Probabilistic evaluation of the computational uncertainty in ultimate ship hull strength prediction

Li, Shen and Benson, Simon; Soares, C. Guedes and Santos, T.A., eds. (2021) Probabilistic evaluation of the computational uncertainty in ultimate ship hull strength prediction. In: Developments in Maritime Technology and Engineering. Taylor & Francis, PRT, pp. 567-573. ISBN 9781003216582 (

[thumbnail of Li-Benson-DMTE-2021-Probabilistic-evaluation-of-the-computational-uncertainty-in-ultimate-ship-hull-strength-prediction]
Text. Filename: Li_Benson_DMTE_2021_Probabilistic_evaluation_of_the_computational_uncertainty_in_ultimate_ship_hull_strength_prediction.pdf
Accepted Author Manuscript

Download (1MB)| Preview


Simplified progressive collapse method (Smith method) is codified in the Common Structural Rule (CSR) to calculate the ultimate bending capacity of ship hull girders. However, several benchmark studies have demonstrated a notable uncertainty in predicting the progressive collapse and ultimate limit state of ship hull girders, which is primarily attributed to load-shortening curve (LSC) of local structural components adopted by different participants. In this regard, this paper employs a probabilistic approach to assess the uncertainty of ultimate ship hull strength prediction caused by the critical features of structural component's LSC, e.g. ultimate compressive strength. Probability distribution of ultimate compressive strength estimation of stiffened panels is developed based on a dataset generated by different empirical formulae and nonlinear finite element method. An adaptable LSC formulation, with ability to accommodate different compressive strength of local components, is utilised in conjunction with the Monte-Carlo Simulation procedure where the simplified progressive collapse method is employed to complete the ultimate ship hull strength prediction in each sampling. Case study is conducted on a single hull VLCC and it is found that the CSR-based ultimate ship hull strength prediction follows a Weibull distribution when considering the computational uncertainty caused by different LSC models. The results and insights developed from this paper would be useful to improve the reliability-based design of marine structures.