Selecting a modeling approach for predicting remnant fatigue life of offshore topside piping

Abstract

Setting an optimal inspection plan for fatigue critical offshore piping relies on accurately estimating its remnant fatigue life (RFL). Several modeling approaches, such as knowledge-based, model-based, data-driven, fusion techniques etc., have been used to build RFL models in the past. The aim of this paper is to review these approaches and thereby recommend the most favorable approach for building a probabilistic RFL model for offshore piping. Firstly, a brief discussion about the aforementioned approaches is presented. Thereafter, a comparison is made between these approaches. For instance, there is uncertainty in model-based approaches, due to the assumptions of the underlying physical model, which poses substantial limitations on this approach. Conversely, a data-driven approach exploits the monitored operational data associated with the condition of the piping system. Fusion technique combines the features of the former two approaches and is recommended to build a model for estimating the RFL of offshore piping.