the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
07 Jan 2021
07 Jan 2021
Distributed wind measurements with multiple quadrotor UAVs in the atmospheric boundary layer
- 1Deutsches Zentrum für Luft- und Raumfahrt e.V., Institut für Physik der Atmosphäre, Oberpfaffenhofen, Germany
- 2Deutscher Wetterdienst, Meteorologisches Observatorium Lindenberg – Richard-Aßmann-Observatorium, Lindenberg, Germany
- 1Deutsches Zentrum für Luft- und Raumfahrt e.V., Institut für Physik der Atmosphäre, Oberpfaffenhofen, Germany
- 2Deutscher Wetterdienst, Meteorologisches Observatorium Lindenberg – Richard-Aßmann-Observatorium, Lindenberg, Germany
Abstract. In this study, a swarm of quadrotor UAVs is presented as a system to measure the spatial distribution of atmospheric boundary layer flow. The big advantage of this approach is, that multiple and flexible measurement points in space can be sampled synchronously. The algorithm to obtain horizontal wind speed and direction is designed for hovering flight phases and is based on the principle of aerodynamic drag and the related quadrotor dynamics. During the FESST@MOL campaign at the Boundary Layer Field Site (Grenzschichtmessfeld, GM) Falkenberg of the Lindenberg Meteorological Observatory - Richard-Aßmann-Observatory (MOL-RAO), 76 calibration and validation flights were performed. The 99 m tower equipped with cup and sonic anemometers at the site is used as the reference for the calibration of the wind measurements. The validation with an independent dataset against the tower anemometers reveals that an average accuracy, regarding the root mean square deviation, of σrms< 0.3 m s-1 for the wind speed and σrms, ψ < 8° for the wind direction was achieved. Furthermore, we compare the spatial distribution of wind measurements with the swarm to the tower vertical profiles and Doppler wind lidar scans. We show that the observed shear in the vertical profiles matches well with the tower and the fluctuations on short time scales agree between the systems. Flow structures that appear in the time series of a line-of-sight measurement and a two-dimensional vertical scan of the lidar can be observed with the swarm and are even sampled with a higher resolution than the deployed lidar can provide.
Tamino Wetz et al.
Status: final response (author comments only)
-
RC1: 'Comment on amt-2020-471', Anonymous Referee #1, 25 Feb 2021
General Comments
In this study, the ability of quadrotors to measure wind speed and wind direction through a relationship with the pitch angle is discussed. While the methodology has already been shown to work well in previous studies, the novel part is the use of multiple quadrotors simultaneously. The authors demonstrate clearly, through comparison with in-situ anemometers along a 99-m tower and a wind lidar, that the use of ‘swarms’ of drones has a large potential for atmospheric boundary layer studies. The research is presented well, but there are some comments and concerns that need to be addressed.
Specific Comments
1) The word ‘swarm’ in the context of drones is used a lot throughout the manuscript without defining it. Does it mean two or more drones? Or is there another minimum threshold that he authors define as ‘swarm’?
2) line 83: Explain why the flight kinematics of a multirotor with four rotors is more simple than multirotor with more rotors. This may not be obvious to the average reader of this article.
3) line 85: The argument of choosing a ‘racing’ drone is not very convincing. There are many settings in the Pixhawk software that one can change/tune to make the multirotor more agile. It is not necessary to select a racing drone in that sense.
4) Page 4: the argument about weight being below the threshold for permits/license requirements, is that still true for the new EU regulations? Or does Germany perhaps still have their own specific regulations?
5) To make a fair comparison between quadrotors vs other multirotors with more than 4 rotors, one should also include some argument against quadrotors. For example: one may argue that the ability to respond to changes in wind speed is better for more than 4 rotors, and that in case of motor failure of one of the rotors, having more than 4 rotors would be desirable. More arguments can be thought of and should be included.
6) line 97; include some details of the relevant sensors, including IMU, GPS, HYT271, etc. e.g. accuracy, precision, etc. Best to do that in separate table.
7) line 104: Did the authors compare wind directions from quadcopter in weathervane mode, vs. quadrotor pointing in one direction and calculating wind direction from pitch and roll (a more common methodology in previous studies)? I would think that in conditions with quite turbulent, varying wind directions, as often encountered in the lower ABL, the weathervane mode may not be optimal.
8) line 107: is there a reference to the software “QGroundControl”. Who developed it, how many drones can be controlled by the software? Some more details of this software need to be included.
9) As for air traffic regulations/permits, is one single operator allowed to control more than one quadrotor? If so, is there a limit?
10) line 130: Sentence a bit unclear. I think that at least, you have to replace “tips of booms”, “to tips of three booms”. I understand that at each level, there are three booms and each of these booms has one anemometer. Is that correct? Is that also the case for the wind vanes at the two heights?
11) The sampling rate of anemometers need to be mentioned somewhere. Later on in the paper, I believe it is mentioned that only 1-minute cup data were available which I was a bit surprised about.
12) line 137; 76 flights with how many quadrotors each? Give a number or range of numbers here.
13) line 141: ‘hovering for a certain time’. Be more specific. Give at least a time range in minutes.
14) line 147: “A safety distance of 20 m to the tower was chosen”. That seems quite large, given the possible spatial variability of winds and lack of eddy coherence on those scales. How does that distance translate into a distance from cup anemometer and/or vane? Also, was there a test done to estimate the minimal distance to avoid disturbance of anemometer measurements by the drone?
15) Page 8, Eq. 9-11: Quite a few variables/parameters are not defined, including p,q,r, and angles phi and theta.
16) Section 5: in the general methodology, I am unclear if the pitch/wind speed relationship is based on time-averaged data or on the higher frequency (1Hz for sonic, 1 minute for cup anemometer) data. This needs to be clarified somewhere, and compared with methodology in previous studies.
17) line 231: “Once the offset is determined”. It is not clear how the offset is determined. Please explain.
18) line 233: Given the large choice of available minimization approaches, can you say a few words about how the solution depends on the specific minimization approach used?
19) line 234: “time-averaged wind speed”. Since the time average changes from flight to flight and from drone to drone (or is it exactly 10 minutes?), I think there is an issue with a fair comparison here, and the time averaging needs to be normalized somehow. The comparison between cup wind speed and drone wind speed improves with the time averaging (as with other comparisons between independent sensors) and the impact of a varying time-average should be removed in some way. One possibility is to split up all the data in e.g. 1 or 2 minute averages.
20) Table 2: similar comment to previous one: unless accuracy from each drone is calculated with identical time averaging periods, the comparison is not entirely fair. At least some discussion of this issue needs to be included.
21) line 260-262: can the authors include the analytic relationship they have derived, perhaps in an Appendix? For comparison with past and future studies, this would be very helpful.
22) line 277: “Flight #31 is selected as the calibration flight as it has an average wind speed of 6 ms−1”. Why is 6 m/s average so good to aim for, for a calibration flight? Isn’t it more important that the wind speed has both a large average AND a large variance. Otherwise, there may be a potential bias for covering relationship between pitch and the larger wind speeds only.
23) line 291: what typical values or range of values did you encounter for the “offset calibration of the yaw angle” for these drones?
24) Figure 7: Several outliers (at both high and low wind speeds) for the UAV data are clearly visible. Can you discuss?
- AC1: 'Reply on RC1', Tamino Wetz, 29 Mar 2021
-
RC2: 'Comment on amt-2020-471', Anonymous Referee #2, 01 Mar 2021
The paper describes the procedure to use multiple drones for free field measurements. It is well written and includes many details on the drones itself as well as on the methodology for calibrating the drones. Measurements for different flight patterns for the drones are compared to measurements from a met mast with cup and ultrasonic anemometers as well as lidar systems. Even though the analysis is not at its limit, the results are already very promising. Nevertheless I do have a few questions and comments.
- The term „swarm“ suggests that the drones are somehow communicating and that one drone-path depends on the path and reaction of the other drone, which is not the case. I would suggest to use another term (unfortunately I do not have a better idea).
- Would it be possible to use perform the calibration in the a wind tunnel und laminar wind flow? Even though the presented calibration function seem to be linear I can imagine that the ambient turbulence and gusts in the wind field might result in an overshoot in the control system which can bias the parameters. In the paper the authors used the 10 minutes averaged data right away — what happens if they use lets say 2 minutes averaged data or 5 minute averaged data for their calibration? Will that increase the error? By looking at shorter time windows the amount of calibration data will automatically increase and maybe will give additional insight in the dynamic response of the drones.
- The authors mention that they are interested in capturing small-scale structures with an array of drones, which is a very nice idea. Doing that, they should say something about the smallest scale they can resolve with the drones, which is about 0.25m for a single drone, but what about the arrangement of multiple drones? What is the minimum spacing between the drones in horizontal and vertical direction so that the drones do not „feel“ the effect of the neighbouring drone? I can imagine that this could be an issue in the vertical direction due to the downwash of the drones. Can the authors comment on that?
- figure 7, b) and c): the vertical positions of the cup and the quadrotors should be marked and maybe separated. It is hard to identify e.g. seven time series in figure 7 c) and it is hard to see which measurement represents which range in vertical direction.
- AC2: 'Reply on RC2', Tamino Wetz, 29 Mar 2021
Tamino Wetz et al.
Tamino Wetz et al.
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 264 | 128 | 10 | 402 | 6 | 4 |
- HTML: 264
- PDF: 128
- XML: 10
- Total: 402
- BibTeX: 6
- EndNote: 4
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1