Reliability Analysis of Wake-induced Riser Collision
MetadataShow full item record
- Institutt for marin teknikk 
Reliability-based design of a riser array requires accurate estimation of extremes of the stochastic relative motion process between pairs of risers in order to avoid collision. Hydrodynamic interference is considered to be one of the most critical drivers in relation to riser collision. This thesis develops a practical approach for assessment of the collision probability between flexible risers subjected to combined current and waves, by accounting for the hydrodynamic interference. The focus of the thesis is on a pair of flexible risers in tandem. Thus, only wake interference is considered. Two semi-empirical static wake models which account for the wake interaction between individual risers in steady current, i.e. the Blevins model and the Huse model, are introduced and modified in order to be applied in the near wake field. These models are further implemented in the finite element software Riflex in order to calculating the drag force acting on the downstream riser. Due to the high nonlinearity, time domain analysis is required. However, it is impractical to simulate several collision events which require long simulation time. Thus, the riser collision process, represented by the normalized minimum distance between the risers, is formulated as a random process so that extreme value analysis can be applied. In total, six methods for prediction of the extreme value distribution are evaluated, and the results are compared in terms of accuracy, the required number of simulations and the selection of the thresholds. Finally, an approach for estimating the collision probability is developed by considering the uncertainties associated with a number of basic parameters. The selected variables are supposed to significantly affect the riser behavior, i.e. drag coefficient , current velocity , response amplitude operator , platform horizontal offset and riser diameter . The relationship between these variables and the responses are described by using the response surface method. Reliability analysis is conducted by both first/second order reliability methods and Monte Carlo simulation technique. The results emphasize the importance of considering the inherent uncertainties in relation to riser clearance assessment.