Pipe laying projects are getting more challenging due to increasing operational depth of pipeline. The increasing depth requires more of the strength and load bearing capability of the pipeline and its girth welds. The assessment method used for checking the integrity of flaws in girth welds is called Engineering Critical Assessment (ECA) and is based on fracture mechanics (FM) formulas. A complete ECA covers both static and fatigue loading encountered over the multiple phases of the pipe laying process. The current ECA methodology for both static and fatigue load produces strict requirements that may cause unnecessary repairs of the girth welds. From within the industry it is now wondered if ECA produces (over) conservative results.

The current static ECA methodology uses a worst case (deterministic) load scenario as input to determine the maximum allowable initial flaw depth (MAIF) of a girth weld which is then directly used as NDT criterion for that specific pipeline project. The proposed methodology in this research uses a reliability based approach where the probability of failure (PoF) of the pipeline/girth welds is governing in order to determine the NDT criteria. By (conservatively) assuming girth welds in a pipeline form an independent series system, the required PoF of the pipeline ($10^{-4}$) can be converted to the individual required PoF of a girth weld ($10^{-8}$). This together with stochastic input variables provides the required information to determine the NDT criterion based on the required PoF of the pipeline.

The available data to determine the best probability density functions (PDFs) of the stochastic variables used in the reliability based approach is limited, introducing uncertainty in the accuracy of the results. However when using the limited data the two models already provide NDT criteria in the range of 5.05mm-5.20mm, while the deterministic approach dictates a NDT criteria of 4.45mm. A difference of 0.55mm-0.75mm is substantial in terms flaw assessment criteria. Experts on welding and flaw assessment within HMC estimate the impact of the increase in NDT criterion to result in one less offshore shift of 12 hours, potentially saving up to hundred of thousands of Euros.

From the research performed it may be concluded it is possible to provide NDT criteria on basis of a probabilistic ECA, it is found the current ECA for static loading on external weld flaws might indeed be conservative. Therefore using a probabilistic approach towards ECA for both static and fatigue loading can benefit future projects by reducing repairs and cost.