You are on page 1of 8

Journal of Materials Processing Technology 223 (2015) 305–312

Contents lists available at ScienceDirect

Journal of Materials Processing Technology


journal homepage: www.elsevier.com/locate/jmatprotec

Enhancement of the fatigue strength of underwater wet welds by


grinding and ultrasonic impact treatment
Wenbin Gao a , Dongpo Wang a , Fangjie Cheng a,∗ , Caiyan Deng a , Yang Liu a , Wei Xu b
a
Tianjin Key Laboratory of Advanced Joining Technology, School of Materials Science and Engineering, Tianjin University, Tianjin 300072, People’s Republic
of China
b
Offshore Oil Engineering Co., Ltd., Tianjin 300072, People’s Republic of China

a r t i c l e i n f o a b s t r a c t

Article history: Grinding + underwater ultrasonic impact treatment (UUIT) is proposed to improve the fatigue property
Received 9 January 2015 of underwater wet welded joints. Using single grinding or UUIT technique cannot improve the fatigue
Received in revised form 12 April 2015 life significantly while the combination of the two can obtain great enhancement. The microstructures in
Accepted 13 April 2015
the weld metal contain proeutectoid ferrite, side-plate ferrite and acicular ferrite and the microstructure
Available online 23 April 2015
in coarse grain heat affected zone (CGHAZ) is dominated by martensite. Compressive residual stress
decreases the effective stress intensity factor range and extends the fatigue life. The transverse residual
Keywords:
stresses on the surface of as-welded joint are compressive due to the fast cooling effect of water and it is
Underwater wet welding
Ultrasonic impact treatment
less in weld metal (WM) than in CGHAZ, which results in the initiation of fatigue crack at WM.
Fatigue © 2015 Elsevier B.V. All rights reserved.
Microstructure
Residual stress

1. Introduction obtained at shallow water exceeded the requirements for AWS D3.6
Class A Weld. Perez et al. (2003) added nickel to oxide electrode
Underwater wet shielded metal-arc welding is widely applied in covering and showed that 2% (wt.) nickel could obtain the opti-
basic repairs of offshore structures because of its versatility and low mal Charpy impact values. In addition, Rowe et al. (2002) studied
cost compared with other repair techniques. Rowe and Liu (2001) the influence of adding ferro-alloy additions on the quality of wet
revealed that the main problems of this technique were related welds.
with the presence of water around the electric arc, which caused However, it should be emphasized that published research on
higher cooling rate, arc instability, and large amounts of oxygen the fatigue behaviour of underwater wet welds is still rare. More-
and hydrogen in the arc atmosphere. over, there are almost no reports on the application of underwater
For underwater wet welding, most previous experiments have ultrasonic impact treatment (UUIT) to strengthen the fatigue per-
focused on the common mechanical properties or the metallurgi- formance of underwater wet welds. Therefore, the current research
cal aspects of the welds obtained with different compositions of the on this field is particularly significant.
electrode rod, coating, or waterproof materials. Pessoa et al. (2006) Haagensen and Maddox (2006) showed that burr grinding,
found that porosity reduced along the multi-pass underwater wet tungsten inert gas (TIG) dressing, hammer peening, and needle
welds and samples extracted from the end of the welds showed peening were the most commonly applied post weld treatments
higher strength and ductility. Jia et al. (2013) studied the spectro- for improvement of fatigue behaviour of welded joints. Yildirim and
scopic analysis of the arc plasma of flux-cored arc welding in air and Marquis (2012) showed that ultrasonic impact treatment (UIT) is
under water. They found that the two sets of spectrum signals were considered an ideal method because of its easy operation, high effi-
largely similar and a unique peak at 656.2793 nm existed in the ciency and low cost. TIG dressing cannot be directly applied under
underwater spectrum suggesting that H atoms became involved. water. So in this study burr grinding and UUIT were utilized to
Santos et al. (2012) developed an oxyrutile electrode for wet weld- improve the fatigue behaviour. The microstructure, residual stress
ing combining the good operability of rutile electrodes and the low and fracture morphology were also discussed.
diffusible hydrogen content. The mechanical properties of welds
2. Experimental procedure

∗ Corresponding author. Tel.: +86 022 27405889; fax: +86 022 27405889. The base metal was DH36 steel plate with a thickness of
E-mail address: chfj@tju.edu.cn (F. Cheng). 12.7 mm. The filler material was an E7014 underwater electrode

http://dx.doi.org/10.1016/j.jmatprotec.2015.04.013
0924-0136/© 2015 Elsevier B.V. All rights reserved.
306 W. Gao et al. / Journal of Materials Processing Technology 223 (2015) 305–312

Table 1
Chemical compositions of materials used in this study, wt.%.

Material C Si Mn Al Nb V Ti O CEa

Base metal 0.13 0.18 1.4 0.02 0.02 0.05 0.012 – 0.373
Filler material 0.073 0.31 0.46 0.006 – – 0.012 0.042 –
a
CE = C + Mn /6 + (Cr + Mo + V ) /5 + (Ni + Cu ) /15.

Table 2 a Microhardness Tester MHV2000. Then, cross-sectional morpholo-


Mechanical properties of materials used in this study.
gies and fracture surfaces of the joints were observed by OM and
Material Yield strength Tensile Elongation rate SEM.
(MPa) strength (MPa) (%)

Base metal 375 510 30 3. Results


Filler material 476 523 8.0
3.1. Microstructure

Fig. 3 depicts typical OM and SEM images showing the


microstructures of the underwater wet welded joint. The
microstructure of the base metal (BM) consists of a banded struc-
ture with fine elongated ferrite (F) and pearlite (P) grains (Fig. 3a).
As shown in Fig. 3c and e, the columnar microstructures in the weld
metal (WM) are characterized by proeutectoid ferrite (PF), side-
plate ferrite (SPF) and a small amount of acicular ferrite (AF). The
PF can nucleate at the boundaries or at the inner microdefects in
the austenite grains. In the SPF, the proeutectoid phase separates
inside the grains along certain crystallographic planes. SPF sharply
decreases the toughness of the weld metal.
The microstructures of the coarse-grained heat affected zone
(CGHAZ) (Fig. 3d and f) include lath martensite (M) containing dif-
ferently oriented packets with several fine parallel laths and some
Fig. 1. Geometric characteristics of specimen (all dimensions in mm). retained austenite. Since CGHAZ could not be totally protected by
the bubbles, the rapid cooling effect of water leads to the genera-
with a diameter of 3.2 mm and a waterproofing of wax. The tion of quenching microstructure. In fact, Johnson (1997) reported
chemical composition and mechanical properties are listed in that the cooling time t8/5 (from 800 ◦ C to 500 ◦ C) at the point 1mm
Tables 1 and 2, respectively. distant from the fusion line on the plate surface did not exceed
The underwater joints were welded in the 3G down position 2 s. The lath martensite of CGHAZ has a very high hardness, which
(AWS D3.6 M:2010) in a tank. The water temperature and water makes it easy to generate hydrogen-induced cracks due to the large
depth were 20 ◦ C and 2 m, respectively. The welding parame- amounts of hydrogen in the joint.
ters were as follows: arc voltage of 24–26 V, welding current of
153 A and welding speed of 300 mm/min. Each fillet weld included 3.2. Fatigue behaviour
three passes. 45 non-load-carrying transverse cruciform speci-
mens (Fig. 1) were prepared. Specimens were divided into four The fatigue testing data were analysed according to the IIW sta-
groups: as-welded, burr grinding, UUIT and burr grinding + UUIT. tistical method described by Hobbacher (2008). The S–N curves
Burr grinding was carried out according to the Recommendations were fitted based on the following formulae:
on Post Weld Improvement of Steel and Aluminium Structures m
Cm = N() (1)
of IIW described by Haagensen and Maddox (2006). As shown in
Fig. 2a, the UUIT system consists of four parts: an ultrasonic (20 kHz) where Cm represents 50% survival probability calculated from the
frequency generator, an ultrasonic impact gun, ultrasonic horn and mean on the basis of two-sided tolerance limits at the 75% level; N
an air compressor. The radius of the needle tip is 2 mm. The UUIT is the number of cycles to failure;  is the nominal stress range at
process was conducted in a water container, as shown in Fig. 2b. the weld toe; m is the characteristic value. The statistical results are
Air was continually compressed into the gun to protect it from listed in Table 3. The characteristic values m of the four S–N curves
the water. The treatment parameters were as follows: a frequency range from 5.51 to 7.44, which are all larger than the value (m = 3)
of 20 kHz, vibration amplitude of 25 ␮m, treating speed of about suggested in the fatigue design criteria of the IIW.
1.0–1.5 m/min, and treatment coverage of 200% (treatment cover- Fig. 4 shows the S–N curves of the four groups of speci-
age is defined as the ratio of the area covered by the shot impacts mens. Table 4 shows the fatigue strength data corresponding to
to the complete surface of the treated sample). 2 × 106 cycles. It is 135 MPa for the as-welded specimens, which
Fatigue tests were conducted under a constant amplitude ten- is increased by burr grinding, UUIT and burr grinding + UUIT by
sile load with a stress ratio (R) of 0.1 at room temperature. A
300 kN HF fatigue testing machine of CIMACH GPS 300 was utilized. Table 3
An OLYMPUS GX51 optical microscope (OM) and a scanning elec- Statistical results of fatigue test data.

tron microscopy (SEM) of Hatchi S-4800 were used to observe the Status Stress m Cm Fitted equation
microstructure of the joint after being polished and etched by nital ratio R
solution (4% HNO3 , 96% C2 H5 OH by volume). The standard sin 2 As-welded 0.1 5.51 1.259 × 1018 1.259 × 1018 = N()5.51
X-ray diffraction technique was used to measure the residual stress. Burr grinding 0.1 6.26 1.381 × 1020 1.381 × 1020 = N()6.26
The material removal in the depth direction was achieved by elec- UUIT 0.1 7.44 1.389 × 1023 1.389 × 1023 = N()7.44
Grinding + UUIT 0.1 6.86 2.375 × 1022 2.375 × 1022 = N()6.86
trolytic polishing. The microhardness values were measured using
W. Gao et al. / Journal of Materials Processing Technology 223 (2015) 305–312 307

Fig. 2. (a) UUIT system. (b) UUIT in a water container.

Table 4 crack propagates along the line, which is perpendicular to the load-
Fatigue strength of underwater welded specimens.
ing direction.
Status Fatigue strength (2 × 106 )  (MPa) Improvement (%)

As-welded 135 – 3.4. Fracture morphologies


Burr grinding 162 20
UUIT 182 35
Grinding + UUIT 217 61
Fig. 6 shows the typical fatigue fracture surface. The fatigue frac-
tography consists of three different zones: fatigue crack initiation,
propagation and the final fracture zone. Since fatigue crack initiates
at the weld metal, the propagation zone contains three parts: weld
20%, 35% and 61% to 162 MPa, 182 MPa and 217 MPa, respectively. metal, HAZ and base metal (Fig. 6a). The crack initiates from the
FAT80 is the standard of non-load-carrying cruciform joint with edge imperfections at weld toe (Fig. 6b). In the propagation zone
fillet plates, and 80 is stress range (unite: MPa) corresponding to of weld metal, fatigue striations are not observed and the area has
2 × 106 cycles according to IIW. Clearly, the underwater joint can a nature of transcrystalline fatigue fracture with clear secondary
reach the standard of FAT80. cracks (Fig. 6c). Secondary cracks can decelerate the crack growth
Huo et al. (2005) have reported that UIT can enhance the FAT by consuming the energy of crack tip. As shown in Fig. 6d, the
values of air-welded joints by about 70%. Obviously, UUIT does not propagation zone of CGHAZ appears at around 0.5 mm away from
achieve the same degree of improvement. Molten metal is unable the edge, which is consistent with Fig. 5. This region is very rough
to flow to the edges of the molten pool so undercut often appears owing to a large volume of facets and can be characterized by brittle
in the weld metal near the fusion line. For the specimen with seri- intergranular and transcrystalline fracture feature. One hydrogen-
ous undercut defect, due to the hard martensite microstructure in induced crack perpendicular to the fusion line can be seen. The grain
CGHAZ, UUIT cannot completely eliminate the defects and induce size in CGHAZ grows heavily and grain boundary weakens owing to
compressive stress on the surface layer. Therefore, fatigue crack a long time of overheating, which results in intergranular fracture.
still initiates at the untreated part. In the propagation zone of base metal, fatigue striations and some
In short, using single burr grinding or UUIT technique cannot secondary cracks are observed (Fig. 6e). The striation directions are
improve the fatigue life significantly while the combination of the perpendicular to the direction of the crack growth. The fracture sec-
two techniques can obtain great enhancement under both higher tion of instantaneous destruction region (Fig. 6f) is characteristic of
and lower applied stresses. This mainly results from the combined dimple, in which the second phase particles exist, indicating that
effect of the remove of weld defects by burr grinding and the intro- the fracture is ductile.
duction of compressive residual stress by UUIT. The mechanism will
be explained in further detail.
3.5. Residual stress

3.3. Fatigue crack initiation sites Fig. 7 shows the residual stress distribution of the weld metal
near the fusion line (potential crack initiation site) along the depth
Fig. 5 shows the fatigue crack initiation sites and propagation direction of the four types of specimens. The residual stress at the
paths of the four types of specimens. The cracks all initiate at the surface layer of as-welded specimen is −106 MPa and the com-
weld metal close to the fusion line. It should be noted that this pressive stress along the depth direction quickly decreases and
location does not correspond to the deepest point of the groove. turns into tensile. The depth of compressive stress is just more
The mechanism will be illustrated in Section 4.2.2. Then, the fatigue than 100 ␮m. For burr grinding specimen, the residual stress at the
308 W. Gao et al. / Journal of Materials Processing Technology 223 (2015) 305–312

Fig. 3. OM and SEM images of microstructures of the welded joint: (a) base metal, (b) fusion line, (c) and (e) columnar microstructure in the weld seam, and (d) and (f)
microstructure of CGHAZ.

surface layer is 85 MPa and the tensile stress is introduced by the


grinding process. Due to the huge heat produced by grinding, the
metal in the ground surface layer contracts sharply and generates
the compressive plastic deformation. So the tensile residual stress is
present in the surface layer. The residual stress distribution along
the depth direction is similar to that of as-welded specimen. For
UUIT and grinding + UUIT specimens, no distinct difference is seen.
The compressive stress zones are both extended to the depth of
about 1200 ␮m, and the maximum compressive residual stresses
(about −320 MPa) are both in the near surface layer.

3.6. Microhardness

Fig. 8 shows the average microhardness values of weld metal


and CGHAZ at the distance of 50 ␮m from the surface. It can be seen
that for joints without UUIT treatment, the microhardness of weld
metal and CGHAZ is 290 HV0.1 and 350 HV0.1 , while for UUIT treated
joints, it increases by 48% and 40% to 430 HV0.1 and 490 HV0.1 ,
respectively. Due to the plastic deformation by UUIT, the grains
Fig. 4. S–N curves of underwater-wet-welded joints of DH36 as measured by the on the surface layer become elongated and have a great deal of
nominal stress (R = 0.1).
dislocation. Microhardness is very sensitive to dislocation density,
so the surface layer after UUIT has high hardness.
W. Gao et al. / Journal of Materials Processing Technology 223 (2015) 305–312 309

Fig. 5. Fatigue crack initiation sites.

4. Discussion stress concentration factor because of the largest notch radius of


3.2 mm. Kf of as-welded joint is 3.66 and it is reduced by 34%,
4.1. Stress concentration 28% and 34% using grinding, UUIT and grinding + UUIT, respectively.
Therefore, increasing transition radius of weld toe can significantly
The fatigue crack initiation life is affected considerably by the decrease the stress concentration and increase the number of cycles
weld toe profile. In order to evaluate the improvement of UUIT for crack initiation.
and burr grinding by the modification of weld toe profile, the con-
cept of fatigue notch stress concentration factor Kf is applied and
Radaj et al. (2006) showed that it could be calculated by Peterson’s 4.2. Residual stress
equation:
4.2.1. Residual stress of underwater wet welded joint
1 + (Kt − 1)
Kf =   (2) Yanagida and Koide (2006) applied a water shower behind the
1 + a∗ / welding torch to reduce tensile residual stress in the welded region.
Fast cooling, which leads to large temperature difference in the
where Kt is the stress concentration factor at the weld toe, the criti-
plate, can promote compressive stress on the surface. During cool-
cal distance a* is Peterson’s material constant,  is the notch radius
ing, the surface metal contracts more rapidly than the inner metal
(the unit is in mm).
and the contraction is prohibited by the inner metal, leading to
For cruciform joint, Radaj et al. (2006) showed that Kt can be
tensile stress in the surface. Suppose that there is no flexural defor-
calculated by the following equation according to Lawrence on the
mation between the two parts of different temperatures. Tan (2000)
basis of finite element method.
showed that the thermal stress can be calculated by the following
  H 5/3 1/2  t 1/2
 1/4 equation:
Kt = 1 + 0.35 tan 1 + 1.1 (3)
L 
[(˛2 − ˛1 ) (T2 − T1 )]
1 = −2 =   (5)
where  is the angle of weld toe, H and L are the sizes of the two 1/E1 + 1/E2
weld legs, t is the thickness of the plate.
For steel, Radaj et al. (2006) showed that a* could be calculated where  1 and  1 are the thermal stresses; ˛1 and ˛2 are coefficients
by the following equation according to Lawrence. of linear thermal expansion; T1 and T2 are the temperature; E1 and
10, 870 E2 are the modulus of elasticity. Davies et al. (2015) showed that for
a∗ = (4) DH36 steel, the coefficient of linear thermal expansion decreases
(b )2
slowly as temperature drops while modulus of elasticity decreases
where  b is the material ultimate tensile strength at the notch. quickly. Therefore, the large gap of temperature could lead to large
Quenched microstructure and work hardening by UUIT would have thermal stress. Due to the lower yield strength at high temperature,
an influence on Kf . Since  b is difficult to be measured directly, it can the surface metal easily tends to yield and generate the tensile plas-
be obtained by the conversion of hardness values near the surface tic deformation, which results in the compressive residual stress
shown in Fig. 8 according to the standard of ISO 18265:2013. during the further cooling to room temperature (seen in Fig. 9a).
The calculation results of Kf are shown in Table 5. It can be seen In addition, as shown in Fig. 9b, the HAZ width near the surface
that the joint treated by grinding has the smallest fatigue notch is much smaller than that in the inner metal, which indicates the
310 W. Gao et al. / Journal of Materials Processing Technology 223 (2015) 305–312

Fig. 6. Fracture morphology of specimen treated by burr grinding + UUIT and tested at the stress range of 160 MPa: (a) macroscopic fracture surface, (b-e) locally magnified
views of regions A–D, respectively, marked in Fig. 6a, and (f) fracture surface of instantaneous destruction region.

intensive water quenching effect and corresponds to the explana- (2010) studied the transverse residual stress distribution in a pipe
tion in Fig. 9a. girth weld under rapid cooling conditions by finite element analy-
Biallas (2013) showed that it is the transverse stress component sis method and found that compressive residual stress existed near
that directly contributes to fracture driving and has a significant weld surface, which corresponds to the test results in this study.
effect on crack growth threshold. As shown in Fig. 10, for underwa- Therefore, underwater wet welding has a potential to produce com-
ter as-welded butt joint, the transverse residual stress distribution pressive residual stresses in the surface of the joint.
along the line perpendicular to the weld centre line is measured
by X-ray diffraction technique. It can be found that the transverse
residual stresses in WM and HAZ both are compressive. The com- 4.2.2. Effect of compressive residual stress on fatigue property
pressive stress on the surface of CGHAZ can reach −203 MPa while UUIT can introduce compressive residual stress on the surface
that of WM is −120 MPa. The cooling rate in WM is slower due to layer, which can explain the significant improvement of fatigue life
the protection of water bubbles than that in CGHAZ. So the com- by grinding + UUIT. The ultimate stress is the superposition of the
pressive stress on the surface is lower in WM than in CGHAZ. Dong external tensile stress and the compressive residual stress. Zhang

Table 5
Fatigue notch stress concentration factor Kf .

Status Transition radius  (mm) Tensile strengtha  b (MPa) Kf Decline percentage (%)

As-welded 0.8 1385 3.66


Burr grinding 3.2 1350 2.40 34
UUIT 2.4 1613 2.62 28
Grinding + UUIT 3.2 1595 2.41 34
a
Converted from hardness values shown in Fig. 8 according to the standard of ISO 18265:2013.
W. Gao et al. / Journal of Materials Processing Technology 223 (2015) 305–312 311

Fig. 7. Residual stress profiles along the depth direction.


Fig. 10. Residual stress distribution on the surface of as-welded joint as a function
of distance from the weld centre line.

et al. (2015) showed that the effective stress intensity factor Keff
can be calculated by the following equation:

Keff = Kw + Krs (6)

where Kw is stress intensity factor of applied load, Krs is the resid-


ual stress intensity factor. For UUIT joint, Krs is mathematically
negative. The compressive residual stress is much larger than
the minimum stress in a cycle at R = 0.1. As a result, the mini-
mum effective stress intensity factor is mathematically negative
(Keff,min = Kmin + Krs < 0). Compressive stress would not arise crack
opening, so Keff,min equals to zero. Then, the effective stress inten-
sity factor range Keff and the effective stress ratio Reff can be
calculated by the following two equations:

Keff = Keff, max − Keff, min = Kmax + Krs − 0 < Kmax − Kmin = K

(7)

Keff, min
Fig. 8. Average micro-hardness at the distance of 50 ␮m from the surface.
Reff = =0 (8)
Keff, max

Fig. 9. (a) Mechanism of compressive residual stress distribution on the weld surface. (b) Cross-sectional image of weld joint.
312 W. Gao et al. / Journal of Materials Processing Technology 223 (2015) 305–312

Due to the compressive residual stress, Keff is less than the machines and processes for local dry and wet underwater welding)
nominal stress intensity factor range K, and the effective stress sponsored by Offshore Oil Engineering Co., Ltd.. The grant number
ratio Reff equals to zero. Consequently, the micro-crack initiation is Z12JZFNY151. Great thanks are given to Jun Cao and Wei Xu from
is retarded notably and the fatigue crack growth is hindered dis- Offshore Oil Engineering Co., Ltd.
tinctly.
In the present paper, all the fatigue cracks are found to initi- References
ate in the weld metal near the fusion line where there is not the
greatest reduction in the plate thickness. This can be explained by America Welding Society, 2010. Underwater Weld Code, AWS D3.6M:2010. America
Welding Society.
the residual stress difference between HAZ and WM. As shown in Biallas, G., 2013. Effect of welding residual stresses on fatigue crack growth thresh-
Fig. 10, large compressive residual stresses in CGHAZ significantly olds. Int. J. Fatigue 50, 10–17.
reduce Keff and make CGHAZ uneasy to be the initiation site. Davies, C.M., Ahn, J., Tsunori, M., Dye, D., Nikbin, K.M., 2015. The influence of pre-
existing deformation on GMA welding distortion in thin steel plates. J. Mater.
Eng. Perform. 24, 261–273.
5. Conclusions Dong, P., 2010. Fitness-for-service assessment of underwater welds in offshore
structures. In: International Workshop on the State of the Art Science and Reli-
ability of Underwater Welding and Inspection Technology, Houston.
On the basis of all the obtained results, the following conclusions Haagensen, P.J., Maddox, S.J., 2006. Recommendations on post weld improvement
can be drawn: of steel and aluminium structures. In: IIW Document XIII-1815-00.
Hobbacher, A., 2008. Recommendations for fatigue design of welded joints and
components. In: IIW, doc. XIII-2151r4-07/XV-1254r4-07.
1. The microstructure of the weld exhibits heterogeneous features. Huo, L., Wang, D., Zhang, Y., 2005. Investigation of the fatigue behaviour of the
The columnar microstructures in the weld seam contain large welded joints treated by TIG dressing and ultrasonic peening under variable-
volumes of proeutectoid ferrite and side-plate ferrite and small amplitude load. Int. J. Fatigue 27, 95–101.
ISO, 2013. Metallic materials—conversion of hardness values. In: ISO 18265: 2013
volumes of acicular ferrite. However, in CGHAZ, the microstruc- (E). ISO.
ture is dominated by lath martensite. Jia, C.B., Zhang, T., Maksimov, S.Y., Yuan, X., 2013. Spectroscopic analysis of the arc
2. The fatigue strength corresponding to 2 × 106 cycles of the as- plasma of underwater wet flux-cored arc welding. J. Mater. Process. Technol.
213, 1370–1377.
welded non-load-carrying cruciform joints is 135 MPa, which is Johnson, R., 1997. The Effect of Water Temperature on Underbead Cracking of Under-
increased by burr grinding, UUIT and grinding + UUIT by 20%, water Wet Weldments. Naval Postgraduate School, Monterey, Master’s Thesis.
35% and 61% to 162 MPa, 182 MPa and 217 MPa, respectively. Pessoa, E.C.P., Bracarense, A.Q., Zica, E.M., et al., 2006. Porosity variation along multi-
pass underwater wet welds and its influence on mechanical properties. J. Mater.
Burr grinding + UUIT can significantly improve the fatigue life of Process. Technol. 179, 239–243.
underwater wet welded joint by combining the strengths of the Perez, F., Liu, S., Smith, C., Rodriguez, E., 2003. Effect of nickel on toughness of under-
two techniques. water wet welds. In: 22nd International Conference on Offshore Mechanics and
Arctic Engineering, June 8–13, 2003, Mexico.
3. Compressive residual stress decreases the effective stress inten-
Rowe, M., Liu, S., 2001. Recent developments in underwater wet welding. Sci. Tech-
sity factor range Keff and the effective stress ratio Reff equals nol. Weld. Joining 6, 387–396.
to zero at R = 0.1, which significantly extends the fatigue life. The Rowe, M.D., Liu, S., Reynolds, T.J., 2002. The effect of ferro-alloy additions and depth
transverse residual stresses of as-welded joint in WM and HAZ on the quality of underwater wet welds. Weld. Res., 156–166.
Radaj, D., Sonsino, C.M., Fricke, W., 2006. Fatigue Assessment of Welded Joints by
are compressive due to the fast cooling effect of water and the Local Approaches, second ed. Woodhead Publishing, Cambridge, pp. 110–114.
former is less than the later. Santos, V.R., Monteiro, M.J., Rizzo, F.C., et al., 2012. Development of an oxyrutile
4. The fatigue cracks of the four types of specimens all initiate at the electrode for wet welding. Weld. J. 91, 319–328.
Tan, W., 2000. Mechanical Basis of Chemical Equipment, first ed. Tianjin University
weld metal close to the fusion line because of the difference of Press, Tianjin, pp. 183–184.
transverse compressive residual stresses between WM and HAZ. Yanagida, N., Koide, H., 2006. Effect of water-shower cooling during welding on ten-
sile residual stress improvement in multi-layer welding of austenitic stainless
steel plates. Mater. High Temp. 23, 251–262.
Acknowledgments Yildirim, H.C., Marquis, G.B., 2012. Fatigue strength improvement factors for high
strength steel welded joints treated by high frequency mechanical impact. Int.
The authors acknowledge the financial support of Marine Sci- J. Fatigue 44, 168–176.
Zhang, X.Q., Li, H., Yu, X.L., Zhou, Y., Duan, S.W., Li, S.Z., Huang, Z.L., Zuo, L.S., 2015.
entific Research Project (Investigation on the key technology of Investigation on effect of laser shock processing on fatigue crack initiation and
underwater welding and inspection processes-developing special its growth in aluminum alloy plate. Mater. Des. 65, 425–431.

You might also like