Experimental investigation of vortex-induced vibration of long marine risers

Abstract

There is considerable disparity between predictions of marine riser vortex-induced vibration (VIV) fatigue damage, and often the agreement between computer models and observed VIV-related damage is inaccurate by orders of magnitude. Resulting problems for deepwater riser design are the need for large safety factors on fatigue damage predictions and the use of expensive vortex-suppression devices (e.g. helical strakes). Understanding is especially limited for long risers, which may be excited in multiple and higher modes. Norwegian Deepwater Programme (NDP)—a group of oilfield licensees in Norway—has commissioned experiments on riser models over a range of scales and current conditions in order to improve the ability to predict VIV. Recently, testing of a model with an L/ D (length-to-diameter ratio) of 1400 was conducted at Marintek's Ocean Basin in Trondheim. The riser was tested without VIV suppression and with various strake arrangements. Testing was performed on an innovative test rig that could simulate uniform and linearly sheared currents, with a composite fibre model that featured a dense array of high-quality instrumentation. In-line and cross-flow responses were considered important with respect to fatigue. Indeed, this study reinforces recently published results that in-line fatigue damage is as severe as cross-flow fatigue damage. However, industry analysis approaches generally ignore in-line damage due to VIV. The findings also indicate that the response character of a bare riser can be quite distinct from that of a riser partially or fully covered with helical strakes. In addition, while helical strakes of different types can be effective in mitigating VIV fatigue of long risers, their performance is dependent on their geometry. Finally, the results suggest that a key consideration in VIV fatigue design is the length of suppression coverage and the nature of the flow to which the bare section of the riser is exposed.