A mobile drone-based anemometer system for time-resolved atmospheric boundary layer surveys

Modern commercial wind turbines have average diameters of ~120 m, maximum diameters of ~200 m, and cost millions of dollars each. Atmospheric boundary layer surveys, together with corresponding numerical simulations, are valuable for determining the optimal placement of both large and small wind turbines within large heterogeneous wind farms (especially onshore, in regions with hills). This complex "big data" optimization problem is a matter of economics, balancing the cost of the land to be used and the infrastructure required (including the costs of installing the windmills themselves and the electrical network that connects them, as well as the maintenance of the entire system as it ages) with the average energy that the wind farm produces over a year, during which wind conditions change substantially.

The present research focuses on obtaining distributed atmospheric boundary layer surveys relevant for both wind farm optimization and scientific research. One approach is to build, over a region of interest, scores of 200 m towers with anemometers mounted, and to collect wind data over time. However, that only gives vertical wind profile data at the (sparse) locations of the towers themselves.

Commercial aircraft, whose dynamics are well known, are capable of accurately computing winds aloft as they fly, by comparing airspeed with groundspeed. We investigate here (see Soltaninezhad et al 2025 and Ilyas et al 2025 for recent surveys of related work) doing something similar, called the wind-arc method, leveraging repeated flights of inexpensive drones equipped with cm-level-accurate RTK GPS units.

Our initial feasibility experiments focus on calibrating the attitude of a standard quadcopter executing an accurate position hold to the corresponding (variable) wind field in which it is flying, and quantifying repeatability. With appropriate further calibration, the general approach should extend directly to the case of flying a drone over a preset ground track leveraging feedback control. Both steady and time-varying winds may be determined over an extensive area with this approach, and several drones may be flown simultaneously, in coordinated formations, to determine the vertical profile and horizontal variation of the prevailing wind.