Extreme Value Analysis of Riser Systems for large Water Depths
MetadataShow full item record
- Institutt for marin teknikk 
In relevant Norwegian regulations, the characteristic response is defined as a given annual probability of exceedance. Effective and accurate methods of estimating the response for the required annual probability levels are consequently a vital part of the design. The thesis provides an overview of the riser systems that are currently available, along with a description of the main loads acting on the systems. Further, methods for predicting extreme response are presented. The most consistent method of predicting the characteristic extreme response of a complicated nonlinear system is the all sea-states approach. However, due to it being computationally very demanding, alternative less demanding methods such as the contour plot method have been developed. In this thesis, both methods will be used to determine the 100-year return response value of a riser system consisting of a steel catenary riser connected to a semi-submersible at 1000 meters depth. The value derived from the all sea-states method will serve as a benchmark response level in order to determine the necessary percentile level for the short-term response of the critical sea-state along the contour plot. Additionally, a comparative study of environmental parameters derived from in situ wave measurements versus hindcasting of meteorological data was conducted. The pre-processing of the environmental data and post-processing of the response analysis was executed using MATLAB. The modeling and time domain simulations of the riser systems were conducted using the software package SIMA/Riflex. The required percentile level to account for short-term variability was discovered to be 0.997, which is significantly higher than the recommended NORSOK value of 0.9. The high percentile level is likely due to the large amount of variability carried by the short-term extreme value distribution relative to the slowly varying environmental characteristics, due to the high degree of nonlinearity in the system. The vessels pitch period coinciding with the high-energy wave period is likely to have caused large dynamic effects on the system, consequently increasing the percentile level substantially. Note, the value given by NORSOK is not necessarily representative for the system, but merely meant as a guidance value. The short time step required in the analysis to obtain a stable numerical solution for the worst sea-states also indicates pronounced nonlinear characteristics. The comparative study of in situ wave measurements and hindcast data indicates a slightly higher Hs-peak for the hindcast based contour plot, and a significantly higher Tp-peak for the in situ wave measurements. Further comparative studies are necessary to draw any conclusions, but the hindcast data seem to give sufficiently accurate estimate of the contour in the more critical areas around the high-energy waves.