TLP SPRINGING RESPONSE: RELIABILITY AGAINST EXTREMES AND FATIGUE

Vertical springing motions of a tension leg platform (TLP) under random wave loads are studied. Nonlinear diffraction results are used to include second-order sum-frequency loads. Statistics of platform motions and accelerations are obtained. Moment statistics are found through eigenvalue analysis, which recasts the second-order response as a quadratic sum of independent normal random variables. Rates of extremes and fatigue damage are then estimated analytically from a non-Gaussian (Hermite) random process model. These results are weighted over North Sea scattergrams to find long-term extreme and fatigue statistics. In predicting extreme springing responses, it is found useful to consider several design seastate: one with extreme Hs ("wave height driven") and another with resonant second-order effects ("wave period driven"). These are found to provide good agreement with critical seastates found by full FORM analysis. For the Snorre TLP model, fatigue damage rates are underestimated by 60% if second-order effects are neglected. Far greater accuracy is found if second-order effects are retained in response rms alone: non-Gaussian effects cause only 5% additional damage. It is also found that second-order springing effects on fatigue can be considerably greater for more flexible systems.