Feature Articles—October 2009 Issue

Assessing Fatigue Life of Welded Joints On Offshore Constructions
Experimental Fatigue Analyses of Fillet-Welded Joints Under Multiaxial Loading in Ships and
Offshore Structures

By Huseyin Ozden
Professor
and
K. Turgut Gursel
Professor
Mechanical Engineering Department
Ege University
Izmir, Turkey

Welded joints are the critical weak points of structures made of materials like cast iron due to the existence of micro pores and notches in welds. Under normal operating conditions, the welded joints in ships and offshore constructions are mostly loaded by aperiodic, multiaxial forces and moments and therefore must meet high structural durability requirements under these stresses. However, few studies have dealt with determining the fatigue life of fillet-welded structures under multiaxially complex loads. In all of this research, multiaxial superposed loads were analyzed experimentally and computationally under different criteria, particularly regarding the influence of load combinations on structural durability.

However, it must be said that numerous influences on structural durability, such as complicated geometries, different materials and manufacturing methods, and complex load conditions, make experiments at construction sites expensive and time-consuming. The lifetimes of construction units determined by computing procedures can deviate strongly from those determined in tests. The obtained results often apply to a test series in each case.

For a computational fatigue life assessment, both the loading and bearing capacity of a structure must be available. With a suitable computational method and using a hypothesis of damage accumulation, fatigue life cycle proof can be obtained.

There are different concepts for the consideration of multiaxial loading conditions in fatigue life determination. For proportional loading conditions, the conventional strength hypotheses can be selected. However, in loading conditions with temporarily variable main stress directions, an integrating damage hypothesis is suggested in which the interaction of all loads in the cross-section planes of a volume element is used to determine fatigue life.

No method truly provides a reliable, generally accepted determination of fatigue life for multiaxial loading cases, however. Substantial uncertainties exist in the definition of the damage parameters as well as in the accumulation of damage. Additionally, there are many computational concepts, hypotheses and sets of rules for fatigue life estimation as well as for structural durability computations, which often cause confusion. This article presents the results of tests and analyses in many diagrams taking the International Institute of Welding (IIW) recommendations into consideration. Primarily simple fatigue life estimations were performed for different welded joints. The data obtained can be used for computer-aided fatigue life estimations.

Tests at Different Welded Joints
In order to create a database for the computation of the lifetime of fillet-welded joints under uni- and multiaxial loading, extensive investigations and single-step (i.e., Woehler tests) and structural durability tests were performed at the Institute of Plant Engineering and Fatigue Analysis of the Technical University of Clausthal in Germany.

This study analyzed the research institute’s test data extensively, and the results were made available for computational lifetime estimations. For the investigations performed, the fine-grain structural steels S460-EN 10297 and S460M-EN 10 113-3—frequently used in machine constructions—were selected. The examined construction samples included fillet-welded joints of a plate to a pipe, which were made with and without joint preparation.

The samples were preheated to approximately 160° C before welding. The residual stresses in welds were relieved primarily with thermal treatment (for approximately 150 minutes at a maximum temperature of 540° C). The experiment was designed so that the bending and torsion loading could be introduced separately (i.e., during the introduction of torsion, no bending stress was applied to the critical areas of the fillet-welded joints and vice versa). As a uniform damage criterion for the tests, a crack through the sample was specified.

For the investigations, load functions of GAU99-E4 and GRO5-E5 with stress ratios of R=-1 and R=0 were used, respectively, with a partial extent of 5x10⁴ cycles and an irregularity factor of I=0.99.

Test Results and Discussion
The results of single-step and multiaxial loading tests on the joints, dimensioned according to the IIW rules, were compared. Thus, these results can be used as a basis for the computational lifetime estimation.

The breaking points of the structural durability lines can be determined on the basis of the IIW rules. Currently, no recommendations exist for combined multiaxial loading of fillet welds.

For pure bending and combined loading with a larger amount of bending, the inclination of the straight line of a Woehler curve (with constant amplitude loads) was k=3, and a breaking point between 10 and 10⁸ life cycles was selected. For lines of a Gassner curve (with variable amplitude loads), the same inclination (k=3.0) was selected, and the point halfway between the intersection points of the structural durability lines of the Gassner and Woehler lines was selected as the breaking point.

For fillet-welded structures loaded with pure torsion, k=5 was selected as the inclination of the Woehler curve, and the breaking point was 108 life cycles. Using the same procedure as mentioned above for the Gassner line, an inclination of k=5 and the intersection point of 108 life cycles was obtained. Fundamentally, the intersection point life cycles can be variable.

During combined loading of bending and torsion with a larger amount of torsion, the inclination coefficient was determined to be k=4, and the breaking point was between 107 and 108 cycles.

This procedure provides rather accurate estimation values for the structural durability of fillet-welded constructions. Nevertheless, perfect accuracy is not to be expected.

During changing shear and normal stress, a frequency difference increases the structural durability properties slightly, according to the shear stress hypothesis. Maximum structural durability occurs in a frequency shift between normal and shear stress of lxy=0.5, and it is approximately five percent higher than in loading with the same frequency. The frequency shift can have both a strength-increasing and a strength-reducing effect.

The results showed that in multiaxial loading, the phase difference between the stress components is an important parameter. The variation of the stress components in time is defined by equations 1, 2 and 3.

The Woehler and Gassner fatigue strength curves show similar characteristics for both weld shapes, i.e., HV- and HY-fillet welds (HV-fillet welds are with joint preparation, HY-fillet welds are without joint preparation).

The local stress values, types and forms as well as the combination of the loading crucially affect the fatigue life of fillet-welded structures and often cause a strong reduction in life cycles. The superposition of bending and torsion with a phase difference of 90° (between normal and shear stress) for R=-1, and especially R=0, results in a reduction of the fatigue life in comparison to bending and torsion without phase difference.

In the multiaxial loading tests, the lifetimes of samples dimensioned for operating conditions are clearly longer than the life cycles of those used in the single-step tests, if the average magnitudes of stress amplitudes are not very different.

The fillet-welded joints usually fail in joint transition notches (weld joint to pipe). In the case of pure torsion and in superposition of torsion and bending with mainly torsion, the fillet-welded joint fails in different places with unequal crack geometry. With higher nominal stresses, cracks initiate from the root of the weld.

During a lower loading, irregular fracture surfaces (zigzag surfaces) develop, and the weld joints usually fail in the joint transition notching areas.

This study and other preliminary investigations show that the weld shape does not have a dominating influence on structural durability under loads of bending, torsion, and bending and torsion in each case.

Due to dynamic loading, all materials change their mechanical characteristics under environmental influences, such as corrosion and radiation, in the course of time, and they eventually fail. Continuous straight lines of structural durability running parallel to the time axis with an accurately determined breaking point do not usually exist.

Depending on the magnitude and form of the loading, there is a finite fatigue life that is strongly dependent on the condition of materials, construction and sea conditions.

The breaking point is more or less an arbitrary assumption that applies to the test conditions or is suggested by experience.

This is proven by the fact that the breaking point is differently defined by experts.

Furthermore, for the joints of welded structures, an experimentally and computationally accurate fatigue life determination is still very difficult for the following reasons: there remains a lack of sufficient data and experience from experimental investigations, there are numerous variables in welding and structural durability analyses, and the structural durability and damage mechanisms of the welded joints are not reproducible, even by using the same adjusting variables.

In steel constructions such as ships and offshore structures, the fillet-welded joints are extremely vulnerable under combined, irregular, dynamic loads. These joints usually function as notches and, thus, have a smaller deformation and energy absorption capacity until failure. Therefore, the stress peaks occurring due to a certain plastic deformation in the notches and/or failure-critical areas cannot be relieved in dynamic loading. Furthermore, the production and welding imperfections still contribute to the failure of materials in the critical transition notching areas, which rise with increasing weld bath volumes. These areas cause high stress concentrations that contribute to cracking.

Conclusions
In marine constructions, cracks mostly begin in welded joints, especially in fillet-welded ones. This research shows that the weld shape does not have a dominating influence on the structural durability under loads of bending, torsion, and bending and torsion in each case, meaning it is generally possible to avoid cost-intensive HV-fillet welds in marine constructions.

These test results show the clear influence of combined loading on the structural durability of fillet-welded joints. The numerous variables and random influences during welding make accurate life estimation difficult; however, local stress values, types and forms, as well as the combination of different types of loading, crucially affect the fatigue lifetime of fillet-welded structures and often reduce it drastically.

Acknowledgments
The authors would like to thank the Institute of Plant Engineering and Fatigue Analysis of the Technical University Clausthal, Germany, for performing these tests.

---

Huseyin Ozden be-gan his professional career at the fatigue strength branch in Ay-val?k, Turkey, but completed his Ph.D. in Germany. He is currently a professor in the mechanical engineering department of Ege University.

K. Turgut Gursel studied naval architecture and marine engineering at Istanbul Technical University in Turkey and at the University of Ham-burg in Germany. He completed his Ph.D. at the Berlin University of Technology. Currently, Gursel is employed as a professor in the mechanical engineering department of Ege University.

Sea Technology is read worldwide in more than 110 countries by management, engineers, scientists and technical personnel working in industry, government and educational research institutions. Readers are involved with oceanographic research, fisheries management, offshore oil and gas exploration and production, undersea defense including antisubmarine warfare, ocean mining and commercial diving.