Wed, 13 March, 2024
A joint that experiences cyclic loading in service is expected to eventually crack by fatigue. The question is, how long will it take for fatigue cracking to occur?
This question is important in industries that use welded fabrications or bolted joints that are subject to cyclic loading, such as the wind industry and the oil and gas industry.
Wind farms contain many wind turbines mounted on substructures containing welded joints and bolted flanges, and the design life of the substructure is controlled by the fatigue performance of the joints, typically the welds.
The large size of wind substructures means that welds and bolts are being used in applications that are outside of the guidance in the current fatigue design documents. The standard welding procedure qualification tests (e.g. side bend, Charpy impact tests and tensile tests) don’t provide any information on the fatigue performance of the welds. Specific fatigue tests are needed for that.
TWI’s expertise in designing experimental programmes, designing and making bespoke test fixtures, interpreting fatigue test results, and our large load capacity test machines were all used to run a recent CRP project, investigating the fatigue performance of large (M72) bolts that are used in wind turbine ring flanges.
A fixture suitable for testing M72 bolts was designed and built, and tests were carried out on TWI’s largest test frame, with a 2500kN capacity. The bolts tested were hot-dip galvanised, M72 bolts with rolled threads, from two manufacturers. The results were compared to the S-N curves and size corrections in current fatigue design standards. One set of tests was carried out with a high mean load to simulate the high pre-tension present in the bolts in service. Since the bolts’ pretension may relax in service, the set of tests also investigated the effect of lower levels of pre-tension on fatigue performance.
The results confirmed that the fundamental fatigue performance of M72 galvanised bolts could be described by the DNV Class G and BS 7608 Class X-20% S-N curves, provided that the reference size used in the size correction equation was 25mm (as used in DNV RP C203 and BS 7608). The stress measured in the bolts tested with different levels of pretension allowed a relationship between applied pretension, applied service stresses and bolt stress cycle to be derived. The modified Goodman correction was shown to be suitable to account for the effect of different mean stresses in the bolts, in order to predict fatigue life over a range of bolt pretensions and maximise the efficiency of the joint.
Another example which demonstrates TWI’s impact on industry in the area of fatigue of welded joints is the fatigue classification of single sided girth welds in pipes for the oil and gas industry. Since the consequences of failure of oil pipelines and risers are so great (in terms of environmental impact and loss of production), operators need confidence in the actual fatigue performance of pipeline welds made using new welding procedures. This has led the oil and gas industry to routinely perform full-scale resonance fatigue tests on pipeline welds made using new procedures, in order to generate S-N curves describing the actual fatigue life of their welds.
Based on the results from these tests carried out at TWI over many years, fabricators and pipeline operators have been able to optimise their single sided girth welding techniques. TWI’s analysis of the resulting data has shown that the fatigue performance of current industry standard single sided welds has improved. Our analysis of a large database of these tests results in 2011 showed that it was possible to justify an increase of the fatigue class of single sided girth welds, if the root mismatch is limited to 1mm, from class F to class E. This increase in a fatigue design class increases the allowable operational life of the welds to increase by a factor of 1.65; e.g. from 20 years to 33 years.
This example shows how fatigue testing can help to demonstrate that the actual fatigue performance of present day industry standard welds may be significantly higher than the S-N curves in the current fatigue design documents suggest, as these aim to be a worst-case, lower bound for design. Without fatigue testing, if a fabricator cannot guarantee that pipe alignment can be controlled sufficiently to limit the mismatch to 1mm or less, and if the weld mismatch cannot be measured and proved to be 1mm or less, the designer must assume that the fatigue class of the welds is F rather than E, and the safe operational life of the pipe will have to be limited to a factor of 1.65 less than might otherwise be demonstrated. If fatigue tests were carried out on representative welds and the results showed that they had a higher fatigue class, this significant benefit can be used by the designer to safely extend operating life.
Fatigue testing provides confidence in the actual fatigue performance of a fabricated joint. Once this is known, the likelihood of a weld cracking by fatigue prematurely in service can be quantified and an application specific safe life can be defined. With the latest welding technology, fatigue testing often demonstrates that the fatigue performance of welds made using a particular procedure is higher than calculated using the fatigue design classes, which are based on historical worst case results. If a welding procedure can be shown to have performance in line with a higher class, the designer can use that higher fatigue class to extend operational lifetimes, or increase the limit of allowable stresses on the structure. Both of these things can result in dramatic cost savings.
It is only by detailed analysis of the loading conditions that the welded structure will experience, coupled with a bespoke programme of fatigue testing on representative welds and components, that designers can accurately answer the question, “how long will it take for fatigue cracking to occur?” with confidence.
Please contact us, at the email address below, to find out more…
- Carol Johnston (Consultant)