In-Line Vibrations of Flexible Pipes

Abstract

A semi-empirical prediction tool for pure in-line vortex-induced vibrations is under development. The long-term goal is to be able to realistically model the dynamic behavior of free spanning pipelines exposed to arbitrary time dependent external flows at low velocities. Most VIV programs operate in frequency domain, where only steady currents and linear structural models can be simulated. In contrast, the proposed model predicts hydrodynamic forces as function of time, enabling a time integration scheme to solve the equation of motion. Non-linear time domain simulations allow for modelling of excitation from non-steady currents. In addition, non-linear effects such as soil-pipe interaction, varying tension, and response dependent material, stiffness and damping properties may be included in the analysis, when combining the hydrodynamic force model with a structural non-linear finite element model. Hydrodynamically, the proposed prediction tool consists of the general Morison equation plus two vortex shedding forcing terms. The latter two are able to synchronize with the structural motion for a given frequency band, to induce vibrations in lock-in regimes. In this paper, the proposed pure in-line VIV model is compared to the frequency domain model VIVANA and DNV Recommended Practice, simulating experiments with a model-scale flexible pipe exposed to current velocities at which cross-flow vibrations have not yet developed. A few experimental data points are included in verifying the performance of the newly developed time domain model. The effect of changing empirical coefficients in the vortex shedding forcing terms, and allowing only one of the terms to excite structural vibrations during a simulation, is numerically investigated. A goal is to obtain increased understanding of how the proposed time domain model performs when simulating VIV of a flexible pipe, which is more complex than that of an elastically mounted rigid cylinder since several natural frequencies and corresponding modes might be excited.

Time domain simulation of riser VIV in current and irregular waves

“…The cross-flow vortex shedding force coefficient is taken to be C v,y = 0.85, which is lower than the default value proposed for the pure cross-flow model, but reasonable for such low Reynolds number [19]. The in-line vortex shedding force coefficient is C v,x = 0.8, which is the same value as used by Ulveseter and Saevik [24], for pure in-line vibrations. Default values of the cross-flow synchronization range are utilized, i.e.…”

“…Pure in-line vibrations are not modelled here, since all the following case studies are outside of this regime. However, it is to be noted that a similar synchronization model as stated above has successfully been used for pure in-line vibrations [23,24].…”

Report of Committee III.2: Fatigue and fracture

“…CFD VIM or VIV results are validated versus experimental tests (Maximiano et al, 2017) that do not always predict accurately in situ observations (Koop et al, 2016). For risers, the VIV phenomena are mostly determined based on the empirical model (Ulveseter & Saevik, 2017, Yin et al, 2017 using the Morison equation, but the hydrodynamic parameters are directly calibrated based on tests measurements in a basin (Yin et al, 2015). The results from testing on regular waves with the results of a simulation of a Multi-Body System representing a floating wind turbine, loaded using Computational Fluid Dynamic software were compared in (Beyer et al, 2015).…”