External stress-corrosion cracking of pipelines is a serious problem for the gm transmission industry. Longitudinal cracks initiate on the outside surface of the pipe and link up to form flaws that, in some cases, can lead to pipe rupture. This paper presents a model that quantities the effect of stress-corrosion cracking on pipe failure stress. The model is an extension of those that have been developed for oil and gas pipelines and considers both flow-stress and fracture toughness dependent failure modes. A methodology also is presented to calculate the remaining life of a pipeline containing flaws of known size.
The first incident of stress-corrosion cracking (SCC) on natural gas pipelines occurred in the mid 1960?s and hundreds of failures have occurred since that time. A characteristic of this form of failure is the presence of patches of? up to hundreds of longitudinal cracks on the outside surface of the pipe. In some cases, these small cracks link-up to form long shallow flaws that can lead to pipe rupture. Appropriate engineering procedures and models are required to evaluate the structural integrity of pipelines with stress-corrosion flaws. This paper describes a model that has been developed to evaluate the effect of SCC or other forms of localized attack on structural integrity. The current model k an extension of previous models that were developed to predict failure stress levels of line pipes-with local wall thinning. Those previous models account only for the hoop stress induced by internal pressure. The current model accounts for axial stress, excluding membrane bending stress, as well as hoop stress in evaluating flaws. Cracks are evaluated by means of inelastic fracture mechanics (FM) procedures that employ the J integral. The IFM procedures were adapted from those that have been used to evaluate the structural integrity of steam piping; 7-12 they were modified to account for stress-corrosion cracking instead of seep cracking. The current model is implemented by means of computerized calculations because of the complexity of the numerical analyses required for its application.