| Popis: |
Due to an on-going project's specificities, running time domain analyses using wave and vessel motions timeseries has been required. However due to numerical limitations, preserving time consistency between wave elevation and vessel motions led to unacceptable simulation run time (several days per analysis). To solve this, consistency between wave elevation and vessel motions was discarded. The present paper has two objectives. First, validating the use of a timeseries to represent vessel motions in combination with wave timeseries that are inconsistent from a timestamp perspective (i.e. phasing between vessel motions and wave elevation is incorrect), in a shallow water context (33 m) where direct wave loading on the risers contributes as much as vessel motions. Second comparing the use of a Gumbel distribution versus an empirical distribution to determine levels of non-exceedance. The following methodology was used: Time domain simulations using both consistent and inconsistent vessel motions/wave elevation were performed. Statistical processing was then performed to derive design values from each type of analysis (consistent and inconsistent). Results from each approach were then compared to assess the impact of wave inconsistency on design variables. In a second step, time domain analyses were performed with 20 seeds. Results from these analyses were processed using both a Gumbel distribution and an empirical approach. A comparison between the results obtained with each approach was then performed. The performed analyses put in evidence: Losing the consistency between wave elevation and vessel motion generates an error that lies within the statistical error related to the randomness of the wave process. On that basis, provided a sufficient number of seeds (set at 20 in the frame of this study) is considered and that a proper statistical processing is performed, not preserving the time consistency between vessel motions and wave elevations is deemed acceptable for riser design, even in shallow waters where direct wave loading on the risers cannot be neglected. In addition, the comparison between a Gumbel distribution and empirical distribution provides matching results for the system modelled on the considered project. |
