An efficient importance sampling method for long-term fatigue assessment of deepwater risers with time domain analysis

A floating production system is exposed to many sea conditions over its lifetime. Consequently, fatigue design of a riser system is challenging, owing to the need to consider the long-term wave condition, characterized by the joint statistics of the significant wave height and peak wave period. The problem becomes computationally intractable if dynamic analysis is performed in the time domain, which is often necessary due to nonlinearities. This paper outlines a new efficient approach for long-term riser fatigue analysis, in conjunction with time domain simulations. The time series of an irregular wave elevation is widely simulated as the sum of regular wave components with random amplitudes and phase angles. Importance sampling is the basis of the proposed approach; however, it is not exploited for variance reduction. Instead, this technique allows one to simulate the wave amplitudes from a distribution different from the original one. This implies that following a dynamic analysis, the results for many sea states can be obtained by simply altering the importance sampling weighting function. The proposed approach is enhanced by a succession of additional techniques to reduce the sampling variability. Since importance sampling is not an approximate method, but an alternative means of simulation, the predicted long-term fatigue damage is unbiased. The standard error is estimated via bootstrapping. (C) 2016 Elsevier Ltd. All rights reserved.