Failure from a flaw in a girth weld of a pipeline may be induced when vertical buckling occurs due to temperature changes or other causes. Although conventional assessment of pipeline girth weld flaws has tended to qualify weld flaws as fixed values, they are treated as random variables in this study. Monte Carlo simulation is applied to estimate the conditional failure probability and field data of weld flaws detected by AUT (Automated Ultrasonic Testing) are utilized to improve the precision of the simulation.
Offshore pipelines provide one of the safest and most reliable engineering structures for the transportation of oil and gas from one place to another. In recent gas pipeline projects, ECA (Engineering Critical Assessment) has often been carried out to decide the acceptance criteria for the weld defects. The ECA is usually conducted according to the acknowledged standards like BS 7910 and the most conservative design condition is considered. As a non-destructive inspection, Automated Ultrasonic Testing (AUT) is commonly used when the acceptance criteria derived from ECA are applied to pipeline projects. At the same time, the reliability-based approach is also considered to be efficient both for designs of newly constructing pipelines and for reassessments of traditionally designed pipelines. This is because it has two advantages. First, the reliability-based approach enables uncertainties to be handled in a rational and logical way in design and assessment. Second, while decisions are never clear and perfect, it provides a more rational basis for decision making than a purely deterministic analysis. The object of this paper is to analyze the data for weld flaws statistically and to quantify the safety of pipelines using the failure probability. Height, length and location of the flaws, the accuracy of flaw detection by AUT and the material properties were regarded as random variables and Monte Carlo simulation was carried out to calculate the conditional probability of failure in the special situation of upheaval buckling due to thermal changes or other causes.