Analysis of Corrosion Fatigue Crack Growth in Welded Tubular Joints

Abstract

ABSTRACT An improved fracture mechanics model for fatigue crack growth in welded tubular joints is developed. Primary improvements include the use of a wide-ranged equation for the fatigue crack growth rate properties and the incorporation of the influence of local weld toe geometry into the stress intensity factor equations. The latter is shown to explain the dependence of the fatigue life on the size of tubular joints. Good agreement between predicted and measure fatigue lives of full-scale joints tested in air further supports the applicability of the fracture mechanics approach to offshore structures. Although the model should also be applicable to corrosion fatigue, additional input data and verification testing are needed under these conditions. Factors which could improve the model are discussed. INTRODUCTION In general, the fatigue life (Nf) of structures includes cycles required for crack initiation (Ni) plus cycles required for crack propagation (Np):(Mathematical equation available in full paper) However, the relative contributions of Ni and Np to the fatigue life are specific to the structure and depend on fabrication procedures, stress history and acuity of metallurgical and geometric stress concentrations. Generally, for large, welded structures the fatigue life is dominated by crack propagation. For example, data from full-scale tests of welded tubular joints demonstrate that "engineering-sized" cracks of 1 mm to 3 mm are present at less than 10 percent of the fatigue life; thus, Equation 1 reduces to(Mathematical equation available in full paper) Consequently, linear elastic fracture mechanics represents a physically realistic approach to evaluating the fatigue life of offshore structures. The primary advantage of this approach is that it provides a rational procedure for assessing the influence of defects on the fatigue performance of these structures. Furthermore, once adequate analysis procedures are developed and verified, they can be combined with measured fatigue crack growth properties to predict the influence of weld geometry, joint configuration and loading mode on fatigue life, thereby reducing the need for costly, full-scale fatigue testing. Although past efforts have been made to apply fracture mechanics to offshore structures, most of these treatments have been oversimplified, particularly with regard to the formulation of crack-tip stress intensity solutions and/or crack growth rate equations. The objective of the current study was to develop a fracture mechanics model with improvements in these two elements. Stress intensity factor solutions are approximated for welded tubular joints by combining existing analytical solutions to account for size and shape of surface cracks, local stress gradients at the weld toe, joint geometry and loading mode. Crack growth rate equations are formulated so as to be applicable over a wide range of growth rates; parameters in these equations are determined for key loading and environmental variables. Initial flaw size, as well as a "worst-case" notch, are non arbitrarily established as a function of material and weld-toe stress concentration factor using the concepts given in Refs (3-6). In order to evaluate the model, predicted fatigue lives are compared with available data from fatigue tests on full-scale tubular joints.