Rare event risk analysis – application to iceberg collision

Arif, Mohammad; Khan, Faisal; Ahmed, Salim; Imtiaz, Syed

doi:10.1016/j.jlp.2020.104199

Cited by 10 publications

(9 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For focus on the tail event, instead of using any random values, the Hill and the SmooHill estimators have been used to estimate the possible parameter range. The 95% confidence interval is estimated for both the Hill and the SmooHill estimate, as explained in; 19 and it has following steps:…”

Section: Bayesian Inferencementioning

confidence: 99%

“…The iceberg risk estimation process follows the same as those in Sulistiyono et al 17 and Arif et al 19 The risk is defined as:…”

Section: Risk Estimationmentioning

confidence: 99%

“…[14][15][16][17][18] In such cases, the EVT based methodology fails to capture most extreme events, and as an alternative, a rare event is considered as a heavy tail event. A tail based approach has been proposed by Arif et al, 19 and considering the 1000 year return period to capture most extreme scenarios. A similar approach was also proposed by Clauset and Woodard 20 The Monte Carlo method 21,22 and the Mixer model 23 have been considered in past research for rare event modelling.…”

Section: Introductionmentioning

confidence: 99%

“…25 To address this issue the Hill 26 and SmooHill 27 estimator are implemented to estimate the tail index. 19 However, all point-based methods does not provide a probability distribution over the possible range of the model's parameter and does not incorporate prior information. 24 To overcome this isse, Bayesian inference has shown some potential in recent research.…”

Section: Introductionmentioning

confidence: 99%

“…28 and a POT based Heavy Right Tail distribution (HRTD) is implemented similar to. 19,20 Bayesian inference is used to estimate the distribution parameter, and finally, the extreme risk is estimated by using the model's prediction of extreme iceberg speed.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Evolving extreme events caused by climate change: A tail based Bayesian approach for extreme event risk analysis

Arif

Khan

Ahmed

et al. 2021

Proceedings of the Institution of Mechanical Engineers, Part O:

Self Cite

View full text Add to dashboard Cite

Natural hazards are of significant concern for engineering development in the offshore environment. Climate change phenomena are making these concerns even greater. The frequency and extent of natural hazards are undesirably evolving over time; so risk estimation for such events require special consideration. In most cases the existing extreme models (based on the extreme value theory) are unable to capture the changing frequency and extremeness of natural hazards. To capture the evolving frequency and extremeness of natural hazards and their effects on offshore process operations, an advanced probabilistic approach is proposed in this paper. The approach considers a heavy right tail probability model. The model parameter is estimated through the Bayesian inference. Hill and the SmooHill estimators are used to evaluate the lowest and highest exponent of the probability model. The application of the approach is demonstrated through extreme iceberg risk analysis for the Jeanne d’Arc basin. This study shows climate change or global warming is causing to appear a significant number of icebergs every year in the study area. Offshore structures are often designed to withstand the impact of 1 MT icebergs weight; however, the study observes large icebergs (10 MT weight) are sighted in recent years (14% of the total number of cited icebergs for the period of 2002–2017). As a result, the design philosophy needs to be revised. The proposed risk-based approach provides a robust design criterion for offshore structures.

show abstract

Section: Bayesian Inferencementioning

confidence: 99%

“…The iceberg risk estimation process follows the same as those in Sulistiyono et al 17 and Arif et al 19 The risk is defined as:…”

Section: Risk Estimationmentioning

confidence: 99%