Degradation limit state model for structural reliability screening of ageing fixed offshore platforms

Abstract

Ageing fixed offshore platforms are growing in numbers worldwide. Operators choose to extend the platforms lives beyond the original design to improve economic viability and increase profitability. The structural integrity of these platforms is affected by various degradation factors throughout their service lives. However, a limited number of comprehensive studies have been conducted on the relationship between reserve strength ratio, probability of failure and return period with multiple degradation factors. This study aims to develop a comprehensive regression model of reserve strength ratio, probability of failure and return period by considering marine growth, corrosion and subsidence. Calculating reserve strength ratio, probability of failure, and return period are fairly time consuming. The presence of the proposed model is to provide a quick reference and immediate results of the remaining life of the fixed offshore platform in the occurrence of degradations, thus minimising the usage of industry resources. It is also expected that degradation effects over time will be predicted accurately. The development of the degradation limit state model adopted structural reliability assessment, which has been widely used in the oil and gas industry to determine the probability of failure and return period of offshore structures. The assessment can provide a higher confidence level that is required by regulators and stakeholders. This study includes the effects of wave height at the collapse of the platform caused by wave-in-deck. The wave-in-deck load has been calculated based on the silhouette method introduced by International Organization for Standardization. The degradation limit state models considered have 0 m, 2 m, 4 m, 6 m and 8 m subsidence. Each of them with 0 mm, 3 mm, 6 mm, 9 mm and 12 mm corrosion depth has been studied separately. The model has been developed using both single and linear multi regression method. The proposed models are then validated with a platform of similar configurations. Based on the validation results of single regression method, the lowest accuracy for reserve strength ratio was 94.9 %, while the probability of failure and return period were 56.6 % and 69.7 %, respectively. Despite that, the variations are acceptable since both probability of failure and return period values conform with the standard industry requirements. However, the results for 5 m subsidence shows very low accuracies, hence not recommended to utilise subsidence value that has not been considered in model development. For linear multi regression method, the lowest accuracy for reserve strength ratio has been 92.1 %. However, both probability of failure and return period shows very low accuracies, therefore, not recommended to be utilized by industry. It has been found that although the analysis model used for validation had a similar configuration, the overall platform surfaces were different, which in turn gave different platform responses. This eventually led to differences in the probability of failures and return periods. Careful consideration is expected prior to adopting the proposed model as it is to be used with platforms, which have similar platform configurations, structural member sizing, water depth and metocean data. The accuracy and effectiveness of the proposed model will generally assist operators in the industry in decision-making and more importantly, in outlining the action items for business risk management in which marine growth, corrosion, and subsidence are expected to occur.