A Model to Estimate the Failure Rates of Offshore Pipelines

Abstract

Pipelines are subjected to several threats which can cause failure of the line, such as external impact, mechanical defects, corrosion and natural hazards. In particular, offshore operations present a unique set of environmental conditions and adverse exposure not observed in a land environment. For example, offshore pipelines located near harbor areas and in major shipping lanes are likely to be exposed to the risk of damage from anchor and dropped object impact. Such damage may result in potential risk to people and the environment, and significant repair costs. Quantitative Risk Assessment (QRA) is a method which is often used in the oil and gas industry to predict the level of risk. In QRA calculations the frequency of an incident is often assessed by a generic failure frequency approach. Generic failure frequencies derived from local incident databases are largely used in pipeline risk assessments. As a result, risk assessments for offshore pipelines may not reflect accurately operational experience for a specific pipeline or region of operation. In addition, a better understanding of the causes and characteristics of pipeline failure should provide important information to improve inspection and maintenance activity for existing pipelines and to aid in selection of design criteria for new pipelines. This paper presents an analysis of the failure data from various pipelines databases to see if there is a common trend regarding failure rates, and failure-rate dependence on pipeline parameters. A breakdown of the causes of failure has been carried out. The effect on failure frequency of factors such as pipeline age, location, diameter, wall thickness, steel grade, burial depth, and fluid transported have been investigated and are discussed. The objective of this paper is to provide a guideline for the determination of failure frequency for offshore pipelines and to describe a new model developed for use within BP for this purpose. This model uses historical databases and predictive methods to develop failure frequencies as a function of a range of influencing parameters.