NONLINEAR SHIP LOADS AND SHIP FATIGUE RELIABILITY

Stochastic methods are described here to predict the effects of nonlinear ship loads on fatigue accumulation in random seas. These are found to be capable of predicting the net fatigue damage over many ship response cycles, at a fraction of the cost of direct time-domain analysis. It is also shown how these nonlinear ship load models can be used within a full reliability analysis of ship components. This uses first-order reliability methods to estimate not only the net effect of all uncertainty sources, but also the relative contribution of each (e.g., load vs. material property variability). In our ship load application, nonlinear effects are found to increase the relative impact of load (and wave) modeling and its uncertainty. The ship load analysis for fatigue is based on a new concept, the "narrow-band transfer function" (NTF) method. The basic idea is to apply the nonlinear time-domain ship load analysis to a limited set of regular, sinusoidal waves. This establishes the narrow-band transfer function; i.e., the ship load amplitude from nonlinear analysis as a function of the wave amplitude and period. Stochastic process theory is used both to select which regular waves (i.e., heights and periods) to use, and to decide how these results should be weighted in predicting load statistics in an irregular sea.