Results are presented from two different case studies related to extreme response estimation of tension leg platform (TLP) tendons in stochastic waves. The first investigation is devoted to the influence from non-linearities when analysing stresses m large-diameter tendons. For this case both axial and bending stresses are important, and a non-linear coupling between the two components exists. The results indicate, however, that this nonlinearity can be neglected even under extreme environmental conditions. The second pan of the paper deals with correlation between first and second order surge motions and how extreme' tendon response is influenced by this correlation. Time domain simulations indicate that the two response types are correlated. It is also concluded that by taking correlation into account, the predicted extreme response will be reduced compared to results based on uncorrelated motion components.


When dealing with TLP tendon design one has to estimate lifetime extreme stresses as a pan of a conventional Ultimate Limit State check. Excluding operational aspects that certainly are important, the problem is still quite complex involving environmental statistics and response calculation. In principle, lifetime extreme response must be found from long-term statistics of the response parameter of concern. It is, however, quite laborious to establish this kind of probability distributions, which means that rational short-cuts need to be taken. The long-term statistical problem is hence apparently reduced to the far less complicated short-term problem. The present paper is not a complete discussion on TLP tendon extreme response estimation, but reports results from limited studies of two aspects of this task. For both cases it is assumed that a "design storm approach" can be used. What then is needed is to perform an adequate load effect analysis for the actual seastate and to estimate extreme values from these results.

This content is only available via PDF.
You can access this article if you purchase or spend a download.