Fatigue Assessment of Aluminium Welded Box-stiffener Joints in Ships

Abstract

One of the trends in ship building is the building of larger and faster ships. However, increasing demand for speed has forced the ship designers to search for alternative materials to reduce the weight of the ship without compromising strength. Aluminium alloys, once properly designed, can achieve this goal by reducing the weight of structural components by over 50% compared to those made of normal carbon steel. Another advantage of aluminium alloys is that its corrosion resistance. In a sea environment, steel has been found to corrode about 100 times faster than aluminium alloys under the same corrosive environment.

As the size of aluminium ship increases, fatigue has become one of the main design criteria. In order to effectively make use of the high strength/weight ratio of aluminium alloys, the cross section of the structural components as well as the joint design should be optimized to provide sufficient fatigue strength while still maintain acceptable fabrication cost. Extruded hollow stiffeners have been proven to be able to provide excellent light and stiff panels for aluminium ships.

Box stiffener is such a cross section profile investigated in the present study. The box stiffener is usually oriented longitudinally by interacting with a large number of web frames by fillet welds. These welded joints may be exposed to larger fatigue damage due to its significant stress variations experienced during the service of the ships. In addition, welding of these joints is quite costly since there are a lage amount of them and it is difficult to apply automatic welding due to complex profile. Any simplification in the welding procedure would cause a significant benefit to the ship builders.

Three similar web frame/box stiffener connections are considered in these studies, by varying the opening shape on the web fame and accordingly the welding procedure. The objective is to achieve sufficient fatigue strength while still maintaining an economically efficient joint. The first connection (denoted as Alt-1) tries to make a relatively larger opening on the web frame for easy assembling and more space for welding. The second and third alternatives (denoted as Alt-2 and Alt-3) are identical in the opening on the web frame while differs on welding procedure only. Interrupted welding is performed for Alt-2 to save both the human resource and weld material while continuous welding is required for Alt-3 for a simple welding procedure. Alt-3 also provides a water tight solution due to the continuous welding.

It is found that the change of cutting shapes on the web frame as well as the corresponding weld procedure has great impact on both the static and fatigue properties of the joints. Extra welding around the web frame may cause a high stress concentration at the weld toe and therefore imply an unacceptable fatigue damage. However, a simple weld can be a good candidate to achieve sufficient fatigue strength. Butt weld is usually used to join two pre-fabricated units to obtain sufficient overall strength of the joined single piece structure. However, difficulty may arise for such units when the box stiffeners are presented as the longitudinal stiffeners due to relatively large shrinkage and distortion introduced in the fabrication process including welding in the adjacent area. A special type of lap joint, denoted as box stiffener lap joint is designed to provide a joining solution for the box stiffeners. By enabling the inside profile of the lapping plate to be identical to the outside profile of the box stiffener, the joining of two box stiffeners is just to weld the intersection lines between the lapping plate and box stiffener. Only the wide flanges of the box stiffener require a conventional weld. Cracking from the weld root should be avoided to ensure sufficient fatigue strength.

Good fatigue performance is observed for the box stiffener lap joint. The weld leg length is found to be a determining factor that affects both the fatigue strength and fatigue crack initiation site. Weld leg length between 6.0 and 7.5 mm is suggested to avoid high stress concentration at the weld toe. Moreover, this design reduces the likelihood of fatigue cracks to initiate from the most complex manual welding area.

Information about fatigue analysis of welded aluminium structures is quite limited in the open literature. This thesis deals with fatigue assessment methods of welded aluminium plate structures based on both numerical analysis and experimental tests. Various stress based fatigue assessment methods such as the nominal, structural and notch stress approaches have been applied to the investigated joints.

The nominal stress range approach can not provide a satisfactory prediction of the fatigue strength of novel welded joints because the local geometrical details including structural geometry and weld geometry are implicitly embedded in the nominal design SN curves. Laboratory tests would hence be needed to establish nominal stress approaches. More refined local approaches including the structural stress range approach and the notch stress range approach have been developed for a more accurate assessment of the fatigue strength. The main challenge for these local approaches is to find suitable stress calculation methods in the vicinity of the weld and thereby the uncertainties that determine the fatigue strength of the joint can be quantitatively reflected in a local stress concentration factor. Several of these methods are applied in the study of novel types of welded aluminium joints in this thesis. The choice of design SN curves must be consistent with the way by which the design stress range is calculated. Fatigue test data support the use of a structural stress design SN curve of fatigue class 40 for the fatigue assessment of aluminium fillet welded in-plane bracket connections when the structural stress has been derived from the extrapolation methods. It has been shown that a structural stress design SN curve of fatigue class 44 for the fatigue assessment of aluminium lap joints can be used.

Another important issue is to investigate the effect of the local weld geometry such as the weld toe angle, weld toe radius, and weld leg length on the value of the structural/notch stress and accordingly on fatigue behaviour. The weld parameters of the joints have been measured in the laboratory and the statistical measured values are used in the numerical finite element models (FEM).

Improving fatigue strength of critical joints is found to be necessary late in the design process or during the service time of a structure to extend the whole service life of the structure. The effect of weld toe grinding on the stress concentration at the weld toe is studied by finite element analysis (FEA). Fatigue tests show that the fatigue life in terms of number of cycles to failure is nearly doubled by simply grinding the weld toe of the joint. However, the standard deviations are not so much affected as the fatigue strength. The effect of weld parameters such as the grinding depth has been found to play a decisive role in fatigue life improvement.