Abstract
Oscillating Surge Wave Energy Converters (OSWECs) aim to harness the energy in ocean waves using a buoyant flap with a bottom hinge that oscillates in response to surge forces. To interpret OSWEC performance in real-word conditions, foundational knowledge is required from simulation and laboratory testing with scaled models. In laboratory experiments to characterize an OSWEC’s hydrodynamics, we need to accurately relate the hydrodynamic response to control actuation. These relations can be confounded by non-linear time lags between the commanded control actuation and the application of this actuation to the flap.
We are in the process of developing an experimental OSWEC. The power take-off is emulated by a regulating motor and gearbox, hydrodynamic forces on the flap are measured by a pair of six-axis load cells, encoders measure flap and motor position, and up to ten pressure sensors can measure flap pressure. Experiments aimed to characterize hydrodynamics and test control strategies will require precise control of flap position. However, the gearbox and shaft coupler in the driveline may introduce differences in rotational position between the motor shaft and the flap. These nonlinearities could arise from backlash or coupler flexion, but the behavior is not adequately documented in manufacturer component specifications. Characterizing this potential nonlinear behavior is a necessary initial step for commissioning the OSWEC.
We designed a test fixture to evaluate positional nonlinearity in the driveline under varying torsional loads. The benchtop fixture secures the following components in-line in the same order as in the experimental OSWEC: a servo motor, gearbox, rigid shaft coupler, encoder, six-axis load cell, bearing, flexible shaft coupler, and a magnetic particle brake. In an inversion of OSWEC operation, the motor simulates hydrodynamic forcing from the flap as a result of interaction with a regular wave profile and the particle brake simulates motor resistance applied to the driveline in response. Because of this, the particle brake resistive torque is specified as a constant linear damping scheme proportional to the shaft rotational velocity. The load cell measures the torque applied by the particle brake, while the encoder position of the motor is compared to the encoder position of the shaft to characterize nonlinearity from the gearbox and rigid coupling. For a range of oscillation amplitudes and resistive torques, we evaluate motor and shaft position alongside driveline torque for each test, defining a shaft lag as the difference between motor and shaft position.
We find that shaft lag is minimal for all experimental conditions evaluated, with 95th percentiles below 3% of the oscillation amplitude. We found a positive correlation between shaft lag and root mean square (rms) driveline torque. The magnitude of lag is similar in both rotation directions, and it has negligible cycle-to-cycle variation. These results indicate that the motor controller estimates for shaft position with no corrections applied will be within a half of a degree of actual flap position during regular operation, and corrected motor controller estimates for shaft position will be within a quarter of a degree. We expect that this positional variation is sufficient for future study of OSWEC hydrodynamics and optimal control schemes.