In offshore steel structures engineers often face the problem of assessing the criticality of existing hot spots to predict the remaining lifetime and thus to develop sound reliability-based inspection programs. One problem with such an approach is that the past fatigue conditions cannot be appropriately modeled, and the degree to which damage has accumulated in hot spot areas cannot be consistently modeled. This paper shows a practical methodology for predicting the remaining fatigue life of hot spots by using a probabilistic fracture mechanics approach and shows how in general the results can be used in reliability-based inspection programs.

At present, the best practice for modeling remaining fatigue life and identifying hot spots in existing offshore steel structures is the traditional S-N approach (Miner's rule). The uncertainties in the S-N approach for existing structures are significant, and it is often impossible to take into account the stress cycles to which the structure has been subjected in the past, especially in cases where the structure has been strengthened or modified. This results often in very low fatigue life, sometimes even lower than the actual age of the structure. The S-N approach has thus only limited use for identifying measures and for establishing risk- and reliability-based inspection plans for such hot spots in existing structures.

The present paper shows how a probabilistic fracture mechanics approach can be used to analyze the existing hot spots in an offshore steel structure and how reliability-based inspection planning can be established based on these results. The general approach is to postulate cracks of specific length and depth that match the threshold of known inspection techniques, to have cracks located in specific directions at the hot spot locations, and then to predict the crack growth and failure probability of the postulated cracks.