Influence of the Mean Load on the Fatigue Performance of Mooring Chains

Abstract

Whilst it is known that mean stress has an effect on the fatigue endurance of steel components, this effect is not considered when designing mooring system components. The S-N and T-N fatigue design curves for mooring chain in the standards are based on tests carried out at a single mean load, which is 20% of the chain minimum breaking load (MBL), and these curves are used to compute the damage of all load cycles regardless of their mean value. Lately it has been found that the effect of the mean load can be larger than probably expected, and that mooring chains exhibit a significant increase of fatigue capacity when cyclically loaded at reduced mean load. In the majority of the floating units, the pre-tension of the moorings without environmental loads is below or well below 15% of the chain MBL, and most, if not all, of the in-service damage is produced at mean loads below 20% MBL. This in practice results in additional conservatism to the fatigue life computed using the existing design curves. Some deepwater units, subjected to high pre-tension level, can experience some or relevant damage occurring at mean loads above 20% MBL, which would be underestimated with the present design approach. The paper provides an insight on the effect of the mean load on the fatigue endurance of mooring chains and quantifies this dependency based on a large number of fatigue tests carried out on different chain diameters between 70 and 171 mm, different grades, and different mean loads ranging between 7% and 20% of the MBL of the tested chains. The Smith-Watson-Topper (SWT) mean stress correction model is used to transform the stress state of the tested chains into different stress states associated to different mean loads. Then regression analyses are performed and correction functions derived for the design curves of both S-N and T-N approaches to account for the mean load while keeping the same confidence of the existing curves.

Influence of Mean Tension on Mooring Line Fatigue Life

“…Although a mean load effect on fatigue capacity was suspected based on results at an early stage, it was hard to conclude at the time due to large uncertainties with respect to chain condition (corrosion and surface pits), different chain qualities and manufacturers, and statistical uncertainty. As more full-scale tests have been performed with an increasing awareness of mean load effects, combined with increasing knowledge sharing across initiatives, the general consensus is now that mean load effect on fatigue damage is significant [3][4][5].…”

“…Existing fatigue design curves for offshore mooring chain (e.g. DNVGL-OS-E301 [6]) are based on fatigue tests performed at a mean load of 20 % of minimum breaking load (MBL) [4,5], and are used for fatigue design check regardless of actual mean loads. For systems operating at mean loads below 20 % MBL, this may imply a need for costly replacements at an earlier stage than strictly required.…”

“…Despite the consensus on significant mean load effect, a main challenge has been to quantify its practical importance. Fernández et al [5] explore and discuss use of mean load correction models for mooring chain, and conclude that a Smith-Watson-Topper (SWT) model is preferred over alternative methods like Gerber and Goodman. The SWT model is defined as [5,7]:…”

“…The study is based on numerical, time-domain simulations, using environmental conditions defined from a series of hindcast data. A parameterized S-N design curve suggested by Fernández et al (2019), incorporating a Smith-Watson-Topper mean stress correction model, is applied for fatigue damage calculation and compared to results for the S-N design curve prescribed by current standards.For the semi-submersible unit considered there is negligible difference in basing the correction on 3-hour mean load compared to the mean load of individual stress cycles, due to small low frequency tension variations. On this basis, a single correc-…”

“…The study is based on numerical, time-domain simulations, using environmental conditions defined from a series of hindcast data. A parameterized S-N design curve suggested by Fernández et al (2019), incorporating a Smith-Watson-Topper mean stress correction model, is applied for fatigue damage calculation and compared to results for the S-N design curve prescribed by current standards.…”