Abstract
Hydrodynamic characterization of floating bodies is a necessary component of wave energy converter (WEC) design and optimization. Many WEC modeling and design tools utilize boundary element method (BEM) simulations to compute hydrodynamic forces. These simulations make linear potential flow assumptions (i.e., incompressible and irrotational flow, small amplitude motion), which have varying accuracy for different WEC floats and operating conditions. Prior studies have experimentally validated BEM for specific geometries, but this validation set is small in comparison to the geometric options for point absorber floats.
To expand the range of validation cases, we considered four, contrasting 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”). For each geometry, excitation coefficients in heave were simulated in WAMIT and Capytaine. For comparison, we conducted wave excitation experiments in a wave flume for a range of incident wave frequencies and amplitudes. Experimental excitation coefficients were derived from the first harmonic of the measured heave force and wave elevation signals.
We observe discontinuous nulls in BEM excitation coefficients for the diamond and hat geometries, even when accounting for irregular frequencies. Such discontinuous nulls have been documented in prior numerical work for hat type geometries and our experimental results validate the nulls for both geometries. Subsequent analysis identified the source of the null as diffraction and Froude-Krylov forces of equal amplitude and opposite phase. Such features could be potentially exploited to reduce excitation in certain sea states with variable-geometry hulls. Similarly, moonpool floats (“ring”) are known to exhibit spikes in excitation forces at the internal resonance frequency, which manifests as a vertical spout in the experimental moonpool. The measured excitation force also exhibits a spike at the moonpool resonant frequency, although smaller than the spike predicted by BEM. Overall, we observe generally good agreement between both BEM codes and experiments, provided that BEM solver options are appropriately configured. An open question from this study is the presence of higher harmonics in experiments that are required to accurately reconstruct time-domain forces (i.e., max/min forces, time history). Further work is required to understand if these are merely experimental artifacts (e.g., reflections from the channel walls) or dynamics relevant to float design that cannot be captured by the linear assumptions required for BEM.