Advancing the damage tolerance of laser beam welded steels using crenellation technique

Abstract

Fatigue life extension, via the retardation of a growing crack under cyclic loading conditions, has been an important issue for designers and operators to reduce the cost and to improve the structural safety of welded structures. The methods to be employed for retarding fatigue crack propagation, needless to say, should ideally be knowledge based and practical (i.e. easy to implement, cost effective and should not lead to weight increase etc.). In this study, a novel approach based on modifying the distribution of the stress intensity factor (SIF) in plates using thickness variations is presented. By varying thickness via the introduction of ribs, “crenellations”, on one side or surface of the welded or non-welded plates, it was shown that this novel technique can contribute to the fatigue crack propagation life of plates substantially. Furthermore, crenellations can potentially lead to weight reduction for a given target fatigue crack propagation life as well. This research programme aimed to gain a fundamental understanding of the mechanisms involved in the damage tolerance behaviour of crenellated parent material and laser beam welded steel plates which operate, in principle, under cyclic loading. Particular emphasis was given to describe the crack growth retardation mechanism using the crenellation technique. The “crenellations” were introduced on the surface of the steel sheets and plates of medium grade and high-strength steels with and without laser beam welds. Experiments were conducted to study and investigate the fracture toughness, fatigue response and fatigue life of crenellated plates. In parallel, numerical analyses were carried out to compute the actual SIF for the combinations of crenellations in order to correlate the fatigue crack growth data. Investigations focused on fatigue crack propagation (FCP) under constant amplitude loadings, with cracks growing within (in butt joints) and towards (in fillet joints) the fusion zone. A number of crenellation configurations was investigated and the fatigue crack growth data were correlated using SIF. It was shown that the edge regions -where the thickness sharply changes from thin to thick or vice versa- and crack length have a substantial effect on the retardation and acceleration of the fatigue cracks. Factors such as sudden thickness variation and crack lengths -where plasticity can be substantial depending on the load level- affect crack retardation. However, it was shown that the evaluation of the correct SIF with respect to thickness and crack length conveniently describes fatigue cracking in these crenellated plates and allows the behaviour of these plates under cyclic loads to be compared with the behaviour of the reference, as-received plates subject to the same conditions. Based on the present tests and analyses, it is expected that an improvement in the fatigue crack propagation life of crenellated plates can be as much as 50 - 120 % depending on the configuration machined on the parent material and fillet welded plates, and crack arrest is very likely to occur depending on the crenellation configuration machined on the butt welded plates. The roles of local microstructural properties, residual stress profiles and K-distribution variations (due to crenellations) were discussed. The governing micro-mechanism(s) of failure of the plain and stiffened panels were identified and compared. The regions -where there is a rapid thickness change-exhibit complex crack fronts which contribute to the retardation of the crack under cyclic loading conditions. However, the overall cracking mechanism is the same as the mechanism observed in the un-crenellated, reference plates. Additionally, a numerical damage model based on strain energy was used to predict fatigue crack growth in the crenellated plates. In general, the results were conservative. The comparisons presented are used to identify the future trends in the development of numerical techniques that should be developed to complement testing. This is imminent especially for the cases, where large components with various features need to be tested. Given the difficulty and expense of running such industrial scale test programmes, numerical methods will be crucial for the optimization of the number of tests that may be necessary. This work and the conclusions presented are novel, and it was shown that no investigations exists in the open literature and on-going publicly funded projects, that considers both experimental and FE-methods to study the effect of crenellations. Furthermore, no previous work -using both extensive experimental and numerical programmes- has been reported on the role of crenellations on steels and their weldments. The results of this thesis are expected to have a significant impact on the existing knowledge and design as well as methods to extend fatigue life of the welded structures where crack propagation plays an important role.