant tumescent test applying a stress ratio of R = 0.1. The abort criterion for the fatigue tests is total rupture and the run-out level is set to a number of 50 million load cycles.

4 Fatigue test results

In the IIW recommendation [1], additional fatigue strength modifications such as the thickness factor f(t) are well defined. The thickness correction is only valid in case of the nominal stress approach and not in local methods. In the nominal stress approach, the fatigue strength is defined for a reference plate thickness of t ref = 25 mm. For thicknesses greater than this value, a reduction of the fatigue strength by the factor f(t) is recommended [1]. In the same way, a benign thinness effect might be considered, which however needs to be verified by experimental tests. Previous investigations [6–8] showed that such a benign thinness effect is applicable to thin-walled joints if the weld process quality is high as it is in this in case of semi-automated manufacturing of the samples. An overview of the nominal stress results is given in Figs. 5, 6, and 7. Each subfigure outlines the S/N curves of the experimental test results for the base material as well as the as welded and the HFMI-treated condition for each weld specimen type.

Fig. 5

figure 5

Nominal S/N curves for S355

Full size image

Fig. 6

figure 6

Nominal S/N curves for S690

Full size image

Fig. 7

figure 7

Nominal S/N curves for S960

Full size image

The fatigue assessment in the finite life region is performed using the evaluation procedure according to [12] by assuming a log-normal distributed test data. To determine the endurance stress level in the high-cycle fatigue region, the arcsine square root transformation [13] is applied at constant number of 50 million load cycles to determine the fatigue limit. Both methods use a probability of survival of P S = 97.7 % to determine the corresponding FAT class. In case of the fatigue assessment in the high-cycle fatigue region, a recommended shallower slope of k = 22 is used for the high-cycle fatigue assessment.

The black solid lines in Figs. 5, 6, and 7 mark the structural detail dependent recommended S/N curves, which are enhanced by the introduced benign thinness factor. It can be seen that the as-welded fatigue test samples are still above these limits which show the basic applicability of the benign thinness factor for the examined high-quality manufactured joints in the nominal stress approach.

To display the observed technical crack initiation mode and its dependency on seam geometry and HFMI treatment, an accompanying fracture surface analysis for each specimen was carried out (Fig. 8). In general, two typical zones with the fatigue fracture (A) and the rupture area (B) are recognisable. For the butt and T-joint specimens, multiple crack initiation appears along the weld toe line both for the as welded and the HFMI-treated condition.

Fig. 8

figure 8

Fracture surface examples [14]

Full size image

A detailed look at the fracture surfaces of the longitudinal attachments shows crack initiation at the end-of-seam caused by the high local notch effect and a major change in stiffness at this region. Most specimens failed in the weld toe region, but a minor number of HFMI-treated specimens cracked at the bottom side of the base material within the heat-affected zone which are further on not considered for the local fatigue assessment. Thereby, the HFMI post-treatment leads to a shift of the critical area from the weld toe to the heat-affected base material and a subsequent increase in endurable number of load cycles. In this case, the geometric notch effect is removed and subsequently the metallurgical notch is essential for the fatigue behaviour of the welded joint. To maintain an utmost conservative approach, HFMI-treated specimens with technical crack initiation significantly apart from the treated weld toe, e.g., in the heat-affected base material, were not considered in the depicted fatigue results.

Comparisons of the nominal stress fatigue test results lead to the following statements:

For the T and butt joint, there is only a minor enhancement of the fatigue behaviour due to the HFMI treatment compared to the as welded condition. This is due to high quality of the weld process which includes use of optimised weld process parameters [6, 7].

The most significant increase in fatigue is obtained for the longitudinal attachment, caused by the small, highly stressed volume at the weld toe whereat, the HFMI treatment is very effective.

In the HFMI-treated condition, the achievable FAT stress range reaches about 85 % of the base material fatigue behaviour in case of the common construction steel S355. But for the two high-strength steels, S690 and S960, the HFMI-treated fatigue behaviour of the investigated thin-walled joints is almost equal to the base material fatigue limit.

In the high-cycle fatigue region, the post-treatment resulted in a significant enhancement compared to the finite lifetime regime. In the nominal stress approach, the HFMI treatment bonus factors seem to be nearly independent of the geometry-dependent stress concentration factor K t as the fatigue behaviour is close to (S355) or almost equal (S690 and S960) to that of the base material.

5 Fatigue assessment of HFMI-treated joints

5.1 Nominal stress approach

In [3], an extensive analysis of published experimental data on the fatigue strength of HFMI-treated welded joints describes a procedure to consider the increased fatigue behaviour due to higher base material yield strength in the nominal stress concept. Previous investigations in [15] indicate that a slope of k = 5 fits the HFMI-treated data best. A previous proposal [15, 16] of a design-based method for HFMI improved joints up to 690 MPa yield strength defines a strength magnification factor k y . The factor depends on the strength correction α for yield strength after HFMI treatment, the reference yield strength f y,0 = 355 MPa and the yield strength of the applied base material f y , see Eq. 1.

ky=α⋅(fy−fy,0)/fy,0

(1)

Based on the procedure given by [3], each fatigue test result (ΔS i , N f,i ) can be transformed to two million load cycles using a slope of k = 5 by Eq. 2.

ΔS∗i=((ΔSi)k⋅Nf,i/(2⋅106))1/k

(2)

To get a unified fatigue assessment independent of the investigated material strengths, the fatigue test result ΔS i is transferred into a yield strength corrected nominal stress range ΔS i C by multiplication with the evaluated mean strength ΔS m,A , see Eq. 3. The yield strength dependency is taken into account by the strength magnification factor k y . The evaluation procedure was done in accordance to the statistical methods summarised in [15].

ΔSCi=(ΔSm,A)ky⋅(ΔS∗i)1−ky

(3)

The value of the strength magnification factor can be solved by minimization of the standard deviation σ N of the yield strength corrected nominal fatigue test results. To evaluate all three specimen types as a single dataset, the derived fatigue strength values are normalised by dividing both sides of Eq. 3 with ΔS m,A , reflecting the individual joint specific structural detail [3].

(ΔSCi/ΔSm,A)=(ΔS∗i/ΔSm,A)1−ky

(4)

Figure 9 depicts the achieved normalised stress range results achieved by the magnification correction factor f y for all three specimen types. A value of α = 0.277 leads to a minimum standard deviation of σ N = 0.203 for the three investigated weld geometries and the three base material combinations. The normalised fatigue strength at two million load cycles lead to a mean value of ΔS m (2e6) = 1.75 and ΔS k (2e6) = 1.45 for P S = 97.7 %, which is in good agreement with the published results [3].

Fig. 9

figure 9

Normalised nominal stress results with material strength correction (α = 0.277)

Full size image

5.2 Notch stress approach

In the course of the local fatigue assessment, a calculation of the notch stress concentration factors K t at the weld toe for the three investigated specimen types is performed. The modelled geometries depict the technical cross-section of the seam, local deviations from the design geometry are not considered. In the investigated numerical calculations, at least four elements were applied at the 45° weld toe region [17]. Figure 10 depicts the results of the finite element analysis in the form of contour plots of the three investigated specimen geometries using a reference radius of 1 mm.

Fig. 10

figure 10

Numerical evaluation of stress concentration factor K t

Full size image

To determine the fatigue strength of HFMI-treated joints by the notch stress approach, it is recommended in [17] to model the real weld toe radius by an increased value of plus 1 mm and apply the derived notch stress concentration factor within the notch stress approach. This procedure is preferable for relatively sharp notches and has not been thoroughly verified.

An alternative procedure for assessing post-weld-treated joints using the notch stress approach is investigated in [18] and maintains the reference radius of r ref = 1 mm in addition to use a higher FAT class due to the post-weld treatment. Hence, the same stress concentration factors (r ref = 1 mm) for the untreated (as welded) and HFMI-treated joints are applicable for the assessment. In Fig. 11, the local master S/N curves for the as welded and the HFMI-treated condition (base material S690) are depicted for further discussion.

Fig. 11

figure 11

Local master S/N curves (S690); a as welded, b HFMI treated

Full size image

In the as-welded condition, the investigated specimens show a similar notch stress fatigue range with a relatively small scatter band of T S = 1:1.28 (σ N = 0.173). The evaluated master S/N curve is above the recommendation, especially in the high-cycle fatigue region, due of the meticulous manufacturing process. However, the local notch stress assessment of the HFMI-treated joints shows that the longitudinal attachment is significantly above the other joints due to the high effectiveness of the post-treatment for this particular joint. The influence of the stress concentration factor K t is remarkable, which leads to an improper scatter band of T S = 1:2.30 (σ N = 0.348). Summing up, the local notch stress approach based on a reference radius of r ref = 1 mm is well applicable in the as-welded condition but not for the HFMI-treated joints without further modification.

To improve the assessment of HFMI-treated joints by the notch stress concept, an alternative model is introduced as follows. Figure 12a depicts the principle procedure of the suggested method to assess the local notch stress S/N curve for HFMI-treated joints. The local HFMI notch stress model can be described by three characteristic points:

Fig. 12

figure 12

a Basic principle of HFMI notch stress model, b reference HFMI master S/N-curve for f y = 355 MPa and K t = 1.6

Full size image

1.

The base point stress range Δσ B at N B = 2 · 103 in the low-cycle fatigue region is influenced by the base material yield strength f y only.

2.

A reference HFMI master S/N curve for f y = 355 MPa and K t,r = 1 mm = 1.6 with a slope of k 1,min acts as reference for the calculation of the slope k 1 in the finite-life fatigue region by application of a power law dependency of the yield strength f y and stress concentration factor K t,r = 1 mm.

3.

A fixed transition knee-point of one million load cycles is suggested for simplification. The recommended, shallower slope of k 2 = 22 is used in the high-cycle fatigue section.

To assess the HFMI notch stress model as a single reference master S/N curve, the fatigue test data points within study and in addition, data from literature [5, 10, 14, 16, 19–25] are related to the shallowest notch stress condition of K t = 1.6, which occurs in case of butt joints. Finally, the different material classes are unified to 355 MPa yield strength by the material strength magnification factor k y . The observed master S/N curve in Fig. 12b shows a very small scatter band of σ N = 0.173 and proofs therefore the basic applicability of the novel HFMI notch stress model.

To apply the introduced HFMI notch stress model based on the assessed reference master HFMI notch stress S/N curve, it is necessary to consider only three adjustment factors as follows:

The strength magnification factor k y considers the increase of the fatigue strength due to the yield strength f y of the base material (Eq. 1). As shown in the nominal stress assessment, a value of α = 0.277 is applicable for the yield strength correction after HFMI treatment.

To consider different stress ratios R, a strength magnification adjustment k R is recommended in [16]. Equation 5 depicts the correlation.

kR=1.075−0.75⋅Rfor0.1≤R≤0.5kR=1forR

(5)

Finally, a new slope magnification adjustment factor k k considers the increase in slope due to the HFMI treatment. The slope magnification adjustment depends on the stress concentration factor K t,r = 1 mm and the material yield strength f y . This two-parametric dependency of this new factor is shown in Fig. 13a. As an example, the notch stress concentration factor for the investigated longitudinal attachment shows a value of K t,r = 1 mm = 2.55. For the S690 high-strength steel, this leads to a slope magnification adjustment factor of k k = 1.57 and further on the estimated slope in the finite lifetime region for this specific joint type is calculated as k 1 = k k · k 1,min = 1.57 · 3.7 = 5.8.

Fig. 13

figure 13

a Slope magnification factor k k in dependence of K t,r = 1 mm and f y ; b slope k in dependence of normalised stress concentration factor (S690)

Full size image

Figure 13b depicts the dependency of the slope k of the normalised stress concentration at the weld toe for S690 welded joints in HFMI-treated condition. Thereby, a power law fits the experimental data with a low scattering. In the same way the influence of the base material on the slope magnification factor k k can also be described in dependence of the material yield strength f y , related to a reference yield strength of 355 MPa.

Further on, the conservative applicability of this new HFMI notch stress model is shown for the experimental test results of the mild construction steel S355 and the high-strength steel S690 in Fig. 14a,b. The calculation is based on the master HFMI notch stress S/N curve in Fig. 12b, considering the derived FAT class of 225 MPa and the slope value k 1,min = 3.7 in the finite-life region as starting point.

Fig. 14

figure 14

Application of HFMI model for different base materials; a S355, b S690

Full size image

As expected, the given values of this master HFMI S/N curve are close to the recommended notch stress S/N curve with a reference radius of r ref = 1 mm. Firstly, the FAT class is enhanced due to the use of high-strength steel by the strength magnification bonus factor 1 + k y , see Eq. 1. Secondly, the influence of the stress ratio affects the HFMI notch stress FAT value by the magnification factor k R , see Eq. 5. Thirdly, the slope k 1,min of the HFMI notch stress S/N curve in the finite life region has to be multiplied by the two-parametric slope magnification adjustment factor k k , see the drawn power law dependency in Fig. 13a. Finally, the evaluated finite life S/N curve is valid up to one million cycles. Above this constant HFMI notch stress transition knee point, a shallower slope of k 2 = 22 is applied for the high-cycle fatigue region.

The experimental fatigue test data points in Fig. 14a,b are clearly above the evaluated S/N curves implying a conservative approach. This can be attributed to the high quality of manufacturing of the thin-walled, single-layered specimens within this series. As shown in Fig. 12b, the developed HFMI notch stress method seems to be widely applicable for different joints and materials. The experimentally investigated HFMI data points from the authors are given in Appendix 1, whereas Appendix 2 sums up the properties of the data points taken from literature.

6 Conclusions

Thin-walled longitudinal attachments, T-joints, and butt welds were dynamically tested at a stress ratio of R = 0.1 to determine the influence of the HFMI post-treatment on fatigue. The investigated materials are a common construction steel S355 and two low-alloyed high-strength steels S690 and S960. The plate thickness in each case was 5 mm. The weld toe region was HFMI-treated without static prestressing during the conditioning. The following conclusions have been reached in this work:

As exemplified in the fatigue test summary section, the base material stress range can be reached for the high-strength steel joints with HFMI treatment. All fatigue test data points are above the IIW recommendation, even if a bonus factor of 1.5 for HFMI treatment and a benign thinness bonus factor are applied.

The nominal stress assessment of HFMI-treated joints was done in accordance to the procedure recently introduced in [3]. The examined test results validated the applicability of the method, leading to a strength correction factor α = 0.277.

A novel model to assess the notch stress HFMI fatigue behaviour is introduced. It consists of a material strength dependent base point at 2,000 cycles a material strength and notch stress concentration factor dependent slope in the finite life region, a fixed transition knee point at one million load cycles, and a second fixed shallower slope up to the endurance limit. By evaluation of extensive fatigue datasets, an HFMI notch stress master S/N curve was assessed. Based on this master S/N curve and the novel slope magnification factor k k , a local S/N-curve can be derived by application of the joint type-dependent notch stress concentration factor K t , to be determined by numerical assessment using a reference radius r ref = 1 mm, and the base material yield strength f y . It is shown that this new method is applicable to unify the notch stress HFMI fatigue test data into a notch stress master S/N curve which implies both material and joint type dependency.

References

Hobbacher A (2009) IIW recommendations for fatigue design of welded joints and components, WRC Bulletin 520. The Welding Research Council, New York

Google Scholar

Haagensen P.J. and Maddox S.J. (2009) IIW recommendations on post weld improvement of steel and aluminum structures, IIW-Doc. XIII-2200r3-07, revised

Yildirim HC, Marquis GB (2012) Fatigue strength improvement factors for high strength steel welded joints treated by high frequency mechanical impact. Int J Fatigue 44:168–176

Article

Google Scholar

Marquis GB (2010) Failure modes and fatigue strength of improved HSS welds. Eng Fract Mech 77:2051–2062

Article

Google Scholar

Leitner M, Stoschka M (2012) Influence of steel grade on the fatigue strength enhancement by high frequency peening technology on longitudinal fillet weld gusset. J Eng Technol 1(3):80–90

Google Scholar

Stoschka M, Leitner M, Fössl T, Posch G (2012) Effect of high-strength filler metals on fatigue. Weld World 56(3/4):20–29

Article

Google Scholar

Stoschka M, Leitner M, Posch G, Eichlseder W (2013) Effect of high-strength filler metals on the fatigue behaviour of butt joints. Welding World 57(1):85–96

Article

Google Scholar

Barsoum Z (2011) Fatigue design of welded structures—some aspects of weld quality and residual stresses. Welding World 55(11–12):2–11

Article

Google Scholar

Pitec GmbH. http://www.pitec-gmbh.com/

Statnikov ES, Muktepavel VO, Blomqvist A (2002) Comparison of ultrasonic impact treatment (UIT) and other fatigue life improvement methods. Weld World 46:28–39

Article

Google Scholar

Sonsino CM (2009) Effect of residual stresses on the fatigue behaviour of welded joints depending on loading conditions and weld geometry. Int J Fatigue 31:88–101

Article

Google Scholar

ASTM International. Standard practice for statistical analysis of linear or linearized stress–life (S-N) and strain–life (ε-N) fatigue data, designation: E739-91, reapproved 1998

Dengel D, Harig H (1980) Estimation of the fatigue limit by progressively-increasing load tests. Fatigue Fract Eng Mater Struct 3(2):113–128

Article

Google Scholar

Leitner M, Stoschka M, Schanner R, Eichlseder W (2012) Influence of high frequency peening on fatigue of high-strength steels. FME Trans 40(3):99–104

Google Scholar

Yildirim HC, Marquis GB (2012) Overview of fatigue data for high frequency mechanical impact treated welded joints. Weld World 57(7/8):82–96

Article

Google Scholar

Weich I (2008) Fatigue behaviour of mechanical post weld treated welds depending on the edge layer condition (Ermüdungsverhalten mechanisch nachbehandelter Schweißverbindungen in Abhängigkeit des Randschichtzustands). Technische Universität Carolo-Wilhelmina, Doctorate Thesis

Google Scholar

Fricke W. (2008) Guideline for fatigue assessment by notch stress analysis for welded structures, IIW-Document XIII-2240r1-08/XV-1289r1-08

Pedersen M.M., Mouritsen O.Ø., Hansen M.R. and Andersen J.G. (2010) Experience with the notch stress approach for fatigue assessment of welded joints. Swedish Conference on Light Weight Optimized Welded Structures, pp. 122–133

Kuhlmann U, Dürr A, Bergmann J, Thumser R (2006) Fatigue strength improvement for welded high strength steel connections due to the application of post-weld treatment methods (Effizienter Stahlbau aus höherfesten Stählen unter Ermüdungsbeanspruchung), Forschungsvorhaben P620 FOSTA. Verlag und Vertriebsgesellschaft GmbH, Düsseldorf

Google Scholar

Kuhlmann U, Gunther H (2009) Experimentelle Untersuchungen zur ermüdungssteigernden Wirkung des PIT-Verfahrens. Versuchsbericht, Universität Stuttgart, Institut für Konstruktion und Entwurf

Google Scholar

Lihavainen VM, Marquis G, Statnikov ES (2004) Fatigue strength of a longitudinal attachment improved by ultrasonic impact treatment. Welding World 48(5–6):67–73

Article

Google Scholar

Maddox S. J., Dore M. J. and Smith S. D. (2010) Investigation of ultrasonic peening for upgrading a welded steel structure, IIW-Document XIII-2326-10

Mori T., Shimanuki H. and Tanaka M. (2011) Effect of UIT on fatigue strength of web-gusset welded joints considering service condition of steel structures, IIW-Document XIII-2376-11

Okawa T., Shimanuki H., Funatsu Y., Nose T. and Sumi Y. (2011) Effect of preload and stress ratio on fatigue strength of welded joints improved by ultrasonic impact treatment, IIW-Document XIII-2377-11

Wang T, Wang D, Huo L, Zhang Y (2009) Discussion on fatigue design of welded joints enhanced by ultrasonic peening treatment (UPT). Int J Fatigue 31:644–650

Article

Google Scholar

Download references

Acknowledgments

Financial support by the Austrian Federal Government (in particular from Bundesministerium für Verkehr, Innovation und Technologie and Bundesministerium für Wirtschaft, Familie und Jugend) represented by Österreichische Forschungsförderungsgesellschaft mbH and the Styrian and the Tyrolean Provincial Government, represented by Steirische Wirtschaftsförderungsgesellschaft mbH and Standortagentur Tirol, within the framework of the COMET Funding Programme is gratefully acknowledged.

Special thanks are given to the Austrian Research Promotion Agency (FFG), who founded the research project by funds of the Federal Ministry for Transport, Innovation and Technology (bmvit) and the Federal Ministry of Economics and Labour (BMWA), and to all the industry partners for the supply of material and the fabrication work done.

Author information

Authors and Affiliations

Mechanical Engineering, Department Product Engineering, Montanuniversität Leoben, Leoben, Austria

M. Leitner, M. Stoschka & W. Eichlseder

Corresponding author

Correspondence to M. Leitner.

Additional information

Doc. IIW-2413, recommended for publication by Commission XIII "Fatigue of Welded Components and Structures".

Appendices

Appendix 1

S355

S690

S960

Δσn [MPa]

N [−]

Δσn [MPa]

N [−]

Δσn [MPa]

N [−]

Butt joint (HFMI treated)

400

46,788

600

34,108

600

51,391

350

130,700

500

69,884

500

126,047

350

117,144

500

88,929

500

114,960

325

300,466

400

277,182

400

396,154

325

210,170

400

209,163

400

290,788

325

166,587

350

385,506

350

951,175

300

450,641

350

232,637

325

17.888,401

300

146,755

300

448,500

300

5e7

300

390,968

300

606,954

275

366,638

300

513,201

250

765,948

275

5e7

225

1.108,615

225

1.384,740

200

3.490,632

175

5e7

T-joint (HFMI treated)

400

45,762

600

24,590

700

38,777

400

40,219

500

79,746

700

33,922

350

127,859

500

87,332

600

68,315

350

129,313

400

208,053

600

78,002

300

331,773

400

179,853

500

93,213

300

292,273

350

453,600

500

121,296

300

357,196

350

776,715

500

135,449

300

285,857

325

1.201,199

500

116,512

250

467,515

300

5e7

400

368,640

250

569,303

400

264,182

225

1.159,771

375

481,572

225

1.224,537

325

5e7

200

5e7

Longitudinal attachment (HFMI treated)

400

11,347

550

26,654

700

18,272

350

23,192

500

50,082

600

32,266

350

28,142

500

47,592

600

36,696

325

43,619

450

97,218

500

136,787

325

45,176

450

155,664

500

85,693

300

123,131

400

199,496

400

276,312

300

105,499

400

183,912

350

630,259

275

261,661

375

288,633

350

557,678

250

549,371

350

681,242

350

804,910

250

834,344

350

380,182

325

1.073,244

250

532,668

275

5e7

300

7.658,700

238

605,922

275

5e7

225

5e7

Appendix 2

figure a

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Leitner, M., Stoschka, M. & Eichlseder, W. Fatigue enhancement of thin-walled, high-strength steel joints by high-frequency mechanical impact treatment. Weld World 58, 29–39 (2014). https://doi.org/10.1007/s40194-013-0097-4

Download citation

Received02 October 2012

Accepted21 August 2013

Published13 September 2013

Issue DateJanuary 2014

DOIhttps://doi.org/10.1007/s40194-013-0097-4

Share this article

Anyone you share the following link with will be able to read this content:

Provided by the Springer Nature SharedIt content-sharing initiative

Keywords

High frequency mechanical impact treatment (HFMI)

Fatigue tests

Notch stress approach

High strength steels

Download PDF

Working on a manuscript?

Avoid the common mistakes

Abstract

Introduction

Objectives

Experimental work

Fatigue test results

Fatigue assessment of HFMI-treated joints

Conclusions

References

Acknowledgments

Author information

Additional information

Appendices

Rights and permissions

About this article

Advertisement

Access provided by Technische Universität Graz Universitätsbibliothek

1 Introduction

In the finite lifetime area, the fatigue behaviour of welded high-strength steel joints is mostly beneficial due to the high yield strength of the base material. In regard to the high-cycle fatigue zone, the notch topography, microstructure in the heat-affected zone, and the residual stress state have a significant influence on the fatigue lifetime. According to the International Institute of Welding (IIW) recommendation [1], the fatigue life is independent of the yield strength of welded steel components. However, additional high-frequency mechanical impact (HFMI) treatment offers the possibility to increase the fatigue life, especially for high-strength steels. To assess the local fatigue strength of welded and HFMI post-treated high-strength steel joints experimentally, fatigue tests using three different thin-walled (t eff = 5 mm) joints are investigated. Figure 1 illustrates the investigated specimen types, ranging from butt joint, root surface in grinded condition, over T-joint to longitudinal attachment.

Fig. 1

figure 1

Investigated joint types

Full size image

Two low-alloyed high-strength steels S690 and S960, and for comparative purposes a common construction steel S355, with a sheet thickness of 5 mm are used as different base materials. All plates were sand blasted before welding. For the as-welded condition, the structural detail-dependent nominal stress range is defined in [1] as FAT, characteristic fatigue strength at two million cycles considering a probability of survival of 97.7 %. Figure 2 depicts the nominal stress FAT classes of the three probed joints.

Fig. 2

figure 2

Recommended nominal stress FAT values for the investigated steel joints [1]

Full size image

In the nominal stress concept, the fatigue strength enhancement for higher strength steel welds (f y > 355 MPa) improved by hammer or needle peening is expressed by a bonus factor of 1.5 [1]. This factor is applied to the recommended stress range. An overview of the existing post-weld treatment methods, their application, and the proposed benefit in fatigue is given in [2]. Recent research results [3–5] observed that the fatigue strength of improved HFMI-treated welds increases with material yield strength. In [3], an extensive study including 228 experimental test data points showed that the fatigue benefit can be expressed by an increase of 12.5 % for every 200 MPa increase in the material strength, choosing f y,0 = 355 MPa as base material strength reference.

2 Objectives

In this contribution, an evaluation of the fatigue behaviour of HFMI-treated joints using the nominal and the notch stress concept for the investigated joints is presented. The goal is to assess methodically the benefit of the HFMI treatment for thin-walled high-strength steels. The work packages undertaken in this study are:

Experimental fatigue tests to ascertain the influence of the base material yield strength and the stress concentration factor at the weld toe on the fatigue behaviour of as welded (without post-treatment) and HMFI-treated steel joints.

Application of an existing procedure [3] to consider HFMI in the nominal stress concept and verification of the own experimental data compared to the published results.

Proposal of a novel S/N model to consider the HFMI treatment by the notch stress approach. Validation of this HFMI-enhanced notch stress approach by extensive experimental data.

3 Experimental work

Preliminary investigations [6–8] concerning the fatigue testing of welded joints, showed that a ratio 10:1 between width and thickness of the base plate encourages technical crack initiation in plate centre and a semi-elliptical crack growth through the plate thickness. After welding and cooling down to room temperature, half of each specimen lot is additionally post-treated by HFMI. The post-treatment procedure is shown in Fig. 3 for the three investigated joints.

Fig. 3

figure 3

High-frequency mechanical impact treatment

Full size image

The HFMI treatment is done using the PIT (Pitec) [9] device. For the pneumatic actuation energy, a common industrial air pressure of 6 bar (0.6 MPa) is needed. Comparative tests showed that a sufficient post-treatment quality can be achieved when the treatment velocity is in between v = 20–30 cm/min and at an operating frequency of 90 Hz. The radius of the hardened pin used in this study is R = 2 mm, which fits to the weld seam size of the investigated specimens [10]. Post-treatment is applied on the welded specimens without additional static prestressing or former cyclic damage. Due to the HFMI treatment, the pulsed transaction of the hardened pin rounds out the curvature of the weld toe region and thereby reduces the geometric stress concentration factor at the weld toe. In addition, compressive stresses are induced which counteract the superposition of residual weld and external load applied tensile stresses [11].

A further effect is an increase of the local hardness. For the evaluation of the local hardness distribution in the vicinity of the weld toe, a Vickers macrohardness measurement device with a test load of 3 kg weight is used. A grid of 0.25 mm spacing in horizontal and vertical direction is applied for the measurement. Figure 4 depicts the hardness mappings on a S960 longitudinal attachment in as-welded and HFMI-treated condition. In case of the HFMI-treated specimen, a slightly increased local hardness value of 360 HV3 in the area beneath the treated weld toe was found. In the untreated condition, a hardness value of about 340 HV3 was measured in the heat-affected zone.

Fig. 4

figure 4

Hardness measurements (HV3) for S960 longitudinal attachments

Full size image

In total, over 300 specimens were tested with a constant tumescent test applying a stress ratio of R = 0.1. The abort criterion for the fatigue tests is total rupture and the run-out level is set to a number of 50 million load cycles.

4 Fatigue test results

In the IIW recommendation [1], additional fatigue strength modifications such as the thickness factor f(t) are well defined. The thickness correction is only valid in case of the nominal stress approach and not in local methods. In the nominal stress approach, the fatigue strength is defined for a reference plate thickness of t ref = 25 mm. For thicknesses greater than this value, a reduction of the fatigue strength by the factor f(t) is recommended [1]. In the same way, a benign thinness effect might be considered, which however needs to be verified by experimental tests. Previous investigations [6–8] showed that such a benign thinness effect is applicable to thin-walled joints if the weld process quality is high as it is in this in case of semi-automated manufacturing of the samples. An overview of the nominal stress results is given in Figs. 5, 6, and 7. Each subfigure outlines the S/N curves of the experimental test results for the base material as well as the as welded and the HFMI-treated condition for each weld specimen type.

Fig. 5

figure 5

Nominal S/N curves for S355

Full size image

Fig. 6

figure 6

Nominal S/N curves for S690

Full size image

Fig. 7

figure 7

Nominal S/N curves for S960

Full size image

The fatigue assessment in the finite life region is performed using the evaluation procedure according to [12] by assuming a log-normal distributed test data. To determine the endurance stress level in the high-cycle fatigue region, the arcsine square root transformation [13] is applied at constant number of 50 million load cycles to determine the fatigue limit. Both methods use a probability of survival of P S = 97.7 % to determine the corresponding FAT class. In case of the fatigue assessment in the high-cycle fatigue region, a recommended shallower slope of k = 22 is used for the high-cycle fatigue assessment.

The black solid lines in Figs. 5, 6, and 7 mark the structural detail dependent recommended S/N curves, which are enhanced by the introduced benign thinness factor. It can be seen that the as-welded fatigue test samples are still above these limits which show the basic applicability of the benign thinness factor for the examined high-quality manufactured joints in the nominal stress approach.

To display the observed technical crack initiation mode and its dependency on seam geometry and HFMI treatment, an accompanying fracture surface analysis for each specimen was carried out (Fig. 8). In general, two typical zones with the fatigue fracture (A) and the rupture area (B) are recognisable. For the butt and T-joint specimens, multiple crack initiation appears along the weld toe line both for the as welded and the HFMI-treated condition.

Fig. 8

figure 8

Fracture surface examples [14]

Full size image

A detailed look at the fracture surfaces of the longitudinal attachments shows crack initiation at the end-of-seam caused by the high local notch effect and a major change in stiffness at this region. Most specimens failed in the weld toe region, but a minor number of HFMI-treated specimens cracked at the bottom side of the base material within the heat-affected zone which are further on not considered for the local fatigue assessment. Thereby, the HFMI post-treatment leads to a shift of the critical area from the weld toe to the heat-affected base material and a subsequent increase in endurable number of load cycles. In this case, the geometric notch effect is removed and subsequently the metallurgical notch is essential for the fatigue behaviour of the welded joint. To maintain an utmost conservative approach, HFMI-treated specimens with technical crack initiation significantly apart from the treated weld toe, e.g., in the heat-affected base material, were not considered in the depicted fatigue results.

Comparisons of the nominal stress fatigue test results lead to the following statements:

For the T and butt joint, there is only a minor enhancement of the fatigue behaviour due to the HFMI treatment compared to the as welded condition. This is due to high quality of the weld process which includes use of optimised weld process parameters [6, 7].

The most significant increase in fatigue is obtained for the longitudinal attachment, caused by the small, highly stressed volume at the weld toe whereat, the HFMI treatment is very effective.

In the HFMI-treated condition, the achievable FAT stress range reaches about 85 % of the base material fatigue behaviour in case of the common construction steel S355. But for the two high-strength steels, S690 and S960, the HFMI-treated fatigue behaviour of the investigated thin-walled joints is almost equal to the base material fatigue limit.

In the high-cycle fatigue region, the post-treatment resulted in a significant enhancement compared to the finite lifetime regime. In the nominal stress approach, the HFMI treatment bonus factors seem to be nearly independent of the geometry-dependent stress concentration factor K t as the fatigue behaviour is close to (S355) or almost equal (S690 and S960) to that of the base material.

5 Fatigue assessment of HFMI-treated joints

5.1 Nominal stress approach

In [3], an extensive analysis of published experimental data on the fatigue strength of HFMI-treated welded joints describes a procedure to consider the increased fatigue behaviour due to higher base material yield strength in the nominal stress concept. Previous investigations in [15] indicate that a slope of k = 5 fits the HFMI-treated data best. A previous proposal [15, 16] of a design-based method for HFMI improved joints up to 690 MPa yield strength defines a strength magnification factor k y . The factor depends on the strength correction α for yield strength after HFMI treatment, the reference yield strength f y,0 = 355 MPa and the yield strength of the applied base material f y , see Eq. 1.

ky=α⋅(fy−fy,0)/fy,0

(1)

Based on the procedure given by [3], each fatigue test result (ΔS i , N f,i ) can be transformed to two million load cycles using a slope of k = 5 by Eq. 2.

ΔS∗i=((ΔSi)k⋅Nf,i/(2⋅106))1/k

(2)

To get a unified fatigue assessment independent of the investigated material strengths, the fatigue test result ΔS i is transferred into a yield strength corrected nominal stress range ΔS i C by multiplication with the evaluated mean strength ΔS m,A , see Eq. 3. The yield strength dependency is taken into account by the strength magnification factor k y . The evaluation procedure was done in accordance to the statistical methods summarised in [15].

ΔSCi=(ΔSm,A)ky⋅(ΔS∗i)1−ky

(3)

The value of the strength magnification factor can be solved by minimization of the standard deviation σ N of the yield strength corrected nominal fatigue test results. To evaluate all three specimen types as a single dataset, the derived fatigue strength values are normalised by dividing both sides of Eq. 3 with ΔS m,A , reflecting the individual joint specific structural detail [3].

(ΔSCi/ΔSm,A)=(ΔS∗i/ΔSm,A)1−ky

(4)

Figure 9 depicts the achieved normalised stress range results achieved by the magnification correction factor f y for all three specimen types. A value of α = 0.277 leads to a minimum standard deviation of σ N = 0.203 for the three investigated weld geometries and the three base material combinations. The normalised fatigue strength at two million load cycles lead to a mean value of ΔS m (2e6) = 1.75 and ΔS k (2e6) = 1.45 for P S = 97.7 %, which is in good agreement with the published results [3].

Fig. 9

figure 9

Normalised nominal stress results with material strength correction (α = 0.277)

Full size image

5.2 Notch stress approach

In the course of the local fatigue assessment, a calculation of the notch stress concentration factors K t at the weld toe for the three investigated specimen types is performed. The modelled geometries depict the technical cross-section of the seam, local deviations from the design geometry are not considered. In the investigated numerical calculations, at least four elements were applied at the 45° weld toe region [17]. Figure 10 depicts the results of the finite element analysis in the form of contour plots of the three investigated specimen geometries using a reference radius of 1 mm.

Fig. 10

figure 10

Numerical evaluation of stress concentration factor K t

Full size image

To determine the fatigue strength of HFMI-treated joints by the notch stress approach, it is recommended in [17] to model the real weld toe radius by an increased value of plus 1 mm and apply the derived notch stress concentration factor within the notch stress approach. This procedure is preferable for relatively sharp notches and has not been thoroughly verified.

An alternative procedure for assessing post-weld-treated joints using the notch stress approach is investigated in [18] and maintains the reference radius of r ref = 1 mm in addition to use a higher FAT class due to the post-weld treatment. Hence, the same stress concentration factors (r ref = 1 mm) for the untreated (as welded) and HFMI-treated joints are applicable for the assessment. In Fig. 11, the local master S/N curves for the as welded and the HFMI-treated condition (base material S690) are depicted for further discussion.

Fig. 11

figure 11

Local master S/N curves (S690); a as welded, b HFMI treated

Full size image

In the as-welded condition, the investigated specimens show a similar notch stress fatigue range with a relatively small scatter band of T S = 1:1.28 (σ N = 0.173). The evaluated master S/N curve is above the recommendation, especially in the high-cycle fatigue region, due of the meticulous manufacturing process. However, the local notch stress assessment of the HFMI-treated joints shows that the longitudinal attachment is significantly above the other joints due to the high effectiveness of the post-treatment for this particular joint. The influence of the stress concentration factor K t is remarkable, which leads to an improper scatter band of T S = 1:2.30 (σ N = 0.348). Summing up, the local notch stress approach based on a reference radius of r ref = 1 mm is well applicable in the as-welded condition but not for the HFMI-treated joints without further modification.

To improve the assessment of HFMI-treated joints by the notch stress concept, an alternative model is introduced as follows. Figure 12a depicts the principle procedure of the suggested method to assess the local notch stress S/N curve for HFMI-treated joints. The local HFMI notch stress model can be described by three characteristic points:

Fig. 12

figure 12

a Basic principle of HFMI notch stress model, b reference HFMI master S/N-curve for f y = 355 MPa and K t = 1.6

Full size image

1.

The base point stress range Δσ B at N B = 2 · 103 in the low-cycle fatigue region is influenced by the base material yield strength f y only.

2.

A reference HFMI master S/N curve for f y = 355 MPa and K t,r = 1 mm = 1.6 with a slope of k 1,min acts as reference for the calculation of the slope k 1 in the finite-life fatigue region by application of a power law dependency of the yield strength f y and stress concentration factor K t,r = 1 mm.

3.

A fixed transition knee-point of one million load cycles is suggested for simplification. The recommended, shallower slope of k 2 = 22 is used in the high-cycle fatigue section.

To assess the HFMI notch stress model as a single reference master S/N curve, the fatigue test data points within study and in addition, data from literature [5, 10, 14, 16, 19–25] are related to the shallowest notch stress condition of K t = 1.6, which occurs in case of butt joints. Finally, the different material classes are unified to 355 MPa yield strength by the material strength magnification factor k y . The observed master S/N curve in Fig. 12b shows a very small scatter band of σ N = 0.173 and proofs therefore the basic applicability of the novel HFMI notch stress model.

To apply the introduced HFMI notch stress model based on the assessed reference master HFMI notch stress S/N curve, it is necessary to consider only three adjustment factors as follows:

The strength magnification factor k y considers the increase of the fatigue strength due to the yield strength f y of the base material (Eq. 1). As shown in the nominal stress assessment, a value of α = 0.277 is applicable for the yield strength correction after HFMI treatment.

To consider different stress ratios R, a strength magnification adjustment k R is recommended in [16]. Equation 5 depicts the correlation.

kR=1.075−0.75⋅Rfor0.1≤R≤0.5kR=1forR

(5)

Finally, a new slope magnification adjustment factor k k considers the increase in slope due to the HFMI treatment. The slope magnification adjustment depends on the stress concentration factor K t,r = 1 mm and the material yield strength f y . This two-parametric dependency of this new factor is shown in Fig. 13a. As an example, the notch stress concentration factor for the investigated longitudinal attachment shows a value of K t,r = 1 mm = 2.55. For the S690 high-strength steel, this leads to a slope magnification adjustment factor of k k = 1.57 and further on the estimated slope in the finite lifetime region for this specific joint type is calculated as k 1 = k k · k 1,min = 1.57 · 3.7 = 5.8.

Fig. 13

figure 13

a Slope magnification factor k k in dependence of K t,r = 1 mm and f y ; b slope k in dependence of normalised stress concentration factor (S690)

Full size image

Figure 13b depicts the dependency of the slope k of the normalised stress concentration at the weld toe for S690 welded joints in HFMI-treated condition. Thereby, a power law fits the experimental data with a low scattering. In the same way the influence of the base material on the slope magnification factor k k can also be described in dependence of the material yield strength f y , related to a reference yield strength of 355 MPa.

Further on, the conservative applicability of this new HFMI notch stress model is shown for the experimental test results of the mild construction steel S355 and the high-strength steel S690 in Fig. 14a,b. The calculation is based on the master HFMI notch stress S/N curve in Fig. 12b, considering the derived FAT class of 225 MPa and the slope value k 1,min = 3.7 in the finite-life region as starting point.

Fig. 14

figure 14

Application of HFMI model for different base materials; a S355, b S690

Full size image

As expected, the given values of this master HFMI S/N curve are close to the recommended notch stress S/N curve with a reference radius of r ref = 1 mm. Firstly, the FAT class is enhanced due to the use of high-strength steel by the strength magnification bonus factor 1 + k y , see Eq. 1. Secondly, the influence of the stress ratio affects the HFMI notch stress FAT value by the magnification factor k R , see Eq. 5. Thirdly, the slope k 1,min of the HFMI notch stress S/N curve in the finite life region has to be multiplied by the two-parametric slope magnification adjustment factor k k , see the drawn power law dependency in Fig. 13a. Finally, the evaluated finite life S/N curve is valid up to one million cycles. Above this constant HFMI notch stress transition knee point, a shallower slope of k 2 = 22 is applied for the high-cycle fatigue region.

The experimental fatigue test data points in Fig. 14a,b are clearly above the evaluated S/N curves implying a conservative approach. This can be attributed to the high quality of manufacturing of the thin-walled, single-layered specimens within this series. As shown in Fig. 12b, the developed HFMI notch stress method seems to be widely applicable for different joints and materials. The experimentally investigated HFMI data points from the authors are given in Appendix 1, whereas Appendix 2 sums up the properties of the data points taken from literature.

6 Conclusions

Thin-walled longitudinal attachments, T-joints, and butt welds were dynamically tested at a stress ratio of R = 0.1 to determine the influence of the HFMI post-treatment on fatigue. The investigated materials are a common construction steel S355 and two low-alloyed high-strength steels S690 and S960. The plate thickness in each case was 5 mm. The weld toe region was HFMI-treated without static prestressing during the conditioning. The following conclusions have been reached in this work:

As exemplified in the fatigue test summary section, the base material stress range can be reached for the high-strength steel joints with HFMI treatment. All fatigue test data points are above the IIW recommendation, even if a bonus factor of 1.5 for HFMI treatment and a benign thinness bonus factor are applied.

The nominal stress assessment of HFMI-treated joints was done in accordance to the procedure recently introduced in [3]. The examined test results validated the applicability of the method, leading to a strength correction factor α = 0.277.

A novel model to assess the notch stress HFMI fatigue behaviour is introduced. It consists of a material strength dependent base point at 2,000 cycles a material strength and notch stress concentration factor dependent slope in the finite life region, a fixed transition knee point at one million load cycles, and a second fixed shallower slope up to the endurance limit. By evaluation of extensive fatigue datasets, an HFMI notch stress master S/N curve was assessed. Based on this master S/N curve and the novel slope magnification factor k k , a local S/N-curve can be derived by application of the joint type-dependent notch stress concentration factor K t , to be determined by numerical assessment using a reference radius r ref = 1 mm, and the base material yield strength f y . It is shown that this new method is applicable to unify the notch stress HFMI fatigue test data into a notch stress master S/N curve which implies both material and joint type dependency.

References

Hobbacher A (2009) IIW recommendations for fatigue design of welded joints and components, WRC Bulletin 520. The Welding Research Council, New York

Google Scholar

Haagensen P.J. and Maddox S.J. (2009) IIW recommendations on post weld improvement of steel and aluminum structures, IIW-Doc. XIII-2200r3-07, revised

Yildirim HC, Marquis GB (2012) Fatigue strength improvement factors for high strength steel welded joints treated by high frequency mechanical impact. Int J Fatigue 44:168–176

Article

Google Scholar

Marquis GB (2010) Failure modes and fatigue strength of improved HSS welds. Eng Fract Mech 77:2051–2062

Article

Google Scholar

Leitner M, Stoschka M (2012) Influence of steel grade on the fatigue strength enhancement by high frequency peening technology on longitudinal fillet weld gusset. J Eng Technol 1(3):80–90

Google Scholar

Stoschka M, Leitner M, Fössl T, Posch G (2012) Effect of high-strength filler metals on fatigue. Weld World 56(3/4):20–29

Article

Google Scholar

Stoschka M, Leitner M, Posch G, Eichlseder W (2013) Effect of high-strength filler metals on the fatigue behaviour of butt joints. Welding World 57(1):85–96

Article

Google Scholar

Barsoum Z (2011) Fatigue design of welded structures—some aspects of weld quality and residual stresses. Welding World 55(11–12):2–11

Article

Google Scholar

Pitec GmbH. http://www.pitec-gmbh.com/

Statnikov ES, Muktepavel VO, Blomqvist A (2002) Comparison of ultrasonic impact treatment (UIT) and other fatigue life improvement methods. Weld World 46:28–39

Article

Google Scholar

Sonsino CM (2009) Effect of residual stresses on the fatigue behaviour of welded joints depending on loading conditions and weld geometry. Int J Fatigue 31:88–101

Article

Google Scholar

ASTM International. Standard practice for statistical analysis of linear or linearized stress–life (S-N) and strain–life (ε-N) fatigue data, designation: E739-91, reapproved 1998

Dengel D, Harig H (1980) Estimation of the fatigue limit by progressively-increasing load tests. Fatigue Fract Eng Mater Struct 3(2):113–128

Article

Google Scholar

Leitner M, Stoschka M, Schanner R, Eichlseder W (2012) Influence of high frequency peening on fatigue of high-strength steels. FME Trans 40(3):99–104

Google Scholar

Yildirim HC, Marquis GB (2012) Overview of fatigue data for high frequency mechanical impact treated welded joints. Weld World 57(7/8):82–96

Article

Google Scholar

Weich I (2008) Fatigue behaviour of mechanical post weld treated welds depending on the edge layer condition (Ermüdungsverhalten mechanisch nachbehandelter Schweißverbindungen in Abhängigkeit des Randschichtzustands). Technische Universität Carolo-Wilhelmina, Doctorate Thesis

Google Scholar

Fricke W. (2008) Guideline for fatigue assessment by notch stress analysis for welded structures, IIW-Document XIII-2240r1-08/XV-1289r1-08

Pedersen M.M., Mouritsen O.Ø., Hansen M.R. and Andersen J.G. (2010) Experience with the notch stress approach for fatigue assessment of welded joints. Swedish Conference on Light Weight Optimized Welded Structures, pp. 122–133

Kuhlmann U, Dürr A, Bergmann J, Thumser R (2006) Fatigue strength improvement for welded high strength steel connections due to the application of post-weld treatment methods (Effizienter Stahlbau aus höherfesten Stählen unter Ermüdungsbeanspruchung), Forschungsvorhaben P620 FOSTA. Verlag und Vertriebsgesellschaft GmbH, Düsseldorf

Google Scholar

Kuhlmann U, Gunther H (2009) Experimentelle Untersuchungen zur ermüdungssteigernden Wirkung des PIT-Verfahrens. Versuchsbericht, Universität Stuttgart, Institut für Konstruktion und Entwurf

Google Scholar

Lihavainen VM, Marquis G, Statnikov ES (2004) Fatigue strength of a longitudinal attachment improved by ultrasonic impact treatment. Welding World 48(5–6):67–73

Article

Google Scholar

Maddox S. J., Dore M. J. and Smith S. D. (2010) Investigation of ultrasonic peening for upgrading a welded steel structure, IIW-Document XIII-2326-10

Mori T., Shimanuki H. and Tanaka M. (2011) Effect of UIT on fatigue strength of web-gusset welded joints considering service condition of steel structures, IIW-Document XIII-2376-11

Okawa T., Shimanuki H., Funatsu Y., Nose T. and Sumi Y. (2011) Effect of preload and stress ratio on fatigue strength of welded joints improved by ultrasonic impact treatment, IIW-Document XIII-2377-11

Wang T, Wang D, Huo L, Zhang Y (2009) Discussion on fatigue design of welded joints enhanced by ultrasonic peening treatment (UPT). Int J Fatigue 31:644–650

Article

Google Scholar

Download references

Acknowledgments

Financial support by the Austrian Federal Government (in particular from Bundesministerium für Verkehr, Innovation und Technologie and Bundesministerium für Wirtschaft, Familie und Jugend) represented by Österreichische Forschungsförderungsgesellschaft mbH and the Styrian and the Tyrolean Provincial Government, represented by Steirische Wirtschaftsförderungsgesellschaft mbH and Standortagentur Tirol, within the framework of the COMET Funding Programme is gratefully acknowledged.

Special thanks are given to the Austrian Research Promotion Agency (FFG), who founded the research project by funds of the Federal Ministry for Transport, Innovation and Technology (bmvit) and the Federal Ministry of Economics and Labour (BMWA), and to all the industry partners for the supply of material and the fabrication work done.

Author information

Authors and Affiliations

Mechanical Engineering, Department Product Engineering, Montanuniversität Leoben, Leoben, Austria

M. Leitner, M. Stoschka & W. Eichlseder

Corresponding author

Correspondence to M. Leitner.

Additional information

Doc. IIW-2413, recommended for publication by Commission XIII "Fatigue of Welded Components and Structures".

Appendices

Appendix 1

S355

S690

S960

Δσn [MPa]

N [−]

Δσn [MPa]

N [−]

Δσn [MPa]

N [−]

Butt joint (HFMI treated)

400

46,788

600

34,108

600

51,391

350

130,700

500

69,884

500

126,047

350

117,144

500

88,929

500

114,960

325

300,466

400

277,182

400

396,154

325

210,170

400

209,163

400

290,788

325

166,587

350

385,506

350

951,175

300

450,641

350

232,637

325

17.888,401

300

146,755

300

448,500

300

5e7

300

390,968

300

606,954

275

366,638

300

513,201

250

765,948

275

5e7

225

1.108,615

225

1.384,740

200

3.490,632

175

5e7

T-joint (HFMI treated)

400

45,762

600

24,590

700

38,777

400

40,219

500

79,746

700

33,922

350

127,859

500

87,332

600

68,315

350

129,313

400

208,053

600

78,002

300

331,773

400

179,853

500

93,213

300

292,273

350

453,600

500

121,296

300

357,196

350

776,715

500

135,449

300

285,857

325

1.201,199

500

116,512

250

467,515

300

5e7

400

368,640

250

569,303

400

264,182

225

1.159,771

375

481,572

225

1.224,537

325

5e7

200

5e7

Longitudinal attachment (HFMI treated)

400

11,347

550

26,654

700

18,272

350

23,192

500

50,082

600

32,266

350

28,142

500

47,592

600

36,696

325

43,619

450

97,218

500

136,787

325

45,176

450

155,664

500

85,693

300

123,131

400

199,496

400

276,312

300

105,499

400

183,912

350

630,259

275

261,661

375

288,633

350

557,678

250

549,371

350

681,242

350

804,910

250

834,344

350

380,182

325

1.073,244

250

532,668

275

5e7

300

7.658,700

238

605,922

275

5e7

225

5e7

Appendix 2

figure a

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Leitner, M., Stoschka, M. & Eichlseder, W. Fatigue enhancement of thin-walled, high-strength steel joints by high-frequency mechanical impact treatment. Weld World 58, 29–39 (2014). https://doi.org/10.1007/s40194-013-0097-4

Download citation

Received02 October 2012

Accepted21 August 2013

Published13 September 2013

Issue DateJanuary 2014

DOIhttps://doi.org/10.1007/s40194-013-0097-4

Share this article

Anyone you share the following link with will be able to read this content:

Provided by the Springer Nature SharedIt content-sharing initiative

Keywords

High frequency mechanical impact treatment (HFMI)

Fatigue tests

Notch stress approach

High strength steels

Download PDF

Working on a manuscript?

Avoid the common mistakes

Abstract

Introduction

Objectives

Experimental work

Fatigue test results

Fatigue assessment of HFMI-treated joints

Conclusions

References

Acknowledgments

Author information

Additional information

Appendices

Rights and permissions

About this article

Advertisement

Original language | English |
---|---|

Pages (from-to) | 29-39 |

Number of pages | 11 |

Journal | Welding in the World |

Volume | 58 |

DOIs | |

Publication status | Published - 2014 |

Externally published | Yes |