Riser Analysis Methods: Comparison with Measured Field Data

Abstract

ABSTRACT Predictions of riser responses using three time domain riser analysis methods were compared to full scale data. The transfer function, significant wave and random wave methods were used to predict response in the wave active zone and lower portion of an instrumented production riser. Simultaneous measurements of vessel motion, wave height, and riser stresses and rotations were obtained from the Dorada instrumented production riser system, located in the Spanish Mediterranean. Hydrodynamic interactive effects of the vessel hull on the fluid flow field around-the riser have been approximately accounted for. Fluid flow was found to be altered-by as much as 40 percent when considering fluid/vessel interaction. Predicted lower riser dynamic stress response by each method correlated well with measured data. Response at the splash zone, however, is over predicted by the random wave method and under predicted by the transfer function method. With increasing storm severity, the random method tends toward more accurate prediction of measured splash zone bending stresses. Conversely, poorer agreements in transfer function method predictions occur with increasing storm severity. INTRODUCTION Dynamic analysis of marine risers falls into three basic categories:Regular wave time domain analysis,Random wave time domain analysis, andFrequency domain analysis A cimplified hybrid approach can be used for the simulation of riser response in an irregular (random) seaway by combining the methods of (1) and (3) above. In this procedure, a series of regular wave analyses are used to develop response transfer functions which combine with wave spectra in the frequency domain to yield response spectra. This transfer function method is a standard, expedient approach to riser design, but is thought to be less accurate than other methods. A theoretically more precise random vibration analysis is performed by method (Z) above. This approach involves decomposition of the wave surface spectrum and subsequent generation of a "synthetic" wave. Previous studies have compared full scale field measurements from drilling risers to random wave, (Reference 4 and 5) and regular wave (Reference 5) simulations. Measured splash zone bending response was related to bending stress standard deviation using random wave analysis, and bending stress standard deviation using random wave analysis, and bending stress transfer functions using random wave analysis. Riser model tests (Reference 3) have also been conducted. Measured bending stress was compared to that predicted using regular wave analysis. These studies tend to indicate an overestimate of measured data using the random wave approach, and an underestimate of measured data using the regular wave approach. The present study compares analytical riser responses from regular wave, transfer function and random wave approaches to field measurements from an instrumented production riser. Due to their size, the production risers are significantly drag-dominant. Therefore, the effect of nonlinear hydrodynamic drag on riser responses obtained using the various analysis approaches can be assessed. Riser responses near the sea bottom and wave active zone are compared on a spectral and significant basis.