Optimization Design and Atomization Performance of a Multi-Disc Centrifugal Nozzle for Unmanned Aerial Vehicle Sprayer

Abstract

The nozzle is a crucial component in unmanned aerial vehicle (UAV) sprayers. The centrifugal nozzle offers unique advantages; however, there is a scarcity of published research regarding the structural parameters, spraying parameters, and practical applications specifically for UAV spraying. Furthermore, there is a need for UAV-specific nozzles that demonstrate high efficiency and excellent atomization performance. In this present study, a multi-disc centrifugal nozzle (MCN) capable of controlling droplet size was designed and optimized. The droplet size spectra with different atomizing discs were tested, and indoor and field tests were conducted to investigate the atomization and spray deposition characteristics of the MCN. It was found that the MCN with six atomizing discs with a curved groove, a disc angle of 120°, and a disc diameter of 77 mm demonstrated better atomizing performance. The volume median diameter was 96–153 μm, and the relative span was 1.0–1.3. Compared with the conventional hydraulic nozzle, this nozzle increased the effective spray swath width from 2.5–3.0 m to 4.0–5.0 m and promoted the average deposition rate by 132.4% at a flying height of 1.0 m and a flying speed of 3.0 m/s, which tends to raise the operation efficiency by four to five times. This study can provide a reference for the design and optimization of centrifugal nozzles for a UAV sprayer and the selection of operating parameters in aerial spraying operations.

1. Introduction

Unmanned aerial vehicle (UAV) sprayers have experienced rapid development worldwide in recent years, especially in China and other Asian countries. China’s plant protection operations are carried out in complex terrain, on small plots of land, where efficient plant protection machinery is not popular. Traditional pesticide application methods suffer from low operational efficiency and poor application effectiveness, leading to issues such as personnel exposure to toxins, pesticide residues, and environmental pollution [1,2]. With the advantages of safety, convenience, high mobility, high operational efficiency and applicability to various complex terrains, UAV sprayers have experienced nearly 10 years of rapid development and practical application in China [3,4,5]. As of 2023, China’s UAV sprayer market has reached more than 200,000 units, with an operating area of about 1.42 million ha [6]. Notably, UAV sprayers produced by Chinese companies, represented by DJI and XAG, have been exported to many countries in Asia, Europe, and North and South America. By the end of June 2024, UAV sprayers have already treated more than 500 million hectares of farmland around the world [7]. Global policies are gradually opening up to UAV sprayers, showing that the use of UAV sprayers for pesticide application has significant advantages over ground-based application equipment in terms of safety and other aspects [8,9].

Over the past five years, the application of the UAV sprayer in rice, corn, wheat and other crops in China has become more reliable, with the result that there is more and more related research, including the determination of application parameters, the use of adjuvants to enhance performance, and the evaluation of spraying efficacy [10,11,12,13,14]. With improvements in UAV flight control technology and the demands of practical application, the operational scenarios for UAV sprayer application have been gradually expanded to fruit trees and other economic crops [15,16]. In order to meet the needs of fruit tree application, the load capacity of the UAV sprayer has been significantly increased, and the existing main models can carry a load of up to 50–60 L. In addition, studies have shown that in application by UAV sprayer on fruit trees, fine droplets are mainly deposited on the lower and middle parts of the canopy [17]. The droplets with narrow droplet spectra are concentrated in the droplet size, which makes it easier to control the volume of droplets and reduce droplet drift [18]. Hence, the demand for high flow rate, fine droplets, narrow droplet spectra, adjustable droplet size has increased. Under the conditions of high-volume liquid application, elevated pesticide concentrations, and the mixing of multiple pesticides, the hydraulic nozzles with small pore sizes used for pesticide application are prone to clogging. This not only diminishes operational efficiency but can also lead to potential damage from the pesticides. Consequently, the hydraulic nozzles in UAV sprayers are increasingly being replaced by centrifugal nozzles.

The process of centrifugal atomization entails the uniform distribution of a pesticide liquid to the periphery of an atomizing device through the application of high-speed centrifugal force. Consequently, the liquid is propelled away from the device’s edge. The combined effects of friction and shear from the surrounding air then disperse the liquid into a homogeneous mist of fine droplets [19]. This technology is widely applied across various industries, including agriculture. In pesticide spraying, centrifugal nozzles have the unique capability to atomize high-viscosity liquids, emulsions, and suspensions, resulting in a narrowly defined droplet spectrum. Furthermore, the droplet size can be controlled by adjusting the rotational speed of the atomizing disc, offering significant advantages over traditional hydraulic nozzles. Since the 1980s, researchers have continued to explore the atomization characteristics of centrifugal nozzles. Bode et al. investigated the droplet distribution pattern, droplet size and droplet drift of the Micromax centrifugal nozzle at certain speeds, flow rates, heights and mounting angles [20]. Derkson et al. investigated the droplet size, distribution uniformity and drift potential of centrifugal nozzles at different rotational speeds and flow rates, and found that the uniformity of centrifugal nozzle droplet distribution was related to the operating parameters and spray liquid [21]. Alock et al. proposed an energy equation to predict the exit velocity of centrifugal nozzle droplets and conducted a comparative analysis of the droplet distribution patterns of rotating cup, inverted rotating cup, and multilayer planar atomizing disc centrifugal nozzles. The findings revealed that the droplet distribution from multilayer centrifugal nozzles was notably more homogeneous and demonstrated a broader range of variation [22]. Guo et al. investigated the static spraying characteristics of centrifugal nozzles with a variable spraying system. The findings indicated that the rotational speed of the centrifugal nozzle exerted a considerable influence on the droplet size and the distribution of droplet deposition [23]. Gao et al. used UAV sprays to control wheat sucking pests and figured out that deposition density and control efficacy with centrifugal nozzles were superior to those with hydraulic nozzles [24]. Crause et al. evaluated the effect of flight altitude and rotational speed of the centrifugal nozzle of the UAV sprayer on the effectiveness of fertilizer application to coffee plants and obtained the optimal operating parameters [25].

As the use of centrifugal nozzles in UAV sprayers continued to rise, researchers designed UAV-specific centrifugal nozzles and evaluated their spray characteristics. Zhou et al. studied how the structural parameters of rotary cup atomizers for UAV sprayers influenced atomization properties [26]. Yang et al. developed a dual-atomizing centrifugal nozzle for UAV spraying and investigated its structural optimization and atomization performance across different rotational speeds and flow rates. Their research revealed that factors such as rotational speed, flow rate, tooth shape, and the number of teeth significantly influenced the nozzle’s atomization performance [18]. Hu et al. designed a centrifugal electrostatic spray system for UAV sprayers that was mounted on a UAV for testing. The system was found to improve penetration and increase droplet deposition and uniformity on the crop [27]. Wang et al. designed an aerial electrostatic spraying system for UAVs and mounted it on a UAV sprayer for pear tree spraying tests. The results showed that the system could promote deposition in the lower canopy of pear trees [28]. In designing centrifugal nozzles for UAV sprayers, while some theoretical studies have been conducted, most commercially available nozzles are developed by individual manufacturers without a unified standard to guide their production and usage. This lack of standardization results in significant variability in the atomization characteristics and operational performance of these nozzles. In addition, there is a lack of research on the relationship between the flow rate of the centrifugal nozzle, the speed of the motor, the characteristics of the atomizing disc, and the resulting droplet spectrum, as well as research on controlled droplet technology. Consequently, some centrifugal nozzles performed inadequately in field applications, even producing results worse than those of traditional hydraulic spraying systems, thereby failing to meet their intended design and operational objectives.

In this present study, a multi-disc centrifugal nozzle (MCN) was designed and optimized for UAV sprayers. Furthermore, the effects of both structural and operational parameters on the atomization characteristics of the nozzle were investigated, and the relationship between spray parameters and atomization performance were clarified. In addition, practical spraying tests were conducted by integrating the nozzle into a six-rotor UAV sprayer. The results of this research are expected to provide reliable support for the design and scientific application of centrifugal nozzles in UAV sprayers.

2. Materials and Methods

2.1. Design of Multi-Disc Centrifugal Nozzle

At present, the applied volume of UAV spraying for field crops and fruit tree crops is normally 15–45 L/ha and 75–300 L/ha, respectively, and previous studies showed that the deposition performance and control efficacy were better at the corresponding volumes [11,29]. Compared with hydraulic nozzles, centrifugal nozzles have better deposition effects [30], and centrifugal nozzles have become a trend in UAV application.

In this present study, a centrifugal nozzle with multi grid-type atomizing discs (Figure 1a) was designed, and its structure is shown in Figure 1b. The nozzle used a central liquid supply method with a brushless hollow shaft external rotor motor driving the atomizing disc to reduce the influence of the centrifugal nozzle inlet position on the uniformity of droplet deposition distribution from the nozzle. In addition, the atomization disc was designed as a structure with a narrow top and a wide bottom and a downward opening, and the liquid conducting grooves were distributed on the outside of the atomization disc. The liquid is atomized through the liquid-conducting grooves on the outside of the atomization disc; it not only acquires an initial horizontal velocity but also gains a vertical velocity, which makes the droplets reach the canopy faster and reduces the risk of drifting. The inside of the atomizing disc is smooth without grooves, which can play the role of baffle and make the liquid spread evenly. When the flow rate is large, the single-layer grid format nozzle droplet spectrum tends to be wide. In this regard, to accommodate the demands of high-flow atomization, the nozzle was engineered to allow overlap between the top and bottom layers of the multi-layer atomizing disc structure. The liquid delivery pipe features multiple outlets at various locations on the atomizing disc. As the spray flow rate increases, liquid enters the atomizing disc through a hollow shaft and is distributed among the different atomizing discs for effective atomization.

Figure 2a shows the outer rotor brushless motor with hollow axle. There are four 3 mm diameter screw holes in the bottom of the outer rotor, 8 mm from the center of the motor’s hollow shaft, and the four points are co-circular and evenly dispersed around the disc. The outer rotor has a diameter of 35.0 mm, a height of 8.0 mm, a motor KV of 2020, a no-load voltage of 7.4 V, and a no-load current of 2.4 A. The stator hollow shaft in the motor is the infusion tube, the inner diameter of the infusion tube is 2.0 mm. In order to install different layers of spray discs, the part of the infusion tube that extends into the spray discs was designed as a detachable structure (Figure 2b). The bottom of the infusion tube has an internal thread and the detachable infusion tube has an external thread, which can be screwed and fixed to the stator of the brushless motor. Grid-type atomizing disk is designed as a narrow top and wide bottom structure, Figure 2c are the outside and inside structure of the atomizing disk, a single atomizing disk shape for a certain angle and diameter of the conical disk. As Figure 2c, the top layer atomizing disk 1 and the bottom layer atomizing disk 3 overlap the top and bottom to get a single layer of centrifugal nozzle, in the middle of the 1 and 3 to increase the atomizing disk 2 can get a two layer of centrifugal nozzle, and so on, to get the centrifugal nozzle in different layers as Figure 1a. The top layer of the atomizing disk is actually a baffle plate, the middle and bottom layer of the outside of the atomizing disk for the grid-type liquid guide groove, that is, the outside of the atomizing disk of Figure 2c concave-curved grooves, liquid guide groove to play the role of dividing the film, the inside is a smooth surface. Except for the bottom disc, the center of the middle and top discs has circular holes for the liquid delivery tube to pass through. There are four 3 mm diameter screw holes in the bottom of each disc and in the outer rotor of the motor. The discs are secured to the motor rotor with four M3 hex bolts. For the sake of preventing the set screws from reaching inside the motor and entangling the coils, copper washers were attached to the screws to provide clearance between the motor coils and the discs.

2.2. Droplet Size Measuring System

The droplet size measuring system is shown in Figure 3, with an electric diaphragm pump (12 V, 0~3 L/min) (EFT Electronic Technology Co., Ltd., Hefei, China) supplying the test system, and the flow rate was measured by a digital display flow meter (Digmesa AG, Ipsach, Switzerland). The wireless control system for MCN includes a 3S brushless electronic speed controller (ESC) (Shenzhen Hobbywing Technology Co., Ltd., Shenzhen, China), a Futaba R2008SB wireless receiver and a Futaba T14SG wireless remote controller (Futaba Corp., Yotsukaido, Japan). The radio receiver and the brushless ESC were connected to the centrifugal nozzle. Then the remote control, the power and paired receiver were turned on, and then the remote control was used to remotely control the motor speed of the centrifugal nozzle after successful pairing. The motor speed was measured by UT372 non-contact handheld laser digital tachometer (Unidux Technology Co., Ltd., Guangdong, China), with a range of 0~99,999 rpm, a test distance of 5~20 cm, and an adjustable sampling frequency of 4~200 Hz.

Droplet size spectra were measured using a laser particle size analyzer (Spraytec, Malvern Panalytical Ltd., Malvern, UK) according to ISO 25358 standard [31]. The spray solution was tap water and the droplet size measurement lasted for 10 s and was repeated five times for each nozzle. The final droplet size parameters, 10th percentile diameter (D_v10), volume median diameter (VMD, D_v50), 90th percentile diameter (D_v90), and relative span (RS) were obtained for further analysis.

2.3. Structural Optimization of MCN

The 3D printing technology was used to produce different structures of grid-type atomizing discs. Its main specifications are as follows: the thickness of the atomizing disc is 8 mm, the distance between the screw holes and the center of the disc is 8 mm, the number of liquid guide tanks is 45, and the inner diameter is 25 mm. The range of the test flow rate was 0.6~1.6 L/min, with a gradient of 0.2 L/min, and the rotational speed of the discs was 6000, 7000, 8000, 10,000, and 12,000 rpm, respectively. The droplet size spectra of nozzles with different structures were tested at different flow rates and disc speeds, and the best structured atomizing discs were selected. Liquid guide groove shape, disc angle, disc diameter and the number of layers in the atomizing disc can have an impact on the centrifugal nozzle atomization performance. To this end, the atomizing disc’s liquid guide groove was designed for three kinds: the line groove, curved groove and no groove; the disc angle was designed for 120°, 135°, 150°; the disc diameter was designed for 52 mm, 66 mm, 77 mm; the number of layers for the disc was designed for 1~6 layers.

2.4. Spray Deposition Distribution of MCN

The experiment was conducted at the Center for Chemical Application Technology, China Agricultural University, Beijing, China. Deionized water was used as the spray solution, a 9 cm diameter petri dish was used as the droplet collector, the test flow rate was selected as 1.0 L/min, and the rotational speed of the centrifugal nozzle was 10,000 rpm. The centrifugal nozzle was mounted on a height-adjustable rack, and the deposition distributions of the centrifugal nozzle were tested by adjusting the height of the rack to four heights, namely, 0.5, 1.0, 1.5, and 2.0 m. As shown in Figure 4, a coordinate system was established with the center of the turntable projected on the floor, and petri dishes were placed in front, back, left, right, left front, left back, right front, right back, a total of eight directions in sequence, with a measurement width of 1.5 m in each direction and a distance of 0.15 m between the petri dishes. Ten petri dishes were placed, and the mass of droplet deposition was weighed using an electronic balance to record the mass of droplet deposition after 5 min of spraying each time.

2.5. Field Spray Test

2.5.1. Features of Unmanned Aerial Vehicle Sprayer

The UAV sprayer used in field tests was an M8A electric octocopter drone produced by Beijing TT Aviation Technology Co. Ltd. The UAV was equipped with a spray boom and 6 hydraulic hollow-cone nozzles TR80-0067 (Lechler GmbH, Metzingen, Germany). The boom length is 3.18 m, and the nozzle distances are 0.98 m or 0.55 m. The UAV is equipped with a flight control system with RTK differential positioning function, and the flight speed of the UAV can be controlled to an accuracy of 0.10 m/s. The layout of the UAV’s wingspan structure is a circular symmetrical structure, with a symmetrical motor axis distance of 1630 mm, a rotor diameter of 760 mm and a rotor width of 2390 mm; eight symmetrically arranged brushless motors are the power source for the corresponding rotors attached to the motors, and the flight control system controls the rotational speeds and directions of each rotor, thus, by combination, providing upward lift to the fuselage, controlling the direction of movement of the fuselage and maintaining the balance of the fuselage.

2.5.2. Test Site and Methodology

The test was conducted in Machikou town, Changping district, Beijing, China (40°11′30″ N; 116°10′10″ E) with an open lawn and clear and windless weather conditions. The test setup is shown in Figure 5. Fifteen droplet collectors were arranged perpendicular to the UAV flight path, which were symmetrically distributed along the axis of the UAV flight path, and organized from left to right according to the forward direction of the UAV as No. 1 to No. 15, with No. 8 in the middle, No. 3 to No. 13 with a spacing of 0.5 m, and the remaining 4 sampling points at a distance of 1.0 m from the adjacent points, with a total width of the sampling area of 8 m. The Testo350-XL (Testo SE & Co. KGaA, Schwarzwald, Germany) environmental analyzer sensor was placed away from the UAV flight area and at a height of 2.0 m above the ground to measure ambient wind speed, temperature, and relative humidity. At each sampling point, one sheet of water-sensitive paper (WSP, Syngenta Crop Protection AG, Basel, Switzerland) and one sheet of polyethylene droplet collector were attached with double-ended clamps. The droplet collectors were placed 1.0 m above the ground to simulate the top of the crop canopy.

There were two types of nozzle arrangements for the UAV. The hydraulic nozzle arrangement was based on the configuration of the UAV itself, which was equipped with six TR80-0067 nozzles, with a boom/rotor = 1.33. The centrifugal nozzle arrangement was as follows: the length of the boom was 1.9 m, the boom/rotor = 0.84, and the two centrifugal nozzles were located at each end of the boom. The flight height (from the droplet collector) was set to 1.0 m. The flight speed was set to 1.0 and 3.0 m/s. There were four sets of tests, and each set of tests was repeated three times. At the end of each flight, the WSP and polyethylene droplet collectors were collected, numbered, placed in self-sealing bags, and returned for analysis.

2.5.3. Meteorological Conditions

The meteorological conditions during the test period are shown in

Table 1. Flight parameters and environmental parameters in each test.

Test No.	Nozzle Model	Flight Velocity/(m·s−1)	Temperature/°C	Humidity/%	Wind Speed/(m·s−1)
T1	TR80-0067	1.0	30.50 ± 0.30	29.00 ± 1.12	1.37 ± 0.26
T2	TR80-0067	3.0	32.45 ± 1.18	27.90 ± 2.07	1.35 ± 0.35
T3	MCN	1.0	33.60 ± 0.22	23.45 ± 1.80	1.50 ± 0.20
T4	MCN	3.0	34.10 ± 0.52	24.93 ± 1.61	1.27 ± 0.19

. The average natural wind speed was 1.2–1.5 m/s, the average temperature was 30–35 °C. The environmental meteorological parameters were in accordance with the requirements of the ISO 24253-1 standard [32] for field spray droplet deposition tests.

2.6. Sample and Data Analysis

Based on the Dv₁₀, the Dv₅₀ and the Dv₉₀ measured by the laser particle size analyzer, the RS of the droplets can be calculated using Equation (1):

$$\mathrm{RS}={\displaystyle \frac{{\mathrm{D}\mathrm{v}}_{90}-{\mathrm{D}\mathrm{v}}_{10}}{{\mathrm{D}\mathrm{v}}_{50}}},$$

(1)

where RS is the relative span; Dv₉₀ is the 90th percentile diameter; Dv₁₀ is the 10th percentile diameter; Dv₅₀ is the volume median diameter.

A self-sealing bag containing filter paper was filled with 20.0 mL of deionized water, and the samples were shaken at 200 rpm for 5 min using an orbital shaker (TS-1000, Kylin-Bell Lab Instruments Co., Ltd., Haimen, China) to ensure that the tracer was completely dissolved in the water. The tracer concentration in the samples was measured by setting the absorbance wavelength of the iMark Microplate Absorbance Spectrophotometer (Bio-Rad, Hercules, CA, USA) to 505 nm. Based on the absorbance of each sample, the amount of spray deposition per unit area can be calculated using Equation (2).

$${\mathsf{\beta }}_{\mathrm{d}\mathrm{e}\mathrm{p}}={\displaystyle \frac{{(\mathsf{\rho }}_{\mathrm{s}\mathrm{m}\mathrm{p}\mathrm{l}}-{\mathsf{\rho }}_{\mathrm{b}\mathrm{l}\mathrm{k}})·{\mathrm{F}}_{\mathrm{c}\mathrm{a}\mathrm{l}}·{\mathrm{V}}_{\mathrm{d}\mathrm{i}\mathrm{l}}}{{\mathsf{\rho }}_{\mathrm{s}\mathrm{p}\mathrm{r}\mathrm{a}\mathrm{y}}·{\mathrm{A}}_{\mathrm{c}\mathrm{o}\mathrm{l}}}},$$

(2)

where ${\mathsf{\beta }}_{\mathrm{d}\mathrm{e}\mathrm{p}}$ is the spray deposition in μL·cm⁻²; ${\mathsf{\rho }}_{\mathrm{s}\mathrm{m}\mathrm{p}\mathrm{l}}$ is the tracer reading of the sample; ${\mathsf{\rho }}_{\mathrm{b}\mathrm{l}\mathrm{k}}$ is the tracer reading of the blanks (collector + dilution water); ${\mathrm{F}}_{\mathrm{c}\mathrm{a}\mathrm{l}}$ is the calibration factor; ${\mathrm{V}}_{\mathrm{d}\mathrm{i}\mathrm{l}}$ is the volume of dilution liquid for the tracer from the collector in L; ${\mathsf{\rho }}_{\mathrm{s}\mathrm{p}\mathrm{r}\mathrm{a}\mathrm{y}}$ is the spray concentration, or amount of tracer solute in the spray liquid sampled at the nozzle in g·L⁻¹; and ${\mathrm{A}}_{\mathrm{c}\mathrm{o}\mathrm{l}}$ is the projected area of the collector for capturing spray deposition in cm².

The WSP samples were scanned using an EPSON DS-1610 (Seiko Epson Corporation, Suwa, Japan) scanner to produce photographs that were imported into the macro DepositScan programmed in ImageJ software V1.38x (National Institutes of Health, Bethesda, MD, USA) to determine the droplet deposit density (deposits/cm²).

All statistical analyses were performed using SPSS Statistics for Windows 10 (IBM Corp., Armonk, NY, USA) and all figures were plotted using Origin 2022 (Origin Lab Corporation, Northampton, MA, USA). Dunn’s multiple comparison test within a multifactorial ANOVA was employed to assess whether there were significant differences in VMD and RS across varying rotational speeds, flow rates, and nozzle structural parameters (p = 0.05) [33], and Duncan’s test in a one-way ANOVA was used to compare whether there was a significant difference between droplet deposition for different nozzles (p = 0.05).

3. Results

3.1. Structure Optimization of MCN

3.1.1. Groove Shape

Three types of atomizing discs—featuring a 135° angle, a diameter of 77 cm, and fluid-conducting grooves designed to be straight, curved, or absent—were selected to create a four-layer centrifugal nozzle. Droplet size tests were then conducted at various flow rates and rotational speeds of the atomizing discs. As shown in

Table 2. Droplet size tests were performed on multi-disc centrifugal nozzles with different grooves at different flow rates and atomizing disc speeds to obtain the volume median diameter (Dv₅₀) and relative span (RS). The significance results obtained from multifactorial analysis of variance (ANOVA) were performed on Dv₅₀ and RS.

Information Value	Dependent Variable	DF	p Value	Sig
Groove shape	Dv50	2	0.4605	NS
	RS	2	2.72476 × 10−4	*
Flow rate	Dv50	5	0.06162	NS
	RS	5	1.90524 × 10−5	*
Rotational speed	Dv50	4	3.99405 × 10−14	*
	RS	4	0.15183	NS

Note: * indicates that the difference between the totals is significant at the p < 0.05 level, NS indicates that the difference between the totals is not significant at the p < 0.05 level. DF: degree of freedom.

, according to the results of multifactorial ANOVA, the groove and flow rate had no significant effect on Dv₅₀ and a significant effect on RS, and the atomizing disc rotational speed had a significant effect on Dv₅₀ and no effect on RS. As shown in Figure 6, according to Dunn’s multiple comparison test, the RS of the droplets from the curved groove nozzle was significantly different and relatively smaller than the RS values of the other two nozzles. Comparing the distribution of RS from the nozzles with three different grooves, the RS of the curved groove nozzle and the linear groove nozzle were less affected by the flow rate. Furthermore, with the change in flow rate, the change in RS was relatively more concentrated. Therefore, the atomization performance of the nozzle with a curved groove was the most effective among the designs tested.

3.1.2. Disc Angle

Three kinds of curved groove atomizing discs with a diameter of 77 cm and a disc angle of 120°, 135° and 150° were used to explore the effect of disc angle on the droplet spectra, and the number of layers of MCN atomizing discs during the test was four. As shown in

Table 3. Droplet size tests were performed on multi-disc centrifugal nozzles with different disc angles at different flow rates and atomizing disc speeds to obtain the volume median diameter (Dv₅₀) and relative span (RS), and the significance results obtained from multifactorial analysis of variance (ANOVA) were performed on Dv₅₀ and RS.

Information Value	Dependent Variable	DF	p Value	Sig
Disc angle	Dv50	2	0.5651	NS
	RS	2	6.38205 × 10−7	*
Flow rate	Dv50	5	0.19688	NS
	RS	5	0.00773	*
Rotational speed	Dv50	4	3.7681 × 10−13	*
	RS	4	0.65482	NS

Note: * indicates that the difference between the totals is significant at the p < 0.05 level, NS indicates that the difference between the totals is not significant at the p < 0.05 level. DF: degree of freedom.

, according to the results of multifactorial ANOVA, disc angle and flow rate had no significant effect on Dv₅₀ and a significant effect on RS, and atomizing disc rotational speed had a significant effect on Dv₅₀ and no effect on RS. As shown in Figure 7, according to Dunn’s multiple comparison test, the RS of the droplets at the disc angle of 150° was significantly higher than the RS of the droplets at the disc angle of 120° and 135°. Moreover, there was no significant difference in the RS of the droplets at the disc angles of 120° and 135°. However, when comparing the RS distributions of the droplets under different spray parameters, the RS distribution of the droplets was more concentrated when the disc angle was 120°. Therefore, the MCN with an atomizing disc angle of 120° had the best atomizing performance.

3.1.3. Disc Diameter

Three kinds of curved groove atomizing discs with a disc angle of 150° and disc diameters of 52, 65 and 77 mm were selected for the test, with four layers of atomizing discs. As shown in

Table 4. Droplet size tests were performed on multi-disc centrifugal nozzles with different disc diameters at different flow rates and atomizing disc speeds to obtain the volume median diameter (Dv₅₀) and relative span (RS), and the significance results obtained from multifactorial analysis of variance (ANOVA) were performed on Dv₅₀ and RS.

Information Value	Dependent Variable	DF	p Value	Sig
Disc diameter	Dv50	2	1.61312 × 10−4	*
	RS	2	0.02862	*
Flow rate	Dv50	5	0.0832	NS
	RS	5	9.40613 × 10−6	*
Rotational speed	Dv50	4	1.14237 × 10−9	*
	RS	4	8.13246 × 10−5	*

Note: * indicates that the difference between the totals is significant at the p ≤ 0.05 level, NS indicates that the difference between the totals is not significant at the p ≤ 0.05 level. DF: degree of freedom.

, according to the results of multifactorial ANOVA, the flow rate did not have a significant effect on the Dv₅₀ of the droplets. In contrast, both the disc diameter and the rotational speed of the atomizing disc had a significant impact on the Dv₅₀. Additionally, all three factors significantly influenced the RS of the droplets. As shown in Figure 8, according to Dunn’s multiple comparison test, when the disc angle was 77 mm, the average Dv₅₀ was the smallest, and the difference from the other two groups was significant; when the disc angle was 52 mm, the RS of the droplets was the largest, and the difference in the RS of the droplets when the diameter of the atomizing disc was 77 mm was significant. As shown in Figure 8a, the Dv₅₀ of the droplets of the atomizing disc with a disc angle of 77 mm was about 125 μm on average, and its Dv₅₀ could reach at least 80 μm, which may meet the requirement of small and medium particle sizes in practical applications. In addition, Figure 8b shows that the RS of the droplets was minimized when the disc diameter was 77 mm. Therefore, the MCN with 77 mm diameter atomizing disc had the best atomizing performance.

3.1.4. Disc Number

The curved groove atomizing disc with a disc angle of 150° and a disc diameter of 77 mm was selected to carry out the droplet atomization performance test for the different layers in the nozzle; the number of layers in the atomizing discs to be tested was one to six. As shown in

Table 5. Droplet size tests were performed on centrifugal nozzles with different numbers of atomizing discs at different flow rates and atomizing disc speeds to obtain the volume median diameter (Dv₅₀) and relative span (RS), and the significance results obtained from multifactorial analysis of variance (ANOVA) were performed on Dv₅₀ and RS.

Information Value	Dependent Variable	DF	p Value	Sig
Atomizing disc number	Dv50	5	0.02268	NS
	RS	5	7.85068 × 10−4	*
Flow rate	Dv50	5	0.20192	NS
	RS	5	2.06622 × 10−11	*
Rotational speed	Dv50	4	4.262 × 10−31	*
	RS	4	1.66184 × 10−9	*

Note: * indicates that the difference between the totals is significant at the p < 0.05 level; NS indicates that the difference between the totals is not significant at the p < 0.05 level. DF: degree of freedom.

, according to the results of multifactorial ANOVA, the number of layers in the atomizing disc and the flow rate did not have a significant effect on the Dv₅₀ of the droplets; the rotational speed of the atomizing disc had a significant effect on the Dv₅₀, and all three factors had a significant effect on the RS of the droplets. As shown in Figure 9, according to Dunn’s multiple comparison test, the RS of the droplets was relatively smaller and more concentrated when the number of layers in the atomizing disc was six, and the RS of the droplets of the single-layer atomizing disc was the largest on average. Hence, it turned out that the MCN with six layers of atomizing discs had the best atomizing performance.

3.2. Atomization Characteristics of MCN

3.2.1. Atomization Characteristics

According to the optimization results, the MCN composed of six overlapping layers of atomizing discs with curved grooves, an angle of 120°, and a diameter of 77 mm was used to test its droplet size spectra at different atomizing disc rotational speeds and flow rates. As shown in

Table 6. Droplet size tests were performed on multi-disc centrifugal nozzles with optimal configurations at different flow rates and rotational speeds to obtain the volume median diameter (Dv₅₀) and relative span (RS), and the significance results obtained from multifactorial analysis of variance (ANOVA) were performed on Dv₅₀ and RS.

Information Value	Dependent Variable	DF	p Value	Sig
Flow rate	Dv50	5	0.34339	NS
	RS	5	2.28396 × 10−4	*
Rotational speed	Dv50	4	1.44708 × 10−4	*
	RS	4	0.59701	NS

Note: * indicates that the difference between the totals is significant at the p < 0.05 level, NS indicates that the difference between the totals is not significant at the p < 0.05 level. DF: degree of freedom.

, according to multifactorial ANOVA, flow rate has a significant effect on RS and no significant effect on Dv₅₀; atomizing disc speed has a significant effect on Dv₅₀ and no significant effect on RS. According to the results shown in the graphs in Figure 10, as the speed of the atomizing disc increases, the droplet size decreases, as shown in Figure 10a. At each speed, the distribution of droplet size with flow rate changes are more concentrated; in practice, we can choose the speed of the atomizing disc according to the need. As shown in Figure 10b, when the flow rate is less than 1.2 L/min, the flow rate does not have a significant effect on the RS of the droplets; however, when the flow rate continues to increase, there is a notable rise in the RS of the droplets. Combined with the above analysis, the best application flow rate below 1.2 L/min was selected for practical applications.

The optimized MCN droplet size was combined with the rotational speed and flow rate from the multiple linear regression analysis to obtain the relationship between its Dv₅₀ and rotational speed ω and flow rate q model as Equation (3). It turned out that in the range of 0.6~1.6 L/min flow rate and 6000~12,000 rpm rotational speed, when the MCN works, the rotational speed increases, the droplet size decreases, and the droplet size increases with the increase in flow rate, and the coefficient of determination R² is 0.965, which is a better fitting effect.

Dv₅₀ = −0.00697ω + 20.23q + 164.60, R² = 0.965

(3)

3.2.2. Deposition Distribution

According to the test results for the MCN in the cloud diagram of the height of the droplet deposition distribution, in Figure 11, the atomization area exhibits an approximately hollow conical distribution. In this region, droplet deposition first increases and then decreases as one moves from the center outward, with the deposition area primarily consisting of a ring radius of 1.05 m. At each height, the spray width of a single nozzle measures 2.1 m. When the height of the nozzle above the collector is set at 0.5 m, the amount of droplet deposition within the spray width is significantly greater than at the other three heights; however, the uniformity of deposition within the effective spray width is poor. When the nozzle is 1.0 m, 1.5 m and 2.0 m from the collector, the droplet deposition distribution is almost identical in other positions except for the central part, which is slightly different, and the deposition uniformity is relatively good.

3.3. Spray Performance of UAV Sprayer with MCN

A comparison of the droplet deposition of the two nozzles under different flight speeds is shown in Figure 12,

Table 7. Average droplet deposition of 2 models of nozzles with effective spray swath under different flight velocities by unmanned aerial vehicle (UAV).

Flight Velocity/(m·s−1)	Deposition (μL·cm−2)
	Hydraulic Nozzle	Multi-Disc Centrifugal Nozzle
1.0	0.197 ± 0.027 b	0.305 ± 0.061 a
3.0	0.071 ± 0.010 b	0.165 ± 0.026 a

Note: Data in the table are mean ± standard deviation, and different lowercase letters after the numbers in the same row indicate significant differences at the p < 0.05 level by Duncan’s new complex polarity test.

. Under different flight parameters, the deposition of both nozzles shows the regularity of deposition more below the UAV’s track, and the deposition gradually decreases as the distance to both sides increases. The highest deposition in the deposition volume graph, in Figure 12, was shifted to the right side of the track, probably due to the crosswind. The faster the UAV flight speed, the lower the deposition amount; at the same flight speed, the MCN deposition amount was higher than the hydraulic nozzle, and the closer to the UAV course directly below, the greater the difference between the two. The deposition amount when the UAV was installed with the MCN to operate at a flight speed of 3 m/s was similar to that when the hydraulic nozzle was installed to operate at a flight speed of 1 m/s, which indicates that the deposition performance of the MCN is good. In addition, the operational efficiency can be improved by about three times from the perspective of flight speed with the same deposition effect. As shown in Table 7, the MCN can significantly increase the average droplet deposition within the effective spray width (54.8% and 132.4%, respectively) and enhance the droplet deposition effect at both flight speeds at a flight height of 1.0 m from the top of the droplet collector.

The range of droplet deposition density above 15/cm² is deemed as the effective spraying range. The effective spraying ranges of UAVs equipped with hydraulic nozzles and MCN at different flight speeds can be obtained as shown in Figure 13. At a flight altitude of 1.0 m, the effective spraying ranges of UAVs equipped with hydraulic nozzles are about 2.5–3.0 m. In terms of MCN, the effective spraying range decreases from 5.0–5.5 m to 4.0–4.5 m with the increase in flight speed, but it is still significantly higher than that of the hydraulic nozzle.

4. Discussion

4.1. Structural Optimization and Atomization Characteristics of MCN

It has been demonstrated that the structure of the centrifugal nozzle has a significant effect on its atomization characteristics [18,34], a pattern that is also evident in the results of this study. As shown in Table 2, Table 3, Table 4, Table 5 and Table 6, the groove shape, angle of the atomizing disc, and number of layers in the atomizing disc primarily affect the atomization performance of the MCN by influencing its RS [22]. While the impact on droplet size is not significant in this test, the diameter of the atomizing disc does have a marked effect on Dv₅₀. Specifically, as the diameter of the disc increases, the Dv₅₀ of the droplet size decreases. Furthermore, when there is a substantial difference in disc diameters, the RS tends to decrease with increasing disc diameter. This phenomenon indicates that the multilayer atomizing disc cannot change the droplet size, and the disc diameter has a significant effect on the droplet size, which is in line with the existing studies [35]. This phenomenon may be attributed to the increase in the diameter of the atomizing disc, which results in a larger circumference. A greater number of droplets can be dispersed across the atomizing disc, facilitating rapid atomization. This design effectively addresses the issue of excessive liquid droplets accumulating at the edge of the atomizing disc, which can hinder quick atomization before they are ejected from the disc. Compared with other nozzle structures, with the curved groove, smaller angle of the atomizing disc, larger diameter of the disc and a six-layer atomizing disc, its RSs are significantly reduced. The essence is to increase the path of liquid droplets moving inside the nozzle so that they are more uniformly dispersed inside the nozzle, resulting in a reduction in droplet RS. There is no significant difference in the droplet size of the MCN at different flow rates, and the flow rate mainly affects the atomizing performance of the MCN by influencing its relative span. The rotational speed of the atomizing disc mainly affects the droplet size [23,27,34]. As the rotational speed of the atomizing disc increases, the droplet size decreases. Therefore, if we want to develop a larger flow rate under the atomization performance of the centrifugal nozzle, it can be achieved by increasing the droplet movement path inside the nozzle; if we want to reduce the droplet size, it can be achieved by increasing the circumference of the atomizing disc; this can also be explored to make the atomizing disc speed increase. Nonetheless, the power consumption of the method is smaller.

We optimized the structure of the designed MCN. The results showed that this was the centrifugal nozzle with curved groove, atomizing disc angle of 120°, and an atomizing disc diameter of 77 mm. Also, six atomizing disc layers gave the best atomizing performance. As shown in Figure 10a, this nozzle has the characteristics generally found in centrifugal nozzles, i.e., the nozzle speed increases and Dv₅₀ decreases [23,27,34]. In the 6000–12,000 rpm nozzle speed range, its Dv₅₀ minimum was up to 95 μm, maximum up to 155 μm. In existing studies, researchers have designed centrifugal nozzles with a Dv₅₀ mostly below 120 μm, and it has been verified that the designed centrifugal nozzles have better deposition performance at this droplet size [28,36]. As shown in Figure 10b, in the range of flow rate 0.6–1.6 L/min, the droplet distribution of the MCN is more concentrated, and the RS is kept in the range of 1.0–1.3, which is less affected by the flow rate change, and the atomization performance is good. According to Equation (3), the droplet size of the MCN under different flow rates and rotational speeds can be effectively predicted; as it is difficult to accurately measure the droplet size when the UAV is operating in the field, and according to the relationship model, the droplet size can be regulated to guide the UAV to apply pesticide in the field, thus the purpose of controlling the droplet size can be achieved.

4.2. Deposition Characteristics of MCN

According to Figure 11, the MCN single nozzle deposition distribution is approximated as a hollow cone table distribution; the amount of droplet deposition from the center outward first increases and then decreases. At a nozzle height of 0.5 m, the droplet deposition in the hollow cone area is significantly higher than in other spray width positions; however, the deposition uniformity is poor. This is likely due to the low nozzle height, which causes the droplets to deposit on the collector before they can fully disperse. Conversely, when the nozzle height exceeds 1.0 m, the uniformity of deposition within the spray pattern improves and remains relatively consistent across varying nozzle heights. In this regard, it is recommended that when using the nozzle, the UAV flight height should be higher than 1.0 m. According to its indoor deposition distribution characteristics, the installation of the nozzle needs to consider the reasonable intersection of the deposition area. To this end, the installation position of this nozzle on the stick of the UAV needs to be further studied to achieve the optimal deposition effect. Chen et al. optimized the number of nozzles and the position of nozzles for the six-rotor crop protection UAV by CFD simulation [37]. The results showed that the deposition coverage was increased by 23.25% and the drift was reduced when the nozzles on both sides were shifted about 50 mm into the rotor blades. This present study tentatively shows that installing the nozzles at suitable positions within the rotor area of the UAV can improve the deposition coverage and reduce the drift of the UAV application. Related research will have more practical significance in the field of UAV application technology in the future.

According to Table 7 and Figure 13, the deposition performance of the MCN was significantly better than that of the TR nozzle in the field deposition test, which was the same as that of Gao et al. [24]. The effective spraying width of the UAV application can be increased from about 3.0 m to about 5.0 m with only two nozzles installed, which not only increases the effective spraying width, but also significantly increases the average droplet deposition within the effective spraying width at the same flight speed, and enhances the effect of droplet deposition, as analyzed in the design of the atomizing disc of the MCN with a narrow top and a wide bottom and a downward opening, which contributes to faster downward deposition of the droplets. According to Figure 12, the deposition curve when the UAV is installed with the MCN operating at a flight speed of 3 m/s is similar to that achieved when the UAV is installed with the TR nozzle operating at a flight speed of 1 m/s. According to the results in Table 7 and Figure 13, the installation of the MCN can improve the efficiency of the UAV application operation by about four to five times under the approximate deposition effect.

The MCN designed in this present study has a narrow droplet spectrum and commendable atomization performance. Compared with the TR80-0067 nozzle, it has better deposition performance and can significantly improve field operation efficiency. Figure 10b shows that the RS increases significantly as the flow rate increases from 0.6 L/min to 1.6 L/min, so it is necessary to develop centrifugal nozzles that can still maintain an ideal atomization performance at high flow rates. In this paper, we only studied the atomizing performance of the MCN in the flow rate range of 0.6–1.6 L/min, which cannot provide a reference for the application to fruit trees under high flow rate, and so we will test the atomizing performance of the MCN under a higher flow rate and make further reasonable optimizations of its structure according to the results.

5. Conclusions

Comprehensive previous research was carried out in order to improve the performance of nozzle droplet atomization and deposition under high flow rate. An MCN for UAV spraying that can control droplet size was designed and evaluated for its atomization and deposition performance. The optimization results showed that the centrifugal nozzle with a curved groove, an atomizing disc angle of 120°, an atomizing disc diameter of 77 mm and six layers of atomizing discs had the best atomizing performance. When the flow rate is 0.6–1.6 L/min and the rotational speed of the atomizing disc is 6000–12,000 rpm, the droplet distribution of the MCN is concentrated, the RS is kept in the range of 1.0–1.3, the atomization performance is favorable., and its Dv₅₀ is 95–155 μm. In addition, the relationship model between its Dv₅₀ and operating parameters: Dv₅₀ = −0.00697ω + 20.23q + 164.60 (R² = 0.965), with a good fitting effect, can predict the droplet size under the specified flow rate and rotational speed, and achieve its purpose of controlling the droplet size. The deposition characteristics of the MCN were determined through field tests. At a flight height of 1.0 m and two flight speeds (1.0 m/s and 3.0 m/s), the installation of the MCN can improve the effective spraying width and the amount of droplet deposition, which can increase the efficiency of the UAV’s field application operation by about four or five times. The MCN has good atomization performance and its deposition characteristics are better than the TR nozzle, which can greatly improve the operation efficiency.

The centrifugal nozzle droplet size test system established in this study can be applied to test the atomization characteristics of different centrifugal nozzles. Meanwhile, the results of the study can provide a reference for the design optimization of centrifugal nozzles for UAVs as well as the selection of operating parameters in actual operation. Future research is expected to explore the atomization performance of the MCN at higher flow rates, further optimizing its structure for fruit tree application scenarios at high flow rates, and provide recommendations for the design of centrifugal nozzles for UAVs.