VIV Model Test Data Comparison With Shear7 V4.5

The prediction of riser and tendon fatigue damage due to vortex-induced vibration (VIV) remains an active area of research in the offshore industry. In 2003, ExxonMobil performed VIV testing on a 10-m long, 20-mm diameter model in an effort to better understand the mechanics of VIV response of a long flexible pipe. Measured results from these tests, summarized in terms of response frequency, strain, and damage, were published in a series of papers in OTC 2004. Due to the dense array of instrumentation (17 cross flow and 35 inline stations along the riser span) and the use of strain gages in the experiments, the 2003 ExxonMobil data allows for a direct estimation of fatigue response and provides an excellent benchmark for validation of VIV predictions. This paper extends our previous work by comparing the measured test results to simulations of the test conditions (for example, model properties, boundary conditions, and current profiles) using the widely used VIV prediction tool, Shear7. We compared measured and predicted response in terms of a “damage index” for bare, fully straked, and partially straked risers. The damage index, defined as response frequency times the third power of RMS strain ( f×erms 3 ), is proportional to the fatigue life and thus can be used as a basis to assess the accuracy of predictions. For the comparison work presented in this paper, ExxonMobil has utilized a latest version of Shear7 (version 4.5) which provides added functionality and increased user control over analysis assumptions through assignment of key input parameters. In order to investigate the influence of the analysis assumptions on the predictions, a matrix of 64 distinct input parameter sets was defined. The study indicated that the new “time-sharing” model in general can generate prediction results with reasonable bias and scatter for bare risers. However, prediction errors for straked risers are still high even when the most favorable parameters are selected for the analysis. These comparisons highlight the continuing need for improved formulations with smaller prediction bias and scatter and for high quality benchmarking data and benchmarking methods for objective assessment of VIV prediction accuracy.Copyright © 2008 by ASME