Bilge Keel Roll Damping: Combining CFD and local velocities:

by M.J. Van Kampen

Institution: Delft University of Technology
Year: 2015
Keywords: bilge keels; roll damping; CFD; FPSO; riser balcony
Record ID: 1254867
Full text PDF: http://resolver.tudelft.nl/uuid:d4c717e8-8090-41ee-a6a0-6c7a243ea836


As hydrocarbon supplies dwindle, technology develops and long-term hydrocarbon prices rise it is becoming more and more economical to develop hydrocarbon fields offshore. An FPSO vessel can be favored as it is flexible, quickly commissioned and cost-effective, be used at all water depths and does not require additional pipelines to shore. To predict the roll motion of an FPSO traditional techniques such as the Ikeda-Tanaka-Himeno (ITH) method are no longer sufficient due to aberrant dimensions and shapes of bilge keels and riser balconies. The goal of this thesis was to provide a practical method to evaluate the roll damping and motions of an FPSO with aberrant bilge keels and/or riser balconies in (ir)regular waves. For this the ITH method was modified in three manners: 1. by extending current formulations with out-of-phase terms, 2. by obtaining relevant coefficients from 2D CFD simulations in forced roll oscillations, and 3. by using linear potential theory to obtain local velocities consisting not only of rigid body velocities, but also radiated wave velocities, incoming wave velocities and diffracted wave velocities. This new methodology was compared to forced oscillation and regular wave experiments performed by MARIN of the model scale Glas Dowr FPSO. Forced oscillations were reproduced satisfactory after a correction for additional heave and sway motions was applied. Regular wave results were compared to measurements, simulations using experimentally derived damping coefficients, ITH method damping coefficients, local velocity-based ITH method coefficients and using the proposed methodology. Results were good when compared to simulations based on measured damping coefficients but inconclusive when compared directly to measured roll amplitudes. Reasonable agreement at low to medium wave amplitudes and an underestimation at high wave amplitudes were obtained. The underestimation at high amplitudes is faulted to the linear increasing hull pressure coefficient while it is more likely to become saturated at higher local velocities. It is concluded that the combination of CFD and local velocities yield promising results and is more flexible than the traditional ITH method. A more thorough validation should be performed against data at various frequencies, hulls, keels and wave amplitudes before application becomes feasible.