Abstract
The use of acoustic Doppler current profilers (ADCPs) for the characterization of flow conditions in the vicinity of both experimental and full scale marine hydrokinetic (MHK) turbines is becoming increasingly prevalent. The computation of a three dimensional velocity measurement from divergent acoustic beams requires the assumption that the flow conditions are homogeneous among all beams at a particular axial distance from the instrument.
In the near wake of MHK devices, the mean fluid motion is observed to be highly spatially variable as a result of torque generation and energy extraction. We examine the performance of ADCP measurements through modeling of a virtual ADCP (VADCP) instrument in the wake of an MHK turbine resolved using unsteady computational fluid dynamics (CFD). This is achieved by sampling the CFD velocity field at equivalent locations to the sample bins of an ADCP, and performing the coordinate transformation from beam coordinates to instrument coordinates, and finally to global coordinates.
The stream-wise velocity deficit and tangential swirl velocity caused by the moving rotor lead to significant misrepresentation of the true flow velocity profiles by the VADCP. The most significant errors were observed in the transverse (cross-flow) velocity direction, with a relative error of 17% of the incident velocity. The relative RMS error over the rotor plane was calculate to be less than 2% of the incident velocity by a distance of x/D ≥ 2,x/D and x/d ≥ 6 for the stream-wise, transverse and vertical directions, respectively. For the default instrument orientation, A maximum error in the V-velocity occurs when the instrument is located directly behind the turbine, and a maximum W-velocity error is found at a lateral offset y/D = 0.20. For an instrument directly behind the MHK device, the maximum error in the U-velocity and V-velocity occurs at a yaw angle of Ø ≈ 30° and Ø ≈ 60°.