Subscribe to our newsletter to receive the latest news and events from TWI:

Subscribe >
Skip to content

Optimisation of SENT test specimen design

   
Phillipa Moore

TWI Ltd, Granta Park Great Abington, Cambridge, CB21 6AL, UK

Paper presented at Eurojoin 9. 20-22 May 2015. Bergen, Norway.

1. Abstract

As the implementation of Single Edge Notch Tension (SENT) tests extends beyond the J R-curves for pipeline girth welds described in DNV RP F108, and into a new British Standard for SENT testing, BS 8571, it is necessary to identify the optimum specimen design for a general SENT test. In this paper, as well as reviewing recent literature on SENT testing, a series of experimental SENT tests are described, which were carried out to evaluate the effect of different specimen parameters on the R-curves produced. The findings show that SENT specimens of BxB cross section design are becoming more widely used in addition to 2BxB, and that this specimen design gives higher R-curves. It is also feasible to use both surface and through-thickness notching for these specimens. However, when using the unloading compliance method for R-curves, a side-groove depth of 5% each side gives the best balance of straight crack shape without penalising the R-curve toughness.

This paper also comments on the use of threaded-ends for SENT specimens, and the ease of testing with this specimen design. An SENT specimen with threaded ends can be tested within a temperature-controlled chamber using established tensile testing grips for threaded-end specimens. When testing at particularly low temperatures, this avoids the risk of yielding in the specimen arms, and improves the temperature control compared to using locally-applied cooling around the notch. A method to friction weld threaded ends onto SENT specimens is described.

2. Introduction

The single edge notched tension (SENT) fracture toughness test specimen is rapidly gaining acceptance and is being more widely used as a substitute for the deeply-notched single edge notched bend (SENB) specimen for assessing the integrity of pipeline girth welds, and other steel structures. Generally, SENT tests result in higher fracture toughness being measured than with equivalent SENB tests, since the tensile loading mode gives lower constraint at the crack tip. The lower constraint means that a higher fracture toughness may be measured. Nonetheless, research has shown that the constraint in an SENT specimen is closer to, but still conservative in relation to that in a pipe girth weld. This means that the fracture toughness measured from SENT tests, in the same way as SENB tests, still conservatively represents the fracture toughness of such welds in service. As SENT tests are being carried out more often, it is important to standardise the fracture toughness testing of SENT specimens, in particular for specimens notched into the weld metal and heat affected zone of girth welds, which often exhibit the lowest fracture toughness in a steel fabrication.

A test standard for SENT testing has recently been issued by British Standards as BS 8571:2014 [1], and the work presented in this paper has been carried out to support the recommendations made to BSI for the development of this standard. BS 8571 is intended to be a general test standard, equivalent to established SENB test standards, and therefore a wide set of test conditions needed to be tested and evaluated. This means that fracture toughness in terms of crack tip opening displacement (CTOD) and well as J-integral will be calculated, that single-point fracture toughness as well as R-curves can be determined, and that validity limits are given to confirm whether the test has been carried out satisfactorily.

Prior to BS 8571, most guidance on SENT test methodology was taken from DNV RP F108:2006 [2], which provides a relatively limited set of test conditions; only multiple-specimen J R-curves. Other SENT testing methods from literature describe methods to generate J and CTOD R-curves using unloading compliance techniques [3-5] and the direct-current potential drop technique [6]. These procedures have also been reviewed to compare the unloading compliance functions [7], and one of the unloading-compliance methods was evaluated through a round-robin [8]. Published research on the practical issues of performing SENT tests has also discussed the instrumentation of the crack mouth [9, 10].

This paper systematically reviews other practical aspects of SENT specimens design. A series of SENT tests has been performed to investigate the effect of various specimen design parameters on the fracture toughness obtained from the tests, in order to identify the optimum ‘standard’ specimen designs for BS 8571.

3. Optimisation Test Programme

Several aspects of SENT specimen design were reviewed in this work. Initially a literature review of recent research in SENT testing was carried out to identify which aspects of SENT specimen design were well established, evaluated and reported in the public domain. It was then possible to identify those areas which required further validation, and it was these which were investigated in the experimental programme subsequently carried out.

The aspects of most interest for research were:

  • The effect of notch depth on SENT R-curves
  • The difference between 2BxB and BxB geometry
  • The use of through-thickness notched specimens as well as surface notched specimens
  • The optimum side groove depth for unloading compliance R-curve tests
  • The need for local compression for weld specimens
  • The potential to use threaded-ends for temperature-controlled tests

The material used for the majority of the mechanical testing was a multi-pass TIG girth weld in 22in OD 19.1mm wall thickness pipe, welded in the 5G position. The weld macrosection is shown in Figure 1. The parent pipe had a yield strength of 568MPa and UTS of 615MPa. The weld metal had a yield strength of 543MPa and UTS of 652MPa. The SENT tests were all carried out using clamped ends, and fracture toughness in terms of J was calculated using the equations given in DNV RP F108 [2], while crack tip opening displacement, CTOD, was calculated from a double clip gauge [11]. When R-curves have been generated the values of CTOD and J have not been corrected for crack extension.

Figure 1 Macro section of the multi-pass TIG girth weld in 22in OD 19.1mm wall thickness pipe.
Figure 1 Macro section of the multi-pass TIG girth weld in 22in OD 19.1mm wall thickness pipe.

4. Effect of notch depth

The notch depth of an SENT specimen, expressed as the ratio of the crack length, a to the specimens width, W, can have a significant effect on the fracture toughness obtained. A shallower notched specimen has a lower constraint than a deeply notched specimen. Therefore, SENT specimens with a shallower notch depth (a/W) ratio often give a higher SENT R-curve than more deeply notched specimens, particularly once the crack extension exceeds 0.5mm. This trend is well reported in a number of papers where specimens with a/W of 0.2 to 0.3 were compared to those with a/W of 0.4 to 0.5 [12-14].

For pipeline girth weld assessments, SENT R-curves are generated to assess the significance of girth weld fusion flaws in pipes, and shallower notched SENT specimens (with a/W of 0.25 to 0.35) more closely matches the constraint of a typical flaw in a pipe during installation. The length of the crack in the SENT specimen is also sometimes intended to be a similar actual size to the flaws in the pipeline welds being assessed. Tang et al [12] recommend that for pipelines, a notch depth (a/W) ratio should be equal to the a/t ratio (largest permissible flaw divided by pipe wall thickness) plus 0.1, while Tan et al [15] consider that a/W of 0.35 in a SENT specimen can essentially match that of all pipelines with axial cracks. DNV RP F108 [2] permits a/W ratios of 0.2 to 0.5 for SENT specimens, while Cravero and Ruggieri [3] offer solutions for unloading compliance of SENT specimens valid for a/W ratios between 0.1 and 0.7.

Xu et al [16] show by using numerical modelling that the R-curves from deeper notch SENTs with a/W 0.5 are more conservative in relation to the equivalent cracked pipe R-curve than SENT specimens with a/W of 0.3. This supports the conclusion that using more deeply notched SENT specimens could be overly conservative.

The findings from the literature indicate that using a notch depth a/W of 0.35 to 0.5 would be the preferred SENT specimen configuration to generate consistent and conservative data, where a justification was not available to use a shallower notch depth for a particular case. Based on the findings of this literature review, it was not considered necessary to further investigate experimentally the effect of notch depth in SENT specimens. When the a/W ratio is less than 0.3 the R-curve generated is significantly higher than for a deeper notch depth, therefore, for the experimental work reported in this paper a/W of 0.4 was used for all the SENT tests.

5. Specimen dimensions

5.1 Origins of SENT specimen design

SENT specimen cross section dimensions are expressed in terms of the specimen thickness x width, BxW. For a surface notched SENT specimen, W is the material thickness, see Figure 2. The dimension ‘B’ is usually equal to W or 2W (often referred to as BxB or 2BxB specimens respectively). DNV RP F108 [2] recommends that 2BxB (over-square) plain-sided SENT specimens should be used to generate a multiple specimen J R-curve, but permits size reduction to BxB to maintain sufficient specimen thickness from pipe with significant curvature. Tang et al [12] compared 2BxB SENT specimens with BxB specimens that had side-grooves. The results demonstrated that toughness measurements obtained using a BxB configuration with side grooves were similar to those using a 2BxB configuration without side grooves; however, the specimens with side grooves and BxB geometry had the benefit that they facilitated even crack growth along the crack front.

BxB specimens are described more often than 2BxB specimens in recent literature; BxB specimens were the test configuration in [4, 12, 13, 17, 18], whereas 2BxB specimens were referred to in [19, 20]. Cravero & Ruggieri [3] were alone in using a Bx2B SENT specimen for their modelling and validation work, but made no explanation for their choice. The literature suggested that there is an increasing interest in using BxB SENT specimens, with potential benefits, but that more validation would be needed to give sufficient confidence for including a range of permitted SENT specimen designs in BS 8571.

5.2 Experimental 2BxB and BxB R-curves

Triplicate sets of SENT tests were carried out on specimens using the unloading compliance method at room temperature. The method implemented the compliance functions from [3] with a rotation correction factor from [4]. The first three tests used SENT specimens with an over-square (2BxB) cross section with a ‘width, W’ equal to 14mm and ‘thickness, B’ equal to 28mm. These were all surface notched into the weld centreline to an a/W ratio of 0.4. This is the SENT specimen design given in DNV RP F108 to generate multiple-specimen J R-curves. The ‘DNV’ specimen does not have side grooves (considered necessary only when the unloading compliance method is used to generate an R-curve rather than multiple specimens). Therefore two of these three specimens were tested using the unloading compliance method without side grooves to be equivalent to this ‘DNV’ specimen design. However, the literature review showed that when unloading compliance is used with SENT specimens side grooves are recommended, and so one of these first three specimens was side grooved to 5% depth each side. These first three 2BxB R-curves were then compared with a further set of three BxB SENT specimens, which were also surface notched into weld metal with a/W of 0.4, and all three were side grooved to 5% depth each side. When side grooves were used, the equations to calculate J and compliance were based on the effective thickness, Beff, which is calculated from the original thickness, B and the net section thickness after side gooving, BN, by square-rooting their product; i.e. Beff = √(BxBN).

Figure 2 SENT test specimen with a BxB square section design (a), and an SENT test specimen instrumented with a double clip gauge, tested at room temperature to generate a CTOD and J R-curve (b).
Figure 2 SENT test specimen with a BxB square section design (a), and an SENT test specimen instrumented with a double clip gauge, tested at room temperature to generate a CTOD and J R-curve (b).

The R-curves for this comparison are shown in Figure 3, where the 2BxB SENT tests are identified as W01-01 to W01-03 (W01-03 was the side-grooved specimen), and the BxB specimens are identified as W01-04 to -06. The J R-curves from the 2BxB specimens give consistently lower R-curves than the BxB specimens, even when they are not side grooved. The 2BxB specimen that had side grooves showed the lowest of this set of R-curves, as would be expected, due to side grooves increasing the constraint as discussed later in this paper. The CTOD R-curves show less effect of specimen type on the R-curves produced with side-grooved BxB specimens and plain sided 2BxB specimens giving similar R-curves. The side-grooved 2BxB specimen W01-03 was the lowest of this set.

Later in the test programme, data was generated on side-grooved 2BxB specimens surface notched into the parent metal and into the HAZ, and also on side-grooved BxB specimens which had been through-thickness notched into parent metal and into the HAZ, which allowed the specimen designs to be compared further. These specimens were W01-10 to -21. The reason why surface and through-thickness notched specimens could be considered as equivalent was also evaluated in this work; this is discussed in the next section.

The results of the parent metal and HAZ tests again showed the same trend, with the BxB specimens showing higher R-curves than 2BxB specimens is seen again (Figures 4 and 5). The HAZ notched specimens for both specimen designs showed more scatter within the three results, reflecting the narrow region being sampled and variety of microstructures in the HAZ, compared to parent and weld metals.

Figure 3 Room temperature J (a) and CTOD (b) R-curves for surface notched SENT specimens notched into the weld metal, for specimens of 2BxB design and BxB design. All the specimens were side grooved by 5% each side, except specimen W01-01 and W01-02.
Figure 3 Room temperature J (a) and CTOD (b) R-curves for surface notched SENT specimens notched into the weld metal, for specimens of 2BxB design and BxB design. All the specimens were side grooved by 5% each side, except specimen W01-01 and W01-02.
Figure 4 Room temperature J R-curves for surface notched SENT specimens notched into parent metal, for specimens of 2BxB design and BxB design.
Figure 4 Room temperature J R-curves for surface notched SENT specimens notched into parent metal, for specimens of 2BxB design and BxB design.
Figure 5 Room temperature J R-curves for surface notched SENT specimens notched into the heat affected zone (HAZ), for specimens of 2BxB design and BxB design.
Figure 5 Room temperature J R-curves for surface notched SENT specimens notched into the heat affected zone (HAZ), for specimens of 2BxB design and BxB design.

6. Notch orientation

Tang et al [12] reported on a parametric study of SENT notch orientations, looking at weld centreline root and cap surface notches, along with through-thickness notches. They found that through-thickness notches could result in uneven stable crack growth across the notch front through the weld, but the R-curves measured were very similar to those notched from the outside diameter (OD) surface. The internal (ID) notches gave slightly higher R-curves than specimens with all the other notch orientations. Tang et al [12] show that through thickness notches or external surface notches give similar R curves for the same welds.

Since only a limited number of references discussing SENT notch orientation were found in the published literature, this was the second area of experimental validation which was carried out in this programme of work. A set of three specimens were tested, again BxB specimens notched into the weld metal (also a/W 0.4 and side grooved the same as specimens W01-04 to -06), but this time through-thickness notched instead of surface notched, and these specimens identified as W01-07 to -09. These R-curves were compared to the results from W01-04 to -06 to determine whether there was a difference between surface and through-thickness notched specimens in the weld metal, for the same BxB specimen design. The surface notched specimen is intended to sample the same weld bead all across the whole notch front, but there is likely to be more scatter between nominally similar tests as the precise location of the notch tip will vary within a set of specimens. In contrast, a through-thickness notch will sample every weld bead through the weld thickness, which means it is sampling a variety of microstructures, but each test within a set will be testing a similar set of microstructures, and hence might be expected to show less scatter between the fracture toughness tests. This effect can be seen to some extent in Figure 6.

In this work, for the same material being tested there was no significant difference in the R-curves obtained from surface notches or through-thickness notches, as seen in Figure 6. This weld had been fabricated using the same welding process throughout (as shown in the macro in Figure 1), so the root would not show significantly different properties to the fill passes in this case.

Figure 6 Room temperature J R-curves for BxB SENT specimens notched into the weld metal with both surface notches and through thickness notches.
Figure 6 Room temperature J R-curves for BxB SENT specimens notched into the weld metal with both surface notches and through thickness notches.

7. Notch Sharpness

The requirement for fatigue pre-cracked notches in SENT specimens, as opposed to simply using electro-discharge machined (EDM) notches, was not investigated in this work, because it was the subject of a separate study, which was published in 2013 [21]. The conclusions were that R-curves generated using fatigue pre-cracked SENTs showed slightly lower R-curves than when using EDM-notched SENTs, but that both R-curves converged after a level of tearing greater than around 1mm. Although there was little difference between the fracture toughness determined at the maximum load for EDM and fatigue precracked specimens, at temperatures low enough for the fatigue pre-cracked specimens to start to show transition behaviour and low fracture toughness, the EDM-notched specimens were still showing fully ductile behaviour, resulting in significant difference in the fracture toughness values measured (Figure 7).

It was recommended that EDM-notched SENT specimens should only be used for the assessment of fracture toughness determined from the maximum load in the load-displacement curve, i.e. upper shelf behaviour. If it cannot be shown that the material is on the upper shelf then precracking must be used for SENT specimens to avoid over-estimating the fracture toughness.

For a general SENT fracture toughness testing standard it cannot be presumed that the fracture behaviour will already be known, and fatigue pre-cracked specimens are required in BS 8571. All the SENT tests discussed in this paper are therefore on fatigue pre-cracked specimens.

Figure 7 Fracture behaviour of EDM-notched and fatigue pre-cracked SENT specimens over a range of temperatures [21].
Figure 7 Fracture behaviour of EDM-notched and fatigue pre-cracked SENT specimens over a range of temperatures [21].

8. Side Grooves

8.1 Side grooves in SENT specimens

Much of the published research on SENT testing has concentrated on ductile materials and the generation of tearing resistance curves. Several alternative methods for measuring unloading compliance in SENT specimens have been developed by teams in Brazil, Canada and the USA [3, 4, 5, 7]. When the unloading compliance technique is used in SENB specimens it is standard to machine grooves along the sides of the specimen in the plane of the crack propagation to locally increase the constraint, in order to prevent the formation of shear lips and ensure that stable ductile tearing proceeds in a straight and flat manner. For SENB specimens it is common to side-groove to a depth of 10% of the specimen thickness on each side.

Cravero and Ruggieri [3] used side grooved specimens with a groove depth of 10% of the specimen thickness each side for their unloading compliance test method. However Shen et al [4] used 5% each side as the groove depth. The 5% side groove each side was also successfully used by Park et al [14] and Tang at al [12] to give even crack growth in R-curve tests, and is adopted in the ExxonMobil procedure [5].

Shen et al [22] also published a numerical analysis of the effect of side grooves on clamped SENT specimens. The depth of the SENT specimen side grooves affects the crack-tip constraint, which is highest at the centre of the thickness for a plain sided specimen and near the root of the side grooves for side grooves equal to or greater than 10% total depth [22]. The highest J integral for side grooves equal to or less than 10% total depth (ie 5% each side) is at the centre of the specimen, but at the root of the side groove for side grooves 20% total depth (10% each side). Shen et al [22] concluded that the J-resistance of a SENT specimen with 20% side grooves may be overly conservative and recommended that the optimum depth for side grooves lies between 10% and 20% total depth (between 5 and 10% depth each side). As a result of this, in subsequent papers a side groove depth of 7.5% each side has been preferred by Canadian researchers [13, 23, 24], and adopted by researchers from the University of Gent [6].

8.2 Influence on the R-curve

The effect of side groove depth was also investigated experimentally in this work to verify the conclusions from [22], in order to make confident recommendations on SENT side-grooving for the British Standard BS 8571. The first specimens tested in this work (in Figures 3, 4 and 5) had been side grooved to 5% depth each side as the default value chosen based on the literature.

But a new set of tests included specimens side-grooved to different depth levels, and again using both surface notched specimens and through-thickness notched specimens. Six further BxB SENT specimens notched to a/W of 0.4 into the weld metal were tested; two were left plain sided (i.e. 0% side grooves), one was side grooved to 3% depth each side, two to 7% and one to 10% each side. This set of tests was repeated for both surface notched and though-thickness notched specimens. The results were then be compared to the equivalent data for 5% side grooves from earlier tests. All these differently side-grooved R-curves are compared for surface notched specimens in Figure 8 and for through-thickness notched specimens in Figure 9.

Figure 8 J R-curves from surface notched BxB SENT specimens with side grooves between 0% and 10% of the specimen depth each side.
Figure 8 J R-curves from surface notched BxB SENT specimens with side grooves between 0% and 10% of the specimen depth each side.
Figure 9 J R-curves from through-thickness notched BxB SENT specimens with side grooves between 0% and 10% of the specimen depth each side.
Figure 9 J R-curves from through-thickness notched BxB SENT specimens with side grooves between 0% and 10% of the specimen depth each side.

For both notch orientations, the specimens notched with side grooves up to 5% each side tended to group together, and then the 7.5% deep side grooves each side were significantly lower than this set of R-curves, with 10% side grooves each side giving a lower R-curve again. The results from these tests suggest that there is no effect on the R-curve fracture toughness obtained when using side grooves up to 5% depth each side, as a consequence of the higher constraint of the side groove reducing the fracture toughness. Side grooves greater than 5% will reduce the R-curve that is generated from the test.

8.3 Influence on the crack shape

The intention of side-grooving is to ensure straight stable tearing propagation, so it was also necessary to quantify the shape of the stable tearing on the fracture faces from these tests. As a consequence of side-grooving, the original fatigue pre-crack shape is sometimes improved in terms of crack curvature as well, by machining away the most curved part of the pre-crack at the outer surfaces.

To evaluate the effect of side grooves on the pre-crack straightness, and the straightness of the stable tearing in the specimen, the nine crack length measurements made across the crack front were used to determine the average initial precrack length, a0, [25]. The final crack length after tearing, af, was determined in the same way. None of the nine individual measurements of SENT pre-crack length should differ from the average by more than 20% of the average [20]. Likewise, the difference between the average final crack length measurements at any of the nine points was compared to a target difference of no more than 20% of the average.

In this analysis, instead of simply quoting the magnitude of the difference between the average and the maximum or minimum, it was recorded whether the largest difference was positive or negative. The greatest deviation from the average crack length is usually the measurement closest to the outside surfaces of the specimen, furthest from the centre of the pre-crack. If the largest deviation from the average pre-crack length is negative, it suggests that the edges of the pre-crack bow back behind the pre-crack. If the largest deviation is positive, it suggests that the pre-crack shape is re-curved and that the outer edges of the pre-crack are ahead of the average crack length. By keeping the positive or negative sign associated with the deviation in crack measurements from the average, this provides some information about the crack shape which could be also plotted against the depth of the side groove each side to determine whether there was a correlation. This is shown in Figure 10 for surface notched specimens and Figure 11 for through-thickness notched specimens.

Figure 10 The deviation of the maximum crack measurement across the crack front from the average (the crack curvature) expressed as a percentage of the average crack length, for the fatigue pre-crack and the final crack including stable tearing, for
Figure 10 The deviation of the maximum crack measurement across the crack front from the average (the crack curvature) expressed as a percentage of the average crack length, for the fatigue pre-crack and the final crack including stable tearing, for surface notched specimens.

For the initial crack precrack shape, the difference between the average and minimum/maximum values was consistently negative (backward bowing), and showed a weak improvement with side groove depth. The precracking is performed before side grooving, so deeper side grooves machine off more of the most backward curving part of the precrack at the outer surfaces, making the shape of the remaining precrack as being measured straighter.

For the final tearing crack shape, the difference between the average and minimum/maximum values of was strongly affected by the depth of the side grooves for both surface and through-thickness notched specimens. Without side grooving the difference was negative, similar to the precrack difference. However, by 3% side grooving some values were positive and some negative (depending on notch orientation), and at 5% and above the difference was positive, increasing with side groove depth. Based on the correlations in Figures 10 and 11, the difference exceeded the target of 20% at a side groove depth of 7%, although some 5% side grooved specimens were beyond the 20% limit as well.

The experimental results from this work confirm the recommendation that the side groove depth for SENT specimens should be 5% each side when generating unloading compliance R-curves.

Figure 11 The deviation of the maximum crack measurement across the crack front from the average (the crack curvature) expressed as a percentage of the average crack length, for the fatigue pre-crack and the final crack including stable tearing, for
Figure 11 The deviation of the maximum crack measurement across the crack front from the average (the crack curvature) expressed as a percentage of the average crack length, for the fatigue pre-crack and the final crack including stable tearing, for through-thickness notched specimens.

9. Local compression

The fracture face measurements plotted in Figures 10 and 11 were produced without local compression before fatigue pre-cracking. For weld and HAZ fracture toughness SENB test specimens it is common to apply local compression to relieve some of the residual stresses in the notch location prior to fatigue pre-cracking [25]. This helps ensure a sufficiently straight pre-crack shape. It is anticipated that this technique will give the same benefit in SENT specimens as for SENB specimens. Although SENB specimens are usually compressed to a total of 1% specimen thickness ahead of the notch, for SENT specimens this would cause excessive bending in the specimen (which needs to remain straight for testing). Therefore a total local compression of 0.5% of specimen thickness was applied to the SENT specimens in this work.

Although the fatigue pre-cracks in the 14mm thick material did not appear to be excessively bowed as a consequence of not having had local compression applied, some additional tests were carried out which had been locally compressed before fatigue pre-cracking so show any influence of the technique on the test.

The effect of local compression on the crack curvature of the specimens is shown in Figure 12. The scatter in the curvature of the fatigue pre-crack shapes is significantly reduced when local compression is used, as shown in Figure 12, compared to equivalent specimens without local compression in Figure 11. In the specimens tested in this work, all the fatigue pre-crack shapes were considered to be valid [20] regardless of local compression, and perhaps a stronger effect of local compression would be seen when testing thicker specimens, or specimens which exhibited lower shelf fracture behaviour.

Figure 12 The deviation of the maximum crack measurement across the crack front from the average (the crack curvature) expressed as a percentage of the average crack length, for the fatigue pre-crack and the final crack including stable tearing, for
Figure 12 The deviation of the maximum crack measurement across the crack front from the average (the crack curvature) expressed as a percentage of the average crack length, for the fatigue pre-crack and the final crack including stable tearing, for through-thickness notched specimens using local compression.
Figure 13 R-curves generated from BxB SENT specimens through-thickness notched into the weld metal for specimens using both local compression and without local compression before pre-cracking.
Figure 13 R-curves generated from BxB SENT specimens through-thickness notched into the weld metal for specimens using both local compression and without local compression before pre-cracking.

Local compression seemed to have a detrimental effect on the shape validity of the stable tearing in the specimens. When comparing the tearing shapes in Figures 10 and 11, a 5% side groove each side has about 20% curvature after stable tearing without local compression being applied before pre-cracking. However, an equivalent specimen which has been locally compressed has a stable crack curvature of around 30%. This shows that caution should be applied in the use of local compression, especially where it may not be necessary for the fatigue pre-crack shape such as in surface notched specimens, on specimens less than 15mm thick (the results presented here are for specimens 14mm square).

The R-curves generated for equivalent specimens (through-thickness notched weld metal test specimens without side grooves) prepared both with and without local compression having been applied are shown in Figure 13. When these R-curves were compared there was no noticeable difference in the tearing resistance behaviour, within the scatter of the data. This means that local compression can be used when preparing SENT specimens on welded material without any significant influence on the fracture toughness determined, while improving the fatigue pre-crack shape.

10. Threaded ends

All of the SENT tests reported in the previous sections used straight sided specimens which were clamped in testing jaws as shown in Figure 2b. This was a straightforward testing configuration for room temperature tests, as these had been. However, when SENT tests are to be carried out at temperatures significantly higher or lower than room temperature, in this configuration the cooling or heating medium must be applied locally around the notch since it is not possible to apply simple clamped loading inside a temperature controlled chamber. When testing a specimen with locally applied heating or cooling, the tensile properties will differ along the specimen, and the risk of yielding in the arms of the specimen (for low temperature tests) becomes significant. This would mask the true behaviour of the test specimen at the test temperature.

An alternative is to use threaded ends beyond the SENT specimen length (equal of ten times the width for clamped specimens). The threaded ends can fit inside a standard tensile testing temperature chamber and can be tested at any temperature with ease. Although it is understood that threaded ends are often machined onto SENT specimens is some test laboratories, there is no published literature about this technique. Therefore a feasibility trial was carried out to weld round threaded ends onto square section SENT specimens to consider whether threaded-end SENT specimens could be made that way.

These specimens were made by welding round bar onto the ends of the shorter SENT specimen using a Continuous Drive Rotary Friction Welding (CDRFW) procedure. The round bar was then machined with a thread so it fitted into the appropriate tensile testing grips. The material used for these SENT tests was a 14.3in OD and 24mm wall thickness pipe welded in the 5G position using a mechanised GMAW technique, and was identified as ‘W02’ to distinguish it from the weld used for the previous tests. It had a parent metal yield strength of 504MPa and UTS of 598MPa. The SENT specimens were of size 22mm x 22mm in cross section, and a bespoke holder for the friction welding was made for these specimens by machining a square hole of 22 x 22mm in the middle of a 57mm diameter round steel bar. The holder was about 70mm long and enabled the specimens to be held in the CDRFW machine (Figure 14).

Figure 14 Set-up for friction welding round bar onto the end of square section SENT blanks, which were held in a custom-made holder for the specimen dimensions.
Figure 14 Set-up for friction welding round bar onto the end of square section SENT blanks, which were held in a custom-made holder for the specimen dimensions.

The SENT square centre section had been cut to a length of 230mm, in order to achieve a specimen length close to 220mm (H = 10W) between the threaded ends after manufacture. The square sample was clamped in the non-rotating holder on the moving head, whilst the round extension bars were rotated, via the 3-Jaw chuck. The round bar was 30mm diameter and a 75mm length was welded to each end of the SENT specimen. The challenge was to find material for the round bar which was sufficiently strong to match the notched bar of material being tested without yielding in the grips, while still being weldable. For these reasons, neither mild steel, nor tool steel (common materials available as round bar products) were suitable. A structural steel grade S355J2 was used which had a yield strength of 345MPa minimum, and UTS of 470MPa minimum.

The 11kW friction welding machine provided a spindle rotation speed of 1460 rev/min. Friction welding parameters were selected from experience to produce good joint quality coupled with minimum heat affected zones to reduce any influence on the test sample. The friction force was 600kN and the forging force was 1000kN. This resulted in an axial burn-off of 4mm on each end of the specimen. The specimen at this stage are shown in Figure 15.

Figure 15 Round bar friction welded onto each end of the SENT square section blanks, prior to notching of the specimen and machining threaded ends.
Figure 15 Round bar friction welded onto each end of the SENT square section blanks, prior to notching of the specimen and machining threaded ends.

The round bar ends were then machined with an M30 thread with a 2mm pitch and the SENT specimens were then notched and pre-cracked in the usual way (Figure 16). The nine specimens produced in this way were tested at low temperature satisfactorily, and this design was found to be very easy to test. Being able to soak other specimens in the environmental chamber while carrying out a test sped up the overall testing duration. A macro-section through one of the friction welded ends shows an example of the good fusion quality of the weld produced (Figure 17).

Figure 16 Threaded end SENT specimens after notching and pre-cracking ready to test inside the temperature controlled chamber.
Figure 16 Threaded end SENT specimens after notching and pre-cracking ready to test inside the temperature controlled chamber.
Figure 17 Macro section through a friction weld used to attach threaded ends to some of the SENT specimens.
Figure 17 Macro section through a friction weld used to attach threaded ends to some of the SENT specimens.

Although this aspect of SENT specimen design is only included here at the feasibility stage, it would be useful to perform further research on the aspects of threaded end design, with consideration of other methods to generate threaded end specimens, strength matching of the threaded end and the rest of the specimen, limitations on notch depth and other parameters.

11. Summary and Conclusions

Various aspects of the experimental design of SENT specimens were considered in this paper, based on published literature and test data. The experimental work presented is intended to consolidate, validate and extend the current research being undertaken on aspects of SENT specimen design. The results and recommendations will be put to the British Standards committee responsible for BS 8571 for inclusion into future versions of that standard.

The findings show that SENT specimens of BxB cross section design are an acceptable alternative to 2BxB, and that this specimen design gives higher R-curves in the same material. It is also feasible to use both surface and through-thickness notching for BxB specimens.

When using the unloading compliance method for R-curves, a side-groove depth of 5% each side gives the best balance of straight crack shape without penalising the R-curve fracture toughness.

This paper also described a method for using threaded ends instead of simple clamped ends on SENT specimens, particularly if testing at high or low temperatures.

12. Acknowledgements

The work presented here was carried out as part of a joint-industry project funded by Subsea7, Saipem, BP Ltd and TWI Ltd. Thanks to the sponsors of this work for their support throughout the project, and their permission to publish.

Thanks are also given to Phillip Cossey and Peter Sketchley who made significant contributions to generating the experimental results presented here.

13. References

  1.  ‘Method of test for determination of fracture toughness in metallic materials using single edge notched tension (SENT) specimens’, British Standards Institution, December 2014.
  2. DNV ‘Fracture control for pipeline installation methods introducing plastic strain’, Det Norske Veritas Recommended Practice DNV-RP-F108, 2006.
  3. Cravero S and Ruggieri C, ‘Estimation procedure of J-resistance curves for SE(T) fracture specimens using unloading compliance’, Engineering Fracture Mechanics 74 (2007) 2735-2757.
  4. Shen G, Gianetto J, Tyson W, ‘Measurement of J-R Curves using single specimen technique on clamped SE(T) specimens’, in Proceedings of the Nineteenth International Offshore and Polar Engineering Conference, ISOPE2009, Osaka, Japan, June 21-26, 2009.
  5. ExxonMobil, ‘Measurement of Crack-Tip Opening Displacement (CTOD) fracture resistance curves using single-edge notched tension (SENT) specimens’, ExxonMobil Upstream Research Company, September 2010.
  6. Verstraete M, Hertelé S, Van Minnebruggen K, Denys R & De Waele W, ‘UGent guidelines for SENT testing’, in Proceedings of the Pipeline Technology Conference, Ostend, Belgium, 7-9 October 2013.
  7. Zhu X-K, McGaughy T, Orth F & Jennings J, ‘Review of Low-Constraint Fracture Resistance Testing with SENT specimens’, in Proceedings of the 33rd International Conference on Ocean, Offshore & Arctic Engineering, OMAE2014, San Francisco, USA, June 8-13 2014.
  8. Tyson W, Gianetto J, ‘Low-Constraint Toughness Testing: Results of a Round Robin on a Draft SE(T) Test Procedure’, in Proceedings of the ASME 2013 Pressure Vessels and Piping Conference PVP2013, Paris, France, July 14-18 2013.
  9. Weeks T, McColskey J, Read D & Richards M, ‘Fracture Toughness Instrumentation Techniques for Single-Specimen Clamped SE(T) tests on X100 Line Pipe Steel: Experimental Setup’, in Proceedings of the Pipeline Technology Conference, Ostend, Belgium, 7-9 October 2013.
  10. Park D-Y, Gravel J-P, Simha C, Liang J & Duan D-M, ‘Low-Constraint Toughness Testing of Two SE(T) Methods in a Single Specimen’, in Proceedings of the ASME 2014 Pressure Vessels and Piping Conference PVP2014, Anaheim, USA, July 20-24 2014.
  11. Moore P & Pisarski H, ‘Validation of Methods to Determine CTOD from SENT specimens’, in Proceedings ISOPE-2012, The 22nd International Offshore (Ocean) and Polar Engineering Conference, Rhodes, Greece, June 17−22, 2012.
  12. Tang H, Macia M, Minnaar K, Gioielli P, Kibey S and Fairchild D, ‘Development of the SENT test for strain-based design of welded pipelines’, In Proceedings IPC 2010 8th International Pipeline Conference, Calgary, Canada, 27 Sept.-1 Oct, 2010.
  13. Wang E, Zhou W, Shen G, Duan D, ‘An experimental study on J(CTOD)-R Curves of single edge tension specimens for X80 steel’, in Proceedings of the 9th International Pipeline Conference IPC2012, Calgary, Alberta, Canada, September 24-28, 2012.
  14. Park D-Y, Tyson R, Gianetto J A, Shen G, Eagleson R S, ‘Evaluation of fracture toughness of X100 pipe steel using SE(B) and clamped SE(T) single specimens’, in Proceedings of the 8th International Pipeline Conference IPC2010, Calgary, Alberta, Canada, September 27 - October 1, 2010.
  15. Tan J P, Wang G Z, Xuan F Z and Tu S T, ‘Correlation of creep crack tip constraint between axially cracked pipelines and test specimens’, International Journal of Pressure Vessels and Piping, 98 (2012) 16-25.
  16. Xu J, Zhang Z I, Østby E, Nyhus B, Sun D B, ‘Constraint effects on the ductile crack growth resistance of circumferentially cracked pipes’, Engineering Fracture Mechanics 77 (2010) 671-684.
  17. Drexler E, Wang Y Y, Sowards J W and Dvorak M D, ‘SE(T) [single edge notch tension] testing of pipeline welds’, In Proceedings IPC 2010 8th International Pipeline Conference, Calgary, Canada, 27 Sept.-1 Oct, 2010.
  18. Nyhus B, Østby E, Knagenhjelm H O, Black S and Røstadsand P A, ‘Fracture Control – Offshore Pipelines. Experimental studies on the effect of crack depth and asymmetric geometries on the ductile tearing resistance’, in Proceedings 24th International Conference on Offshore Mechanics and Arctic Engineering (OMAE 2005), Halkidiki, Greece, June 12-17, 2005.
  19. Østby E, Sandvik A, Levold E, Nyhus B, Thaulow C, ‘The effects of weld metal mismatch and crack position on the strain capacity in SENT specimens in an X65 material’, in Proceedings ISOPE-2009, The 19th International Offshore (Ocean) and Polar Engineering Conference, Osaka, Japan, June 21−26, 2009.
  20. Malpas A, Moore P and Pisarski H, ‘Crack front straightness validity in SENT specimens’, In Proceedings ISOPE-2012, The 22nd International Offshore (Ocean) and Polar Engineering Conference, Rhodes, Greece, June 17−22, 2012.
  21. Moore P, ‘The Effect of Notch Sharpness on the Fracture Toughness Determined from SENT Specimens’ in proceedings, OMAE 2014, 33rd International Conference on Ocean, Offshore and Arctic Engineering, San Francisco, California, USA, June 8-13, 2014.
  22. Shen G, Tyson W, Gianetto J, ‘Effect of Side Grooves on Compliance, J-Integral and Constraint of a Clamped SE(T) Specimen’ in proceedings ASME 2010 Pressure Vessels and Piping Division/K-PVP Conference (PVP2010) July 18–22, 2010, Bellevue, Washington, USA.
  23. Pussegoda L N, Tiku S, Park D-Y, Tyson R, Gianetto J A, ‘J-Resistance results from multi-specimen and single specimen surface notched SEN(T) geometry’ in Proceedings of the 9th International Pipeline Conference IPC2012, Calgary, Alberta, Canada, September 24-28, 2012.
  24. Park D-Y, Gravel J-P, Simha C, Liang J, Duan D-M, ‘Low-constraint toughness testing of two SE(T) methods in a single specimen’, in proceedings ASME 2014 Pressure Vessels and Piping Conference, Anaheim, California, USA, July 20-24 2014.
  25. BSI, ‘Metallic materials – Method of tests for the determination of quasistatic fracture toughness of welds’, BS EN ISO 15653:2010.

For more information please email:


contactus@twi.co.uk