Abstract
In general, fatigue tests are conducted with a constant amplitude load to estimate the fatigue crack growth characteristics. However, vessels have variable amplitude loads because various forces are applied to them. So it is essential to consider variable amplitude loads and a constant amplitude load together. The existing data shows that fatigue crack propagation is retarded when the single or multiple tensile overloads are applied and they affect fatigue life. But there are no results of overload effect on weld metal (WM) and heat effect zone (HAZ) about Al 5083-O.
The Challenge of Variable Loading in LNG Carriers
Engineers designing Liquefied Natural Gas (LNG) carriers face a complex challenge: these massive ships aren't subjected to steady, uniform loads. Instead, they battle variable sea states and thermal expansion/contraction cycles. Standard fatigue life predictions often rely on "constant amplitude" loads, which are easier to test but don't reflect reality.
Crucially, existing research shows that a single tensile overload can create a large plastic zone at a crack tip, which actually hinders further crack growth—a phenomenon known as the "retardation effect." While this effect is well-documented for base metals, there has been a significant data gap regarding how it behaves in the Weld Metal (WM) and the Heat-Affected Zone (HAZ) of Al 5083-O, the primary alloy used in independent tanks like the MOSS and IHI SPB types.
Experimental Methodology and the Wheeler Model
The researchers combined experimental material science with mathematical modeling to bridge this gap. Al 5083-O plates were joined via Metal Inert Gas (MIG) welding, and specimens were taken from the Base Metal (BM), WM, and HAZ for testing.
The study compared a "Constant Amplitude" (CA) load (max 11 MPa) against a "Single Overload" (SOL) scenario, where a one-time spike of 22 MPa (double the max load) was applied once the crack reached 27mm. To predict the resulting fatigue life, the team utilized the Wheeler Model, a semi-empirical model that uses a retardation parameter ($C_p$) to adjust crack growth rates based on the size of the plastic zone created by the overload. Three different methods for calculating the plastic zone size—Irwin, Jono Masahiro, and Antune—were evaluated to determine the best fit for this specific alloy.
Key Findings: The Retardation Effect in Welds
The study revealed several critical insights into the behavior of Al 5083-O welds:
- Weld "Under-matching": The weld metal was found to be weaker than the base metal (Yield Strength: 157 MPa vs 166 MPa). This "under-matched" condition makes accurate fatigue life prediction even more vital.
- Enhanced Retardation in Welds: Surprisingly, the retardation effect was more pronounced in the weld areas than in the base metal. The plastic zone created by the overload in the weld metal was 1.04 mm larger than that in the base metal, leading to a stronger retardation effect.
- Irwin's Method Superiority: Among the models tested, Irwin’s method provided the most accurate fit for the experimental crack growth slope, suggesting it is the most reliable approach for plastic zone size calculation in Al 5083-O Wheeler Model predictions.
Case Study: MOSS-Type Tank Prediction
The practical implications of these findings were demonstrated through a theoretical case study of a MOSS-type LNG tank plate.
By accounting for the retardation effect, the predicted fatigue life increased significantly:
- Base Metal: Predicted life increased from $2.19 \times 10^4$ to $6.07 \times 10^4$ cycles (+63%).
- Weld Metal: Predicted life increased from $0.96 \times 10^4$ to $3.99 \times 10^4$ cycles (+76%).
These results suggest that standard models without overload considerations are overly conservative, potentially underestimating the true service life of welded LNG tank structures.
Conclusion and Future Directions
While this study provides a robust framework for predicting fatigue life in Al 5083-O welds under single overloads, several areas remain for future exploration. Real-world maritime environments involve stochastic, multiple overloads that may interact in complex ways. Furthermore, since LNG is stored at cryogenic temperatures (-162°C), future research must investigate how extreme cold affects the overload retardation mechanism. Finally, varying the stress ratio ($R$) could further refine our understanding of crack closure behavior in these critical maritime materials.
Summary updated