C

Student Competition

938068 - Bayesian Damage Prediction of Berm Breakwaters in the Arctic

Primary Author(s)

• 

  Maria Pontiki, MS

  Ph.D. student
  Department of Civil and Environmental Engineering, University of Delaware
  Newark , Delaware

The shrinking sea ice cover in the Arctic zone over the last decades hints fundamental difficulties at assessing the potential threats on coastal structures. A Bayesian probabilistic model for the estimation of probabilities of failure on berm breakwaters was constructed to deal with the limited data in the Arctic and tackle the uncertainties. The model was trained based on historical data from the Alaskan Beaufort Sea, as well as, expert predictions on breakwater damages. The predictive goodness of the model was examined with log-likelihood tests. Sensitivity analysis showed the importance of each forcing variable in the failure probabilities of the structure. A key finding is that the developed model can predict damage intensities on a berm breakwater by incorporating the non-well investigated wave – ice interactions. The study provides a novel and time efficient tool on the probabilistic design of port defense structures that lie in the Arctic region.