Evaluating the applicability of linear hydrodynamics across float geometries for heaving point absorbers

Abstract

Accurate characterization of point absorber wave energy converter (WEC) float hydrodynamics is crucial for WEC design as it informs geometry, power take-off, and controls optimization. If the hydrodynamics of the float are consistent with potential flow assumptions (i.e., incompressible and irrotational flow, small amplitude motion), then accurate hydrodynamic characterization can be completed via boundary element method (BEM) simulations. However, certain geometric features and/or float motion could result in nonlinear hydrodynamics. While BEM has been validated against some float geometries, the validation space is narrower than the potential WEC design space. To this end, we compare hydrodynamic coefficients computed using BEM to those calculated from experiments for multiple float geometries moving in heave, with a goal of identifying geometric features and motion that cause experimental results to diverge from potential flow based results.

For initial testing, we considered four float geometries: a compound cylinder with the larger diameter at the bottom (“hat”), a compound cylinder with the larger diameter at the top (“T”), a cylinder with a moonpool (“ring”), and a revolved diamond (based on the optimal geometry from Edwards and Yue [1], “diamond”). We ran BEM simulations for each geometry using WAMIT. 

We conducted forced oscillation experiments in a tank over a range of frequencies and amplitudes for each geometry. From these, we extracted added mass and radiation damping coefficients. A comparable set of excitation coefficient experiments were performed in a wave basin. A comparison of the experimental and BEM values identifies geometries and test conditions that deviate significantly from BEM predictions, as well as amplitude-dependent non-linearity.

Results show that the added mass coefficients computed from experimental data agree relatively well with the values computed by BEM for the ring, diamond, and T profiles. However, the hat profile’s experimental added mass increases with frequency, which is the opposite of the trend for BEM. For all geometries, the added mass coefficient exhibits limited amplitude dependence. The experimentally determined radiation damping coefficients for most geometries are larger than the value predicted by BEM and some are amplitude dependent. This could be partially explained by the convolution of radiation damping with viscous drag for scale-model experiments. The experimental values of wave excitation coefficients for the T, ring, and diamond profiles have limited amplitude dependence, and reasonably agree with BEM. The experimental excitation coefficients for the hat exhibits amplitude dependence and only agrees with BEM for relatively low amplitude wave motion (Keulegan-Carpenter number ~0.2). Overall, these results demonstrate that the applicability of assumptions about potential flow and linearity vary significantly with WEC float geometry. As such, design optimization that relies solely on BEM may fail to identify truly optimal float profiles.

[1] E. C. Edwards and D. K.-P. Yue, “Optimisation of the geometry of axisymmetric point-absorber wave energy converters,” Journal of Fluid Mechanics, vol. 933, p. A1, Feb. 2022.