[Skip to content]

TWI
Search our Site
.

Testing the waters - shipbuilding laser welds face toughness trials. Part II, the findings...

Isabel Hadley
Isabel Hadley

TWI Bulletin, November/December 2000

After a degree from Cambridge University and a doctorate from Sheffield University, Isabel Hadley worked as a materials/structural integrity engineer in a range of industries, including nuclear power, offshore engineering and steel research. She joined TWI in 1992, working mainly on assessment of fitness-for-purpose of critical welded structures. Isabel is currently manager of the Defect Assessment and Reliability Engineering section of TWI's Structural Integrity Technology Group.

Laser welding has many attractive features, in particular the potential for high joining rate and low distortion. Consequently, as Isabel Hadley concludes here, there is considerable interest in this method in the shipbuilding industry, which has hitherto relied on more conventional arc welding techniques.


Part I, published in October, examined the scope and results of Dr Hadley's work.

Discussion

From the results of the Charpy tests shown previously, it is clear that the shift in behaviour from failure in the weld metal to failure by fracture path deviation (FPD) is related to several variables, including the inherent toughness of the weld, weld strength mismatch (M), weld width, crack tip acuity, side-grooving and specimen thickness. It is also clear that welds which cannot be tested using standard or modified Charpy tests (such as the narrow type N welds introduced in Part I of this work) may nevertheless have good toughness as measured by fracture mechanics tests, whereas welds which are readily tested (in this study the wider type W welds) may show relatively poor toughness. Avoidance of FPD during testing is thus not necessarily desirable if it is achieved at the cost of producing a weld with an inherently less tough microstructure such as the bainitic microstructure of weld W.

Hardening of the parent metal (especially in the 6mm thick plate, where full through-thickness hardening was possible) appears to be particularly effective in preventing FPD in type N welds. Use of side welds also shows potential, provided that the shape of the beads is optimal (as close as possible to a parallel bead) and their placement accurate, so that no untransformed material remains between the side welds and the main butt weld. Consequently the side-weld technique, whilst technically feasible, is not likely to be easily implemented in a shipyard test laboratory.

The large number of results on type N welds provided an opportunity to investigate the relationship between weld width, hardness and FPD in a systematic way, as shown in Fig.1. Each datapoint represents a best estimate of the temperature (T FPD) at which the behaviour changes from failure through the weld metal to failure by FPD. A full Charpy transition curve is required for each datapoint; error bars indicate the approximate uncertainty in T FPD. In some cases, the uncertainty is due to the gap between test temperatures (T 1<T 2), but in others it is due to overlap between the two failure modes (T 1>T 2, where T 1 is the lowest temperature associated with FPD, T 2 is the highest temperature at which failure occurs in the weld metal, and T FPD is the average of T 1 and T 2).

Fig.1. Effects of weld width, overmatch and microstructure on FPD
Fig.1. Effects of weld width, overmatch and microstructure on FPD

The trend towards increasing T FPD as the weld width is increased is clear from the results for the type N weld. Two results for a single-pass weld (M = 2.4) are shown, along with one result for a narrow weld with narrow side welds (total weld width 5.8mm) and another for a narrow weld with wide side welds (total weld width 9.0mm). A similar trend is apparent from the two results for type W welds (M = 1.7); widening of the central weld area by use of side welds raised T FPD from -30 to 80°C or higher.

The effect of overmatch ratio can be seen by comparing the three sets of data for type N welds. All results relate to nominally similar weld metal, with a weld width of 1-2mm and weld metal hardness above 300HV5. The result labelled M = 1.4 relates to a part-hardened 12mm thick parent material, subsequently welded using the same welding parameters as those used for the M = 2.4 specimens. However, the result clearly lies above the M = 2.4 curve. Similarly, the effects of overmatch can be seen by comparing the results of the as-welded sub-size specimen (M = 2.4, T FPD lessequal-196°C) with those of a similar specimen in which the material was first hardened (M = 0.8, T FPD approx 50°C).

Although it is convenient to avoid FPD for the purposes of routine testing, the question of the significance of FPD also needs to be addressed. For example, Sumpter (1999) has suggested that in a genuinely brittle laser weld, FPD will not occur and that, conversely, FPD is indicative of reasonable toughness regardless of the level of strength overmatch. Looking at Fig.2, it is apparent that in cases where FPD is avoided, for example by hardening of the parent material or use of side welds, the weld metal Charpy energy exceeds 47J at -20°C by a considerable margin. Since the weld metal is nominally the same as in the type N control weld, it can be concluded that in this case the Charpy energy of the type N control weld is also adequate. The modifications, by avoiding FPD, apparently allow the inherent toughness of the weld to be measured. The wider issue of whether FPD is in general indicative of good laser weld toughness was not specifically addressed, but was the subject of a separate investigation, results of which will be published later.

Fig.2. Charpy transition curve for N type welds
Fig.2. Charpy transition curve for N type welds

Further evidence of the good toughness of the type N welds, comes from the fracture mechanics tests, which show toughness of CTOD = 0.1mm even at as low as -100°C. Although FPD intervenes at higher temperatures, it seems fair to assume that the fracture toughness at temperatures above -100°C would continue to rise and would be satisfactory under service conditions. Sumpter and Caudrey (1995) have proposed a criterion, based on fracture mechanics arguments, for a minimum dynamic fracture toughness of K Jc = 125MPa rootm at 0°C for ship welds. This is equivalent to a CTOD requirement of 0.15mm and it appears highly likely that the type N welds would readily meet this requirement.

From the work described above, it is clear that the measurement of toughness in the laser welds (especially type N welds) proved to be very indirect, using methods not readily reproducible in a shipyard/test house environment. A more pragmatic approach to use of Charpy specimens for quality control might be to continue use of standard specimens but to test at a temperature below the intervention of FPD in all cases. If the relevant acceptance criteria (eg 47J at -20°C) are met and the absence of FPD confirmed, then the toughness of the weld may be considered satisfactory. The project has shown this is a feasible approach if the chemical composition of the plate material and welding parameters are appropriately controlled. Nevertheless, it is a significant departure from standard Charpy testing procedure, so Phase Two of the work (to be published later) will look at alternatives which do not rely on the avoidance of FPD.

Summary and conclusions

  • Two sets of autogenous laser butt welds were made in 12mm thick C-Mn steel plates, the composition of which was intended to be representative of steels used for laser-welded ships. These were designated type N (narrow-profile, madeat an upper-bound travel speed of 1 m/min.) and type W (wide-profile, made at a lower-bound travel speed of 0.5 m/min).
  • Both types of weld showed evidence of FPD when tested using a standard Charpy specimen notched at the weld centreline. The approximate temperatures of onset of FPD (T FPD) were: <-178°C for type N welds, and -30°C for type W welds.
  • Simple modifications to the Charpy test, such as use of side-grooving, use of a fatigue crack in place of a machined notch, and use of supporting side welds, caused an increase in T FPD. In the case of type N welds, the temperature of onset of FPD in modified specimens (up to -80°C) was still substantially below typical Charpy test temperatures for ship welds (around 0°C), makingthe modifications of limited use for routine Charpy testing. However, the modifications were adequate for prevention of FPD in type W welds.
  • FPD was also observed during fracture mechanics testing. The fracture toughness of type N welds which failed in the weld metal was nevertheless high (>0.1mm at -100°C) compared with suggested published criteria, and wascorrelated with a weld metal microstructure of fine autotempered martensite. Type W welds, which had a bainitic microstructure, showed FPD at higher temperatures, but poorer low-temperature fracture toughness than type N welds. It istherefore apparent that avoidance of FPD in test specimens is not necessarily desirable if it is achieved at the expense of producing a less tough weld metal.
  • The effects of the width of the weld zone and of weld metal hardness overmatch (M) on welds having nominally similar microstructures were systematically investigated by the use of supporting side welds and by hardening the parentmaterial to produce low mismatch welds. Increasing weld width and decreasing mismatch were both associated with an increase in T FPD. The modifications, by avoiding FPD, apparently allow the inherent toughness of the weld to be measured. The generality of this approach was not, however, explored.
  • The most pragmatic approach to routine shipyard testing of laser welds may be to test at the highest temperature possible before intervention of FPD (or at the usual test temperature, whichever is the lower). Two criteria wouldthen be applied: the usual energy criterion and a check that fracture occurs in the weld metal only.

Acknowledgements

This work was carried out under contracts with the Ministry of Defence and Lloyd's Register, whose permission to publish is gratefully acknowledged.

Further reading

ASTM, 1998: 'Draft weld fracture test standard (proposed Annex to E1290)', ASTM Rev 6, August 1998.

British Standards Institution (BSI), 1997: 'BS 7448: Fracture mechanics toughness tests. Part 2. Method for determination of K Ic, critical CTOD and critical J values of weld in metallic materials'.

Goldak J A and Nguyen D S, 1977: 'A fundamental difficulty in Charpy V-notch testing narrow zones in welds', Welding Research Supplement April 1977 119-125s.

Hayes B, Norris I M and Towers O L, 1986: 'Preliminary investigation of a modified Charpy test for the assessment of narrow welds', TWI Members' report 308, July 1986

Kristensen J K and Borggreen K, 1996: 'Evaluation of laser welds in structural steels' International Journal for the Joining of Materials 1996 8 (1), 48-54.

Lloyd's Register, 1997: 'Guidelines for approval of CO 2-laser welding', LR Marine Division, March 1997.

Misawa T, Takasa S, Nakano Y and Yasuda K, 1996: 'Ductile-brittle transition evaluation of laser welded steel metal by means of small specimen impact test' Steel and Copper 82 (8) 1996 (in Japanese).

Sumpter J D G, 1999: 'Fracture toughness evaluation of laser welds in ship steels' European Symposium of Assessment of Power Beam Welds (ASPOW), GKSS, Geesthacht, Germany, 4-5 Feb 1999.

Sumpter J D G and Caudrey A J, 1995: 'Recommended fracture toughness for ship hull steel and weld' Journal of Marine Structures 8 1995, 345-357

TWI, 1996: International conference 'Exploitation of laser processing in shipyards and structural steelwork' Glasgow, May 30-31, 1996.

See also Charpy testing laser welds - the significance of fracture path deviation