Fatigue Failure of Welded Tubular Joints

ABSTRACT The fatigue of welded tubular joints, as found in offshore structures, is explored by developing a qualitative model of fatigue behavior. The model is then used, along with basic concepts from fracture mechanics, to develop an analytical model for the prediction of fatigue failure. Fatigue failure is defined as a distinct change in the joint'sload transfer characteristics. Experimental data is presented to support both the definition and prediction of failure. Introduction The ever expanding use of both stationary offshore oil drilling structures and semisubmersible rigs, composed of space frames formed from planar trusses built of tubular sections, has focused increased attention to the unique problems of welded tubular connections. The random and repetitious nature of wind, ice flow and ocean waves causes the development of alternating loads in the web, or bracing, members of these trusses. Extensive experimental studies have been carried out as to the effect of alternating loads on so-called tubular joints of different configurations typical for offshore construction [1 – 6]*. These studies underscored the importance of stress concentrations and their influence on the fatigue behavior of these joints under cyclic alternating loads. The experimental results provided significant information regarding the life-expectancy of tubular joints and contributed to the development of improved joint designs. It is the purpose of this paper to correlate these experimental findings and to develop a theoretical evaluation of the joint life by incorporating basic principles of fracture mechanics. Although theoretical fracture mechanics concepts have been with us for two decades [7, 8]and extensive applied approaches in the aerospace field for at least a decade [9], it is only recently that these concepts are becoming common to the literature concerned with structural steel applications. For example, there is the consideration of a running crack in steels used for pipe lines by Hahn et al [10]; there is Wessel's [11] examination of plane strain fracture toughness in forging steels used for nuclear pressure vessels; and recently there is some discussion of fracture mechanics with reference to offshore platforms [12]. Up to the present, the lack of interest stemmed from the fact that the use of linear elastic fracture mechanics was dependent upon the material having a high ratio of yield strength to modulus. This was essential so that the material would remain nearly elastic through failure; that is, the plastic zone at the crack tip would have to be small compared to other dimensions such as crack length. Obviously, structural steels did not come under this category since plastic behavior on a macro scale is common prior to failure. However, more recent applications of crack-tip displacement concepts [13, 14]and elastic-plastic analysis [15] have generated interest in this field. It is not the purpose of this paper to consider how elastic-plastic concepts might be utilized to analyse the critical stress intensity, Kcr, at unstable fracture. Rather, it is the application of stress intensity concepts to fatigue crack propagation analysis that is of primary concern.