the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 23 Jan 2024
A Correction Algorithm for Propeller-Induced Airflow and Flight Attitude Changes during Three-Dimensional Wind Speed Measurements Made from A Rotary Unmanned Aerial Vehicle
Abstract. A hexacopter unmanned aerial vehicle (UAV) was fitted with a three-dimensional sonic anemometer to measure three-dimensional wind speed, air temperature, relative humidity, and pressure. To obtain accurate results for three-dimensional wind speeds, we developed an algorithm to correct biases caused by the propeller-induced airflow disturbance, UVA movement, and changes in flight attitude in the three-dimensional wind measurements. The wind measurement platform was built based on a custom-designed integration kit that couples seamlessly to the UAV, equipped with a payload and the sonic anemometer. Based on an accurate digital model of the integrated UAV-payload-anemometer platform, computational fluid dynamics (CFD) simulations were performed to quantify the wind speed disturbances caused by the rotation of the UAV's rotor on the anemometer during the UAV's steady flight under headwind, tailwind, and crosswind conditions. Through analysis of the simulated data, regression equations were developed to predict the wind speed disturbance, and the correction algorithm for rotor disturbances, motions, and attitude changes was developed. To validate the correction algorithm, we conducted a comparison study in which the integrated UAV system flew around a meteorological tower on which three-dimensional wind measurements were made at multiple altitudes. The comparison between the corrected UAV wind data and those from the meteorological tower demonstrated an excellent agreement. The corrections result in significant reductions in wind speed bias caused mostly by the propellers, along with notable changes in the dominant wind direction and wind speed in the original data. The algorithm enables reliable and accurate wind speed measurements in the atmospheric boundary layer made from rotorcraft UAVs.
This preprint has been withdrawn.
-
Withdrawal notice
This preprint has been withdrawn.
-
Preprint
(1133 KB)
-
Supplement
(935 KB)
-
This preprint has been withdrawn.
- Preprint
(1133 KB) - Metadata XML
-
Supplement
(935 KB) - BibTeX
- EndNote
Interactive discussion
Status: closed
-
EC1: 'Comment on amt-2023-248', Ad Stoffelen, 30 Apr 2024
The manuscript is of interest, while I've been informed that the anemometer that is used can not measure 3D velocity. It can not discriminate whether the wind is coming from the front or from below / top. Data that shows this effect is available here: https://amt.copernicus.org/articles/14/1303/2021/amt-14-1303-2021-t01.png . For scientific credibity the authors should address this problem.
Citation: https://doi.org/10.5194/amt-2023-248-EC1 -
AC1: 'Reply on EC1', Yanrong Yang, 13 May 2024
Thank you for the comments. To our knowledge, the image you posted originates from Thielicke et al., 2021. Through careful reading of this article, we have the following responses to your query.
1. Thielicke et al. mentioned in their article that the measurement error of the TriSonica anemometer mounted on unmanned aerial vehicles (UAVs) increases with greater UAV pitch angles. In our study, we conducted both Computational Fluid Dynamics (CFD) simulations and comparative flight experiments, maintaining the UAV's speed within a range of 5-15 m/s. Within this speed range, the pitch angles were consistently between 0-10° during acceleration and relatively steady flight, only exceeding 15° during deceleration. These angles correspond to the low bias range discussed in Thielicke et al.'s study. Moreover, by incorporating the measurement error in the uz direction (approximately 10%) in our analysis, we can further assess the accuracy of wind speed correction for UAVs.
2. In their study, Thielicke et al. reported that the TriSonica device tends to exhibit significant measurement errors in the wind velocity component along the uz direction. In our research, we focused on CFD simulations and comparative flight tests with UAV during level flight at a constant altitude, without considering the UAV's ascent and descent phases. During these level flights, the wind speed measured in the uz direction is relatively low (about -1 ~ 1 m/s). As such, the uncertainty in the uz direction have relatively little impact on our results. Furthermore, by keeping the UAV's tilt angle within 0~10 degrees, the error in the uz direction wind speed was noted to be around -5%. Although this represents a deviation from ideal measurement, it is still within a tolerable accuracy range for our research purposes.
3. The TriSonica device employed in our study was acquired after 2021. By that time, the manufacturer had already implemented a firmware update (version 1.7.0, February 2019). This update further enhanced the measurement accuracy of the TriSonica. The sensor was not the same type as the initial version evaluated by Thielicke et al. in their wind tunnel experiments. Thielicke et al.(2021) stated that “The TriSonica was tested in the wind tunnel in November 2018. After these results were reported to the manufacturer, a firmware update (v 1.7.0, February 2019) addressed the issue of wind shadowing, potentially enhancing the accuracy at zero pitch angle. We had no opportunity to test this firmware in a wind tunnel yet”.
4. This study focuses on the wind speed correctionalgorithm for UAV in a stable flight state, typical of the flight conditions for most large multirotor UAV. To validate the correction algorithm, we conducted validation studies on the TriSonica-UAV combination, first with a self-consistency test by flying box flights at level flights by comparison of data from upwind and downwind flight legs, and more importantly validation through a comparative analysis between data recorded by the TriSonica device mounted on the UAV during stable flight and data from three-dimensional anemometers at multiple altitudes on a meteorological tower. The concordance between our corrected results and the meteorological tower data indicates that the measurement errors associated with the TriSonica-UAV combination to assess wind speeds are within acceptable limits.
5. We conducted measurements of CO2 and CH4emissions from a coking plant in China, employing a combination of the TriSonica-UAV combination, coupled to a mass balance algorithm which rely on accurate 3-d wind measurements. These emissions were subsequently compared with emissions derived from material balance method. The comparison revealed a discrepancy of approximately 6% between the two methodologies. The study has been published in the Atmospheric Measurement Techniques, as shown in Han et al, 2023. Considering that the mass balance algorithm our used principally relies on the wind speed and CO2 measurements, this comparative analysis indirectly validates the relative accuracy of our wind speed measurements. Moreover, it also substantiates the viability of deploying the TriSonica on the UAV for measuring wind speeds.
In our opinion, the studies conducted for the present paper have provided adequate results to show the validity and scientific credibility of the TriSonica-UAV methodology for making wind measurements, noting the larger uncertainty in the uz at large pitch angle. In our revision of the paper, this uncertainty will be added to the discussion. We would like to ask that the paper be sent for review.
Citation: https://doi.org/10.5194/amt-2023-248-AC1
-
AC1: 'Reply on EC1', Yanrong Yang, 13 May 2024
-
EC2: 'Comment on amt-2023-248', Ad Stoffelen, 19 Sep 2024
We are grateful to the authors to attempt publishing in AMT. Unfortunately, as AMT associate editor I've to come to a position where we asked the authors to withdraw their manuscript, following an editor's suggestion.
Unfortunately, we found out over recent months that there is a lack of interest by the AMT community in reviewing the manuscript. It presumably implies a lack of interest in the manuscript by the AMT readership too. Without reviewers we cannot finish the review process.
In addition, one potential reviewer reported back to me that the correction method proposed by the authors cannot work, which indeed is not helpful when searching for reviewers. I did not verify the correction method myself, as usually I need expert reviewers to arrive at such conclusion. In the absence of available and independent expert reviewers I can however not make this assessment and cannot reject the manuscript on this basis. As pointed out, the lack of interested and available reviewers is the main reason to withdraw the manuscript.The authors have accepted this AMT position and plan to withdraw the manuscript.
Nevertheless, we do understand that the search for reviewers took a long time and I must apologize for the state of affairs today on behalf of AMT.Citation: https://doi.org/10.5194/amt-2023-248-EC2 -
AC2: 'Reply on EC2', Yanrong Yang, 19 Sep 2024
Dear Editor,
Could you please inform us about the procedure for withdrawing the manuscript?
Best regards,
Yanrong Yang
Citation: https://doi.org/10.5194/amt-2023-248-AC2
-
AC2: 'Reply on EC2', Yanrong Yang, 19 Sep 2024
Interactive discussion
Status: closed
-
EC1: 'Comment on amt-2023-248', Ad Stoffelen, 30 Apr 2024
The manuscript is of interest, while I've been informed that the anemometer that is used can not measure 3D velocity. It can not discriminate whether the wind is coming from the front or from below / top. Data that shows this effect is available here: https://amt.copernicus.org/articles/14/1303/2021/amt-14-1303-2021-t01.png . For scientific credibity the authors should address this problem.
Citation: https://doi.org/10.5194/amt-2023-248-EC1 -
AC1: 'Reply on EC1', Yanrong Yang, 13 May 2024
Thank you for the comments. To our knowledge, the image you posted originates from Thielicke et al., 2021. Through careful reading of this article, we have the following responses to your query.
1. Thielicke et al. mentioned in their article that the measurement error of the TriSonica anemometer mounted on unmanned aerial vehicles (UAVs) increases with greater UAV pitch angles. In our study, we conducted both Computational Fluid Dynamics (CFD) simulations and comparative flight experiments, maintaining the UAV's speed within a range of 5-15 m/s. Within this speed range, the pitch angles were consistently between 0-10° during acceleration and relatively steady flight, only exceeding 15° during deceleration. These angles correspond to the low bias range discussed in Thielicke et al.'s study. Moreover, by incorporating the measurement error in the uz direction (approximately 10%) in our analysis, we can further assess the accuracy of wind speed correction for UAVs.
2. In their study, Thielicke et al. reported that the TriSonica device tends to exhibit significant measurement errors in the wind velocity component along the uz direction. In our research, we focused on CFD simulations and comparative flight tests with UAV during level flight at a constant altitude, without considering the UAV's ascent and descent phases. During these level flights, the wind speed measured in the uz direction is relatively low (about -1 ~ 1 m/s). As such, the uncertainty in the uz direction have relatively little impact on our results. Furthermore, by keeping the UAV's tilt angle within 0~10 degrees, the error in the uz direction wind speed was noted to be around -5%. Although this represents a deviation from ideal measurement, it is still within a tolerable accuracy range for our research purposes.
3. The TriSonica device employed in our study was acquired after 2021. By that time, the manufacturer had already implemented a firmware update (version 1.7.0, February 2019). This update further enhanced the measurement accuracy of the TriSonica. The sensor was not the same type as the initial version evaluated by Thielicke et al. in their wind tunnel experiments. Thielicke et al.(2021) stated that “The TriSonica was tested in the wind tunnel in November 2018. After these results were reported to the manufacturer, a firmware update (v 1.7.0, February 2019) addressed the issue of wind shadowing, potentially enhancing the accuracy at zero pitch angle. We had no opportunity to test this firmware in a wind tunnel yet”.
4. This study focuses on the wind speed correctionalgorithm for UAV in a stable flight state, typical of the flight conditions for most large multirotor UAV. To validate the correction algorithm, we conducted validation studies on the TriSonica-UAV combination, first with a self-consistency test by flying box flights at level flights by comparison of data from upwind and downwind flight legs, and more importantly validation through a comparative analysis between data recorded by the TriSonica device mounted on the UAV during stable flight and data from three-dimensional anemometers at multiple altitudes on a meteorological tower. The concordance between our corrected results and the meteorological tower data indicates that the measurement errors associated with the TriSonica-UAV combination to assess wind speeds are within acceptable limits.
5. We conducted measurements of CO2 and CH4emissions from a coking plant in China, employing a combination of the TriSonica-UAV combination, coupled to a mass balance algorithm which rely on accurate 3-d wind measurements. These emissions were subsequently compared with emissions derived from material balance method. The comparison revealed a discrepancy of approximately 6% between the two methodologies. The study has been published in the Atmospheric Measurement Techniques, as shown in Han et al, 2023. Considering that the mass balance algorithm our used principally relies on the wind speed and CO2 measurements, this comparative analysis indirectly validates the relative accuracy of our wind speed measurements. Moreover, it also substantiates the viability of deploying the TriSonica on the UAV for measuring wind speeds.
In our opinion, the studies conducted for the present paper have provided adequate results to show the validity and scientific credibility of the TriSonica-UAV methodology for making wind measurements, noting the larger uncertainty in the uz at large pitch angle. In our revision of the paper, this uncertainty will be added to the discussion. We would like to ask that the paper be sent for review.
Citation: https://doi.org/10.5194/amt-2023-248-AC1
-
AC1: 'Reply on EC1', Yanrong Yang, 13 May 2024
-
EC2: 'Comment on amt-2023-248', Ad Stoffelen, 19 Sep 2024
We are grateful to the authors to attempt publishing in AMT. Unfortunately, as AMT associate editor I've to come to a position where we asked the authors to withdraw their manuscript, following an editor's suggestion.
Unfortunately, we found out over recent months that there is a lack of interest by the AMT community in reviewing the manuscript. It presumably implies a lack of interest in the manuscript by the AMT readership too. Without reviewers we cannot finish the review process.
In addition, one potential reviewer reported back to me that the correction method proposed by the authors cannot work, which indeed is not helpful when searching for reviewers. I did not verify the correction method myself, as usually I need expert reviewers to arrive at such conclusion. In the absence of available and independent expert reviewers I can however not make this assessment and cannot reject the manuscript on this basis. As pointed out, the lack of interested and available reviewers is the main reason to withdraw the manuscript.The authors have accepted this AMT position and plan to withdraw the manuscript.
Nevertheless, we do understand that the search for reviewers took a long time and I must apologize for the state of affairs today on behalf of AMT.Citation: https://doi.org/10.5194/amt-2023-248-EC2 -
AC2: 'Reply on EC2', Yanrong Yang, 19 Sep 2024
Dear Editor,
Could you please inform us about the procedure for withdrawing the manuscript?
Best regards,
Yanrong Yang
Citation: https://doi.org/10.5194/amt-2023-248-AC2
-
AC2: 'Reply on EC2', Yanrong Yang, 19 Sep 2024
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 428 | 91 | 362 | 881 | 60 | 47 | 30 |
- HTML: 428
- PDF: 91
- XML: 362
- Total: 881
- Supplement: 60
- BibTeX: 47
- EndNote: 30
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Yanrong Yang
Yuheng Zhang
Tianran Han
Conghui Xie
Yayong Liu
Yufei Huang
Jietao Zhou
Haijiong Sun
Delong Zhao
Kui Zhang
This preprint has been withdrawn.
- Preprint
(1133 KB) - Metadata XML
-
Supplement
(935 KB) - BibTeX
- EndNote