You are herePredicting Fatique Damage Matrices for Floating Structures across Multiple Seastates: the DamMat Routine

# Predicting Fatique Damage Matrices for Floating Structures across Multiple Seastates: the DamMat Routine

This document illustrates the use of the routine DAMMAT. It seeks to facilitate fatigue analysis of floating structure components, such as the tethers of a tension-leg platform. Its output takes the form of damage matrices, which give the mean damage a specific floating platform experiences across a range of steady-state ocean conditions. Separate matrices are produced for a range of assumed fatigue properties specifically, of the exponent *b* which relates the applied stress *S* to fatigue life *N*.

Thus, these matrices conveniently summarize all relevant structural information, including (1) the hull geometry, which governs forces, (2) the tether stiffness, which governs resonant response, and (3) the fatigue behavior , as characterized by the exponent *b*. These matrices do not depend, however, on the long-term relative frequencies of seastate parameters such as significant wave height *H _{s}* , peak spectral period

*T*, and heading. As a result , after a single DAMMAT analysis yields an appropriate damage matrix, long-term damage rates can be estimated for multiple sites by applying appropriate (site dependent) seastate weighting probabilities.

_{p}

DAMMAT also reports separate damage matrices assuming first-order, second-order, and combined first- and second-order diffraction loading. Long-term damage rates are efficiently estimated for each seastate, without time-domain simulation, through the TFPoP routine (Ude et al, 1995). This estimates the variation of both the stress amplitude and the stress cycle rate across different seastates.

Beyond its direct use to estimate long-term fatigue, DAMMAT may also serve as a useful example of how stochastic analysis of floating structure response, via TFPoP, can be automated across multiple seastates. Similar approaches may be useful in other contexts. For example, l00-year extreme responses for floating structures may utilize a shell program similar to DAMMAT, with input seastates derived from an appropriate 100-year *H _{s}-T_{p}* contour (Ude and Winterstein, 1996).