Risk Based Inspection Planning Using Bayesian Updating

Abstract Offshore Structures are typically designed with a target end of field life between 20 and 30 years. Many offshore Structures remain in service while exceeding their designed life which requires robust methods for evaluating and maintaining target risk levels through regular sub-sea inspections. Risk based inspection plans are generated based on strength and fatigue analysis (typically S-N based fatigue assessment) which involves a number of uncertainties which are inherent to the fatigue process. These uncertainties inevitably lead to inconsistencies between analysis predictions and the outcome of underwater inspections. Although the inspection findings are acknowledged, there is generally not much effort made to make use of the inspection findings to optimise future inspection plans. This paper applies reliability updating methods to optimise inspection plans (inspection intervals) using fatigue reliability based on the results of in-service inspections. This methodology forms part of the work requested by ADMA-OPCO during the customisation of Atkins Fleet Management System (FMS) for the application to their Offshore Structures fleets. It has been reviewed and adopted by ADMA-OPCO to replace existing qualitative methods. The methodology works towards optimising their inspection efforts and cost to maintain acceptable target risk levels. The methodology is based on evaluation of the SN fatigue reliability since the SN Curves are implemented in the fatigue design methods outlined in International Standards. A probabilistic model for fracture-based reliability calculation is developed by calibration of initial crack size to match SN based fatigue reliability. This model enables utilisation of the Bayesian updating techniques which allow incorporation of the inspection results in the calculations. Environmental data from Umm Shaif Field (ADMA-OPCO) have been utilised in the calculations along with fatigue design SN Curves from the API RP 2A international standard. The calculations target the acceptance probability of failure levels of a critical component (hotspot) with a target design life of 30 years. In the event that no fatigue cracks are identified through in-service underwater inspections, the existing inspection plans can be revised utilising the developed methodology to optimise the inspection intervals, leading to cost saving and optimisation of the inspection plan. Major in-service inspection campaigns have been executed by ADMA-OPCO where no fatigue cracks were identified. Other operators within the region have also concurred that no fatigue cracks were observed during in-service inspection campaigns, which indicate that the methodology proves effective in optimising inspection costs within the region.