Sound power level properties of hovering UAVs (drones) in a semi- anechoic room Gu-In Oh 1 Bon-Su Koo 2 Jang-Won Lee 3 Dae-Gwan Won 4 Yong-Hee Kim 5 Y’sU Youngsan University 288 Junam-ro, Yangsan, Kyeongnam, South Korea Seung-Soo Lee 6 Korea Conformity Laboratories 7 Nambusunhwan-ro, Seocho-gu, Seoul, South Korea Sang-Ho Kim 7 Konkuk University 120 Neungdong-ro, Gwangjin-gu, Seoul, South Korea

ABSTRACT This paper investigated sound power level characteristics of various unmanned aerial vehicles (UAVs, drones) hovering at 0.5 m in a semi-anechoic room as a preliminary experiment to develop a standardized noise level measurement procedure. Eight types of quadcopter drones were used in the experiment, and its weight with battery ranged 85 g to 5.91 kg. In addition, effects of optional safety guards or low-noise type propeller were evaluated. Firstly, the measurement methods from ISO 3744 and ISO 3746 were considered, however it was not applied due to interference between top microphones and flying UAVs. Therefore, in the experiments, a measurement radius of 3 m and six essential measurement points were selected based on Directive 2000/14/EC for the noise radiation characteristics of equipment used outdoors. A-weighted sound power levels of the tested UAVs were calculated using energy-averaged surface sound pressure levels. As a result, the sound power levels of the tested UAVs ranged 72.8 to 97.5 dBW. The correlation coefficient between the sound power levels and the logarithmic value of the body weights with battery was 0.88. Additionally, the influence of optional parts on noise characteristics of the UAVs were discussed.

1. INTRODUCTION

The unmanned aerial vehicle (UAV, drone) industry, which has begun to be developed for military purposes, is currently the most promising technical field and is widely used to meet private demand

1 gioh@gm.ysu.ac.kr 2 kbs@gm.ysu.ac.kr 3 jwlee@gm.ysu.ac.kr 4 dgwon@gm.ysu.ac.kr 5 yhkim@ysu.ac.kr 6 tmdtnwww@kcl.re.kr 7 woodnol2@korea.kr

for industrial and leisure purposes [1-2]. Unmanned aircraft used in the private sector are collectively referred to as drones. It can be easily found in famous electronic shopping malls and local markets. It is reported that buzzing sounds from drones make people more annoy than the ground vehicles [3]. Although several researchers tried to figure out noise characteristics of drones in various approaches [4-6], there is no relevant standardized procedure to quantify drone noise. Therefore, it is urgent to establish performance standards and management measures to safely manufacture and operate it because it is a small aircraft subject to aviation law and has hazards such as safety accidents, noise from operation, and privacy [1]. ISO/TC20/SC16 deals with the unmanned aircraft systems, and WG5, which was opened in October 2019, is standardizing on testing and evaluation of UAV systems. There were several studies to examine noise characteristics of drones [7-8]. However, no test procedures for noise have yet been established. Since complaints about privacy violations or noise caused by drones are still occurring, standardization of measurement methods and evaluation standards for noise performance needs to be preceded first to revitalize the drone industry.

In this study, sound power level characteristics of UAVs (drones) are investigated in a semi- anechoic chamber with various drone models in terms of a body weight. In addition, survey methods to measure sound pressure levels using sound level meter are suggested for in-situ testing of UAVs. Effects of optional parts such as safeguard, or low noise propeller are discussed.

2. MATERIALS AND METHOD

®

2.1. UAV Specimen Eight types of quadcopter drones were used in the experiment, and its weight with battery ranged 85 g to 5.91 kg. In addition, effects of optional safety guards or low-noise type propeller were evaluated. Figure 1 shows the pictures of the eight quadcopter drones tested in the semi-anechoic condition.

(a) (b) (c) (d)

(e) (f) (g) (h) Figure 1: Pictures of the eight quadcopter drones tested in the semi-anechoic condition with its

weight including battery. (a) Tello; (b) Mavic Mini; (c) Mavic Air; (d) Skydio 2; (e) Mavic 2

Enterprise; (f) Mavic 2 Pro; (g) Phantom 4 Pro V2.0; (h) Custom-made quadcopter with axis

distance of 730 mm.

2.2. Test Methods and Equipment The measurement methods of ISO 3744 and ISO 3746 were considered for deriving sound power level [9]. However, it was not applied due to interference between top microphones and flying UAVs. Therefore, in the experiments, measurement radius of 3 m and 6 essential measurement points were selected based on Directive 2000/14/EC for the noise radiation characteristics of equipment used outdoors [10]. A-weighted sound power levels of the tested UAVs were calculated using energy-averaged surface sound pressure levels. Commission Delegated Regulation (EU) 2019/945 specifies noise test code for the hovering drones at 0.5 m high on reflecting ground in a semi-anechoic room [11]. This procedure follows the requirements on ISO 3744 [9] with ten microphones as shown in Figure 2(a). However, there were some possibilities of physical interference between some unstably hovering drones and the tenth microphone on the top of the measuring hemisphere in a laboratory condition. Accordingly, this study followed the noise test code for measuring airborne noise emitted by equipment for use outdoors with six microphones based on the Directive 2000/14/EC as shown in Figure 2(b) [10].

In addition, survey testing method of sound pressure level using sound level meter was additionally considered for in-situ measurement as shown in Figure 2(c). Basically, hovering height was set to 0.5 m from a reflective ground for laboratory condition. Measurement radius was set to 3 m. All measurement results were expressed in terms of A-weighted sound power levels. As for measurement equipment, real-time FFT analyzer with six condenser microphones (01dB NetdB) and four sound level meters which correspond to classes 1 and 2 (RION NL-52 and NL-42) were employed.

(a) (b) (c) Figure 2: Diagram of the measurement positions for measuring sound power level. (a) Ten microphone positions specified in ISO 3744; (b) Six microphone positions in specified in Directive

2000/14/EC; (c) Four sound level meter positions at 0.5 m height for survey test.

Figure 3: Picture of the semi-anechoic room.

2.3. Experimental Configurations Table 1 shows the experimental configurations using the eight drone models. Variables of propeller type and use of safeguard were considered for effects of optional parts on noise properties. Body weight ranged 49.5 g to 3,270 g without battery and ranged 85 g to 5,910 g with battery. In case of Mavic Mini and Phantom 4 Pro V2.0, effects of safeguard were tested. In case of Tello, safeguard was inseparably installed. Only Phantom 4 Pro V2.0 additionally provided low noise type propeller.

Table 1: The summarized test configurations for quadcopter-type drones hovering at 0.5 m high

No. Drone models Propeller

Use of safeguard

Weight without

Weight with

type

battery (g)

battery (g)

1 Tello Normal Yes 49.5 85

2-1

No 150 245

Mavic Mini Normal

2-2 Yes 200 295

3 Mavic Air Normal No 380 575

4 Skydio 2 Normal No 495 775

5 Mavic 2 Enterprise Normal No 595 885

6 Mavic 2 Pro Normal No 600 900

7-1

No

Normal

915 1,380 7-2 Yes

Phantom 4 Pro V2.0

7-3 Low noise

No

profile

7-4 Yes

Custom-made quadcopter

8

with an axis distance of

Normal No 3,270 5,910

730 mm

3. RESULTS

3.1. Individual Results Figure 4 shows the sound pressure level (SPL) test results of the individual quadcopter drones as a function of one-third octave frequency bands. As shown in Figure 4(a), the Tello showed a peak characteristic of SPL at 50 Hz and 630 Hz around, and dip properties at low frequency bands between 100 to 315 Hz. Mavic Mini showed similar frequency characteristics of SPL to Tello model except for dip properties at 80 Hz around as shown in Figure 4(b) and (c). A comparison between Figures 4(b) and (c) showed that the peak characteristics of SPL were emphasized at 200 Hz when using a safeguard for the Mavic Mini. As shown in Figure 4(d), Mavic Air showed a clear peak frequency at 200 Hz but no significant low frequency component. The Skydio 2 showed similar results to the Mavic Air but middle frequency contents at 400 to 1,250 Hz were emphasized as shown in Figure 4(e). However, the Mavic 2 Enterprise and Mavic 2 Pro model showed relatively flat frequency characteristics of SPL as shown in Figure 4(f) and (g). In case of the Mavic 2 Pro, a deviation of SPL at high frequency bands was observed due to unstable hovering behavior. The Phantom 4 Pro V2.0 showed mountain-shaped frequency characteristics of SPL with a peak at 160 Hz around as shown in Figure 4(h) to (k). As shown in Figure 4(l), the Custom-made quadcopter with an axis distance of 730 mm showed the highest SPL among all tested drones with peaks at 125, 250 and 3,150 Hz and dips at 80 and 200 Hz.

Titit

Titit

(a) (b)

aase8 THLt

(c) (d)

titit

titit

(e) (f)

(g) (h)

titit

titit

(i) (j)

titit

(k) (l) Figure 4: Frequency characteristics of drone noise at 3 m off from hovering at 0.5 m high in semi-

anechoic condition. (a) Case 1; (b) Case 2-1; (c) Case 2-2; (d) Case 3; (e) Case 4; (f) Case 5; (g)

Case 6; (h) Case 7-1; (i) Case 7-2; (j) Case 7-3; (k) Case 7-4; (l) Case 8.

titit

ee eae RR (oor | 2 ik 135 eee Fie Gaeadtguane baadstes

3.1. Sound Power Levels A-weighted sound power levels were calculated from the measured SPL at the six microphones in accordance with the Directive 2000/14/EC. The single number of the calculated A-weighted sound power levels of the tested drones ranged 72.8 to 97.5 dBW. Figure 5 showed the calculated A- weighted sound power levels of the tested drones as a function of body weight without and with battery. Sound power level showed a linear characteristic proportional to the logarithmic values of the body weight for both cases with and without battery weights. Among them, the correlation coefficient was slightly higher when including the battery weight as shown in Figure 5(b). The correlation coefficients between the sound power levels and the logarithmic value of the body weights with battery was 0.88, but 0.87 without battery weight. From the graphs of Figure 5, three drone models of the Mavic Air, the Mavic 2 Pro, and the Mavic 2 Enterprise showed a lower sound power level characteristics than the regression line according to the logarithmic value of the body weight. Therefore, it suggests that reduction of noise level is possible through optimal design even at the same body weight of drones.

In addition, the results averaged from all six microphones showed higher sound power levels by around 1.3 dB than the results averaged from four microphones with the same height as 1.5 m. This is because the propeller was opened vertically in a multicopper-type drone, so even the same measuring radius of 3 m showed higher sound pressure levels at the top points than at four points on the side.

Arweighted sound power leva (8A) ®

(a) (b) Figure 5: A-weighted sound power levels of the tested drones as a function of body weight without and with battery (red figures: averaged from all six microphones, black figures: averaged from four microphones with the same height as 1.5 m). (a) Body weight without battery; (b) Body weight with

battery.

3.2. Survey Tests of SPL at the hovering height Figure 6 shows the relationship between A-weighted sound pressure levels from the four sound level meters at 0.5 m height from the ground floor and horizontally 3 m off from the hovering drone in front, right, left, and rear directions, and A-weighted sound power levels from the six microphones. It was showed very high correlation coefficient values of 0.96 to 0.99 for all orthogonal directions. The results in left and right directions showed relatively higher correlations than those in front and rear directions. Therefore, it seems that it is possible to measure noise levels of the drone using the survey test methods in in-situ field measurements.

Avweghted sound power level (8A) |

(a) (b) (c) (d) Figure 6: Relationship between A-weighted sound pressure levels from the four sound level meters

in front, right, left, and rear directions, and A-weighted sound power levels from the six microphones. (a) Front direction; (b) Right direction; (c) Left direction; (d) Rear direction.

3.2. Effects of Optional Parts on Noise Characteristics Figure 7(a) and (b) shows the effects of optional parts such as propeller profile type and presence of safeguard on noise characteristics. As shown in Figure 7(a), low noise profiled propeller for the Phantom 4 Pro V2.0 yielded a decrease of SPL at higher frequency bands above 1 kHz. In terms of single number quantity, A-weighted sound power level could be decreased by 3.3 to 3.7 dB due to application of low noise type propeller in case of the tested drone. On the other hands, SPL tended to increase slightly due to use of safeguard for the Phantom 4 Pro as shown in Figure 7(b). In terms of single number quantity, A-weighted sound power level could be increased by 0.5 to 0.9 dB due to application of a safeguard in case of the tested drone.

(a) (b) Figure 7: Effects of optional parts on frequency characteristics of SPL for the Phantom 2 Pro V2.0

model. (a) Effect of propeller profile type; (b) Effect of presence of a safeguard. 4. CONCLUSIONS

In this study, A-weighted sound power levels were derived for the eight quadcopter-type drones in a semi-anechoic laboratory condition in accordance with the Directive 2000/14/EC. The measurement methods from ISO 3744 and ISO 3746 were considered, however it was not applied due to interference between top microphones and flying drones. In the experiments, a measurement radius of 3 m and six essential measurement points were selected based on Directive 2000/14/EC for the noise radiation characteristics of equipment used outdoors.

Body weight of the tested drones ranged 49.5 g to 3,270 g without battery and ranged 85 g to 5,910 g with battery. As results, the A-weighted sound power levels of the tested drone ranged 72.8 to 97.5 dBW. Low noise type propeller could decrease noise level by 3.3 to 3.7 dB in comparison to normal type propeller. However, use of safeguard could increase noise level by 0.5 to 0.9 dB. There was strong linear relationship between A-weighted sound power levels and the logarithmic values of the drone’s body weights, especially with battery. Since some drone model showed lower noise levels from the regression line, it reveals that reduction of noise level might be possible through optimal design even at the same body weight of drones. Based on the test results, It is necessary to propose a standardized procedure for noise measurement methods.

so Seana ae a08 sys 0avefeqvney bands

5. ACKNOWLEDGEMENTS

This research was supported by a civil-military Standards Standardization Project with the support of the Defense Agency For Technology and Quality with the funds of the Defense Acquisition Program Administration. (grant number:21-DU-UNI-03) 6. REFERENCES

1. Kim, Y. H. Standardization of Noise Measurement Methods for Unmanned Aircraft, Magazine

of Noise and Vibration , 27 (4) , 11–17, (2017). 2. Torija, A. J., Li, Z., Self, R. H. Effects of a Hovering Unmanned Aerial Vehicle on Urban

Soundscapes Perception, Transportation Research Part D , 78 , 102195 (2020). 3. Christian, A., Cabell, R. Initial Investigation into the Psychoacoustic Properties of Small

Unmanned Aerial System Noise, Proceedings of AIAA Aviation Technology, Integration, and Operations Conference (AVIATION 2017) , Denver, CO, (2017). 4. Rydlewicz, W., Blaszczak, J. R. Experimental Investigations of Acoustic Characteristics of Two

Counter-Rotating Rotors, Mechanics and Mechanical Engineering , 20 (4) , 489–497, (2016). 5. Kloet, N., Watkins, S., Clothier, R., Acoustic Signature Measurement of Small Multi-rotor

Unmanned Aircraft systems, International Journal of Micro Air Vehicles , 9 (1) , 3–14, (2017). 6. Cerniglia, A., Feo, G. D., Marzi, A., Elia, L., Vece, M. Some Experiences of Autogenerated

Noise aboard a Drone, Proceedings of Euronoise 2018 , Crete, Greece, (2018). 7. Zhang, Y., Lee, I. C., Lin, D. Measurement of Noise from a Moving Drone Using a Phased

Array Microphone System, Proceedings of the 2017 Asia-Pacific International Symposium on Aerospace Technology , Seoul, Korea, (2017) 8. Didkovskyi, V., Korzhyk, O. Kozeruk, S., Kozak, A., Kostiuk, R., Liakhevych, S. Noise

Measurement of the Multicopter UAV, 2019 IEEE 5th International Conference Actual Problems of Unmanned Aerial Vehicles Developments (APUAVD) , Kiev, Ukraine, 22–24 October, (2019). 9. ISO 3744, Acoustics — Determination of sound power levels and sound energy levels of noise

sources using sound pressure — Engineering methods for essentially free field over a reflecting plane , International Organization for Standardization, (2010). 10. Directive 2000/14/EC - Noise - Equipment for use outdoors, Annex III: Method of measurement

of airborne noise emitted by equipment for use outdoors , European Union Law, (2022). 11. Commission Delegated Regulation (EU) 2019/945, Unmanned aircraft systems and on third-

country operators of unmanned aircraft systems, Part 13 – Noise test code , (2019).