Checking vulnerability to pure loss of stability in long crested following waves: A probabilistic approach

This paper presents a probabilistic methodology for the analysis of the vulnerability of a ship to the risk of inception of pure loss of stability events. A pure loss of stability failure is modelled as the persistence of the metacentric height below a critical level for a too long time. The metacentric height is modelled as a stationary Gaussian process with a spectrum obtained from the sea elevation spectrum. The time dependent failure index is obtained under the assumption of filtered Poisson process for the occurrence of critical events. The analysis separates cases where the fluctuation of the metacentric height is narrow-band from those where the bandwidth of the spectrum is wide, with an intermediate blending. In case of narrow-band processes appropriate approximate solutions to the problem are provided, while in the wide-band cases an exponential distribution for the persistence time below the critical level is employed. A rational development for the critical persistence time is also provided considering an approximation of the roll dynamics during periods of time where the metacentric height is negative. Monte Carlo simulations are performed to check the developed approximate distributions for the persistence time, and examples of application are provided for a sample ship.