The purpose of this paper is to present a method for efficient and unbiased estimation of the long-term extreme response distribution of a catenary riser. In this approach a computationally inexpensive, nonlinear response predictor is used to estimate the response in all sea states, thus allowing selection of relevant sea states and intervals within sea states for detailed, nonlinear finite element simulations. This method requires significantly less simulation time than the conventional approach with extensive nonlinear simulations of many sea states. The method is applied to a catenary riser case. In an earlier study (Passano and Larsen, OMAE 2006), a strong relation was found between the prescribed vertical motions at the top and the axial force and bending moment near the touch down area. This relation was then used to limit simulations to intervals where the largest maxima values were expected. This gave an estimate of the upper part or tail of the extreme response in the selected sea states. In a later study (Passano and Larsen, OMAE 2007), nonlinear response predictors based on the prescribed vertical motions were established. These were used to estimate the short-term extreme response distribution directly. In the case study of this paper three long-term response distributions are compared: 1) A distribution established directly from the predicted short-term response distributions. 2) A distribution based on response simulated in the relevant time intervals of the relevant sea states. 3) A distribution obtained from extensive nonlinear simulations of the sea states in the scatter diagram. It is found that distribution 2) provides a fast and reliable estimate of 3). This allows a significant reduction in analysis work.

This content is only available via PDF.
You do not currently have access to this content.