Operation limits for temporary riser system are determined according to some probability of exceedance of a relevant variable. Accordingly, consistent statistical analysis and probability modelling of the data is required. The common industry approach is to rely on the classical narrow-banded Gaussian process assumption when considering time series of variables of interest. Thus, the time series peaks are characterized by means of the Rayleigh distribution and the relevant extreme values are estimated based on this. However, non-linearities present in riser systems may yield non-Gaussian (wide-banded) processes, rendering the classical approach inappropriate. In the present work, an approximate and practical method is presented to address above issue. It is demonstrated that the approximate method is capable of consistently estimating the relevant extreme values, even where the classical method comes short.

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