Investigations with a bandwidth measure for fatigue assessment of the Foinaven dynamic umbilical including VIV

Abstract

An innovative use of a method to detect Vortex Induced Vibration (VIV) bandwidth in the curvature data of subsea flexible risers and umbilicals is presented. The parameter epsilon gives a valuable estimate of the bandwidth of signals such as VIV as a single value which may be used to track the behaviour with time and against other measures such as current speed. The method is conveniently based on a peak counting approach originally used in a marine context by Cartwright and Longuet-Higgins. Low epsilon values close to zero indicate a narrow-band process whereas values near unity indicate a broad-band process. Curvature and environmental data were gathered by the Foinaven Umbilical Monitoring System (FUMS) installed on the Foinaven Petrojarl IV floating production unit located in the Atlantic margin, west of Shetland. The VIV frequency range owing to current excitation considered is that in the range 0.2–2 Hz. This paper presents analyses showing the correlation of the epsilon bandwidth parameter with standard deviation of curvature, current speed and spectral analysis results for extended periods measured on the Foinaven dynamic umbilical. Examination of the number of frequencies and their values using spectrograms over long periods enable interesting interpretation of the build-up and decay of the VIV behaviour. These data are also reduced to the predominant peak frequencies, and standard deviation along with epsilon to provide a compact description of the VIV behaviour over 10 min intervals. Clear frequency and amplitude shifts can be observed as VIV varies with current ebb and flow. The behaviour is usefully described using only epsilon, standard deviation, and peak frequencies. The importance of epsilon for fatigue damage assessment is demonstrated through a formulation derived by Wirsching and Light. Strong VIV regimes are often characterized by an intense activity around a particular frequency and the impact of such relatively narrow-band events on the fatigue life of the structure is clearly demonstrated. This paper also presents the relative contributions of mooring, and waves/vessel motions, and VIV to fatigue damage. Although wave excitation remains the main source of fatigue, VIV appears to potentially contribute to a significant part of the overall fatigue damage.