Three-Dimensional Acoustic Travel-Time Tomography for Wind Energy

Understanding turbulent flow around wind turbines is crucial for their design, control, and numerical model validation. Existing remote sensing methods, relying on light or sound wave backscatter from airborne particulate matter, limit the resolution of spatial and temporal observations near solid objects and reflective surfaces.

Acoustic tomography (AT) presents a novel approach to remote sensing by utilizing atmospheric thermodynamic properties and the time-dependent stochastic inversion (TDSI) algorithm. This enables the reconstruction of velocity and temperature turbulence fields within an array of acoustic transducers, surpassing the limitations of existing techniques.

Despite its success in homogeneous flow fields, AT's potential application in wind energy research remains unexplored due to the nonhomogeneous nature of turbulent flow in wind farms. This study aims to employ a nonhomogeneous model and implement TDSI in a three-dimensional (3D) configuration to reconstruct turbulent flow in the wake of a wind turbine.

The Flatirons Campus (FC) at the National Renewable Energy Laboratory (NREL) hosts a 3D acoustic tomography array designed for renewable energy and atmospheric science research. The array consists of eight 10-m towers, each equipped with three layers of speakers and microphones, as well as one tower supporting three sonic anemometers and Vaisala pressure/temperature/humidity probes for reference atmospheric data. A sub-scale 1.2kW SkyStream wind turbine is located within the AT array. To date, 3D acoustic tomography has been limited to masts no taller than 6-m, making the NREL AT array the first of its kind.

Data is recorded in 0.5s frames, allowing precise travel time calculations for TDSI implementation. Preliminary results, comparing reconstructed flow velocities in the 2D configuration with homogenous background flow to measurements obtained from a sonic anemometer, show promising normalized-mean square error (NMSE) and standard deviations (STD) values of approximately 0.5 and 0.03, respectively. This study will present results of the turbulent field reconstruction in the 3D configuration, including the subscale wind turbine wake.