Identification of linear and non-linear multi-modal VIV responses for flexible deepwater risers

Offshore energy exploration has been moving into ever increasing water depths. For floating offshore drilling structures, the riser system is a crucial element. Vortex-induced vibration (VIV) is a major concern for deepwater riser developments, as vortex-induced vibration is a major cause of riser fatigue damage. For deepwater risers, current is the dominant factor causing VIV responses. Due to the increased water depth, deep-water risers have long and flexible structures, so that they have the potential to be subject to very high modes of vibration, i.e. multi-modal VIV. The frequencies, amplitudes and modes of VIV responses are usually the focus of deepwater riser design, as they, along with riser material properties, directly determine the riser fatigue life. In recent years, much effort has been devoted into the investigation of riser VIV response, but there are still many uncertainties, especially for the risers with multi-mode VIV responses in currents. For example, frequency lock-in phenomena and modal resonances are still not fully understood for multi-modal VIV responses; the vibration shapes over riser length and the motion trajectories in the cross-sectional plane for a flexible riser with multimodal VIV cannot be found in the literature. The frequency and mode components contained in the multi-modal VIV responses in both in-line and cross-flow directions have not been published in previous work. -- This research aims to improve the understanding of multi-modal VIV in currents. The research objectives include i) frequency characteristics for multi-modal VIV responses, such as frequency versus current velocity and frequency lock-in phenomenon; ii) amplitude characteristics for multi-modal VIV responses, such as amplitude versus current velocity, amplitude range and amplitude resonance; iii) spectral characteristics for multi-modal VIV responses, such as dominant frequencies, power spectrum versus current velocity and power spectrum versus location on the riser; iv) modal characteristics for multi-modal VIV responses, such as modal distribution, dominant mode and mode versus current velocity; v) modal system parameters for a flexible riser in calm water, including modal mass, modal damping, modal stiffness and non-linear damping; and vi) the correlation between the modal parameters and the VIV responses. -- After a review of the state-of-the-art literature involving VIV investigation, an experimental method was proposed for this research. Based on a prototype riser, a length distorted model riser was designed with similarity of the mass, the bending stiffness and the frequency ratio. Two model riser tests were designed and conducted. The first one was a shaker-excitation test, which was designed to investigate the modal system parameters. A shaker was used to generate an excitation to the riser, and the riser responses were measured. The modal system parameters were estimated from the frequency response functions based on a simplified governing equation for the shaker/riser system. Modal analysis was used to estimate the linear modal system parameters, and Bendat's technique was used to estimate the non-linear damping for the flexible riser. -- Another model riser test was a current-excitation test. This test was designed to investigate the VIV responses in currents. The uniform currents were generated by towing carriage. Sixteen pairs of accelerometers were used to measure the VIV responses at sixteen locations on the riser. Spectral analysis and modal analysis are two major tools to analyze the measured data. -- It was found that the frequencies of the multi-modal VIV responses in both the crossflow and in-line directions follow the Strouhal frequencies for a flexible riser. The measured Strouhal number was about 0.12. The frequency lock-in phenomena occurred in both the in-line and cross-flow directions at some modal natural frequencies. The vibration amplitudes fluctuate from 0.3 to 0.9D (D denotes the riser diameter) in both the in-line and cross-flow directions for the flexible riser tested, and no increasing trend existed as the current velocity increases. The resonances are not strong as the vibration energy is shared by several modes. The VIV responses for a flexible riser contained a number of modal components, but one or two were the dominant modes. -- It was also found that flexible risers have a variety of natural frequencies, which correspond to a variety of mode shapes. The modal natural frequencies depend on the tension. The estimated added mass coefficients Ca depend on the tensions and mode shapes for a flexible riser, ranging from 1.0 to 3.7, and the estimated damping coefficients Cd have a relatively big scatter, ranging from 0.5 to 2.5. There is a degree of correlation between the estimated modal parameters and the VIV responses