The hull geometry of a wave energy converter (WEC) is central to the absorption of wave energy and its conversion to mechanical energy. Reduced-order models for WECs, such as WEC-Sim, rely on boundary element methods to predict excitation, radiation damping, and added mass for a given geometry, but there is limited experimental validation of prediction accuracy. We investigate this in the context of a pitching wave energy converter developed for the 2022 Marine Energy Collegiate Competition.
We consider three distinct edge profiles for an extruded trapezoidal hull envelope: concave, convex, and flat-walled. These profiles were manufactured, instrumented with a motor, encoder, and torque cells, and installed in a wave flume. Radiation, excitation, and free response tests were conducted in which hull torque, pitch, incident waves, and generated waves were measured. From these, the pitch response amplitude operator (RAO) was characterized as a function of wave frequency, as well as excitation, damping, and added mass coefficients.
Results from radiation experiments with each of the hull profiles show that damping torques increase with pitch oscillation frequency and inertial torques decrease with frequency. Excitation torques demonstrate a peak frequency near 0.7 Hz for all three hull profiles. Simulation of the same hull geometries in Ansys AQWA predicts hydrodynamic coefficients with resonant peaks not observed in experiments and excitation torque coefficients with similar magnitude but higher peak frequency than experimentation. In both experiment and simulation, the convex profile has a smaller excitation torque than the other two geometries.
The experimental RAO was determined from test results by two methods: direct extraction from free response tests and a reconstruction via linear superposition from radiation and excitation tests. Comparing the results of these two methods assesses the validity of the linear superposition assumption. We find disagreement between the directly measured and reconstructed RAO values for some frequencies. To compare with simulation, the RAO was also calculated from the amplitude of pitch motion generated in WEC-Sim. Overall, we find that simulations predict a similar order of magnitude response to experiments, but significant points of disagreement are found for all three hull topologies.